The Pdramo Vegetation of Ecuador:
the Community Ecology, Dynamics
and Productivity of Tropical
Grasslands in the Andes.
by
Paul Michael Ramsay
A thesis submitted for the degree of
Philosophiae Doctor of the University of Wales.
December 1992
School of Biological Sciences, University of Wales, Bangor,
Gwynedd, LL57 2UW.
i
Dedicated to the memory of Jack
Higgins, my grandfather.
"... a naturalist's life would be a happy one
if he had only to observe and never to
write."
Charles Darwin
ii
Table of Contents
Preface
AcknoW ledgements
vii
ix
Summar y
Resumen
Chapter 1. Introduction to the Ecuadorian P6ramos
Ecuador
The Pâramos of the Andes
Geology and Edaphology of the Paramos
Climate
Flora
Fauna
The Influence of Man
1
2
2
6
8
11
14
14
Chapter 2. The Community Ecology of the
Ecuadorian P6ramos
17
18
n
Introductio
Methods
Results
The Zonal Vegetation of the Ecuadorian Paramos
Discussion
20
36
51
64
Chapter 3. Plant Form in the Ecuadorian Paramos
77
Section I. A Growth Form Classification for the
Ecuadorian Paramos
78
Section II. The Growth Form Composition of the
Ecuadorian Pâramos
Introduction
Methods
Results
Discussion
94
95
97
107
Section III. Temperature Characteristics of Major Growth
Forms in the Ecuadorian PSramos
n
112
Introductio
Methods
Results
Discussion
113
118
123
III
Table of Contents
iv
Chapter 4. Aspects of Plant Community Dynamics
in the Ecuadorian Pgramos
Introduction
Methods
Results
Discussion
131
132
133
140
158
Chapter 5. An Assessment of Net Aboveground
Primary Productivity in the Andean Grasslands of
Central Ecuador
Introduction
Methods
Results
Discussion
165
166
169
177
189
Chapter 6. A Greenhouse Study of Competition
Between Three Andean Grasses at Two Regimes
of Water Availability
Introduction
Materials and Methods
Results
Discussion
196
197
198
201
212
Chapter 7. Overall Discussion
216
References
227
Photographic Plates
244
Appendix 1. Vascular Plant Species
255
Appendix 2. Example Calculation of Chi
Square Expectation for Transition Probabilities
272
Preface
the beginning of the nineteenth century, Alexander von Humboldt and Aime
AtBonpland
travelled amongst the Ecuadorian Andes. The 'Avenue of the Volcanoes', as Humboldt described this section of the Andes, proved an excellent outdoor
laboratory, generating ideas which have become the foundation for many aspects of
plant ecology today. Their plant collection still remains the basis for plant taxonomy
in the region (evidenced by the number of species in this work with the authority of
"H.B.K.", the abbreviated names of Humboldt, Bonpland and Kuntze). Richard
Spruce was collecting in the high pdramo grasslands above the Ecuadorian forests
around the time of Humboldt's death. Soon afterwards, the golden age of plant collecting in South America had begun to decline.
Since then, a number of trips have been made to the Ecuadorian paramos by collectors, but their work has tended to be small-scale (by comparison with the efforts
of Humboldt et al.) and concentrated in the more accessible regions. More recently,
scientific interest in the paramos has increased. A project to produce a Flora of Ecuador is now in progress, based on international research coordinated in Scandinavia.
Taxonomically, therefore, the paramo flora is relatively well-known, though in the
absence of a complete guide to the flora, obtaining identifications still requires
lengthy research in herbaria. However, other aspects of paramo vegetation have
been neglected. Lately, the Centro de Investigaciones EcolOgicas de los Andes Tropicales (CIELAT) based at the Universidad de los Andes in Mêrida has concentrated
on Venezuelan paramo vegetation, with particular emphasis on environmental conditions, productivity, population and reproductive ecology. A co-operative research
programme involving Colombia and the Netherlands (Investigaciones de Ecosistemas Tropandinos —ECOANDES) has produced major contributions to our knowledge of the community composition and biogeography of the Colombian paramos.
Despite this intense effort elsewhere in the Northern Andes, the paramos of Ecuador have not been subjected to the same degree of detailed study as those in neighbouring countries. The country is well-known as a centre of biological diversity and
this has perhaps led research programmes away from the highlands to the speciesrich forests. The miserable climate at high altitudes cannot have helped the case for
the paramo in this respect (locals have coined the term parameando to mean "It's
raining").
During the course of three student expeditions to Ecuador I had the opportunity
to make extensive observations in the high altitude grasslands of the Andes. The
work was of necessity broad-based: so little is known about paramo vegetation that it
was difficult to plan the research programme with any confidence. Access to remote
areas of paramo was difficult to assess from the reference sources available. A low
budget, coupled with these access difficulties and the terrible weather conditions,
limited the technical equipment it was possible to use (warm, rainproof clothing,
camping equipment, food and plant presses constituted a heavy pack without additional scientific equipment).
Preface
vi
However, a number of projects were carried out and will hopefully act as a catalyst
in encouraging further research in this unique ecosystem. The paramo is an ideal
place to study plant responses to environmental gradients and a set of interesting climatic conditions with pronounced daily rather than seasonal cycles. It is also an environment where, in the future, biological monitoring of the greenhouse effect might
be possible via the response of plant communities to the changing climate.
Paul M. Ramsay,
December 1992.
Acknowled ements
First and foremost, I would like to thank Mr. Ralph Oxley for proposing me for a
studentship at the University of Wales, Bangor, and for his supervision. The programme owed much to his advice and encouragement, not only academically but also
in the planning and execution of the three expeditions which formed the basis for
this thesis. I was fortunate to have him for company (albeit briefly) on one of these
expeditions.
My involvement in the field of Ecuadorian botany was the result of a chance liaison with Adrian Barnett — at the time, a fellow student at the University of Wales,
Bangor. His work on the small mammal populations of the Ecuadorian Andes led
me to consider a botanical excursion to Ecuador. Adrian's help in the crucial, early
stages of my exposure to expedition organisation and logistics was invaluable. The
U.C.N.W. El "CAJAS" Expedition to Ecuador in 1985 was the result.
The work was supported by a College Studentship maintenance grant from the
University of Wales, Bangor. The fieldwork for the research presented in this thesis
was funded by many organisations, from trust funds to academic institutions, from
large multinationals to local commercial companies. The donations ranged from 125
to /1,124. I would like to express my gratitude to the following organisations for their
support:
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Overseas Development Administration
Royal Botanic Gardens, Kew
University of Wales, Bangor
Godman Exploration Trust
School of Plant Biology, Bangor
Scarfe Charitable Trust
Royal Geographical Society
ICI Fertilisers
Augustine Courtauld Trust
Minster Agriculture Ltd.
Rio Tinto Zinc
Grafton Ltd.
Busk Charitable Trust
Thomas Kean Advertising Ltd.
British Ecological Society
Anglesey Aluminium
Whitley Animal Protection Trust
Penrhyn Industrial Estate
HRH The Duke of Edinburgh
Pillsbury UK Ltd.
The Rev. Streeter Trust
VII
Acknowledgements
viii
I must also thank all the expedition members that accompanied me on my trips to
the Andes. Particular mention must go to Pete Merrow-Smith for his company, dedication and hard-work on the two-man Paramos '87 expedition on which the core of
the data for this thesis was collected. In addition, he was a member of the earlier
Cajas expedition of 1985 which also included James Bury, Trevor Iszatt, Ralph Oxley
and Paul Thomas. The Sangay '89 team consisted of Lynn Evans, Ian Whitehead, Joe
Creed, Sarah Buckland, Nicky Legg and Michael Bassett.
During my sixteen months in Ecuador, Peter Jorgensen, Patricio Mena, Jaime Jaramillo, Henrik Pedersen, Ximena Rodriguez, Katia Romoleroux, Carmen Ulloa and
all the others at the botany department of the Pontificia Universidad Cat(Mica, Quito,
deserve a very special mention for the help and encouragement they gave — and their
friendship. Others in Quito merit special thanks: Fausto Sarmiento and Miguel
Moreno at the Museo Ecuatoriano de Ciencias Naturales for their co-operation; Dr.
Figueroa, Ministerio de Agricultura y Ganaderia, for the issue of permits for our
fieldwork and exportation of specimens; and Yolanda Kakabadse, Director of FundaciOn Natura for friendly advice. I must also thank Enrique Aguilar in Quito for his repeated help and advice in obtaining permits and information. The Instituto
Geografico Militar were very co-operative and provided maps and aerial photographs.
At the Parque Nacional Sangay headquarters in Riobamba, I received help from
Lcdo. Miguel Mejia, head of the park, and many others including Vicente Alvarez
and the guardeparques Cesar Urefia, Miguel Jaya, Bernado, Mario, Bolivar Coronel
and Jose Vallejo. These latter two also placed their cattle pastures at my disposal for
experimental plots, an offer which I gratefully accepted. My thanks also go to Bolivar's family and all the generous people at the village of Daldal. For help in the productivity study I would also like to thank the people at the Universidad Central and
the Farmacia Rosa (both Riobamba) for the use of their accurate balances for weighing the fertilizers. The Politecnico Nacional, Quito, allowed me the use of their
drying ovens and balances for the harvested grasses.
•"'
Summa
Paramo vegetation was quantitatively surveyed in 192 samples on altitudinal gradients in twelve sites in Ecuador. Thirty-one communities were identified, comprising 348 vascular plant taxa (voucher specimens deposited at Kew and QCA, Quito).
These communities could be assigned to eight general types of paramo vegetation located between the upper forest limit and the snow-line: Shrubby Sub-paramo, High
Altitude Dwarfshrub Pdramo, Tussock Paramo, Bamboo Paramo, Espeletia Paramo,
Cushion Paramo, Rainshadow Desert Paramo and High Altitude Desert Paramo.
Community types were more closely related to altitude than to other variables such
as burning, trampling, grazing and pH.
The species were assigned to ten defined growth form categories. The distributions of these categories in 192 paramo vegetation samples were described. Twelve
growth form communities were identified. Field temperature measurements of plant
parts demonstrated that some plants maintained day and night temperatures several
degrees Celsius above ambient levels.
The effect of fire on cyclical and successional processes within paramo vegetation
were described. Two experimental paramo fires reached temperatures in excess of
400°C in the upper tussock canopy, while the tussock bases were mostly below 65°C.
In a quantitative study, the majority of plant-by-plant replacements soon after a paramo fire did not depart from those expected by chance, though some trends were observed and described.
Field trials in Central Ecuador provided net aboveground grassland community
productivity estimates for five sites. Estimates ranged from 1,359 g m 2 yr-1 at 3,100
m to 512 g m 2 yr-1 at 3,950 m.
In greenhouse experiments, tussock grasses from Calamagrosth spp. at three altitudes in the paramo were grown in a diallel design under two watering regimes. In
both regimes, the grass from the lower altitude yielded more and had a higher
relative yield than that from the higher altitude. RYTs in the wettest treatment lay between 1.2 and 2.5, those of the drier treatments were not greater than 1.0.
ix
Resumen
La vegetaciOn del paramo fue muestreada cuantitativamente en 192 muestras en
un gradiente altitudinal en doce sitios en Ecuador. Treinta y una comunidades fueron identificadas, comprendiendo 348 taxa de plantas vasculares (muestras fueron
depositadas en Kew y QCA, Quito). Esas comunidades pueden ser asociadas a ocho
tipos de vegetaciOn del paramo localizadas entre el limite superior del bosque y la
linea de nieve: Sub-paramo Arbustivo, Paramo Arbustivo Enano de Altura, Paramo
Macolla, Paramo Bambii, Paramo Espeletia, Paramo Almohadillado, Paramo Desertico Seco y Paramo Desertico de Altura.
Los tipos de comunidades estuvieron mas relacionados a la altitud que a otras variables como la quema, pisoteando, pastoreo y pH.
Las especies fueron asignadas a diez categorfas definidas por su forma de creelmiento. Las distribuciones de esas categorfas en las 192 muestras de vegetaciOn del
paramo fueron descritas. Doce comunidades de diferente forma de crecimiento fueron identificadas. Mediciones de temperatura de partes de las plantas en el campo
demostraron que algunas plantas mantienen temperaturas, durante el (ha y la noche,
varios grados celsius por encima de los niveles ambientales.
El efecto del fuego en los procesos cfclicos de sucesiOn y dentro de la vegetaciOn
del paramo fueron descritos. Dos fuegos experimentales en el paramo, alcanzaron
temperaturas superiores a los 400°C en la parte superior del dosel de la macolla,
mientras que en las bases de la macolla fueron menores a 65°C. En un estudio cuantitativo, la mayoria de los remplazamientos planta por planta inmediatamente dèspues
del fuego en el paramo no fueron distintos de aquellas que se esperaba por suerte,
aunque algunas tendencias fueron observadas y descritas.
Pruebas del campo en la parte central del Ecuador provereron estimaciones de
productividad neta de la comunidad de pastos pcima del suelo para cinco sips. El
por
ango de las estimaciones fue desde 1,359 g rn - por alio a 3,100 m y 512 g rn -r
afio a 3,950 m.
En experimentos de invernadero, las gramineas de Calamagrostis spp. a tres altitudes en el 'Aram° fueron crecidas en un diseno "dialler a dos regimenes de regando. En ambos regimenes, el pasto de las altitudes bajas produjeron Inas y tuvieron
mayor cosecha relativo que aquellas de altitudes altas. RYTs en el tratamiento mas
harried() estuvieron entre 1.2 y 2.5, aquellas de los tratamientos secos no fueron
mayores que 1.0.
x
The Pdramo Vegetation of Ecuador:
the Community Ecology, Dynamics
and Productivity of Tropical
Grasslands in the Andes.
by
Paul Michael Ramsay
A thesis submitted for the degree of
Philosophiae Doctor of the University of Wales.
December 1992
School of Biological Sciences, University of Wales, Bangor,
Gwynedd, LL57 2UW.
xi
Chapter 1
Introduction to the
Ecuadorian Paramos
1. Introduction to the Ecuadorian Pâramos
2
Ecuador
situated in north-western South America, is the fourth smallest state on
E cuador,
the continent (Figure 1.1). It derives its name from the equator which passes
through the country which is bounded to the west by the Pacific Ocean, to the north
by Colombia and to the south and east by Peru (though there is considerable dispute
over this border). The total area has been estimated at 269,178 km 2. There are
twenty provinces including the Galapagos Islands and the Banco Central del Ecuador (1990) estimates the population at 10.2 million, about 38 people on average per
square kilometre. The capital city is Quito.
The Andes run approximately north-south, bisecting the country. The western
coastal region is known as the Costa, the Andean uplands as the Sierra, and the eastern lowlands are often referred to as the Oriente or as Amazonas. The Andes consist
of two parallel ranges, the Cordillera Occidental (western range) and the Cordillera
Oriental (eastern range). Many peaks are volcanic and snow-covered. The two ranges are connected by transverse ranges (nudos), rather like rungs in a ladder, with
the main centres of population occupying the depressions between them (hoyas).
Although rich in natural resources, Ecuador has not been able to sustain the high
rates of economic growth it experienced during the early 1970's. The economy was
basically agricultural until extensive exploitation of petroleum deposits in 1972 stimulated industrial development. In 1988, agriculture employed 33% of the workforce
and provided 17% of the Gross National Product (Banco Central del Ecuador,
1990). In 1984, the land use patterns were as follows: forested, 51.1%; meadows and
pastures, 17.0%; agricultural and under permanent cultivation, 9.1%; and other,
22.8%.
Tourism is a growing phenomenon in Ecuador. In 1988 tourism generated US$273
million (Banco Central del Ecuador, 1990). The majority of visitors come from the
Americas, though European tourists constitute a significant proportion. Many visitors are interested in the country's natural resources (mountains, flora and fauna).
Of course, the Galapagos Islands are responsible for attracting a large proportion of
tourists, though their numbers are strictly controlled.
The Paramos of the Andes
paramo is an ancient Spanish word for "an elevated, barren, treeless plateau", then
used to signify the inhospitable plains of Spain (Acosta-Solis, 1984). The term was
brought to South America by the conquistadores and colonialists who applied it to
the exposed grasslands of the northern Andean ranges.
Paramos occupy the vegetation belt between the upper limit of the montane cloud
forest (Ceja A ndina) and the snow-line (Figure 1.2). They occur from Venezuela to
3
1. Introduction to the Ecuadorian Paramos
ESIIHROAS
a
0
A OUITO
PORTOY EJO
3
LATACUNCA
AliBATO
ABAHOYO
CIJAYAOU L
CUENCA
HAW (
.-
ECUADOR
•Capital of
province
A volcano
9 (code)
---,national
boundary
(disputed)
scale
ma.._=n•=n
100
0
Km
Figure 1.1.
Map of Ecuador, showing volcanoes and principal towns and cities. KEY: 1, Volcan Chiles; 2, Volcan Cotacachi; 3, Volcan Cayambe; 4, Volcdn Pichincha; 5, Volcan Antisana; 6, Volcan Sumaco; 7, Volcan Iliniza; 8, Volcan Cotopaxi; 9, Volc6n
Chimborazo; 10, Volcan Tungurahua; 11, El Altar; 12, Volcan Sangay.
Ecuador and northern Peril (where they are called jalcas), with outliers in other parts
of Peril, Bolivia, Panama and Costa Rica (Cleef, 1978). The paramos show similarities to the vegetation of other high elevation tropical environments (Figure 1.3).
These high montane ecosystems are found in the Central Andes (puna), East Africa
(afroalpine or moorland), Malaysia (tropical alpine) and Mexico (zacatal, though
Gomez-Pampa (1973) and Breedlove (1973) speak of 'paramo' vegetation in Veracruz and Chiapas). Some authors advocate the use of paramo to describe all of these
tropical alpine vegetation types (for example, Walter, 1973, and Lind & Morrison,
1974). Monasterio & Vuilleumier (1986) suggest the use of terms such as Andean
paramo, African 'Ammo, Papuan paramo, and so forth. The term tropicalpine has
lately been used to describe all alpine areas in equatorial regions (for example,
Smith & Young, 1987b). Other authors have suggested the use of the term "tundra"
for all treeless regions, north and south of latitudinal tree-lines and above natural altitudinal tree-lines (for example, Holdridge, 1957; Tieszen & Detling, 1983). In this
work, I will use tropicalpine and the more traditional nomenclature for regional
vegetation (paramo, afroalpine, etc.).
1. Introduction to the Ecuadorian Paramos
(A)
A
A
Os
6000
A
Cn
Ca
e
ti
0,1
.1
8
__A__
5000
An„ A A
--,..-- n 1
A ,
NR
Ac
1
s.
N
I
I
I
DRY I DESERT ! DRY
!HUMID I
PUNA ' PUNA
I PUNA I PUNAI
1
TROPCAL
PARAMO
I.
4000
g4000
I
I
I
I
1
• ' -.. — 1
. I
.1•(
I
I
t1
1
1
T
1
v A cAm
1
I
p_3000
-
D
CLOUD
W STEPPE I SHRUBBY \
II SHRUB
FOREST
RES
II STEPPE \
BITVH
I
POLYLEPIS
•
I<
I
2000
I
,
I
„4. _ ._ ...i.
....
I TROPICAL / S
HAIN
FOREST I FOREST l' %o
I
\
DESERT_ I ,L - `
....
7.‘
FOREST
cc
I
II5ii
10°N
0
••
0.0
10
A
A
Ch
, HUMID I
Au
-w
4000r — —
—.
ii\--/
CLOUD
FOREST
P_ 3000
I
.
n
s‘
I
4 \ PUNA
•
TROP CAL
PARAMO
L.I.J
I DRY
•
n
N
t PUNA
, I
. 1
\ I
i-
f- -, n
--I
I
TROPICAL
RAIN FOREST
2000
•
I
1".. .....1
4
•
1000 - - - ." 1
TROPICAL'
SAVANNA'
I
0
HUMID
SAVANNA
TROP CAL
RAIN FOREST
•
Ca
I
L
MONTANE
STEPPE
\
. .
0
• • •
B
O. Cue
20
10
•
AA •
\
..
COW
N
,
...r -- n •••
I PRE
‘1.„
I M ONTANE :
1 FOREST 1
\PARAMO AB
1,),/
''n••,\\
I
.
. .
. ....
....
I STEPPE I TEMPERATE ....."7".
I FOREST
I
I
I
I
I
10°N
I \•
DRY
I TROPICAL
I FOREST
/-
LCW-GRCWING
.nI \
-
<
•
PA
A
AIL
10
50°S
Tu
ACo
..A.
‘,
COLD
40
• •
SC
30
A
6000
\
1 '- -- -
ARA
ceEsTlys,12TTJN_DrikA_
e_FEAVED
20
•
•
(B)
--E..
A
B
i \ x IISC...i.16,:•.,\:‘%
1 ‘. i
iu j vl CORI sei.'1Y4-,\
1_, 1
u¢..: I
TROPICAL
DESERT
-L.,z
S
E
I a I-*-. c `
la`
.r
1000 i I
5000
r.
. \ I \
\
n• I
'`,.,
30
•
Cuz LT LP J
40
50°S
•
BA
•
PA
Figure 1.2.
Vegetation zones of the Andes: (A) Western Andes, (B) Eastern Andes, following
Mann (1968). The zones are divides by dashed lines. Major mountains are shown
at the top of the figure and pricipal cities at the bottom.
Mountain Codes: Ac, Aconcagua; Am, Ampato; B, Balmaceda; Ca, Cayambe; Ch,
Chan': CM, Cerro Muralion; Co, Cotopaxi; Cn, Colon; CP, Cerro de Pasco; D, Domeyco; I, Illampii; II, Ilimani; M, Misti; Ma, Maipo; NR, Nevado del Ruiz; OS, Ojos
del Salado; S, Sangay; T, Tronador; Tu, Tupungato; V, Villarica.
City Codes: B, Bogota; BA, Buenos Aires; C, ConcepciOn; Ca, Caracas; Cue,
Cuenca; Cuz, Cuzco; G, Guayaquil; I, Iquique; J, Jujuy; L, Lima; LP, La Paz; LT,
Lake Titicaca; M, Mendoza; PA, Puerto Arenas; 0, Quito; S, Santiago de Chile.
Other abbreviations used in text: E.G. FOR., Evergreen Forest; SCL. FOR., Sclerophyll Forest.; TEMP SAVANNA, Temperate Savanna.
4
1. Introduction to the Ecuadorian Paramos
5
Figure 1.3.
Regions of the world with high altitude tropical vegetation: 1, Northern Andes
(paramo); 2, Central and Southern Andes (puna); 3, Guyana Highlands (tepuis); 4,
Central American Highlands (paramo and zacatal); 5, East African mountains
(afroalpine moorland); 6, Ethiopian Highlands (tableland); 7, Malaysian Highlands
(tropical alpine grasslands). Adapted from Monasterio & Vuilleumier (1986).
"Paramo" and "Rain Pdramo" appear in Holdridge's (1967) life zone classification
system, but many Ecuadorian paramos are found in the "Moist Forest" and "Wet
Forest" life zones. Though an adequate category is not found in Fosberg (1967),
"Steppe" [2G] would seem to be most appropriate for the majority of paramo vegetation.
The Andean pdramos should not be confused with the paramos templados of high
latitudes in South America (40°S), which show distinctive dissimilarities from equatorial paramos. They are also distinct from Neotropical savannas which have entirely
different floristic, physiognomic and physiological features (Huber, 1987).
Some authors have suggested the limit of distribution of Espeletiinae tribe (Cornpositae) as a guide for the extent of paramo vegetation (for example, Fosberg, 1944).
In Ecuador, only one species of Espeletia is found (Espeletia pycnophylla). One subspecies is restricted to northern areas, such as the paramos of El Angel, Volcdn
Chiles and El Playon de San Francisco, the other to a valley in the Llanganatis (Central Ecuador). Elsewhere, Espeletia is absent, with Puya species becoming more
1. Introduction to the Ecuadorian Paramos
6
dominant. Ecuadorian paramos generally lack Chusquea bamboo species, which is a
dominant plant in more humid paramos.
Acosta-Solis (1984) has estimated that the total area of paramo in Ecuador is
somewhere between 25,000-28,000 km2 , though this is probably an over-estimate,
since he includes the snowcaps and assumes that the lower limit of the paramos is at
3,000-3,300 m. The actual altitude of the upper forest line in Ecuador varies considerably, from under 3,000 m to about 4,000 m, depending mainly on climatic factors and
human interference, though most often the transition zone exists between 3,4003,500 m. The upper limit of the 'Alamo is the snow-line, ranging from 4,400-4,900 m.
Encalada (1986) offers 20,000 km as a more reasonable evaluation of the extent of
the Ecuadorian 'Aram°, and Bonifaz (1981) goes further with an overall estimate of
19,610 km2 , calculating the area of subparamo (3,500-4,000 m) as 13,310 km 2, with
6,300 km 2 of paramo occupying the land above 4,000 m.
Some authors believe the existence of the paramos to be the result of man's activities, in particular burning (Ellenberg, 1979; Laegaard, 1992). Without doubt, clearance of forest has had a major impact on the extent of the lower reaches of the
paramo grasslands. The presence of pockets of quiiivales (woodlãnds of Polylepis
spp. trees) and Gynoxys spp. woodlands amidst the extensive grasslands of paramo
has been used as support for this argument, stating that such pockets of woodland
represent refugia from fire and that without burning, the whole landscape would be
dominated by trees. Simpson's (1979) revision of the genus presents a summary of
the debate. A.P. Smith (1978) carried out field experiments with Polylepis sericea in
the Venezuelan Andes and found that seedlings only survived on rocky talus slopes
and showed 100% mortality over one year in open paramo and bare soil. He concluded that the cause of mortality was an interaction of competition and climatic
stress. Ramsay (1988) also points out that several large islands in lakes in the paramo
of Cajas (southern Ecuador) do not support woodland vegetation. This will be discussed in more detail in the final chapter of this thesis.
Geology and Edaphology of the
Paramos
The
Themassive Andean range owes its existence to the consumption of the oceanic
tectonic plate beneath the continental South American plate (at a rate of
about 6cm per year). It accounts for the crumpling of the stable continental margin
to form the belts of fold mountains that now constitute the eastern ranges of the
Andes, for the birth of the great Andean volcanic cordillera to the west, and for the
continental growth of western South America (James, 1973).
The Ecuadorian highlands or Sierra, according to Baldock (1982), is a composite
mountain belt, formed by two (or more) distinct orogenic episodes. In the Palaeozoic, the Cordillera Oriental (also known as the Cordillera Real) was formed as high
pressures forced fold mountains upwards in a line roughly north-south. This range is
1. Introduction to the Ecuadorian Paramos
7
underlain by metamorphic rocks capped by Cenozoic volcanoes. In the late Mesozoic to Cenozoic, the Cordillera Occidental developed, probably as an island arc
(Baldock, 1982). This range is comprised of Cretaceous-Lower Tertiary volcanic and
volcaniclastic rocks, along with sedimentary rocks which were deposited between the
islands. Many of the higher areas are covered by Neogene volcanics. According to
Hoffstetter (1986), there are eight active volcanoes in Ecuador, and twenty more
have recently become extinct.
The inter-Andean valley is a graben, a fault between two lines of weakness, which
appeared during the Neogene uplift and arching of the Andean mountain belt. This
'Avenue of the Volcanoes', averaging 2,500-3,000 m, becomes morphologically less
distinct to the south, but is still evident geologically (Baldock, 1982).
During the past there have been several glaciations in highland Ecuador. Glacial
features such as moraines, boggy U-shaped valleys, tills, cirques, fluvio-glacial deposits, tarns and glacial lakes, polished bed-rock, roches moutonêes and erratic rocks
are frequently observed, especially in the super-paramo (Cleef, 1981). Repeated glaciations and catastrophic volcanic events have prevented the undisturbed development of soils over much of the paramo area.
There have been few detailed studies of paramo soils. General works dealing with
the soils of South America, have tended to group together the soils of the high
Andes (for example, the "paramo soils" of Beek & Bramao, 1968). Studies of particular countries or regions have used similar general terms or local names for the soils
present.
Sturm (1978) reviewed previous pedological studies in the northern Andes in his
paper on the soil flora and fauna of a Colombian paramo. Cleef (1981) presents an
extensive set of soil data, collected during his characterisation of the paramos of the
Colombian Cordillera Oriental.
In general, the soils of the paramos of the northern Andes are very dark in colour,
acidic, rocky and poorly developed, low in inorganic nutrients but with a high organic
content in the uppermost horizon (Sturm, 1978; Baruch, 1979; Cleef, 1981). The decomposition of the vegetation is slow (reflected in high carbon:nitrogen ratios —
Cleef, 1981), ascribed by Jenny (1948) to the cold temperatures rather than the soil
moisture content. Frei (1958) assigned most Ecuadorian 'Aram° soils to the Black
Andean Soil Group. Soils of this type collected in Ecuador were found by Miller and
Coleman (1952) to be characterised by a relatively high organic content (9-12% dry
weight), high cation exchange capacity, low exchangeable calcium and magnesium,
high concentrations of exchangeable aluminium and a high capacity for phosphate
adsorption, with a pH of 4.1-4.8. However, pH values from 3.8 (Pena Herrera,
quoted in Acosta-Solis, 1984) to 6.2 (Grubb, Lloyd & Pennington, manuscript) have
been recorded in Ecuadorian paramos. Clearly, this variation in pH will alter soil
properties such as cation exchange capacity.
Horizon A can be thick, 30cm to 2m in depth (Acosta-Solis, 1984). Boundaries between horizons are not usually pronounced (Jenny, 1948; Acosta-Solis, 1984).
1. Introduction to the Ecuadorian Paramos
8
A common component of paramo soils is cangagua, a mixed loessal volcanic fallout deposit related to volcanic eruptions and cold, possibly dry periods of glaciation
(Clapperton & McEwan, 1985; Vera & LOpez, 1986). If the soil is deep enough, particularly if cangagua is present in large amounts, there is little influence of the bedrock on the vegetation (for example, the Nevado de Sumapaz, Colombia, is
underlain by limestone but the soil pH is not markedly different than elsewhere;
Cleef, 1981).
In general, these descriptions refer only to the paramo soils beneath tussock grassland at intermediate altitudes. In fact, a mosaic of soil types occur, under the influence of factors such as topography, geological history and altitude. Under very wet
conditions, clays and peat may develop. Though Sturm (1978) found little or no podzolic soils in the paramos, the Cambridge Llanganati Expedition 1969 (1970) found
podzols at 4,200 m in Cerro Hermoso, Ecuador, where a distinct iron pan had
formed 4-10cm below the surface. They also described iron-oxide mottling in soils
with restricted drainage. Furrer & Graf (1978) studied glacial and periglacial phenomena in the higher reaches of the Ecuadorian Andes where stony soils are common, sometimes showing periglacial features such as needle ice, structured soils
(stripes and polygons) and screes.
Climate
From the outset, it is important to differentiate between the alpine climate of tem' perate regions and the tropicalpine environment of the Andes. Tropical climates
do not show as much seasonal variation in temperature and day length as the midand high-latitude climates (Sarmiento, 1986). A characteristic feature of the tropical
climate is that the yearly variation in temperature is not so great as the daily variation. The main climatic pattern is, therefore, the marked circadian cycle rather than
the seasonal pattern. This is the diurnal temperature climate described by Troll
(1968).
The seasonal constancy of temperature holds true for all altitudes. However, temperature decreases on average with increasing altitude with a lapse rate of around
0.6°C per 100 m of altitude. Although there is a constancy of mean temperature, the
minima do vary seasonally. This is particularly important with regard to the number
of frost days.
Rainfall is much more variable than temperature, since it depends on a whole
range of factors relating to the geography of each mountain system. Most tropical regions have two to four seasons with heavy rainfall alternating with dry, almost rainless conditions. The tropical mountains are subject to this rainfall seasonality too, in
all but the wettest and most arid regions. The circulation patterns of the atmosphere
have a direct influence on the climate of the Andes (Eidt, 1968). The Andean range
is not just a watershed between east and west, it is also a climatic dividing line, bisecting the major air masses of the southern hemisphere (Sick, 1969). Sarmiento (1986)
summarized rainfall patterns along the Andean chain. From Venezuela to Central
1. Introduction to the Ecuadorian Paramos
9
Colombia rainfall follows a bimodal pattern, with the main dry season from December to March. Around the thermal equator, between 4°N and 5°N, the rainfall pattern is still bimodal, but the major dry season shifts to the middle months of the year.
In northern Ecuador, the secondary dry season tends to disappear and the climate
becomes more or less two seasonal, showing a pronounced mid-year minimum. This
bi-seasonality is reinforced southward, so that southern Ecuador experiences drier
weather from June to January. The trend culminates in a two-season regime in the
Peruvian Andes. Emphasising this bi-seasonality, Sarmiento (1986) reports that in
the Andes of Ecuador, 70% of the annual total precipitation falls in the wet season,
compared with 80% and 90% for parts of southern Peru and Bolivia, respectively.
Johnson (1976) noted several other climatic gradients along the Andes from Ecuador to Bolivia. Total rainfall decreases smoothly from north to south. In northern
Ecuador, the paramos are quite humid, but in southern Ecuador (at about 2°S) conditions become semi-arid (Johnson, 1976). This is accompanied by unreliability of rainfall. Schwabe (1968) relates this to the position of the Intertropical Front
(responsible for convection rain patterns) which is located near the thermal equator,
and so lies close to Colombia and Ecuador all year round, resulting in rainfall
throughout the year. Further south, away from the continuous influence of the ITF,
rainfall becomes more seasonal. A further trend is the appearance of a cold season
during the southern hemisphere winter as one moves south. However, despite this NS variation, the main axis of environmental variation is E-W (Troll, 1968). This results in climatic zones parallel to the Andes (see Sarmiento, 1986: Fig. 7).
The temperature regime can be governed to a large degree by the rainfall patterns, since rainfall is associated with cloudiness. Tropicalpine environments rely
heavily on direct solar radiation as a temperature input and cloud cover dramatically
reduces insolation. At night, the cloud cover reduces thermal loss via long-wave radiation. Thus, cloudiness buffers temperature variation, restricting the amplitudes of
maximum and minimum temperatures. In this way, seasonality in rainfall leads to
thermoperiodism, an annual cycle with dampened temperature oscillations and
higher night minima during wet seasons and greater temperature fluctuations and
lower night minima during dry seasons (Sarmiento, 1986).
Precipitation is, therefore, of great importance in determining a host of climatic
conditions. Rainfall pattern is dependent on a range of factors, most influential of
which are altitude, topography and geographic position. Weischet (1969) and Lauer
(1976) showed that, in general, there is a maximum amount of precipitation at
middle altitudes, occurring a few hundred metres above the cloud base where drops
begin to form and corresponds to the position of montane cloud forests. Above and
below this altitude, the amount of rainfall steadily decreases.
The Andes are characterised, like other mountain ranges, by irregular topography.
Their geological youth means that slopes are still very steep and level ground is rarely found. As Troll (1968) points out, under such circumstances, meteorological data
collected in one valley may be completely different to that in the next valley: the socalled "Troll effect". Locating meteorological stations in valleys leads to an underestimation of rainfall, and run-off data show that in many mountain areas, as a result of
1. Introduction to the Ecuadorian Pâramos
10
the "Troll effect", precipitation measurements are utterly unrepresentative (Flohn,
1974).
Topoclimates may be more important than regional climates. Geiger (1966, 1969)
and Barry & Van Wie (1974) stress three key topoclimatic factors. Slope aspect and
slope angle modify diurnal temperature and humidity changes through their action
on insolation. This was well-known in the Alps where agriculture was located on
sunny ('adre), south-facing slopes whilst shaded ('ubac), north-facing slopes remained under forest (Garnett, 1937). This is more important in tropical mountains
where diffuse radiation accounts for much less a proportion of the total radiation,
and direct sunlight is consequently more influential. Therefore, steep slopes receive
less radiation than flatter ones. Holland & Steyn (1975) demonstrated that differences between equator-facing and pole-facing slopes (in the absence of cloud) are greatest on steep slopes and at mid-latitudes —and least in equatorial and polar regions.
Their studies suggest that thermal microclimatic differences at tropical latitudes are
generally insignificant. However, these studies did not take into account cloudiness.
In the tropics, E-W slope aspect is more important than N-S slope aspect, mainly as a
result of differences in precipitation on lee- and windward slopes and the effect of
afternoon cloudiness (a common situation) on western-facing slopes (Smith, J.M.B.,
1978; AzOcar & Monasterio, 1979, 1980b). The result is that east-facing slopes receive more sunlight. A third factor is topography, with its relation to catabatic and anabatic winds and night-time inverted temperature regimes (cold valley bottoms).
Winds are usually gentle, but can influence plant growth (Smith, 1972). Strong
winds can be a continual presence locally, where consistent directionality can create
distinct microclimates on leeward and adjacent windward slopes (Smith, 1972). The
absence of strong winds in the paramos has been used to explain why stem rosette
species are dominant in many regions (Cleef, 1978 —reply to Hnatiuk).
The fall in atmospheric pressure with altitude leads to a lowering of the air watervapour pressure at high elevations. This may be limiting to plant growth at these altitudes. High UV-B input is, to some extent, compensated by reduced epidermal UV
transmissivity (Robberecht, Caldwell & Billings, 1980). Barnes, Flint & Caldwell
(1987) provide corroborative evidence for this but suggest that other factors, in addition to the shielding of UV-B radiation by UV absorbing pigments and/or leaf structures, are also involved.
Vegetation types are clearly related to prevailing climatic conditions. Cuatrecasas
(1968) and Monasterio & Reyes (1980) report that a mean temperature of 10°C
roughly corresponds to the climatic boundary between montane and 'Aram° climates, which, according to Sarmiento (1986) corresponds (on wet slopes at least) to
the first appearance of a few days of night frosts (that can occur at any time of the
year). The number of days of frost increases sharply around 3,300 m, which coincides
approximately with the upper limit of montane forests. By 4,500 m, the number of
frost days per year rises to about 100, and the nival limit is usually reached between
4,700-4,900 m. On drier slopes, frosts appear at lower elevations, but the permanent
snow-line may be much higher.
1. Introduction to the Ecuadorian Paramos
11
For a plant to be a successful colonist of the paramos, it must be able to cope with
several climatic factors which present difficulties to plant existence. Sarmiento
(1986) considers freezing temperatures at night and insufficient radiation and sub-optimal temperatures during daylight hours as the main constraints on life in tropical
mountains. Plants may evolve several responses to these factors: modifications in
form, function, behaviour or all three (Sarmiento, 1986). These adaptations will be
discussed in more detail in a later chapter on paramo growth forms. Particularly important to many species is the exploitation of microclimates.
Flora
The
The present 'Aram° vegetation is at least four million years old (Van der Hammen
Cleef, 1986). These authors estimate that the paramo has 30 endemic genera
out of 300 (10%), while in the Cordillera Oriental of Colombia 35% of the species
are endemic (Van der Hammen & Cleef, 1986).
Central to the understanding of paramo vegetation is a knowledge of its development. Van der Hammen & Cleef (1986) present a valuable account of the evolution
of the high altitude vegetation of the northern Andes. About 4-5 million years BP
(before present), lowland and mountain savannas and other tepui-like grasslands,
determined by edaphic and/or climatic (other than temperature) factors, existed
amongst the forest. This "pre-paramo" was made up of floristic elements from Andean, even Sub-Andean forest vegetation. It was to be an important source of
(Neo)tropical and Andean elements of later paramo vegetation types, once the final
upheaval of the Andes took place.
During this great uplift in the Pliocene and Early Pleistocene (some 4-2 million
years BP), night frosts became important above 2,300 m and the flora became
adapted to the new conditions. The "proto-paramo" vegetation was wider in extent
than today's paramo since, despite milder temperatures, the forest line was low
owing to an undeveloped upper Andean forest flora. Some evidence for the existence of a "proto-paramo" flora is provided by pollen analysis. Though poor in genera, some of the taxa are characteristic of the present-day 'Ammo, while others are
now absent. Half of the floral elements were of tropical origin, the remaining 50%
temperate (mostly from the south).
The immigration of temperate species then proceeded more rapidly (both from
the south and north) leading to the appearance of 'Aram° vegetation much as we
know it today. During interglacial times, the paramo had an archipelago-like distribution, similar to that at present. During the glacials, however, the extent of the Oramo resembled that of the proto-paramo, possibly even larger.
Since proto-paramo times (2-0.5 million years BP), there have been 15-20 major
climatic cycles (each of approximately 100,000 years duration), which have displaced
the forest line. Maximum opportunities for the migration of paramo plants were afforded when the upper forest line was around 2,000 m (some 5-10% of the time): the
1. Introduction to the Ecuadorian Paramos
12
area of paramo being increased several times. Immigration to the paramo was also
optimum at this time. For 40% of the time, the paramo occupied its present position
(or slightly higher), and migration and immigration were minimal. For the remaining
50% of the time, the extent of the paramo was intermediate.
The altitudinal movement of vegetation belts was not merely a vertical displacement: the flora had to contend with different soil types (sometimes after a glaciation
there was no soil at all) and wetter/drier conditions. In this way, extinction of elements of the flora could occur when their niches were temporarily unavailable. In
Ecuador and parts of Colombia, further extinctions may have resulted from the continual, often catastrophic effects of volcanic activity.
Tropical alpine environments may, therefore, be thought of as typically insular and
short-lived (Smith & Cleef, 1988), relying on long distance dispersal rather than local
adaptation as the primary source of recruitment to tropical alpine floras. The equilibrium theory of island biogeography (MacArthur & Wilson, 1967) has been used to
calculate the expected number of species for different paramo regions. Vuilleumier
(1970) found good correlation between expected and observed numbers of species
for Andean birds. Simpson (1974) found that plant species diversity conformed to
the theory of island biogeography (though greater correlation was found using 'Aram° area and distance measures at glacial times than those of the present day). It is
probable that the continual displacement of paramo vegetation belts prevents a state
of equilibrium from being reached in all but the most rapidly colonizing groups (such
as birds).
Cabrera (1957) places the Taramo Province' within the Andean Domain', which
in turn is a part of the Neotropical Region'. The most important families in the
Taramo Province' are Compositae (Asteraceae), Gramineae (Poaceae), Cyperaceae, Cruciferae (Brassicaceae), Geraniaceae, Valerianaceae, Bromeliaceae, Caryophyllaceae, Umbelliferae (Apiaceae), Leguminosae (Fabaceae) and Rosaceae.
All paramos merit their amalgamation into a single biotic Province because of similarities in evolutionary history, environmental conditions, fauna, flora and vegetation
(Monasterio, 1980c). However, despite these likenesses, there are clear sub-divisions
within the Province. For example, Monasterio (1980c) differentiates the Province
into three groups: the jalca and Ecuadorian paramos, consisting mainly of tussock
grasses with genera from extra-tropical regions; the paramo of Colombia, with equal
importance of grasses and rosettes; and the Venezuelan paramos, where some dualism exists but the rosettes are more dominant.
Cleef (1981) reports that more than 300 native vascular plant genera are represented in the paramos of the northern Andes (with 260 of them in the Colombian
Cordillera Oriental —comprising about 700 species). There have been numerous collections of plants made in the Ecuadorian paramos, ever since von Humboldt and
Bonpland made their way through the 'Avenue of the Volcanoes" in 1802 (Sandwith,
1926). As a result, the taxonomy of the region is relatively well-advanced, though
plant identification is still only possible by comparing specimens in herbaria as keys
are incomplete or unavailable, despite the excellent efforts of the Flora of Ecuador
project.
1. Introduction to the Ecuadorian Pâramos
13
The literature contains several descriptions of paramo flora in Ecuador (for
example, Heilborn, 1925; Diels, 1934; Acosta-Solis, 1937, 1966, 1984, 1985; Penland,
1941; Drew, 1944; Svensen, 1945; Paredes, 1962; Lojtnant & Molau, 1982; CerOn,
1985). More recently, some quantitative descriptions using methods of Braun-Blanquet have been published (011gaard & Balslev, 1979; Balslev & de Vries, 1982;
Black, 1982; Ramsay, 1988). However, these researches are still far behind progress
in Venezuela and Colombia, where some extensive quantitative comparisons have
been made (for example, Farinas & Monasterio, 1980; Cleef, 1981; Baruch, 1984;
Rangel & Franco, 1985; Franco, Rangel & Lozano, 1986).
Harling (1979) recognizes three types of paramo in Ecuador. Grass paramos or pajonales occupy most ground below 4,000 m. Cushion paramos are found above 4,000
m, as the tussock grasses are replaced by cushion plants. Harling's third category contains the desert pdramos or arenales, which inhabit the higher reaches of the Andean
ranges, where conditions restrict plant growth.
The paramo vegetation of Latin America characteristically shows altitudinal zonation, resulting from progressively higher stress factors with increasing elevation. Cuatrecasas (1934, 1958, 1968) classified the typical Colombian paramo into three belts:
• Sub-paramo (3,000-3,500 m) — the transition zone between
the upper Andean forest and the open paramo, dominated
by bushes of Compositae, Guttiferae and Ericaceae.
• Grassy Paramo (3,500- c.4,100 m) —characterised by tussock grasses (mainly Calamagrostis and Festuca spp.) with
thickets of Hypericum and Senecio vaccinioides. Isolated
woodlands exist in this belt, consisting of trees of the genus
Polylepis. Another feature of this zone is the presence of
Espeletia spp., though this is not usually true of Ecuadorian
paramos.
• Super-paramo (c. 4,100 m and above) — the extreme environmental conditions restrict plant growth and cover is
sparse. Characteristic are Cu/citium spp.
Despite this zonal approach, clear separations between the community types have
not generally been found in quantitative studies, rather a continuum of community
change (Crawford, Wishart & Campbell, 1970, for Perd; Farinas & Monasterio, 1980,
and Baruch, 1984, for Venezuela).
The geological history of the Andes has left the range with a "small-scale mosaic
of types of landscape" (Schwabe, 1968) consisting of high ridges and deep valleys,
clearly showing the influence of glaciation. This high topographic diversity is reflected in a wide spectrum of plant communities with azonal distributions. Cliffs,
rock outcrops, recent moraines, river-banks and waterlogged ground support their
own community types (Grubb et al., unpublished).
1. Introduction to the Ecuadorian P6ramos
14
Alpha diversity (within habitat species richness) appears to be similar in tropical
alpine, temperate alpine and arctic communities (Hanselman, 1975). However,
gamma diversity (regional among mountain species richness) appears to be greater
in tropical alpine communities (evidenced by high levels of endemism and vicarious
species complexes in tropical mountains). This is doubtless a function of the islandlike nature of the 'Aram° regions in a sea of tropical vegetation with pulses of migration during glaciations alternating with periods of speciation in the isolated paramos
when the climate was warmer (Simpson, 1975). In addition, Janzen (1967), Huey
(1978) and Smith (1987) suggest that tropical mountains may be effectively more isolated from each other than are temperate mountains with similar topography. This
has wide implications for understanding the processes of plant biogeography and
evolution in the high Andes.
Fauna
The diversity of paramo habitats supports a variety of animal species. The grasses
' of the paramo provide a living for a number of herbivores such as deer (Odocoileus vitginianus), rabbits (Sylvilagus brasiliensis) and numerous small rodents (16
species in three orders have been collected in the paramo of Cajas, southern Ecuador, by Barnett & Gordon, 1985). These herbivores (in addition to hunting by man)
provide food for several large carnivorous species: puma (Fells concolor), spectacled
bear (Tremarctos ornatus), Andean fox (Dusicyon culpacus) and large birds of prey.
Carrion-feeders include the magnificent condor (Vulturgryphus). All of these large
carnivores, especially the condor, have declined considerably over the last hundred
years as a result of hunting and killing practices of farmers with domestic herbivores.
Whymper (1892) describes very large populations of condors: for Chimborazo he
writes, "When the atmosphere permitted us to look below, we commonly saw a
dozen [condors] on the wing at the same time."; and for Antisana, "A score or more
continually hovered over the pastures." Such sights have long since disappeared.
Twelve species of hummingbird (Trochilidae) inhabit the pdramos (Wolf & Gill,
1986). Cordillera Snipe (Chubbia jamesoni) are common, as are the Paramo Pipit
and members of the genus Cinclodes. Ducks and teals live in some paramo lakes,
along with populations of introduced trout (Salmo gairdnerii and S. trutta).
Descimon (1986) observed a very sharp division in lepidopteran diversity between
the upper montane forest and the paramo. He explained this in evolutionary terms:
the relatively recent uplift of the Andes had not allowed sufficient time for local
species to evolve and take advantage of the paramo habitat and the long distance
from the nearest source of preadapted fauna (in Tierra del Fuego) had prevented immigration from this region.
1. Introduction to the Ecuadorian Paramos
15
The Influence of Man
A lthough the paramos are uninhabited, man's influence is strong there. Human
'communities in the high Andes have been, in general, self-sufficient. Their isolation, coupled with access to a broad resource base, has contributed to this reliance
on subsistence agriculture.
Several decades ago, Ecuadorian agricultural systems in the highlands were among
the most anachronistic in the hemisphere (Haney & Haney, 1989). An agrarian reform programme was launched to improve the lot of the campesinos (peasant farmers) in 1964 but has been only partially successful. Some rural people own small
patches of land (minifundios), which are often incapable of providing even a subsistence living for their owners. In a study of the minifundio community of Santa Lucia
Arriba in highland Ecuador, 80% of households owned less than 1 ha of land or were
landless, though 61% had purchased land over and above what they had inherited
(Forster, 1989). Others work as huasipongueros (tenant farmers) on large haciendas,
in return for which they may receive a house and a small plot of land on which they
may grow crops.
The diet of the campesinos remains essentially traditional, despite commercialisation of the rural economy and changes in production and eating habits (Herrera,
1987). The main foods are barley meal, potatoes, rice, barley grain, beans and wheat
flour, and deficiencies in protein and energy requirements are very common, especially in children (Herrera, 1987).
The campesinos recognize various agricultural zones on the elevational gradient,
summarized for northern Peril by Brush (1976). The paramos are not used for cultivation owing to frequent frosts. Instead, the extensive low-value paja (tussock grass)
provides forage for cattle, horses and sheep. Below the 'Aram° zone is a belt of
tuber cultivation (e.g., potatoes and oca) and below that is a zone of cereal cultivation (e.g., maize and quinoa). The ceja andina provides a source of timber and fuelwood.
The human ecology of the Andes is in a state of flux. In pre-Hispanic times, settlements were higher than at present. The lowering of Andean villages after the Spanish conquest has been attributed to increased dependence on cereal crops under Old
World influence (Brush, 1976). In more recent times, this trend away from the paramos has been reversed. The depletion of local natural resources (particularly fuelwood) as a result of increased population pressure, diminishing isolation and the loss
of self-sufficiency has generated a need for cash crops. Brush (1976) describes livestock as a "living bank account on the hoof", since cattle can be readily converted
into money when required, and the paramo zone has seen an increase in grazing
pressure.
As Brush also points out, the tuber zone has traditionally been the major focus of
subsistence activity in the Andes. Pressure to increase production has resulted in
pushing crops away from their effective limits to their absolute limits in the sub-para-
1. Introduction to the Ecuadorian Paramos
16
mo (where risks of frost damage and disease are higher). Gondard (1988) also identified this trend —of advancing pioneer fringes of agriculture along the edge of natural
vegetation formations —throughout the Andean region of Ecuador.
In a study of perceived stress factors in Ecuadorian Andean campesinos, Stadel
(1989) found that low temperatures at high elevations and isolation were seen as
major worries, but steep slopes and erosion were not. This is perhaps a consequence
of the modern view of maximization of yields, rather than of sustainability.
The treatment of montane forest is a prime example of a non-sustainable approach to local natural resources. Day-to-day survival has forced many communities
to destroy the ceja andina completely in the quest for wood, now leaving them without fuelwood —"the poor man's energy crisis" (Brandbyge 84 Holm-Nielsen, 1987).
In some of the drier paramos, where trees are absent, shrubs are collected for fuel.
This happens, for example, on the slopes of Volcan Chimborazo, where each household collects a horse-load of Chuquiraga jussieui branches every five days or so, for
cooking and heating. A similar practice has been observed in puna vegetation types
in parts of highland Perd and Bolivia (West, 1987). This practice can cause degradation of the ecosystem if the shrubs are over-utilized.
With increasing utilization of the paramos as pastures, burning has become more
widespread in an effort to alter the nutritional value of the paja tussocks by removing
choking dead leaves and stimulating the growth of succulent new shoots. Ellenberg
(1979) puts forward a case for the extreme modification of tropical mountain ecosystems of the Andes by burning and trampling. As he points out, nearly all regions of
the Andean countries are composed of mixtures of ecosystems, representing different stages of landscape history and different levels of human interference.
Despite the hostile climate, the paramos will become increasingly important to Andean peoples as a result of population pressures in a country with the greatest density of humans per cultivable unit and the highest birthrate on the continent.
Chapter 2
The Community Ecology
of the Ecuadorian
Paramos
17
2. Community Ecology of the Ecuadorian Paramos
18
Introduction
high altitude grasslands of the northern Andes of South America are characT he
terised by two great vegetation types, the paramo and the puna. They both cover
similar altitudinal ranges, but are fundamentally distinct, largely in terms of humidity
(Quintanilla, 1983; Acosta-Solis, 1984). The paramos stretch from Central America
to northern Peril and Bolivia, where the drier puna vegetation begins and extends
southwards into Chile (Cleef, 1978).
The paramos of Ecuador are interesting in that they occur towards the southernmost limit of paramo vegetation, and show a trend from humid paramos in the north
to drier paramos in the south (Acosta-Solis, 1984). There is also a similar trend from
the humid vegetation on the eastern Andean range (receiving moisture-laden air
from the Amazon basin) to the drier western range (corresponding to the east-west
climatic trend described by Sarrniento, 1986). There is a complex interaction of these
trends to produce various kinds of paramo vegetation within the same latitude or
along one Cordillera. In the northern province of Carchi (and one site in the eastern
slopes of the Andes in Central Ecuador), there are paramos with vegetation more
typical of Colombia. By contrast, Acosta-Solis (1984) describes various sites in Ecuador with what he considers to be "puna" vegetation, similar to that of Peril (for
example, the Grande Arena! of Volcan Chimborazo).
The community composition of the high altitude grasslands of the Andes has been
studied in various countries. In Venezuela, most research has been centred on the
mountains around Merida (Vareschi, 1970; Monasterio, 1979, 1980a, 1980b; Farifias
& Monasterio, 1980; Baruch, 1984; Ricardi, Briceno & Adamo, 1987). The widest extent of paramo is found in Colombia and a number of researchers have studied its
composition there (Cuatrecasas, 1934, 1958, 1968; Cleef, 1979, 1981, 1983; Sturm &
Abouchaar, 1981; Sturm & Range!, 1985, Rangel & Franco, 1985; Franco, Rangel &
Lozano, 1986; Range! & Lozano, 1986; Range! & Aguirre, 1987). The puna of Peril
has received considerable attention, beginning with the extensive works of Williams
(1941) and Weberbauer (1945) and continuing with overviews such as that by Cabrera (1968) and detailed studies of particular areas (for example, Wilcox eta!.,
1986).
By contrast, the paramos of Ecuador have received relatively little attention. Humboldt and Bonpland were the first to make a serious attempt at collecting Ecuadorian paramo plants in 1802 (Sandwith, 1926). A succession of plant collectors
followed, including Jameson and Spruce, and collections have continued throughout
this century. General descriptions of paramo vegetation have been produced by
Diels (1934), Acosta-Solis (1937, 1966, 1984, 1985), Penland (1941), Drew (1944),
Svensen (1945), Paredes (1962), Lojtnant & Molau (1982) and CerOn (1985). Until
recently, access to many paramo areas has been difficult. Lately, as new roads have
appeared, collecting has been carried out with renewed vigour to complete our
knowledge of plant distributions in the more remote paramos and to provide herbarium material for identification purposes (for example, Holm-Nielsen, 011gaard &
2. Community Ecology of the Ecuadorian Paramos
19
Molau, 1984). This activity has resulted in quite detailed plant inventories of some
paramo areas and the beginnings of a Flora of Ecuador (Harling & Sparre, 1973-).
Very few quantitative descriptions of Ecuadorian paramos have been attempted.
Balslev & de Vries (1982) described the vegetation at a single altitude on Volcan Cotopaxi by means of four 100m2 quadrats. Mutioz, Balslev & de Vries (1985) provided
the same treatment for plots on Volcan Antisana and Black (1982) presented a preliminary account of a detailed study carried out on Volcan Antisana over a number of
years, but as yet no further information is available.
Within any one paramo region in Ecuador, the vegetation can be expected to be
comprised of a number of altitudinally-related vegetation zones. Within this zonal
pattern, other vegetation types may be found which are independent of altitude,
their presence governed by such factors as soil moisture, topography and the like.
This expectation is based on the studies of paramo vegetation in neighbouring Andean countries which have already been cited (but particularly Cleef, 1979) and descriptions of mountainous regions throughout the world (for example, J.M.B. Smith,
1975, and Coe, 1967).
Mills (1975) described plant distribution over an altitudinal transect on Volcan Cotopaxi in terms of presence or absence of species, and this represents the only published study to date which looks at the community composition of an Ecuadorian
paramo in relation to altitude.
Cuatrecasas (1934, 1958, 1968) defined three altitudinal belts of paramo vegetation with reference to the vegetation of Colombia. The transition zone between the
upper Andean forest and the open grassland is dominated by shrubs and grasses, and
Cuatrecasas termed this sub-pAramo. Above the sub-pdramo, he described the grassy paramo or paramo proper, characterized by tussock grasses and giant rosette
plants. Finally, at the highest altitudes, the extreme environmental conditions restrict
plant growth and cover is reduced. This is the super-paramo.
In describing the vegetation types of Ecuador, Harling (1979) also defined three altitudinal belts. Above the montane forest scrub, the grass paramo (pajonal) is located between 3,400 and 4,000 m, dominated by tussock grasses of Calamagrostis,
Festuca and Stipa, with herbs and shrubs from the genera Ranunculus, Lupinus, Gentiana, Halenia, Castilleja, V aleriana, Baccharis, Oritrophium, Chuquiraga, Hypochaeris
and in some cases Espeletia. The shrub and cushion paramo is found between 4,000
and 4,500 m. Tussock grasses are less extensive, being replaced by shrubs, herbs of
various kinds, mats, rosette plants and cushions. Small trees may also occur. Most
characteristic genera of this zone are Chuquiraga, Diplostephium, Baccharis, V aleriana, Cakeolaria, A stragalus, Loricaria, Senecio, Culcitium, W erneria, Oritrophium, Gentiana, Halenia, V iola, Lachemilla, Draba, Bomarea, Jamesonia and Lycopodium. The
most important cushion-forming species are A zorella pedunculata, A zorella aretoides,
A zorella corymbosa, Plantago rigida, Draba aretioides, W erneria humilis and Distichia
tolimensis. At the highest altitudes (above 4,500 m), Harling classifies the sparse
vegetation of xerophytic grasses, alternating with herbs, shrubs, mosses and lichens
as desert paramo. At these altitudes, species such as Ephedra americana, Poa cucullata, Rhopalopodium guzmanii, Lupinus microphyllus, Lupinus smithianus, Not otriche
2. Community Ecology of the Ecuadorian Pijramos
20
pichinchensis, Senecio microdon, Senecio comosus, Culcitium nivale and W emeria rigida are found. Harling also describes the much lower altitude paramo on the western
and southern slopes of Volcan Chimborazo (from about 4,000 m) as desert paramo.
Here he describes a community with scattered clumps of Stipa and a few shrubs and
herbs (for example, Cakeolaria ericoides, A zorella pedunculata, Calandrina acaulis,
Chuquiraga jussieui and Hypochaeris sonchoides).
The divisions between these vegetation zones have been arbitrarily devised, based
on many years' experience of studying paramo vegetation. Cleef (1981) described in
great detail 121 community types of the Colombian Cordillera Oriental, using the
Zurich-Montpellier classification methods. However, there are only a few other
cases of paramo classification using quantitative methodology, these examining the
Venezuelan paramo (Farinas & Monasterio, 1980; Baruch, 1984).
The present study aims to give a quantitative description of the grassy paramo
vegetation of Ecuador, linking the distribution of species and paramo types to environmental variation, including those factors relating to altitude. Azonal vegetation
types, such as bogs or woodlands, were excluded from the study in order to simplify
the relationship between vegetation and altitude.
Methods
Study Sites
Twelve paramo localities were sampled in total: these areas selected to encompass a
range of paramo types from north to south and east to west (Figure 2.1). In northwestern Ecuador, two paramos were sampled. Volcan Chiles straddles the ColombiaEcuador border and maintains an extensive 'Ammo, notable as one of the few in
Ecuador with Espeletia giant rosette plants. The area is very humid and was sampled
from 4,200m (between the summit and the crest of the pass 381un from Tulcan) to
3,600m near Tufino (Figure 2.2). Volcan Cotacachi in Imbabura province supports a
moderately humid paramo. It was studied from the shoulder above the crater at
4,200m to the area just above Laguna Cuicocha at 3,600m (Figure 2.3).
Only one paramo was studied in north-eastern Ecuador. To the north of the pass
on the Quito-Baeza road is a very humid paramo on a lakeland plateau, the Paramo
de Guamani (Figure 2.4). It was sampled from 4,400m on the main jagged ridge to
3,800m in agricultural pâramo below Laguna de Hoyas.
In Central Ecuador, four paramos were sampled in or close to Parque Nacional
Sangay on the Cordillera Oriental. Volcân Tungurahua is still active, its tephra and
ash deposits in this humid area supporting a modified paramo flora. Along the route
to the summit on the northern flank of the volcano, the vegetation was investigated
from 4,300m to 3,900m (Figure 2.5). Further south, the west-facing caldera of El
Altar dominates a more usual paramo vegetation. A transect from 4,200m beneath
21
2. Community Ecology of the Ecuadorian Pâramos
COLOMBIA
Tulcan
Esmereldas
2
•
-%
Quito.....
Latacungae
•Portoviejo
•Tena
Ambr
Guaranda
Babahoya •
Riob
445
• Puyo
gt6
7
Guayaquil •
• Macas
9
Azogues
10
Machala
11
0
12
50 100 150 200
I
• Zamora .1
Loja •
I
km
II Sampling Locality
•
PERU
Major Town or City
National Boundary
Disputed National Boundary
Figure 2.1.
Map of Ecuador showing the location of the twelve study sites. Site codes: 1, Volcan Chiles; 2, Volcan Cotacachi; 3, Paramo de Guamanf; 4, Volcan Tungurahua; 5
& 6, El Altar (2 sites); 7, Paramo de Daldal; 8, Volcan Chimborazo; 9, Paramo de
Zapote Naida; 10, El Area Nacional de RecreaciOn Cajas; 11, Paramo near
Cumbe; 12, Pkamo near Ofia. Land above the 3,000 m contour line is shaded.
2. Community Ecology of the Ecuadorian P6ramos
Figure 2.2.
The location of the sampling transect on Volcan Chiles ( • ). Based on the Instituto Geografico Militar (Quito) map for Tufino. The scale is 1:100,000.
22
23
2. Community Ecology of the Ecuadorian Paramos
otacachi (4,939m)
II..
er
01111111111.111111)/////
,1
I I I
s\f
.. ._ ._
.---
\\
N.
N.
II
.
Crater
,
,
-,
_....
11
s, 11
11
,, o, s
-...,.. s‘..
.... .......
------------,-
---,
-...„
a
--....
111
N---.
N
./,
–S
II
//7 a
-_
--a--
\
N\\\
---.
5.,
—
. .. .-./
01 U1
\ \, \,U(
\\\ k
1 ,-
z'
/
///
/ f I
I /
\
-
,
111
IVI
7 // a11 k
7
/ // al \\
/7
,/
// 7
7
is
/ BE
/ Imo
/
MI \
/
BE
r
Elm \
r II \
r
n\
/
\\
\,
/
/I 1\
PcS
ZHotel
To Apuela
SCALE Km
1
2
3
4
Figure 2.3.
Sketch map showing the location of the sampling transect on Volcan
Cotacachi (R). The dotted area represents permanent snow cover. The scale is
approximately 1:100,000.
2. Community Ecology of the Ecuadorian Paramos
24
1
4,000 m
Q
74
0
lir
'°-
.
AA,
—.ANIMA
41111111
n
111111011
111•nn•n•n••
MINNIMMIII,
oNIIIIIMINNI
rIlIIIIIIVIII
4,328 m
4,200 rn
C
. 4,318 m
4,200 m
III
4,000 m
II
Laguna de Boyeros
.111
III
III
--=
0
n..
\
' >)
11131111111111
•
:.
• ., / ) \
rts
•e=z1r-
4,000
una de Hoyas
4,200 m
I
N
1
4,20o m
P
S
1
l\
SCALE (Km)
2
?
1
4,000 M
%To Papallacta
\
Figure 2.4.
The location of the sampling transect in the Paramo de Guamanf ( • ). Based on
the Institut° Geografico Militar (Quito) map for Oyacachi. The scale is 1:100,000.
25
2. Community Ecology of the Ecuadorian Paramos
To Puyo
Banos
to
Pondoa
s —
.0 ..........
...........
'"
.................
..................
..................
.................
...................
........ A . Volcin Tungura ua
...............................
• .........
.. • .
(5,016 m)
—
_r9ue 4,
acion
erk, j, z
SCALE (KM)
0:25
0:50
0:75
\
Figure 2.5.
The location of the sampling transect on Volcan Tungurahua ( • ). The dotted
area represents permanent snow cover. The scale is approximately 1:25,000.
1.00
2. Community Ecology of the Ecuadorian Paramos
Figure 2.6.
The location of the two sampling transects on El Altar ( • ). The dotted area represents (semi-)permanent snow cover, and the forest is shown to the east. Based on
the PRONAREG-ORSTOM Mapa EcolOgico for Riobamba. The scale is approximately 1:142,800.
26
2. Community Ecology of the Ecuadorian Paramos
27
2. Community Ecology of the Ecuadorian Paramos
28
4,000 m
To Ambato
4,000 m
Vo1c6ri
Carihuairaz
(4,990 m)
.... VoldAn
-Chimborazo
(6,310 m
•.
\
\
N.
s. , ...
...
%
\
%
\
1
n
s
\
n
n
SCALE (Km)
•
%
.
N
?
Figure 2.8.
The location of the sampling transect on Vo!can Chimborazo ( • ). The dotted
areas denote permanent snow cover. Based on the Institut° Geografico Militar
(Quito) map for Chimborazo. The scale is 1:100,000.
S.
2. Community Ecology of the Ecuadorian Paramos
Figure 2.9.
The location of the sampling transect in the Paramo de Cajas (II ). Based on the
Instituto Geografico Militar (Quito) map for Cuenca. The scale is 1:100,000.
29
30
2. Community Ecology of the Ecuadorian Paramos
Zapote Nelda
Transect
3,oeo m
ransect
••
i
/
i
3,000 m
I
I
I
PROVINC
DE LOJA
Figure 2.10.
The location of three sampling transects in southern Ecuador: the ['dram° de Zapote Naida, Cumbe and Oria ( • ). Based on the Instituto Geogrâfico Militar
(Quito) map of Ecuador (at 1: 1,000,000). The scale here is 1: 500,000.
2. Community Ecology of the Ecuadorian Paramos
31
Obispo to the Collanes Plain at 3,800m was the basis for the study in this area (Figure 2.6). On the eastern slopes of the El Altar massif, a much more humid 'Aram°
exists and was studied from 4,300m beneath Chizapucutul to 3,800m beside Laguna
Verde (Figure 2.6). Above the village of Daldal an area of typical agricultural paramo was investigated. The Lomo de Trenzapamba at 4,200m was the highest study
point at this location with the sub-paramo at 3,700m the lowest (Figure 2.7).
On the Cordillera Occidental in Central Ecuador stands Volcan Chimborazo, at
6,310m the highest peak in the country. The famous Grande Arenal ("Great Beach")
was sampled from 4,600m to 4,000m on the northern slopes of the volcano
(Figure 2.8). A highly modified paramo is found here on a sandy substrate in a comparatively dry region.
Finally, four paramos made up the southern section of the study sites. All four
areas were comparatively dry and low-lying. El Area Nacional de RecreaciOn Cajas is
situated to the west of Cuenca on the Cordillera Occidental. It is a lakeland plateau
averaging around 3,800m. The highest elevation studied was at 4,000m on Soldados
above Totorococha, the transect descending from there to 3,400m towards the
treeline above the Rio Mazan forest reserve (Figure 2.9). On the pass between Cuenca and LimOn on the Cordillera Oriental is the Paramo de Zapote Naida
(Figure 2.10). It was studied from 3,500m to 3,200m, though no sampling was conducted at 3,300m because of a belt of shrubby vegetation. A paramo region south of
Cumbe from 3,400m to 3,200m and a small patch of paramo at 3,100m south of Oila
(Figure 2.10) completed the study sites. These two areas were both situated on the
Cuenca to Loja road.
Table 2.1 provides a summary of the study sites. The grid references quoted are for
guidance only, since the studies were conducted along transects rather than at points.
32
2. Community Ecology of the Ecuadorian Paramos
Province
Volan Chiles
Volcan Cotacachi
PAramo de Guamanf
Volan Tungurahua
El Altar (west)
El Altar (east)
Daldal
Volan Chimborazo
PAramo de Zapote Naida
Cajas
Cumbe
Ona
Carchi
lmbabura
Pichincha
Tungurahua
Chimborazo
Chimborazo
Chimborazo
Chimborazo
Azuay/Morona Santiago
Azuay
Azuay
Loja
Altitude (m)
Min
Max
3,600
3,600
3,800
3,900
3,800
3,800
3,700
4,000
3,200
3,400
3,200
3,100
4,200
4,200
4,400
4,300
4,200
4,300
4,200
4,600
3,500
4,000
3,400
-
Latitude &
Longitude
1°49'N 77°57'W
0°35'N 78°20'W
1°15'S 78°12W
1°29'S 78°23W
1°40'S 78°24W
1°43'S 78°25'W
1°48'S 78°32W
1°30'S 77°50'W
3°00'S 78°40'W
2°53'S 79°10'W
3°20'S 79°10W
3°35'S 79°15W
Sampling Dates
20-22 Oct 1987
11-12 Oct 1987
7-8 Oct 1987
28-29 Aug 1987
2-3 Sep 1987
12-13 Aug 1989
30 Oct 1987
25 Oct 1987
21 Sep 1987
12-14 Sep 1987
21 Sep 1987
16 Sep 1987
Table 2.1.
Study sites used in the phytosociological study of the Ecuadorian paramos.
Plant Collection
Since a comprehensive flora of the Ecuadorian päramos has yet to be produced, it
was necessary to collect voucher specimens of plant taxa and cross-reference them
with plants cited in this study. Therefore, in cases where species have not been fullynamed, they are identified by a code number shown in square braces, which permits
cross-reference of the species to voucher specimens via Appendix 1. Thus, for Cerastium sp. [197], the code number 197 in Appendix 1 shows that this species is represented by two voucher specimens numbered 454 and 536.
A collecting licence was obtained from the Ministerio de Agricultura y Ganaderfa,
with the support of the Pontificia Universidad CatOlica del Ecuador (PUCE) and the
Museo Ecuatoriano de Ciencias Naturales (MECN).
During the course of the fieldwork, representative vascular plant material was collected. Ideally, fertile specimens were taken, but occasionally only sterile material
was available. The material was placed in polythene bags until it could be pressed.
The pressing usually took place within one day, though in difficult circumstances up
to three days elapsed before some specimens were pressed.
Plant material was sandwiched between sheets of newsprint or, when available,
proprietary 'flimsies', with a drying unit (consisting of a corrugated aluminium sheet
between two blotting papers) separating specimens. Repeating this pattern, the press
was filled and bound using straps.
Whenever possible, the presses were transported to Quito for drying in the drying
room of the PUCE herbarium. The presses were placed over low-power electric heaters for 3-7 days. On many occasions, however, drying had to be carried out away
from Quito. For this purpose, a wooden frame was constructed to support the press
at the right height and a paraffin pressure stove used as a heat source. This proved
very effective, though the stove was prone to flare up from time to time, requiring
constant vigilance.
2. Community Ecology of the Ecuadorian Paramos
33
When possible, four duplicates were collected for each specimen, and numbered
on collection. Duplicate specimens are housed in the collections at the Royal Botanic Gardens, Kew (K) or the Royal Botanic Garden, Edinburgh (E), the herbarium
of the Pontificia Universidad CatOlica del Ecuador, Quito (QCA) and the Ecuadorian national collection at the Museo Ecuatoriano de Ciencias Naturales, Quito
(QCNE). The remaining duplicate set was deposited with the Ministerio de Agricultura y Ganaderia in Quito.
Preliminary identifications were carried out at the QCA herbarium, but most of
the taxonomic work took place at the herbarium of the Royal Botanic Gardens, Kew.
Many specimens have only been identified to generic level, though it has been
possible in nearly all cases to assign these to more precise taxa without actually naming them (for example, Calamagrostis A, Calamagrostis B, etc.). Where specimens
could not be differentiated with confidence, species aggregates have been formed.
For example, there is a possibility that Bromus pitensis was confused with Bromus lanatus in the field, and therefore this species has been amalgamated into Bromus lanatus aggregate. Similarly, Lachemilla andina has been added to Lachemilla rupestris
aggregate in the analysis.
A complete moss and macro-lichen collection was made for each quadrat and later
used to produce comparative species lists for each paramo area. A portion of the
moss collection was stolen in Ecuador, rendering full comparisons impossible. The
remainder of the mosses and macro-lichens are housed in the British Museum of
Natural History (BMNH), the Royal Botanic Gardens, Edinburgh (E), QCA and
QCNE, but have yet to be determined. The following analyses were, therefore, only
performed on the vascular plant composition of the quadrats.
Sampling Procedure
Only zonal paramo was studied. Azonal vegetation, such as that found in bogs, was
deliberately excluded from the sampling. The vegetation was investigated by means
of 5m x 5m quadrats, randomly located at each 100m of elevation over the range studied (which depended upon the paramo coverage at each site). At each altitude, a
100m transect was established along the contours and the three quadrats located according to random co-ordinates previously generated. This was not always possible,
however, particularly in the super-paramo where the vegetation was sometimes confined to smaller patches (by rocky outcrops, unfavourable conditions or merely by the
small size of the peak). In such cases the horizontal transect was shortened and the
random co-ordinates scaled accordingly.
For each 25m2 sampling unit, a complete list of the vascular plants present was
compiled, along with the corresponding Braun-Blanquet abundance scores (1, <5%;
2, 6-25%; 3, 26-50%; 4, 51-75%; 5, >75% — `r' and ' + ' were not used owing to the
small size of the sample area). The species present were cross-referenced with
voucher specimens in the plant collection.
2. Community Ecology of the Ecuadorian Pâramos
34
Estimates of the coverage (as a percentage) of bare ground and rock cover, including scree, were made for each sample unit. Plant litter coverage was not estimated
because most of the tussock grass material decays while still attached to the plant,
making accurate judgement difficult. The abundance scores for tussocks of this sort
included such standing dead material.
A number of environmental variables were recorded for each quadrat. Altitude
was measured as the mean of two Thommen 6,000m aneroid altimeters set at Quito
observatory (2,818m). It was not possible to account for meteorological changes in atmospheric pressure. Aspect was assigned to 8 compass points (N, NE, E, SE, S, SW, W
and NW) using a prismatic compass and slopes were measured in degrees from the
horizontal using a clinometer. A soil sample was taken from each quadrat (10cm
depth), air-dried and the pH measured with a Whatman pH meter (2:1 ratio of water
to soil by volume). However, some soil samples were stolen along with the mosses,
and therefore pH data is unavailable for the southern sites.
Finally, exposure, burning intensity, grazing intensity, trampling intensity and overall disturbance were estimated for each quadrat on subjective, semi-quantitative
scales from 0 to 5 (where 0 represents the absence of the influence and 5 the highest
influence).
Exposure was judged using the local topography: a sample surrounded by ridges
on all sides was considered to be of low exposure whereas a sample plot on a ridge received a high exposure score. Burning intensity scores were determined from visible
indicators of fire. These included ash deposits, charred remains and the loss of soil
caused by the combustion of its organic material. Indicators of grazing included
visible signs (cropping of vegetation, etc.) and indirect signs (droppings). Trampling
scores were based on the presence of micro-terracing, paths, poaching (hoof-prints),
broken branches of shrubs and cattle-scrapes.
Overall disturbance was recorded in an attempt to deal with the interactive effects
of burning, trampling and grazing, where the combined impact of these variables can
result in higher disturbance than the individual effects might suggest. Thus, a disturbance score was assessed with reference to compounded disturbance, irrespective of
its source.
The majority of the data was collected from August to October 1987, though one
of the sites was visited in August 1989 (Table 2.1).
Data-Handling and Analysis
In order to simplify the dataset, the vegetation samples were classified according
to a polythetic divisive cluster analysis technique. This was achieved using the TWINSPAN algorithm (Hill, 1979), forming part of the VESPAN-II package (Malloch, 1988).
The plant taxa were also classified using this programme, based on their presence in
the stand classes.
2. Community Ecology of the Ecuadorian Paramos
35
Multivariate direct gradient analysis (canonical ordination) was employed to determine the relationship between species distributions and the measured environmental
variables. This combination of regression and ordination was carried out using the
CANOCO programme (ter Braak, 1988). The analysis was performed on all taxa and
all stands, then the TWINSPAN classes were superimposed (as centroids) upon the resulting ordination. The first axis of this ordination and the trace statistic (the sum of
all axes) were tested for statistical significance by means of a Monte-Carlo permutation test (Hope, 1968) with 99 permutations, also part of the CANOCO package.
a-diversity was estimated for each stand using an adapted version of the Simpson
index:
D = x p(p-1)
A(A-1)
where p represents the percentage cover of each species (using the mid-point of its
Braun-Blanquet score), and A is the total coverage of all the species in the quadrat
(that is, II)). This only represents an approximation of the diversity of the stands,
since Braun-Blanquet scores are not 'linear'. Diversity was expressed as the reciprocal of D.
fl-diversity, measuring "the extent of species change along environmental gradients" (Whittaker, 1975), was calculated according to the measure proposed by Wilson & Smida (1984):
PT
= [g(H) + l(H)] .
2a
where g(H) represents the number of species gained along the gradient and l(H) the
number of species lost, a is the average number of species found within the samples.
fl-diversity is essentially the same as MacArthur's (1965) between habitat diversity.
To calculate fl-diversity values, the three replicate quadrats at each altitude were combined (three 25 m2 quadrats becoming one sample of 75 m2).
2. Community Ecology of the Ecuadorian Paramos
36
Results
Twelve paramo regions were studied. The number of quadrats and summary statistics on the number of vascular taxa ("species") recognised for each of these regions
is given in Table 2.2. The number of altitude levels sampled in each study area
ranged from one (3,100 m at Oha) up to seven for several of the sites. In total, 64 altitudes were sampled using 192 quadrats. The 21 stands sampled in the paramo at
Cajas included more species (117) than any of the other study areas, whilst the three
quadrats at Ona yielded just 24 species. On average, a 25 m2 sample of the vegetation contained 21.11 species, though this varied from site to site (only 6.90 species
per quadrat on Volcan Chimborazo to 29.22 species in the paramo at Daldal). The
number of species recorded in the three replicate quadrats at each altitude averaged
29.41. On Volcan Chimborazo, this mean was 10.71 whilst in the paramo around the
crater of El Altar it reached 42.60 species.
No of
Altitude
Levels
Site
VolcAn Chiles
VolcAn Cotacachi
Mum de Guamanf
Volc.in Tungurahua
El Altar (west)
El Altar (east)
Da!dal
VolcAn Chimborazo
Memo de Zapote Naida
Cajas
Cumbe
Ona
Overall
No of
Quadrats
No of
Species in
all Quadrats
Mean No of
Species per
Quadrat
Mean No of
Species per
Altitude
(3 Quadrats)
7
7
7
5
5
6
6
7
3
7
3
1
21
21
21
15
15
18
18
21
9
21
9
3
94
89
97
52
92
71
91
37
47
117
7t
24
22.10
20.10
23.52
17.67
27.54
19.44
29.22
6.90
24.56
22.86
24.78
16.33
31.86
27.43
35.57
25.40
42.60
25.83
36.57
10.71
33.67
36.29
38.33
24.00
64
192
348
21.11
29.41
Table 2.2.
The location and summarised vascular plant composition of 192 paramo quadrats. For each locality the number of altitude levels sampled and the number of
quadrats used are stated. The total number of vascular plant species found in the
stands at each locality are given. The mean number of species found in each
stand and at each altitude level (three quadrats combined) are shown.
Altogether, 348 taxa of vascular plants have been recognised in these sample
stands (Appendix 1). Table 2.3 indicates the composition of the 192 stands in terms
of family, and Table 2.4 shows the genera comprising these families. Almost 20% of
the species found in the quadrats belong to the family Compositae. 25 genera of composites were present in the vegetation samples, the best-represented being Senecio,
Baccharis, Culcitium, Gnaphalium, Diplostephium and Gynoxys.
2. Community Ecology of the Ecuadorian Paramos
37
Number Percentage
of Taxa of all Taxa
in Family
Family
Unidentified to Family
6
1.7
Compositae
Gramineae
Cyperaceae
Lycopodiaceae
Scrophulariaceae
Gentianaceae
Leguminosae
Umbelliferae
Valerianaceae
Ericaceae
Rosaceae
Rubiaceae
Caryophyllaceae
Cruciferae
Geraniaceae
Plantaginaceae
Violaceae
Bromeliaceae
Guttiferae
Hemionitidaceae
Juncaceae
Melastomataceae
Ranunculaceae
lridaceae
Orchidaceae
Lomariopsidaceae
Alstroemeriaceae
Labiatae
Oxalidaceae
Polygalaceae
Alliaceae
Aspleniaceae
Blechnaceae
Campanulaceae
Equisetaceae
Filicopsida
Grossulariaceae
lsoetaceae
Lentibulariaceae
Malvaceae
Melanthiaceae
Onagraceae
Ophioglossaceae
Polygonaceae
Thelypteridaceae
Xyridaceae
69
47
19
15
15
14
12
11
11
10
10
10
8
8
6
6
6
5
5
5
5
5
5
4
4
3
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
19.8
13.5
5.5
4.3
4.3
4.0
3.4
3.2
3.2
2.9
2.9
2.9
2.3
2.3
1.7
1.7
1.7
1.4
1.4
1.4
1.4
1.4
1.4
1.1
1.1
0.9
0.6
0.6
0.6
0.6
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
Total
348
100.0
Table 2.3.
A summary of the 348 recognised taxa in 192 Oramo quadrats. The number of
taxa ("species") in each family is given, along with the percentage of all taxa this
represents. There are 46 families. Six voucher specimens could not be identified
to family level.
The Gramineae accounted for 13.5% of the taxa with 15 genera determined. A grostis, Poa, Calamagrostis and Festuca were the most important genera. The Cyperaceae
2. Community Ecology of the Ecuadorian Paramos
No Family (6 unidentified taxa)
•
Unidentified (6 taxa)
Alliaceae (1 unidentified genus)
•
Unidentified (1 taxon)
Alstroemeriaceae (1 genus)
•
Bomarea (2 taxa)
Aspleniaceae (1 genus)
•
Asplenium (1 taxon)
Blechnaceae (1 genus)
•
Blechnum (1 taxon)
Bromeliaceae (1 genus)
•
Puya (5 taxa)
Campanulaceae (1 genus)
•
Lobelia (1 taxon)
Caryophyllaceae (2 genera)
•
•
Cerastium (7 taxa)
Stellaria (1 taxon)
Compositae (25 genera
+ 2 unidentified taxa)
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Unidentified (2 taxa)
Senecio (8 taxa)
Baccharis (7 taxa)
Culcitium (6 taxa)
Gnaphalium (6 taxa)
Diplostephium (5 taxa)
Gynoxys (5 taxa)
Oritrophium (4 taxa)
Wemeria (4 taxa)
Loricaria (3 taxa)
Erigeron (2 taxa)
Hypochaeris (2 taxa)
Lucilia (2 taxa)
Aphanactis (1 taxon)
Bidens (1 taxon)
Chrysactinium (1 taxon)
Chuquiraga (1 taxon)
Conyza (1 taxon)
Cotula (1 taxon)
Espeletia (1 taxon)
Hieracium (1 taxon)
Perezia (1 taxon)
Sonchus (1 taxon)
Stevia (1 taxon)
Taraxacum (1 taxon)
Vemonia (1 taxon)
Cruciferae (4 genera
+ 2 unidentified genera)
•
•
•
•
•
Unidentified (2 taxa)
Draba (2 taxa)
Eudema (2 taxa)
Cardamine (1 taxon)
Lepidium (1 taxon)
Cyperaceae (5 genera)
•
•
•
•
•
Carex (9 taxa)
Uncinia (4 taxa)
Oreobolus (3 taxa)
Rhynchospora (2 taxa)
Eleocharis (1 taxon)
Equisetaceae (1 genus)
•
Equisetum (1 taxon)
Ericaceae (3 genera
+ 7 unidentified genera)
•
•
•
•
Unidentified (7 taxa)
Disterigma (1 taxon)
Pemettya (1 taxon)
Vaccinium (1 taxon)
Filicopsida (1 genus)
•
Eriosorus (1 taxon)
Gentianaceae (3 genera)
•
•
•
Gentianella (9 taxa)
Halenia (4 taxa)
Gentiana (1 taxon)
Geraniaceae (1 genus)
•
Geranium (6 taxa)
Gramineae (15 genera
+ 2 unidentified taxa)
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Unidentified (2 taxa)
Agrostis (12 taxa)
Poa (10 taxa)
Calamagrostis (5 taxa)
Festuca (4 taxa)
Muhlenbergia (2 taxa)
Paspalum (2 taxa)
St/pa (2 taxa)
Aciachne (1 taxon)
Anthoxanthum (1 taxon)
Bromus (1 taxon)
Cortaderia (1 taxon)
Elymus (1 taxon)
Holcus (1 taxon)
Neurolepis (1 taxon)
Trisetum (1 taxon)
Table 2.4.
Genera in the 192 paramo quadrats. For each family the number of genera (plus
any unidentified specimens, which may or may not represent further genera) are
shown. Within each family, the genera are listed along with the number of taxa assigned to them. In total there are 117 genera in 46 families, with 27 taxa as yet unidentified. [Continued Overleaf]
38
2. Community Ecology of the Ecuadorian Paramos
Orchidaceae (2 genera
+ 2 unidentified genera)
Grossulariaceae (1 genus)
•
Ribes (1 taxon)
Guttiferae (1 genus)
•
•
•
•
Hypericum (5 taxa)
Hemionitidaceae (1 genus)
•
Oxalidaceae (1 genus)
Jamesonia (5 taxa)
•
Iridaceae (2 genera)
•
•
•
Isoetes (1 taxon)
•
•
•
•
•
•
•
•
•
Lupinus (7 taxa)
Vicia (3 taxa)
Astragalus (1 taxon)
Trifolium (1 taxon)
Pin guicula (1 taxon)
Elaphaglossum (3 taxa)
Lycopodium (12 taxa)
Huperzia (3 taxa)
•
Nototriche (1 taxon)
Tofieldia (1 taxon)
•
•
•
•
•
•
Melastomataceae (1 genus
+ 2 unidentified genera)
•
•
Unidentified (2 taxa)
Bra chyotum (3 taxa)
Onagraceae (1 genus)
•
Epilobium (1 taxon)
Ophioglossaceae (1 genus)
•
Ophioglossum (1 taxon)
Bartsia (4 taxa)
Castilleja (4 taxa)
Veronica (3 taxa)
Calceolaria (1 taxon)
Ourisia (1 taxon)
Pedicularis (1 taxon)
Sibthorpia (1 taxon)
Thelypteris (1 taxon)
Umbelliferae (5 genera
+ 2 unidentified genus)
Melanthiaceae (1 genus)
•
•
•
•
•
•
•
•
Thelypteridaceae (1 genus)
Malvaceae (1 genus)
•
Arcytophyllum (4 taxa)
Relbunium (4 taxa)
Galium (1 taxon)
Nertera (1 taxon)
Scrophulariaceae (7 genera)
Lycopodiaceae (2 genera)
•
•
Lachemilla (10 taxa)
Rubiaceae (4 genera)
Lomariopsidaceae (1 genus)
•
Ranunculus (4 taxa)
Anemone (1 taxon)
Rosaceae (1 genus)
Satureja (1 taxon)
Stachys (1 taxon)
Lentibulariaceae (1 genus)
•
Rumex (1 taxon)
Ranunculaceae (2 genera)
Unidentified (1 taxon)
Luzula (3 taxa)
Distichia (1 taxon)
Leguminosae (4 genera)
•
•
•
•
Monnina (2 taxa)
Polygonaceae (1 genus)
Labiatae (2 genera)
•
•
Plantago (6 taxa)
Polygalaceae (1 genus)
Juncaceae (2 genera
+ 1 unidentified genus)
•
•
•
Oxalis (2 taxa)
Plantaginaceae (1 genus)
Sisyrinchium (3 taxa)
Orthosanthus (1 taxon)
Isoetaceae (1 genus)
•
Unidentified (2 taxa)
Altensteinia (1 taxon)
Myrosmodes (1 taxon)
Unidentified (2 taxa)
Azorella (5 taxa)
Eryngium (1 taxon)
Hydocotyle (1 taxon)
Niphogeton (1 taxon)
Oreomyrrhis (1 taxon)
Valerianaceae (1 genus)
•
Valeriana (11 taxa)
Violaceae (1 genus)
•
Viola (6 taxa)
Xyridaceae (1 genus)
•
Xyris (1 taxon)
Table 2.4. (Continued)
Genera in the 192 paramo quadrats.
39
2. Community Ecology of the Ecuadorian Paramos
40
provided 5.2% of the taxa found in the quadrats, with the Lycopodiaceae, Scrophulariaceae and Gentianaceae just below this figure.
Table 2.5 lists the thirty most frequent species occurring in the samples. By far the
commonest species is Calamagrostis sp. [251], which is present in 94.27% of the quadrats, with a mean Braun-Blanquet score between 2 and 3 (5-50% cover). Other common grasses were Paspalum tuberosum (present in 40% of the sample stands), Poa
sp. [262] (26%), A grostis nigritella (20%) and Bromus lanatus (18%). Pemettya prostrata and Disterigma empetnfolium were growing in 56% and 46% of the quadrats respectively.
Species number and name
251
185
79
91
40
97
143
48
72
146
80
64
134
161
88
262
150
130
99
106
29
15
39
139
124
142
244
103
94
153
Calamagrostis sp.
Pemettya sp.
Disterigma empetrifolium
Geranium sibbaldioides
Hypochaeris sessiliflora
Paspalum tuberosum
Etyngium humile
Oritrophium peruvianum
Carex tristicha
Oreomyrrhis andicola
Gentiana sedifolia
Wemeria humilis
Bartsia laticrenata
Hypericum sp.
Halenia weddelliana
Poa sp.
Valeriana bonplandiana
Lachemilla orbiculata
Sisyrinchium jamesonii
Lupinus sarmentosus
Gnaphalium pensylvanicum
Bidens andicola
Hieracium frigidum
Azorella aretoides
Lachemilla rupestris
Azorella pedunculata
Agrostis nigritella
Satureja nubigena
Bromus lanatus
Valeriana microphylla
Family
Gramineae
Ericaceae
Ericaceae
Geraniaceae
Compositae
Gramineae
Umbelliferae
Compositae
Cyperaceae
Umbelliferae
Gentianaceae
Compositae
Scrophulariaceae
Guttiferae
Gentianaceae
Gramineae
Valerianaceae
Rosaceae
lridaceae
Leg uminosae
Compositae
Compositae
Compositae
Umbelliferae
Rosaceae
Umbelliferae
Gramineae
Labiatae
Gramineae
Valerianaceae
Mean Const
B-B Score %
2.7
0.6
0.5
0.4
0.4
0.5
0.4
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.2
0.3
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
94.27
56.25
46.35
44.79
42.19
39.58
38.54
33.85
29.17
28.13
27.60
27.08
26.56
26.56
26.04
26.04
24.48
23.96
22.92
22.92
22.40
21.88
21.86
20.31
19.79
19.79
19.79
19.27
18.23
18.23
N
181
108
89
86
81
76
74
65
56
54
53
52
51
51
50
50
47
46
44
44
43
42
42
39
38
38
38
37
35
35
Table 2.5.
The 30 most frequent species found in the 192 paramo quadrats. The species
number, name and family are shown, along with the mean Braun-Blanquet value,
the frequency (N, out of 192) and the percentage constancy in the sample stands.
The Compositae made an important contribution to the list of most frequent
species: Hypochaeris sessiliflora, Oritrophium peruvianum and W erneria humilis were
all present in over 25% of the sample stands. Geranium sibbaldioides and Dyngium
humile were each found in over 30% of the vegetation samples.
41
2. Community Ecology of the Ecuadorian Paramos
Slope Expos Bare Rock Burn Trampl Graz Disturb pH Cond Cover Diversity
Altitude 2.2+
(3920 m)*
Slope
(24°)
22.4+
***
18.4+
***
4.0+
**
3.2-
Exposure
(Index score 3.2)
0.9+
NS
Bare Ground
(11% cover)
Rock cover
(2% cover)
Burning
(Index score 2.0)
Trampling
(Index score 2.0)
29.6***
20.3***
35.6***
38.2***
11.1+
***
0.6NS
***
8.0+
***
0.0
NS
0.4+
NS
0.0
NS
0.0
NS
0.0
NS
5.8+
**
1.9+
NS
0.2
NS
+1.3+
NS
10.9+
***
23.5***
22.4***
19.8***
25.3***
0.0
NS
0.7NS
11.6***
9.4+
***
0.1NS
6.7***
11.1***
23.9***
***
18.7+
***
1 .7NS
38.3***
0.2NS
5 .4***
4 .4**
5 .6***
3.8+
**
0.4+
NS
7.3***
1.3+
NS
53. 5+
***
70.2+
***
81.6+
***
12.2if**
0.2+
NS
30.8+
***
20.0***
66.2+
***
73 .9 +
***
15.7***
0.1+
NS
32.5+
***
14.4***
83.8+
***
20.1***
0.7+
NS
50.4+
***
19.7***
12.4-
0.3+
NS
40.0+
***
20.6***
5.3+
**
15.9***
0.7+
NS
3.6+
**
3 .8**
Grazing
(Index score 1.8)
Disturbance
(Index score 2.3)
12.5-
***
pH (132 samples only)
(5.2 units)
Conductivity (132 samples only)
(1.2 mS s-1)
24.2-
1.8+
Vegetation Cover
(97% cover)
0.3+
NS
11.4***
Diversity
(8.4)
Table 2.6.
Correlations between environmental variables, vegetation cover and diversity.
Correlations are based on 192 observations for each of the variables (with the exception of soil pH and conductivity with 132 observations). The overall mean for
each of the environmental variables is shown in the left-hand column. The r2
values are given on the first line and the '+' or '-' indicates a positive and negative
correlation respectively. The second line shows if the correlation is significant (*,
p5.0.05; **, p 0.01; ***, p 0.001) or not significant (NS).
The eleven environmental variables which were measured for each quadrat were
subjected to a correlation test (Table 2.6), the distributions of six of these variables
are shown in Figure 2.11. Vegetation cover and species diversity values, derived from
species abundance data for each stand, were also correlated with the environmental
variables in Table 2.6. It is immediately apparent that these environmental variables
are interdependent. All variables except soil conductivity show a significant relationship with altitude. Exposure is greater at higher altitudes (r2 = 22.4%,p < 0.001) and
the cover of bare rock increases (r2 =3.6%,p = 0.008). The amount of bare ground
also rises with altitude (r2 = 18.4%,p < 0.001) and slopes become steeper (r2 =2.2%,
p= 0.038). The soils are less acidic at higher altitudes than those lower down
(r2= 11.1%,p <0.001).
NS
2. Community Ecology of the Ecuadorian P gramos
42
The semi-subjective assessment of overall disturbance was a very good predictor
of burning (r2 = 81.6%,p < 0.001), trampling (r2 = 73.9%,p < 0.001) and grazing
(r2 = 83.8%,p < 0.001). All of these variables were highly correlated with each other.
Clearly, trampling and grazing are linked (r2 = 66.2%,p <0.001) and these forms of
disturbance are concentrated in areas which have been burned (r2 = 53.5%,
p < 0.001; r2 = 70.2%,p < 0.001, respectively). It is reasonable, therefore, to use overall disturbance alone as an indicator of burning, trampling and grazing pressures.
These disturbances show an inverse relationship with altitude (r2 =38.2%,
p < 0.001): disturbance decreases up the elevational gradient.
Vegetation cover, calculated as the sum of the Braun-Blanquet mid-point cover
values for each quadrat, decreases at higher altitudes (r2 = 24.2%,p < 0.001). Species
diversity ("evenness") tends to increase with altitude (r2 = 8.0%,p < 0.001), as dominance by a few species diminishes. Exposure decreases vegetation cover
(r2 = 11.6%,p < 0.001) but increases species diversity (r2 = 11.6,p < 0.001). Vegetation cover is greater on more acidic soils (r2 = 15.9%,p < 0.001), which are more
common on flatter ground (r2 = 5.8%,p = 0.005).
The 192 vegetation samples were classified using TWINSPAN. The dendrogram
showing the divisions leading to the stand classification is presented in Figure 2.12. A
group of three samples was found to be sufficiently different from the other plots to
split away from them at the first division. At the next division, another small group
(containing 15 stands) was separated from the main group of samples. Then, by a
number of successive divisions, the large group of 174 samples was divided into five
groupings of 10, 24, 79, 46 and 15 plots. Ultimately, the classification resulted in 31
stand groups. These end groups may be interpreted as types of paramo vegetation. It
should be remembered that this analysis does not rely on presence or absence of
species alone for the divisions. Instead, it is the combination of species present that
determines the end groupings. Thus, one species may be diagnostic of several
groups, but in each group its association with other characteristic species is unique.
A description for each of these paramo types is provided later.
Figure 2.13 shows the CANOCO ordination biplot of the 348 species plotted
against the first two constrained axes. In the figure, the species are displayed as
points in a two-dimensional subspace (there are many more dimensions that could
be displayed if it were possible). These points represent their approximate optima in
this space, that is, the points where they are most abundant. The environmental variables are shown as arrows: the length of an arrow indicates the magnitude of the
correlation between the variable and the distribution of species, whilst the direction
of the arrow shows the plane in which the variable increases.
This ordination was subjected to a Monte Carlo permutation test to determine the
statistical significance of the relationship between the species distribution and the ordination axes (and, therefore, the environmental variables). The species distribution
in the ordination space was found to be related to the environmental variables with a
high degree of probability (p = 0.01).
Arrows have been drawn on the ordination to show the nature of the relationship
between species distribution and vegetation cover ("Area") and species diversity.
43
2. Community Ecology of the Ecuadorian Paramos
Exposure
100
Burning
Frequency (of 1921
50
80
WO
60
I
,g:',-
40
20-
o
,
1
o
Frequency (of 192)
40
30
20
aW 0
10
All•-
o
23
0
3
Index Score
6
Index Score
Trampling
Grazing
so
Frequency (of 192)
so -
40
20
0
2
3
4
6
0
11
o a 0 El 0
2
1
Index Score
3
Disturbance
60
4
6
Index Score
pH
Frequenc-y (of 192)
20
Fnoquency (of 132)
60
15
40
30
10
20 6
100
1
2
3
4
01
4.3
4.6
I1I1I1I1I1I
4.7
4.9
6.1
6.3
pH value
Index Score
Figure 2.11.
Distribution of six environmental variables in the paramo vegetation samples.
Exposure, burning, trampling, grazing and disturbance were measured on a sixpoint scale (where 0 represents the absence of the influence and 5 the highest
influence) for 192 samples. pH values were measured for 132 samples.
6.5
5.7
44
2. Community Ecology of the Ecuadorian Paramos
0
0
to
•r.
0
•n•••
C—
CC
co
0
ci
•cr
Co
co
0
CO
0
0
r-
cn
N..
3.`
r")
co
>
0
V
Ci
E
Co
C.)
,
..
(15
CL
C
C.3
MI
IIIIIIII••
Co
o
Co
(0
11"..
C.)
C.)
0
C1-)
0
1-n
cni
CD
0
45
2. Community Ecology of the Ecuadorian Paramos
Ax s 2
6
5
4
•
DI turban e.
n
3
•
•
l'• 2 •
•
••
2
•
••
•
dro
••
\e'l ti
A ea
•
2
•
.
•
• •
•
•.• • •
•••,.. • •
IA'
.4•••••••
I.•
n •' •'
••n•• •••
2
3
4
5
6
7
8
9
10
As
A tituce
Figure 2.13.
biplot ordination of the 348 vascular species present in 192 pdramo
stands. Environmental variables are depicted as solid arrows. Passive variables,
which were not used to calculate the ordination, are superimposed as dashed arrows. The axes are divided into standard deviation units (Ai =0.508, 22 = 0.449).
CANOCO
These arrows are passive, in that they were not used to determine the ordination (unlike the environmental variables), but have been superimposed onto it after its creation. For this reason they have been drawn with dashed rather than solid lines.
Most of the species in Figure 2.13 lie within a belt running diagonally from the
upper left portion of the ordination (with negative values for Axis 1 and positive
values for Axis 2) to the lower right part (positive Axis 1 values and negative Axis 2
values). This distribution corresponds well with the environmental variables of altitude, exposure and disturbance, as depicted by their arrows. Those species in the
lower right portion of the ordination are most abundant at higher altitudes, where
disturbance is minimal but exposure is greater. Species diversity was highest in this
region of the biplot. Species located in this part of the ordination include Oritrophium peruvianum, Culcitium ovatum, W erneria humilis, W erneria pumila, Gentianella foliosa, Lupinus purdianus, Lachemilla holosericea, Cerastium sp. [198], A ciachne
flagellifera, A grostis nigritella, A grostis sp. [239], Huperzia hypogoea, Lycopodium sp.
[288], Lycopodium sp. [289], Luzula racemosa, Plantago rigida, V aleriana aretioides,
Bartsia laticrenata and A zorella aretoides.
2.
46
Community Ecology of the Ecuadorian Paramos
Axis 2
Disturbance
Bare
Area
•PCCT HP T
0
VC
2
•
CCD
0 CT
2
1.
ACT•
P Es
5CT04V E.
•CCCD
3
5
•
}-CT
GCCD
ACCD
6
7
CCD
pe
PIDCF:
911,0
l3
A ASCC0tp F
WC
A? A
C)
<<\
to to
<A, c
A titude
Figure 2.14.
biplot ordination of the 31 paramo plant communities. Environmental variables are depicted as solid arrows. Passive variables, which were not used to calculate the ordination, are superimposed as dashed arrows. The axes are divided
into standard deviation units (Ai =0.508, A2=0.449). The centroid of each class is
colour-coded to match that of Figures 2.12 and 2.16.
CANOCO
Species which are more abundant at lower altitudes, where disturbance is greater
but exposure is reduced, are located in the upper left part of the ordination. Species
which are located in this portion of the ordination tend to exist in vegetation of lower
diversity and include Paspalum tuberosum, Poa sp. [266], Poa sp. [267], Stipa sp.
[270], Cluysactinium acaule, Gynoxys bumfolia, Oritrophium peruvianum forma intermedium, Gentianella gracilis, Gentianella hyssopifolia, Orthrosanthus chimboracensis,
Sisyrinchium tinctorium, Pinguicula calyptrata, Tofieldia sessiliflora, Brachyotum ledzfolium, V iola humboldtii, Puya sp. [180], Puya sp. [181], Halenia sp. [187], Cerastium sp.
[197], Jamesonia robusta, Lycopodium clavatum and Oxalis sp. [359].
A number of species lie clustered to the upper right of the diagram. The position
of these species correlates well with the amount of bare ground present, and are
little influenced by the other environmental variables which were measured. These
species were all found on Volcan Chimborazo, and include Chuquiraga jussieui, Erigeron pinnatus, Lucilia radiata, Senecio teretifolius, W enzeria crassum, Gentianella cernua, Lupinus smithianus, V aleriana alypifolia ssp. alypifolia, V aleriana sp. [194],
Geranium sp. [157], Cerastiunz sp. [200], Plantago sp. [301], Stipa sp. [353] and A grostis sp. [348].
SD
Ax's 1
47
2. Community Ecology of the Ecuadorian Pâramos
Of the thirty commonest species in the sample quadrats (shown in Table 2.5), almost all (24) are located around the origin of the ordination. Reference has already
been made to the exceptions.
The species ordination was used to calculate the centroids of the TWINSPAN
classes and to infer their correlation with the environmental variables, vegetation
cover and species diversity. The resulting biplot, showing the ordinated vegetation
classes and the environmental variables, is displayed in Figure 2.14. The mean values
of the environmental variables for each vegetation class are shown in Table 2.7.
Class
Exposure Burning Trampling Grazing Disturb'ce
BS (n =3)
HD (n = 12)
CCCD (n = 1)
CCT (n =3)
DCS (n =5)
FCT (n = 1)
AC (n =3)
AAC (n =3)
WAC (n = 3)
WPC (n = 3)
DTSC (n =3)
DSC (n = 3)
DTC (n =3)
PC (n - 3)
HHCT (n = 6)
HPCT (n = 12)
NB (n =10)
OCT (n = 20)
VCT (n = 10)
SCI (n21)
RCT (n = 4)
PCT (n =4)
ACT (n = 2)
PCCT (n = 24)
PCE (n =6)
VCE (n =6)
GCCD (n - 7)
ACCD (n = 3)
CCD (n =1)
NCCD (n =4)
SD (n =3)
Overall
4.00
4.25
5.00
3.00
4.00
4.00
5.00
4.00
5.00
4.00
3.00
4.00
200
3.00
3.17
2.75
3.40
2.20
2.80
3.14
2.00
3.75
3.00
3.04
3.00
3.50
2.71
3 67
5.00
4.00
2.00
3.19
1.00
2.00
1.00
1.00
0.75
2.00
2.00
1.00
1.00
2.67
4.00
1.80
3.10
3.40
2.52
2.50
2.25
3.00
2.42
4.00
3.50
0.86
1.00
3.00
2.50
3.75
1.90
2.95
3 00
2.90
1.75
2.25
3.00
2.00
4.00
3.00
086
1.00
1.75
1.00
2.00
0.50
1.00
2.05
1.00
1.00
1.00
3.00
2.00
3.50
1.80
2.75
3.00
2.67
1.50
2.25
3.00
2.38
3.00
3.50
1.84
2.00
2.17
4.50
2.40
3.45
3.80
3.14
2.50
2.50
4.00
2.95
4.00
4.00
0.86
1.00
1.00
1.00
2.35
pH
Slope
5.60
5.48
5.70
4.40
4.94
5.00
5.50
5.10
5.10
5.10
28
32
19
15
25
22
20
20
41
5
37
35
15
18
36
34
26
19
22
29
29
25
5
22
17
15
14
20
19
19
5
24
5.30
5.35
5.15
4.90*
5.40*
5.23*
5.05
5.50*
-
4.50
4.55
5.50
5.50
5.70
5.33
5.60
5.19
Bare
Rock
10.2
90.0
0.8
0.8
11.0
1.0
81.7
1.4
23.3
3.0
4.0
2.7
14.0
2.0
9.2
9.9
4.0
2.0
0.4
1.3
18.8
0.8
3.2
3.0
1.0
62.4
68.3
80.0
82.5
80.0
11.1
0.3
5.0
0.5
0.3
3.3
1.8
Table 2.7.
Mean environmental variable values for the 31 pâramo plant communities identified by TWINSPAN. pH values marked with * are derived from fewer values than the
other variables for that class. A blank represents zero, and '-' indicates that no
measurements were made.
Most of the classes are clustered around the origin and, in a similar pattern to that
shown by their constituent species, they extend along a plane from the upper left side
of the ordination to the lower right. This pattern corresponds well with the directions
of the arrows for disturbance, altitude and exposure. It can be inferred, therefore,
that Class PCCT was highly disturbed, of low altitude with a low exposure score. By
comparison, Class AC was composed of plots from exposed, high-elevation situations
with low disturbance.
2. Community Ecology of the Ecuadorian Paramos
Cotacachi
Chiles
Beta Diversity Units
2
48
Beta Diversity Units
2
1.6
1.6
0.6
0.6
6300 2300 8400 1600 8400 0700 1300 3400 4000 4100 4100 4300 4400 4600 4600
6200 3300 3400 3400 3600 8700 3600 61100 4000 4100 4200 4300 4400 4500 4600
Altitude (m)
Altitude (m)
Guamani
Tungurahua
2
Beta Diversity Unite
2
1.6
1.6
1
1
06
0.6
3200 3300 3400 3500 3600 3700 3600 3600 4000 4100 4000 430e 4400 4500 4500
2
Beta Diversity Unite
3200 3300 3400 3600 3600 3700 3600 3000 4000 4100 4. 00 4300 4400 4600 4600
Altitude (m)
Altitude (m)
El Altar (west)
El Altar (east)
2 Beta Diversity Unite
Beta Diversity Units
1.6
1.6
1
1
0.5
0.6
3300 3300 2400 3600 3600 3700 3800 WO 4000 4100 4200 4300 4400 4600 4600
3200 3300 3400 4300 3600 3700 3500 3500 4000 4100 4200 4300 4400 4500 4600
Altitude (m)
Altitude (m)
Figure 2.15. (Continued overleaf)
a-diversity values along the altitudinal gradient in twelve paramo areas.
2. Community Ecology of the Ecuadorian Paramos
Daldal
Chimborazo
Beta Diveraity Units
2
49
2
1.6
Beta Diversity Units
1.6
1
0.5
0.6
0
3200 6800 8400 0800 6800 2700 3600 3300 4000 4100 41100 4300 4400 4600 4600
3100 3300 1400 3600 0600 8700 3200 $1100 4000 4100 4200 8300 4400 4600 4600
Altitude (m)
Altitude (m)
Zap ote Naida
Cajas
Beta Diversity Units
Beta Diversity Units
2
2
16
1.6
06
0.6
3200 1100 3400 6600 3600 3700 3600 3400 4000 4100 4100 4300 4400 4600 4600
3100 3300 3400 3800 3600 0700 2600 1200 4000 4100 4200 4300 4400 4600 4800
Altitude (m)
Altitude (m)
Cumbe
2
Beta Diversity Units
16
1
0.6
it
II
*ill,
0
3100 8300 1400 0600 6600 2700 3800 3100 4000 8100 4300 4300 4400 4800 4000
Altitude (m)
Figure 2.15. (Continued)
fl-diversity values along the altitudinal gradient in twelve paramo areas.
2. Community Ecology of the Ecuadorian Paramos
50
Again following the pattern established in the species ordination, a second cluster
of TWINSPAN classes is located in the upper right of the ordination space, roughly
perpendicular to the plane of the main cluster. These stand groups are characterised
by those species from Volcan Chimborazo, described earlier for this portion of the
ordination, and are correlated with the amount of bare ground present. These outlying groups were very different from those around the origin. Class SD, which split
off from the rest of the stands in the first TWINSPAN division, has been placed furthest from the main cluster of groups. Similarly, Classes GCCD, ACCD, CCD and
NCCD are set apart from the others in the classification as well as in the ordination,
confirming their distinctive composition.
The fl-diversity values for each of the study areas are given in Figure 2.15. Generally, for any one altitudinal gradient, the fl-diversity values continue to increase as the
elevational difference increases. This indicates that plant community composition
changes continuously along the gradient, though the rate of change varies. Almost all
of the fl-diversity curves show a decrease in value at the end of the altitudinal range.
This is an artefact of the fl-diversity formula, where the value depends on the species
lost plus the species gained. At higher altitudes, the number of species lost becomes
almost constant (perhaps even reaching its maximum, with no species in common
with the lowest altitude), but the number of species gained decreases as conditions
reduce plant cover. Thus a reduction in plant cover may result in a decrease in fldiversity. The maximum value of the fl-diversity units obtained is dependent to a
large extent on the altitudinal range covered. Maximum values of 1.5-2.0 units are
found for all sites covering a 600m range and one with a 500m range. The lowest
maximal value (0.65) is produced for Zapote Naida, with samples taken from only
three altitudinal levels.
A total of 31 vegetation types or plant communities have been defined in this
study. The definition of each of these communities is dependent upon assemblages
of species, rather than the presence or absence of key species. This is important, in
that a single species may occur in many different communities, and furthermore, may
play an important part in defining them (but only in conjunction with other species).
The term 'species' is used loosely in this context, since full identification of the
voucher specimens has not yet been achieved. Despite this, some attempt has been
made to define distinct taxa, even where a name has not been determined. In order
to avoid the situation of a 'species', so defined, requiring separation into real species
at a later date, taxa have been defined cautiously. As a result, it is possible that some
taxa may ultimately become merged into one, once full taxonomic studies have been
completed.
Of the commonest species in this study, a number have yet to be fully named.
From an examination of unpublished species lists and herbarium material, it is
possible to speculate upon the identity of some of these. It is likely, for example, that
Calamagrostis sp. [251] is Calamagrostis effusa H.B.K., that Pernettya sp. [185] represents Pernettya prostrata H.B.K., and Hypericum sp. [161] is Hypericum laricifolium
H.B.K. It is also worth noting that four Castilleja sp. (Scrophulariaceae) taxa were
defined in this study, some or all of which may represent Castilleja fissifolia. If all of
these taxa were to be C. fissifolia, then this species would become frequent enough to
2. Community Ecology of the Ecuadorian Paramos
51
merit a place in Table 2.5 of the commonest species in the vegetation samples, with
46 occurrences in the 192 plots (23.96%).
Descriptions of the plant communities derived from the TWINSPAN analysis are described below. Figure 2.16 shows the distribution of these communities in the twelve
study areas. Each description is accompanied by an abbreviation to allow cross-referencing with Figures 2.12, 2.14 and 2.16. The number of sample stands within each
community is also indicated.
The Zonal Vegetation of the Ecuadorian Paramos
Blechnum loxense Shrub Paramo (BS, 3 Stands)
This community was found at 3,900 m, just above the treeline on the flanks of Volcan Tungurahua. The community was dominated by the small (up to 1 m tall) tree
fern, Blechnum loxense, which covered between 50-75% of the surface area. Thelypteris sp. [229] ferns were also strong indicators of this community. Other important
species in this group were Disterigma empetnfolium, Geranium reptans, Calceolaria
ferruginea, Baccharis genistelloides, Gynoxys baccharioides, Pentacalia arbutifolius, Sisyrinchium jamesonia, Luzula gigantea, Relbunium hypocatpium, Oreomyrrhis andicola, Ranunculus sp. [304], Culcitium ovatum and Dyngium humile. A number of other
species were present but did not show a preference for this particular community, including Elaphaglossum sp. [282], A grostis sp. [243], A zorella pedunculata, Erigeron sp.
[333] and Baccharis alpinum each with a cover value of about 5%, and Cukitium ovaturn with over 25% cover. There was no Calamagrostis sp. [251].
The ground sloped 28° and soil pH was measured at 5.6. Exposure was high with
an index score of 4. There were no signs of human influence on the vegetation, direct
or indirect, though a track to the summit, in regular use, passed nearby.
Humid Desert Paramo (HD, 12 Stands)
A rapid change in the plant community occurred between 3,900 m and 4,000 m,
demonstrated by the steepness of the fl-diversity curve between these altitudes on
Volcan Tungurahua (Figure 2.15). By 4,000 m both Blechnum loxense and Thelypteris
sp. [229] had disappeared completely. No single species was dominant on Volcdn Tungurahua from 4,000 m to 4,300 m; instead a low carpet of vegetation was found, characterised by the presence of Lachemilla hispidula, A grostis nigritella, Eriosorus sp.
[288], Baccharis alpinum, Oritrophium peruvianum, Luzula racemosa, Bartsia sp.
[167], A splenium sp. [230], Cukitium ovatum and Erigeron sp. [333], each covering between 5 and 25% of the area. Furthermore, the lower part of this community was
characterised by species such as Hypochaeris sessiliflora, Ericaceae sp. [335], A zorella
pedunculata and A grostis sp. [243], and to a lesser extent by Elaphaglossum sp. [282]
and Lupinus purdianus. On the other hand, Culcitium nivalis (with a cover of more
than 25% at 4,300 m), Ophioglossum crotalophoroides and Cerastium floccosum
tended to occur most often in the upper part of the community. There was no Calamagrostis sp. [251].
2. Community Ecology of the Ecuadorian P6ramos
52
o
0
000
< 0 cn
0
oIts
o
cu
trANI
0
I— 0
0000
< 0 Z
IMO&
ILI IA 0 )—
0 0 0 0
-a
a)
Q_> 0- 0-
eL
E
U) CZ
a) co
C
0 c
C Tts
CC
'—
0 0 0 co
>0 Z
I
"
Ci
d
:::=111
a) c.1
.171) Cn1
0
=
co —
'cTs <
z-c
DUE
0
00
1— 0 <
a a_
-MOM
1110111111.m.s.
Vhar=:::
Cl) o co
0 0 u) 1--
u_ 0 0 0
.1111110WIR
0
-C
ctS
C)
0 CC
U)
CD I— e)
• C.0 CD
CD 0)
E
osco co
.• —
c
w
0 C
+=
M CZ
-C
.c
g -8
Li: a)
>-8
9 9
< 0
12_ c
CO 9 0
z yc , —c5
Z
01-
en 0 0 0
03100
1— 2 F2
ch Co
a) 2 1:2
c.)
15 5 -8
C
0 a, as
_C 0C Icy
775
-0) .6 co
1._ c
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
LO .7
0
a) a)
N
N
•cr .1
cr
"Cr
7-
V)
V)
co co co co
2. Community Ecology of the Ecuadorian Paramos
53
The terrain was fairly constant on the volcanic cone, mostly determined by the
angle of repose of the ash and its erosion by water into gullies. The gradient varied
from 29° to 38° and exposure was high (4 or 5). Any disturbance was low (1) and consisted of occasional trampling damage by mountaineers en route to the summit. The
ash substrate was found to have a pH varying from 5.3 to 5.7, with a mean of 5.5.
Calamagrostis sp. [251] and Chuquiraga jussieui Desert Pbramo
with Cerastium sp. [200] (CCCD, 1 Stand)
Not far beneath the snow-line on Volcdn Chimborazo, a patchy vegetation was
found among the scree at 4,600 m, dominated, like so much of the vegetation on this
mountain, by shrubs of Chuquiraga jussieui and Calamagrostis sp. [251] tussocks
(both less than 25% cover). Indicators of this group were Cerastium sp. [200] and Relbunium croceum, plus cushions of Geranium sp. [157] and a small grey species [178]
with red leaf margins which remains unidentified. W erneria humilis cushions, Hypochaeris sessihflora, A grostis nigritella and Cukitium ovatum were also present.
The slope of the ground was 19°. Exposure was very great (5) and there was no evidence of any disturbance. The pH of the soil beneath this stand was found to be 5.7.
Calamagrostis sp. [251] and Culcitium ovatum Humid Tussock
Grassland (CCT, 3 Stands)
On the Colombian border in the far north of Ecuador, the vegetation at 4,000 m
on Volcdn Chiles was distinguished by the presence of Calamagrostis sp. [251] tussocks (with a cover of 50-75%), with Culcitium ovatum, Lachemilla orbiculata, Ladlemilla pinnata and Geranium sp. [160], all with cover values between 5-25%. Other
constant members of this community were Pentacalia andicola, W erneria pumila,
Carex tristicha,V aleriana plantaginea, Jamesonia sp. [342] and Geranium sibbaldioides.
Other species of importance in this group were Perezia pungens, Lachemilla nivalis,
Niphogeton dissecta and V aleriana bonplandii.
The slope of the ground was 15°. The vegetation was subject to moderate exposure
and disturbance (with index scores of 3 and 2, respectively). Soil pH was measured at
4.4.
Calamagrostis sp. [251] and Festuca sp. [255] Tussock
Grassland (FCT, 1 Stand)
In places at 4,100 m, amongst the Diplostephium rupestre and Calamagrostis sp.
[251] High Altitude Shrub Community on Volcan Chiles (described below), was a
somewhat different vegetation type, lacking the shrubs of Diplostephium rupestre. Instead, Festuca sp. [255] tussocks were co-dominant with those of Calamagrostis sp.
[251] (each with a cover of 25-50%). Between the tussocks, the characteristic species
were Lachemilla rupestris, V aleriana plantaginea, Poa sp. [262] V aleriana ? niphobia,
Oritrophium peruvianum, Pentacalia andicola, Senecio chionageton, W erneria pumila,
Bartsia laticrenata, Cerastium sp. [199], V aleriana sp. [312], Jamesonia sp. [342], A grostis nigritella, A grostis sp. [239], Lachemilla nivalis, Geranium sibbaldioides, Disterigma
empetnfolium and Carex pichinchensis.
2. Community Ecology of the Ecuadorian Paramos
54
The slope of this plot was 22°, exposure was high (4) and there were no signs of disturbance. The soil pH was 5.0.
Diplostephium rupestre and Calamagrostis sp. [251] High
Altitude Shrub Community (DCS, 5 Stands)
Diplostephium rupestre, W emeria humilis and Gentianella foliosa characterised the
plant community at the highest altitudes sampled on Volcan Chiles (4,200 and 4,100
m). A low cover of Calamagrostis sp. [251] tussocks was present, with Jamesonia sp.
[342], A grostis sp. [239], Disterigma empetnfolium, Lycopodium sp. [289], Gentiana
sedifolia, Niphogeton dissecta, V aleriana bonplandiana, A zorella sp. [195], Carex pichinchensis, Geranium sibbaldioides, Lachemilla nivalis, A grostis nigritella, Festuca sp.
[255], small mats of A ciachne flagellifera and cushions of Plantago rigida.
Slope varied from 22° to 28°. No indications of disturbance were found and the exposure index was judged to be 4 in all cases. The soil pH of these plots was 4.9 or 5.0.
High Altitude Diplostephium rupestre Shrub and Cushion
Community (DSC, 3 Stands)
At 4,300 m in the Paramo de Guamani, Ranunculus sp. [304] and Cerastium sp.
[198] were the major diagnostic species, with cushions of Plantago rigida (having a
cover of greater than 50% in one plot) and A zorella corymbosa prominent. Culcitium
ovatum, Lachemilla hispidula, Gentiana sedifolia, Niphogeton dissecta, Eryngium
humile, Castilleja sp. [168], Halenia sp. [189], Ophioglossum crotalophoroides and
Draba sp. [234] were also typical of this community, though to a lesser extent. Diplostephium rupestre shrubs were a constant and conspicuous feature of the community.
Hypochaeris sonchoides, Oritrophium peruvianum, Oreomyn-his andicola, V aleriana
adscendens, V aleriana bonplandiana, A grostis nigritella and Poa sp. [261] were present
in all three plots, and W errzeria humilis, Disterigma empetnfolium and Geranium sp.
[160] were present in two-thirds of the samples.
A slope of 35° was consistent for all three plots. Exposure was high (4). No symptoms of disturbance were encountered in any of the three stands, and the pH was
determined as 5.1.
Tussock and Cushion Paramo with Diplostephium rupestre
(DTSC, 3 Stands)
At 4,200 m in the Paramo de Guamani, one hundred metres below the community
just described and similar to it in many respects, the vegetation differed in one major
way: it was dominated by Calamagrostis sp. [251] tussocks, covering 50-75% of the
sample plots. Co-dominant were shrubs of Diplostephium rupestre. The presence of
Oritrophium hieracioides and Pernettya sp. [185] was also characteristic of this community. W erneria humilis was notable, forming cushions covering 25-50% of the area.
Other distinguishing members of the community were Carex lemanniana, Geranium
sibbaldioides, Luzula racemosa, Satureja nubigena, Bartsia laticrenata, Lycopodium
sp. [189] and Loricaria sp. [334] . Many species present were equally common in the
community described previously at 4,300 m in the Pdramo de Guamani: V aleriana
2. Community Ecology of the Ecuadorian Pâramos
55
bonplandiana, V aleriana adscendens, A grostis nigritella, Poa sp. [261], Disterigma empetnfolium, Geranium sp. [160], Oreomyn-his andicola, Oritrophium peruvianum and
Hypochaeris sonchoides.
The gradient was generally steep at 32°, 34° and 46° for the individual stands. None
of the plots had been recently disturbed and exposure was judged to be moderate
(3). The pH was measured at 5.1.
Distichia muscoides, Azorella corymbosa and Werneria humilis
Cushion Pâramo (DTC, 3 Stands)
On the eastern slopes of El Altar at 4,100 m, a diverse community existed with no
one species dominating. Calamagrostis sp. [251] tussocks varied in cover from 5-15%.
A strong cushion and mat component, consisting of Distichia muscoides, A zorella cogmbosa and W erneria humilis, was responsible for much of the vegetation cover. Between these plants, other diagnostic species were growing: Ranunculus sp. [303],
Baccharis caespitosa, V aleriana adscendens, Lachemilla orbiculata, Luzula racemosa,
Oreomyrrhis andicola, Senecio repens, Oritrophium peruvianum, Isoetes sp. [323], Eleochaeris sp. [363], Nertera granadensis and Lachemilla nivalis. Also present were Pernettya sp. [185], Eryngium humile, A zorella aretoides, Bromus lanatus, Geranium
sibbaldioides, Gentiana sedifolia, Disterigma empetnfolium, Hypochaeris sessiliflora, Sisyrinchium jamesonia, Plantago rigida, Bartsia laticrenata and Hypericum sp. [273],
none of which was indicative of this community.
The slope was the same for all three quadrats in this community (15°). Exposure
was fairly low with an index score of 2. No indications of disturbance were observed
and no pH reading was possible for these plots.
Plantago rigida Cushion P6ramo (PC, 3 Stands)
At 4,200 m on the western slopes of El Altar, near the crater, the vegetation was
dominated by cushions of Plantago rigida, covering 50-75% of the ground surface,
and in one case more than 75%. Associated with this community were tussocks of Calamagrostr's sp. [251], which covered 25-50% of the area. Poa sp. [261],Agrosks sp.
[238], Lachemilla Espidula and Carex pichinchensis were indicative of this community, and to a lesser degree Cukitium ovatum, Loricaria thuyoides, Oritrophium
limnophilum, Halenia weddelliana, Ranunculus peruvianus, Lachemilla sp. [209], A grostis nigritella, Festuca sp. [254], Poa sp. [262] and Uncinia sp. [329]. A zorella aretoides, Bartsia laticrenata, Sisyrinchium jamesonia, Bromus lanatus, Disterigma
empetnfolium, W erneria humilis and Hypochaeris sessiliflora were constant species in
the community, though have no value in defining it. Other important species in the
community were Gentiana sedifolia, Geranium sibbaldioides, Elyngium humile, Pernettya sp. [185] and Hypericum sp. [273].
The slope of these plots was 18°, exposure was 3 and disturbance moderate at 2.
The soil pH was 5.3.
2. Community Ecology of the Ecuadorian Paramos
56
Werneria humilis & Plantago rigida Cushion Paramo (WPC, 3
Stands)
At the highest sampling altitude (4,200 m) on Volcan Cotacachi, cushions of W erneria humilis were co-dominant with Plantago rigida cushions and patches of Lycopodium sp. [289] (each of these species covering between 25-50% of the area).
Calamagrostis sp. [251] was present in two-thirds of the samples, though with less
than 25% cover. A grostis nigritella, Bartsia laticrenata, Geranium sibbaldioides,
A ciachne flagellifera, Oritrophium peruvianum, V aleriana aretioides, Luzula racemosa,
Disterigma empetnfolium and Hypochaeris sonchoides were also present. This community was differentiated from similar cushion paramos by the presence of Loricaria
sp. [334], Calamagrostis sp. [246], Halenia sp. [189], A zorella aretoides, Lachemilla nivalis and Diplostephium rupestre.
The terrain was reasonably flat (4° to 6°). Exposure was high (4) and disturbance
low (1). Soil pH was found to be 5.1.
Werneria humilis & Azorella corymbosa Cushion Péramo (WAC,
3 Stands)
At 4,400 m in the Paramo de Guamani, a similar-looking vegetation to that described above on Cotacachi was found. W erneria humilis was co-dominant, this time
with A zorella corymbosa, their cushions jointly covering more than 50% of the
ground surface. Lachemilla hispidula, Lachemilla holosericea, Cerastium sp. [198],
Oritrophium hieracioides and Oreomyrrhis andicola were also key components of the
vegetation. Other species in the community were Lycopodium sp. [289], A grostis nigritella, Bartsia laticrenata, Oritrophium peruvianum, A ciachne flagelhfera and Geranium
sibbaldioides. There were no tussocks of Calamagrostis sp. [251] in these samples.
These plots were located on very steep slopes, 35° to 45° from the horizontal, and
in a very exposed situation which merited the maximum index score of 5. There were
no visible signs of disturbance. Soil pH was 5.5.
Lachemilla holosericea Cushion P6ramo (AC, 3 Stands)
On the eastern flanks of El Altar at 4,300 m, a thin covering of vegetation lay on
the mineral substrate, with no one species occupying more than about 15% of the
sample plots. Indicators of this vegetation were Luzula racemosa, Lachemilla holosericea and Cukitium aclscendens. Also present were W erneria humilis, A zorella corymbosa, Oreomyrrhis andicola, Huperzia hypogoea, Hypochaeris sessihflora,
Oritrophium hieracioides and Bartsia laticrenata. Once more, Calamagrostis sp. [251]
was absent from these plots.
On a 20° slope, these three stands were located in a very exposed situation and
given an exposure score of 5. No indications of disturbance were observed. Measurement of pH was not possible for these stands.
2. Community Ecology of the Ecuadorian Paramos
57
Aciachne flagellifera &Valeriana aretioides Cushion P6ramo
(AAC, 3 Stands)
100 m below the desert paramo community on the eastern side of El Altar, at
4,200 m, there existed a vegetation dominated by cushion and mat plants. Most indicative of this community were the cushion and mats ofA ciachne flagellifera, Plantago
rigida and V aleriana aretioides, with Geranium sibbaldioides, Baccharis caespitosa and
Poa cucullata. Oritrophium peruvianum, A zorella aretioides, Lachemilla nivalis, Carex
sp. [319] and Eudema nubigena were also characteristic. Huperzia hypogoea, Oreomyrrhis andicola, A zorella corymbosa, Bartsia laticrenata, W erneria humilis, Oritrophium hieracioides and Hypochaeris sessiliflora were present in these stands, but were
not diagnostic of this community.
The slope of the plots was measured at 20° from the horizontal. Exposure was
high, at 4, and no evidence was observed to suggest disturbance of the sites by burning, trampling or grazing.
Calamagrostis sp. [251] Tussock Grassland with Hypochaeris
sonchoides, Halenia sp. [189] and Satureja nubigena (HHCT, 6
Stands)
On Volcan Cotacachi at 4,000 m and in two of the three replicate quadrats at
4,100 m, plus one stand from 4,100 m on Volcan Chimborazo, Calamagrostis sp. [251]
tussocks were almost completely dominant. In most cases the tussocks formed a completely closed canopy about 1-1.2 m above the ground. Associated with the tussocks,
beneath the canopy, were Cerastium danguyi, Satureja nubigena, Halenia sp. [189],
A grostis sp. [242], Bartsia laticrenata, Castilleja sp. [168], Poa sp. [268], Elaphaglossum
sp. [281], Lycopodium sp. [293] and Lycopodium sp. [295]. In addition to these
species, a number of others were present, but were not indicative of the community:
Pernettya sp. [185], Alliaceae sp. [175], V aleriana microphylla, V aleriana rigida, Dyngium humile, Lupinus sarmentosus, Hypochaeris sonchoides and Gnaphalium luteoalbum.
In general, slopes were steep in these plots, between 35° and 46° (though one
stand was found on reasonably flat ground with a slope of only 5°). Exposure was variable, ranging from 2 to 4 with a mean of 3.2. Mean disturbance was 2.2. Measurements of pH were determined at 4.9, 5.5 or 5.6 for these stands, the mean value
being 5.4.
Calamagrostis sp. [251] Tussock Grassland with Hypochaeris
sonchoides, Plantago linearis and Relbunium croceum (HPCT,
12 Stands)
Calamagrostis sp. [251] tussocks were again dominant from 3,600 to 3,900 m on
Volcan Cotacachi, though they were not so dense as those higher on the mountain.
With a cover of 50-75%, the tussocks did not form a completely closed canopy. Plantago linearis, Bidens andicola, Hieracium frigidum, Relbunium croceum, Equisetum bogotense, Paspalum tuberosum, Poa sp. [269] and Gnaphalium coarctatum were
characteristic species between the Calamagrostis sp. [251] plants. Though commonly
2. Community Ecology of the Ecuadorian Pâramos
58
linked with dense Calamagrostis sp. [251] tussocks elsewhere, Pemettya sp. [185],
V aleriana microphylla, V aleriana rigida, Hypochaeris sonchoides and Etyngium humile
were constant members of this community.
These stands grew on slopes which varied from 29° to 37° (mean, 34). Exposure
was moderate with a mean index score of 2.75. Disturbance was high (4 or 5) with a
mean score of 4.5. pH varied from 4.7 to 5.6; the mean value for the twelve stands
was 5.2.
Calamagrostis sp. [251] Tussock Grassland with Sisyrinchium
jamesonia (SCT, 21 Stands)
The remaining stands in the pdramo of Daldal, from 3,800 to 4,200 m, and the
western slopes of El Altar from 3,900 to 4,100 m consisted of a community dominated by Calamagrostis sp. [251] (50-75% cover). Key floristic components of this
community were Hypericum sp. [161], Sisyrinchium jamesoni, Oritrophium peruvianum, Carex pichinchensis, Lupinus sarmentosus, A zorella aretoides, Castilleja sp.
[171], A phanactis jamesonia and Uncinia sp. [329]. In addition, Hieracium frigidum,
Hypochaeris sessiliflora, Carex tristicha, Disterigma empetnfolium, Gentiana sedifolia,
Halenia weddelliana, Geranium multipartitum, Geranium sibbaldioides, Bromus lanatus, Paspalum tuberosum, Ranunculus peruvianum, Lachemilla rupestre, V aleriana
microphylla and Pernettya sp. [185] were present, but did not serve to distinguish this
community from other similar ones.
This large group of plots was found on slopes of 17° to 42°, averaging 28.6°. Exposure was intermediate, with a mean score of 3.1. Disturbance was more variable,
from 2 to 4, though again centred on 3.1. pH measurements extended from 4.8 to 5.4.
The mean pH value for all 21 stands was 5.0.
Calamagrostis sp. [251] Tussock Grassland with Viola humboldtii
(VCT, 10 Stands)
Four plots from 3,500 and 3,600 m in the paramo of Cajas and six more from 3,700
to 3,900 m in the paramo of Daldal, belonged to a community characterised by the
presence of V iola humboldtii. The dominant species was the tussock grass CalamagrostLs sp. [251] (often with a cover greater than 75%) with Paspalum tuberosum a
constant associate. Other species which were consistent members of the community
were Bidens andicola, Carex crinalis, Geranium reptans, A zorella pedunculata, Sibthorpia reptans, Poa sp. [262], Gnaphalium pensylvanicum, Carex tristicha, Disterigma empetnfolium, Halenia weddelliana, Geranium multipartitum, Geranium sibbaldioides,
Ranunculus peruvianus, Lachemilla rupestris, Lachemilla orbiculata, Oreomyn-his andicola, Pemettya sp. [185] and Hypochaeris sessiliflora.
The slopes of these ten plots were very variable, ranging from quite flat terrain
with a slope of only 3° to steep, 37° valley sides. Exposure was moderate, with a mean
score of 2.8, and disturbance was moderate to high (3, mostly 4). The mean disturbance score was 3.8. pH measurements could only be assigned to six of the ten plots in
this community: the mean of these plots was 5.2.
2. Community Ecology of the Ecuadorian Paramos
59
Calamagrostis sp. [251] Tussock Grassland with Oreomyrrhis
andicola and Gnaphalium pens ylvanicum (OCT, 20 Stands)
A plant community dominated by Calamagrostis sp. [251] with Oreomyrrhis andicola, Gnaphalium pensylvanicum and Lachemilla orbiculata was found in three paramo
areas. In the north-east, this community was found from 3,800 to 4,100 mill the Paramo de Guamanf; in central Ecuador it was found at 3,800 m on the Collanes Plain
below the crater of El Altar; and in the south it was found in the paramo of Cajas at
3,400 m and less extensively at 3,500 and 3,600 m.
In these stands, Calamagrostis sp. [251] tussocks were responsible for more than
half the vegetation coverage, often for more than three-quarters of it. Associated
species were Oreomyrrhis andicola, Gnaphalium pensylvanicum, Lachemilla orbiculata, Bidens andicola, A zorella pedunculata, A phanactis jamesonia, A grostis sp. [242],
Poa sp. [261], Poa sp. [262], Satureja nubigena and Eryngium humile.
This community was found on almost flat ground as well as on steep slopes up to
33°. Exposure was moderate with a mean score of 2.2. Disturbance was mostly rated
at 3 or 4, the mean of the twenty stands was 3.5. Once again, soil samples were not
available for one-quarter of the plots. Those that were available had pH values which
varied from 5.1 to 5.6, with a mean of 5.4.
Neurolepis elata Bamboo 13 6ramo (NB, 10 Stands)
This group of stands comprised those plots recorded from 3,800 to 4,000 m on the
eastern slopes of El Altar plus one plot from 4,100 m on Volcan Cotacachi. In the El
Altar stands, the dominant species was the bamboo grass, Neurolepis elata, which
formed dense tussocks covering up to 75% of the plots. Neurolepis was absent from
the Cotacachi community, but the remainder of the flora was similar enough for it to
be joined with those from the eastern flanks of El Altar. Characteristic species of this
community were Culcitium adscendens, Diplostephium hartwegii, Diplostephium glutinosum, Carex lemanniana, Rhynchospora macrochaeta, Disterigma empetnfolium,
Geranium sibbaldioides, Lupinus sarmentosus, Perrzettya sp. [185], A rcytophyllum aristatum, Gentianella sp. [316] and Oritrophium peruvianum. Calamagrostis sp. [251] tussocks were also present, sometimes co-dominant with the bamboo tussocks.
A number of species were common to this community and the Calamagrostis sp.
[251] tussock grassland with Oreomyrrhis andicola and Gnaphalium pensylvanicum
described above: Lachemilla orbiculata, Satureja nubigena, Senecio chionageton, Sisyrinchium jamesoni, Bartsia laticrenata and Eryngium humile.
The stands belonging to this community were located on slopes of 20° to 35°. The
mean slope was 26°. Exposure was fair to high (3 or 4) with a mean of 3.4. The vegetation was moderately disturbed (2 or 3). The mean index score for disturbance was
2.4. Finally, pH was only measured for the single Volcan Cotacachi plot, which was
4.9.
2. Community Ecology of the Ecuadorian Paramos
60
Calamagrostis sp. [251] and Espeletia pycnophylla Tussock
Grassland with Paspalum tuberosum (PCE, 6 Stands)
In the north, on the Colombian border between 3,600 and 3,700 m on Volcan
Chiles, Calamagrostis sp. [251] tussocks covered 50-75% of the ground, but were
themselves covered by a 5-50% cover of Espeletia pycnophylla stem rosettes. Paspaturn tuberosum was a key floristic element, as were Blechnum loxense, Pentacalia stuebellii, Gnaphalium pensylvanicum, Gynatys fulaginosa, Hypochaeris sessiliflora,
Oritrophium hieracioides, Rhynchospora ruiziana, Gentiana sedifolia, Dyngium
humile, Hypericum sp. [161], Bartsia sp. [165], A grostis sp. [241], Hypericum sp. [275]
and Lycopodium clavatum. Other important constituents of the vegetation of these
stands were Oritrophium peruvianum, W erneria humilis, Carex tristicha, Oreobolus
goeppingeri, Halenia weddelliana, Geranium sibbaldioides, Lachemilla rupestris, Lachemilla nivalis, A grostis sp. [240], Pernettya sp. [185], Sibthozpia reptans, Lupinus sarmentosus and Nertera granadensis.
The slope of the plots in this group varied from 7° to 21°, exposure was rated as 3
on the scale from 0 to 5 and disturbance was high at 4. The pH of the plots at 3,700 m
was found to be 4.7, whilst at 3,600 m the soil was the most acidic encountered in the
study with a pH of 4.3.
Calamagrostis sp. [251] and Espeletia pycnophylla Tussock
Grassland with Viola sp. [192] (VCE, 6 Stands)
At 3,800 and 3,900 m on Volcan Chiles, Espeletia pycnophylla was again co-dominant with tussock grass. At 3,800 m the tussock grass species was Calamagrostis sp.
[251] with 50-100% cover. At 3,900 m, however, the Calamagrostis sp. [251] was replaced by tussocks of A grostis sp. [240] with slightly less cover (50-75%). The distinctive floristic elements of this community were V iola sp. [192], Rhynchospora
macrochaeta, Sisyrinchium aff. alatum, Senecio chionageton, Bartsia laticrenata, A zorella aretoides, Diplostephium sp. [233], A grostis nigritella, Poa sp. [262], Jamesonia
pulchra, and A rcytophyllum sp. [305]. Other species were present in this community
in common with the similar stands lower on Volcan Chiles: Lupinus sannentosus, Oritrophium peruvianum, Carex tristicha, Oreobolus goeppingeri, Halenia weddelliana,
Geranium sibbaldioides, Lachemilla rupestris and Penzettya sp. [185].
Once again, slopes were very variable. One of the plots was located on completely
level ground whereas the others were found on slopes between 12° and 20°. Exposure was moderately high with a mean of 3.5 and all of the plots were heavily disturbed (4). As with the paramo below this community, the soils were highly acidic
with values of 4.4 and 4.7.
Calamagrostis sp. [251] Tussock Grassland with Paspalum
tuberosum and Chrysactinium acaule(PCCT I 24 Stands)
In southern areas, a more open tussock grassland existed, with Calamagrostis sp.
[251] tussocks less dominant with a cover of 25-75%. Paspalum tuberosum (often
with a cover of greater than 25%) was indicative of such vegetation, along with Guysactinium acaule, Orthrosanthus chimboracensis, Oreobolus goeppingeri, V aleriana
2. Community Ecology of the Ecuadorian Pijramos
61
bonplandiana, Halenia sp. [188], A grostis sp. [240] and Lycopodium clavatum. Other
members of the community were Hypochaeris sessiliflora, Oritrophium peruvianum
forma intermedium, Geranium sibbaldioides, Dyngium humile, Hypericum sp. [161]
and Pernettya sp. [185].
All of the plots in the paramos of Zapote Naida, Cumbe and Ofia were assigned to
this group by the TWINSPAN analysis, in addition to the three quadrats from 3,700 m
in Cajas.
This was the largest grouping of stands from the analysis and the slopes of the 24
plots varied considerably from 40 to 42°. The index of exposure was between 2 and 4
(mean, 3.0) and disturbance was low in some plots but high in others, the mean was
3.0. Unfortunately, the soil samples from the plots in this community were amongst
those stolen and no pH values are available.
Calamagrostis sp. [251] and Poa sp. [262] Tussock Grassland
(PCT, 4 Stands)
This grouping consisted of the three plots from the highest altitude sampled in the
paramo of Cajas (4,000 m), plus one plot from 3,800 m in the same area. The vegetation was dominated by Calamagrostis sp. [251] tussocks (cover 50-75%). This community was set apart by the presence of Poa sp. [262]. Other consistent species in the
community which were of indicative value included Carex pichinchensis, Halenia weddelliana, Baccharis alpinum, Diplostephium hartwegii, Ranunculus peruvianum, Dyngium humile, V aleriana bracteata, Bartsia sp. [165] and Jamesonia alstonii. Other
members of the community were A rcytophyllum filifonne, Pernettya sp. [185], Lupinus
sannentosus, Oritrophium peruvianum forma intennedium, Hypochaeris sessiliflora,
W erneria humilis and Hypericum sp. [161].
One of the four stands was found on almost level ground, while the remainder occurred on a slope of 31 0 . Exposure was quite high with a mean of 3.8 and disturbance
scored 2 for three of the sites, but 4 for the remaining plot. No pH measurements
were made.
Calamagrostis sp. [251] and Agrostis sp. [243] Tussock
Grassland (ACT, 2 Stands)
The remaining two stands at 3,800 m in the paramo of Cajas were quite similar to
the community classified above as Calamagrostis sp. [251] and Poa sp. [262] tussock
grassland. Approximately half the area was covered by Calamagrostis sp. [251]. This
community differs in that Poa sp. [262] was absent, and a variety of other species
were present: A grostis sp. [243], V aleriana bonplandiana, Lachemilla rupestris, Geranium sibbaldioides, Paspalum tuberosum, Carex tristicha, Gentiana sedifolia, Gentianella hirculus and Oreobolus goeppingeri. However, A rcytophyllum filiforme, Perrzettya
sp. [185], W erneria humilis, Hypericum sp. [161], Lupinus sannentosus and Oritrophium peruvianum forma intermedium were still important elements of the community.
2. Community Ecology of the Ecuadorian Paramos
62
The slope of these two plots was 5 0 , the exposure score moderate at 3 and disturbance high at 4. No pH values were available.
Calamagrostis sp. [251] and Rhynchospora macrocha eta
Tussock Grassland (RCT, 4 Stands)
The plant community at 3,900 m in the paramo of Cajas and one plot from 4,100 m
on Volcan Chimborazo were linked in the TWINSPAN analysis. Dominated by Calamagrostis sp. [251] tussocks (cover 50%), the community contained Festuca sp. [256]
(cover 25-50%), Rhynchospora macrochaeta and Senecio chionageton, Gynoxys miniphylla, Disterigma empetnfolium, Culcitium sp. [232], Pernettya sp. [185], Poa sp.
[262], Jamesonia alstonii and Diplostephium hartwegii.
The slope of three of the four plots was 37° (the remaining plot was on an incline
of just 5°). Exposure was low at 2 and disturbance was moderate (3) to low (1) with a
mean of 2.5. A single pH reading of 5.5 was available for the Chimborazo plot.
Calamagrostis sp. [251] and Chuquiraga jussieui Desert 136ramo
(CCD, 7 Stands)
A sparse, low-diversity vegetation was found at 4,200 m (and less extensively at
4,100, 4,300 and 4,600 m) on Volcan Chimborazo. Consisting of Calamagrostis sp.
[251] tussocks and Chuquiraga jussieui shrubs, very few other species were found,
and in any case were non-selective to this community. Most prominent amongst
these species were Baccharis genistelloides and Geranium sp. [157].
These plots were located on the flanks of Volcan Chimborazo, some of which were
on fairly level ground (5°) whilst others were on steep slopes of 19° to 30°. Exposure
was highly variable: the plot at 4,600 m was rated as very highly exposed with a score
of 5, but as altitude decreased the remaining plots scored 3 or 2. Overall, the mean
was 2.7. Disturbance of these sites was assessed as minimal (mean, 0.9). The mean
pH for this community was 5.5, with a variation from 5.3 to 5.7.
Calamagrostis sp. [251] and Chuquiraga jussieui Desert 136ramo
with Nototriche jamesonia (NCCD, 3 Stands)
The basic Calamagrostis sp. [251], Chuquiraga jussieui, Baccharis genistelloides and
Geranium sp. [157] community on Volcan Chimborazo just described was enhanced
at 4,300 and 4,400 m by the presence of Nototriche jamesonia, V aleriana microphylla,
Culcitium adscendens, Hypochaeris sessiliflora, Gentianella cernua, Cruciferae sp.
[206],A grostis sp. [242] and Plantago sp. [302].
The mean slope of these plots was 20°. Exposure was high, averaging 3.7 and disturbance was low with a score of 1. pH was 5.4 or 5.7 (mean, 5.5).
2. Community Ecology of the Ecuadorian Paramos
63
Calamagrostis sp. [251] and Chuquiraga jussieui Desert
Pàramo, with Agrostis nigritella(ACCD, 1 Stand)
In one plot at 4,600 m on Volcan Chimborazo, a community was found consisting
of a 25-50% cover of Calamagrostis sp. [251] tussocks with Chuquiraga jussieui
shrubs, cushions of Geranium sp. [157] and the herbs Baccharis alpinum, V aleriana
hartwegii and Hypochaeris sessiliflora. It differed from similar vegetation 100-200 m
below in the presence of A grostis nigritella, Culcitium ovatum, Cerastium sp. [200] and
Egngium humile.
This single plot community was found on a slope of 19°, in a very highly exposed location (5) with no signs of disturbance. The soil pH was measured at 5.7.
Calamagrostis sp. [251], Chuquiraga jussieui and Geranium sp.
[157] Desert 13 6ramo (GCCD, 4 Stands)
At 4,500 m, and one plot at 4,400 m, on Volcan Chimborazo the vegetation was
variable in composition. Calamagrostis sp. [251] tussocks covered less than 50% of
the area (and in two of the plots, less than 25%), with Chuquiraga jussieui shrubs, cushions of Geranium sp. [157], Baccharis genistelloides, Baccharis alpinum, V aleriana
hartwegii and Hypochaeris sessiliflora between them. Relbunium croceum, Castilleja
sp. [168], Gentianella cernua, Lupinus smithianus, Erigeron pinnatus, Not otriche
jam esonii, A grostis sp. [242] and Lucilia radians were additional constituents of this
group.
The mean slope of these four plots was 19°. Exposure was high (4) and disturbance
low (1). The mean pH was 5.3, with little variability.
Stipa sp. [253] and Senecio teretifolius Desert 136ramo (SD, 3
Stands)
Tussocks of Stipa sp. [253] covering 25-50% of the ground, interspersed by small
shrubs of Senecio teretifolius and acaulescent rosettes of Plantago sp. [301], characterised the vegetation at 4,000 m on Volcan Chimborazo. Lucilia radians was the only
other species found in these plots, but was not as common as it was higher up the
mountain.
At this location, the terrain was relatively flat, inclined at 5° in an area of reasonable shelter (exposure index score of 2). Few indications of disturbance were noted
and these plots received a score of 1 on the disturbance scale. The pH at this altitude
was found to be 5.6.
2. Community Ecology of the Ecuadorian Paramos
64
Discussion
In total, twelve sites were used to gather the data for this study. These mountain
areas were well distributed throughout the country, and covered all of the major phytogeographical regions of Ecuador. However, owing to the isolated nature of many
'Aram° areas, some regions were somewhat under-represented (in particular, the far
south of Ecuador near the border with Peril, and the outer slopes of the eastern
Andes).
The vegetation of the Ecuadorian paramos was described by means of stratified
random sampling. This is a somewhat different approach to that used by Cleef (1981)
to describe the paramo vegetation of the Colombian Cordillera Oriental. He recorded data from stands which were subjectively chosen as representative of a particular community: vegetation sampling according to the classical method of the
Zurich-Montpellier school. The random approach employed in this study provided a
more objective method for classifying different plant communities. Although it is less
effective at sampling the entire range of vegetation types, a randornised selection of
vegetation samples permits a statistical treatment of the data. However, a totally random procedure would have been impractical over such a large area which included
variable terrain. Therefore, a stratified approach was adopted as a compromise. At
regular 100 m intervals of altitude, a 100 m transect was established perpendicular to
the slope and the location of the samples determined from random coordinates
generated previously. This worked well in most cases, though at higher altitudes
rocky outcrops and cliffs often demanded that the transect be shortened.
Sampling was conducted by means of square quadrats covering an area of 25 m2.
This corresponds well with the minimal areas of 25-35 m2 advocated by Vareschi
(1970) for Venezuelan paramo vegetation and Cleef (1981) for the paramo of the Colombian Cordillera Oriental. Farifias & Monasterio (1980) sampled Venezuelan 'Aram° vegetation with 5 m by 2 m rectangular quadrats which they believed gave
"sufficient floristic information". However, in Ecuador at least, the tussocky nature
of much of the vegetation required a larger sampling area to eliminate variability of
composition resulting from the distribution of tussocks.
The Ecuadorian paramo flora is very similar to that recorded by Cleef (1981), but
with fewer families and genera. This may be accountable to Cleef's wider interpretation of 'zonal paramo' and his additional sampling of azonal bogs and thickets. Despite this, the proportions of families and genera in the Ecuadorian paramos reflect
those found in Colombia. The Compositae is the most important family by far, with
over one-fifth of the recorded genera. The Gramineae family is also highly significant, responsible for 13% of the genera and with a strong influence on the remainder
by virtue of the dominance of some of its members throughout the country. At the
family level, and also to a lesser degree at genus level, the paramo flora described in
this study is similar to that of other mountainous regions: the puna of Peril (Cabrera,
1958), and the mountains of East Africa (Hedberg, 1964) and New Guinea (J.M.B.
Smith, 1977).
2. Community Ecology of the Ecuadorian Paramos
65
It is interesting to note the origins of the Ecuadorian paramo flora and compare
them with the data presented for neighbouring paramo regions. Van der Hammen &
Cleef (1986) provided a check-list of genera of vascular plants for the Colombian
Cordillera Oriental, and assigned each genus to seven geographical floral elements.
Sturm & Rangel (1985) present a similar phytogeographical spectrum for the 130
most important species in the Colombian pdramo flora as a whole. Using the same
approach for the genera presented in Table 2.4, a similar spectrum of origin has been
obtained for Ecuador. These data are summarised in Table 2.8.
Geographical Element
Percentage of all Genera
Ecuador
Paranno Element
Other Neotropical Element
Austral-Antarctic Element
Ho!arctic Element
Wide Temperate Element
Wide Tropical Element
Cosmopolitan Element
Unknown Affinity Element
9
21
10
14
26
3
17
Cordillera Oriental,
Colombia
7
34
9
11
20
10
8
1
130 species,
Colombia
8
30
5
12
7
28
10
Table 2.8.
Phytogeographical spectra of vascular plant genera for the zonal paramos of
Ecuador (this study), the Colombian Cordillera Oriental (Van der Hammen & Cleef,
1986) and the 130 most important species for Colombian paramos as a whole
(Sturm & Rangel, 1985). The latter values are estimated from a graphical source.
The data of both Sturm & Rangel (1985) and Van der Hammen & Cleef (1986)
demonstrate that for Colombia, approximately half of the genera present in the paramos are of (Neo)tropical origin, the other half of temperate origin, with 7 or 8% endemic to the paramos . In Ecuador, the situation is different. Taxa of temperate
origin dominate the Ecuadorian paramos: two-thirds of the genera occurring in the
192 sample stands were of this group. Only one-third were of (Neo)tropical origin.
This is perhaps the consequence of the lower humidity of Ecuadorian paramos and
the more extreme temperatures that are likely to result from this, especially cold
temperatures. Both humidity and temperature regimes in the mountainous regions
of Ecuador are likely to present more of a challenge to developing Neotropical elements of the flora and their range may be expected to be narrower than in Colombia
and Venezuela.
Both Colombian and Ecuadorian paramo floras have been subjected to similar
periods of isolation during glaciations and expansion during warmer periods, and it is
not surprising that endemic genera make up a similar proportion of the flora (9%) in
both regions.
Balslev (1988) looked at the distributions of some Ecuadorian paramo species by
consulting available taxonomic monographs. He found that only 16% of the species
studied were known to occur beyond Peru and Colombia. Of the remainder, almost
half were endemic to Ecuador (40% of the total). However, this study was based on a
2. Community Ecology of the Ecuadorian Paramos
66
limited sample of species and treated all species equally regardless of their rarity.
Balslev also found that the majority of paramo species were trans-Andean and not
confined to one Cordillera.
Despite some general similarities, the Ecuadorian paramos are substantially different from those of Colombia and Venezuela on a number of counts. First of all, in
Ecuador Espeletia species are absent from all but a few paramo regions. These are restricted to the north of the country at the Colombian border, and to one valley in
Central Ecuador (Cuatrecasas, 1986). This genus is a significant member of the paramo flora in Colombia and Venezuela. Fosberg (1944) thought Espeletia so important
that he defined paramo vegetation in terms of its presence.
Secondly, bamboo paramos of Chusquea (formerly known as Swallenchloa) are
generally absent in Ecuador. On the eastern slopes of the Ecuadorian Andes, bamboo grasslands do occur, but are dominated by Neurolepis elata tussocks. This genus
does not dominate the paramo in the same way in Colombia (Neurolepis aristata belongs to the timberline vegetation and locally extends into the sub-paramo — Cleef,
1981). Black (1982) refers to localised areas, which he termed `carrizales', dominated
by 'espadana' ( =Neurolepis) and Swalenoclea (= ?Chusquea) on Volcdn Antisana in
the eastern Andes of Ecuador, but no further details are given. Elsewhere in Ecuador, thickets of Chusquea may be seen (for example, in the paramo of Cajas), but
these are confined to small patches, presumably by topographic and rnicroclimatic
factors.
A further distinction between Ecuadorian paramos and those to the north is that
cushion plants are more abundant in the zonal vegetation of Ecuador. Cleef (1981)
noted the lack of cushion plants in the Colombian superpdramo, but it should be remembered that there is little high altitude vegetation in the Cordillera Oriental
which he studied. Sturm & Rangel (1985) listed more cushion and mat plants for the
Cordillera Central of southern Colombia, which is directly linked to the paramos of
northern Ecuador.
Overall, there appears to be a distinct trend across the paramos of the northern
Andes, noted by Monasterio (1980c). In the far south, the jalca and Ecuadorian paramos are dominated by tussock grasses with genera from temperate regions, whereas
in the Venezuelan pdramos the giant stem rosettes (like Espeletia) are dominant. Between these extremes, the paramos of Colombia are intermediate, with grasses and
rosettes sharing dominance.
It is difficult to compare species diversity in the different paramo regions of the
northern Andes because descriptions are based on vegetation samples of varying
sizes. Species-area effects render direct comparisons impossible. However, in terms
of species evenness (one measure of diversity), the Ecuadorian paramos may be less
diverse than those of Colombia and Peal, since much of the vegetation is dominated
by Calamagrostis tussocks. Elsewhere, co-dominance in paramo vegetation is usual,
with a more open physiognomy. Until more quantitative information is available for
all areas, it is not possible to confirm such speculation on differences or similarities
in terms of species richness.
2. Community Ecology of the Ecuadorian Paramos
67
This study did not include azonal paramo vegetation, such as bogs, thickets or
woodlands. Areas of this kind are common in the paramos of Ecuador, and tend to
be restricted by topographic and possibly edaphic factors. Disturbance, especially
burning also plays a role in their distribution.
Permanently wet areas were typically dominated by cushion and mat plants such as
Distichia muscoides (Juncaceae), Plantago rigida (Plantaginaceae) and Oreobolus obtusangulus (Cyperaceae) and similar communities have been described for other
paramo regions (Cleef, 1978, 1981; Black, 1982). These plants provided a substrate
for the establishment of other species which grew amongst their close-fitting leaves.
Such cushion bogs are an antarctic-montane-tropical type of vegetation and are not
found in the Boreal Zone of the northern hemisphere (Troll, 1968).
In places, thickets of various types were encountered in the paramo (often known
as `chaparrales'). Such thickets were frequently dominated by a single species, most
often Chuquiraga jussieui (Compositae), Loricaria spp. (Compositae), Baccharis spp.
(Compositae), Chusquea spp. (Gramineae) or Brachyotum spp. (Melastomataceae).
Woodlands were found mostly in the more extensive paramos of Ecuador, and
were typically not present on volcanic peaks. Furthermore, they tended to be confined to scree slopes, often beneath sheltering cliffs. They were almost always composed of trees belonging to the genus Polylepis (Rosaceae) in conjunction with
Gynoxys (Compositae). The extent to which these woodlands have been modified by
human influence is still a matter of debate, which will be discussed the final chapter.
The zonal paramo vegetation was found to show pronounced patterns, which appeared to be related as much to regional factors as they were to altitude. Indeed, Furrer & Graf (1978) noted that the lower limit of the paramo was often 300-500 m
higher (and the snow-line some 300 m lower) in eastern p gramos than those in the
west. This was explained by the higher precipitation levels on eastern slopes.
In all, seven main types of p gramo vegetation were recorded in this study.
Shrubby Sub-Oramo
In Ecuador, shrubby paramo vegetation occurs in two distinct zones. The first zone
lies just above the cloud forest where woody vegetation grades into grassland. This is
the sub-paramo referred to by Cuatrecasas (1954, 1958, 1968) and Cleef (1981).
In the Ecuadorian study areas, the sub-paramo appeared little developed. This
contrasts sharply with Lauer's (1979) view that drier paramos are likely to possess a
greater extent of sub-paramo without the ameliorating effects of greater cloud cover.
However, population pressure in Ecuador is greater than that found elsewhere in the
northern Andes, and could be responsible for the absence of sub-paramo. Certainly,
the cloud forest is only locally present in many parts of the Ecuadorian Andes. The
loss of sub-paramo vegetation would probably have accompanied the destruction of
these forests. In most areas, the lower limit of the grassy paramo coincides with more
intensive agriculture, whether as improved pastures or land under cultivation, and
the shrubby sub-paramo no longer exists.
2. Community Ecology of the Ecuadorian Paramos
68
Sub-paramo vegetation was only represented in one of the study areas. The three
replicate stands of Blechnum loxense sub-paramo (BS) at 3,900 m on Volcdn Tungurahua were clearly intermediate between the forest vegetation below and the short
paramo vegetation above. However, the volcanic mineral substrate in this area has
resulted in a sub-pdramo which is not typical of most of Ecuador. There are, of
course, other volcanoes in the country which provide a very similar substrate.
Lojtnant & Molau (1982) described a community dominated by Blechnum loxense on
the humid summit of Volcdn Sumaco in Ecuador. However, the flora of this mountain was considered to be unusual as a result of its isolated nature. Nevertheless, it
does share some species with the plots on Volcdn Tungurahua.
None of the other study areas had a zone of sub-paramo, and as a result there are
no quantitative descriptions of this kind of vegetation for other parts of the country.
In general, sub-paramo vegetation elsewhere in Ecuador was found to be characterised by the presence of Calamagrosth tussocks and associated flora, intermingled
with shrubs of Baccharis, Senecio, Gynoxys, Brachyotum, Escallonia, Hesperomeles,
Miconia, Buddleia, Monnina and Hypericum. Acosta-Solis (1966) also regarded these
genera as important members of the sub-paramo flora of Ecuador.
High-altitude Dwarfshrub Pâramo
A second belt of shrubby vegetation occurs at much higher altitudes, usually above
4,000 m. Cleef (1981) noted this formation in his study of the Cordillera Oriental of
Colombia and explained its existence in terms of a condensation belt at this altitude
which permits the growth of woody vegetation. This zone of high altitude dwarfshrubs was present in most of the Ecuadorian study sites. However, it often inhabited
a very narrow altitudinal range, and many of these communities were missed during
the stratified sampling procedure, where they occurred between the 100 m sampling
levels. However, a number of representative stands were recorded in this study.
Diplostephium rupestre and Loricaria spp. are consistent members of this vegetation type. The tussock and cushion pdramo with Diplostephium rupestre recorded at
4,200 m in the Pdramo de Guamani appears to be the Ecuadorian equivalent of the
high-altitude paramo dwarfshrub vegetation described by Cleef (1981) for Colombia.
However, tussocks of Calamagrostis, which were largely absent from Cleef's community, are dominant here.
One hundred metres higher in the Paramo de Guamani, the tussocks of Calamagrostis are absent, and in physiognomic terms the vegetation matches Cleef's description more closely.
A third high altitude dwarfshrub community was recorded on Volcan Chiles at
4,100 m and 4,200 m. The description corresponds roughly with the community described by Sturm & Rangel (1985) for the adjoining Colombian region of CumbalChiles. However, Loricaria cf. colombiana is absent from the Ecuadorian plots and
cushions of Distichia seem to be replaced by mats of A ciachne.
2. Community Ecology of the Ecuadorian Paramos
69
Tussock PSramo
The most extensive type of paramo vegetation was tussock grassland, almost exclusively dominated by Calamagrostis sp. [251]. In the majority of stands belonging to
this type, the tussocks covered over half of the surface area. Essentially, such vegetation consisted of a patchwork of raised tussocks and the spaces between them. Most
plants occupied the intertussock regions, though some species were more frequent
within the tussocks themselves.
There were altitudinal differences within the tussock paramos, and some regional
differences. The lower reaches of tussock grassland were represented in this study by
a number of communities. Calamagrostis sp. [251] Tussock Grassland with Oreomyrrhis andicola and Gnaphalium pensylvanicum was described for 20 stands in the
paramos of Guamani, El Altar (west) and Cajas. This community occurred between
3,400 to 4,100 m in areas with relatively humid conditions: on the eastern Cordillera
or just above the cloud forest treeline, where rainfall is high.
In central and southern pâramos, where conditions were less humid, the Calamagrostis sp. [251] tussocks were accompanied by a slightly different association of
species. Stands belonging to the Calamagrostis sp. [251] Tussock Grassland with
V iola humboldtii community were representative of such vegetation, and were located in the paramos of Cajas and Daldal between 3,500 and 3,900 m.
On Volcan Cotacachi in the north, from 3,600 to 3,900 m, the lower tussock grass
'Aram° was represented by Calamagrostis sp. [251] Tussock Grassland with Hypochaeris sonchoides, Plantago linearis and Relbunium croceum. Located on the western
Cordillera, this community received relatively little rainfall. Bare ground covered approximately 10% of the surface area on average, and reached as high as 20% in some
stands. It was similar to the paramo vegetation surrounding Laguna Mojanda described by 011gaard & Balslev (1979, Location 35).
In the far south of the country, in the paramos of Ona, Cumbe, Zapote Naida and
Cajas, a widespread community of lower tussock grassland was observed. This community of Calamagrostis sp. [251] Tussock Grassland with Paspalum tuberosum and
Chrysactinium acaule was the least humid of all the lower tussock paramo representatives. The vegetation was shorter in stature than that of more humid areas and
tussock cover was more variable, though usually between 25% and 75% of the total
area.
The TwiNsPAN classification separated those tussock grass communities just described from those higher up the altitudinal gradient. One community occupied an intermediate position between upper and lower tussock paramos on El Altar (west)
and in Daldal: the Calamagrostis sp. [251] Tussock Grassland with Sisyrinchium
jamesonia community was described between 3,800 and 4,200 m.
The upper tussock grass communities were less extensive and this was reflected in
the smaller number of stands of this type which were recorded. In the north, on Volcan Cotacachi at an altitude of 4,000 m, this type of vegetation was represented by
the Calamagrostis sp. [251] and Cukitium ovatum Humid Tussock Grassland corn-
2. Community Ecology of the Ecuadorian Rjramos
70
munity, and at 4,100 m by the broadly similar Calamagrostis sp. [251] and Festuca sp.
[255] Tussock Grassland. The latter community differed mainly by virtue of the codominance of Festuca sp. [255] tussocks.
At 4,000 and 4,100 m on Volcan Cotacachi, above the lower tussock paramo described earlier, there exists a more humid grassland: Calamagrostis sp. [251] Tussock
Grassland with Hypochaeris sonchoides, Halenia sp, [189] and Satureja nubigena. As
with the lower community, bare ground averages about 10% of the surface area. This
vegetation was very similar in composition to that described by 011gaard & Balslev
(1979) for the Paramo de Guamani at 4,000-4,100 m (Location 77).
Finally, three communities in Cajas represent variants of southern upper tussock
pdramos. These were the Calamagrostis sp. [251] and Poa sp. [262] Tussock Grassland (3,800-4,000 m), the Calamagrostis sp. [251] and A grostis sp. [243] Tussock Grassland (3,800 m), and the Calamagrostis sp. [251] and Rhynchospora macrochaeta
Tussock Grassland (3,900 m in Cajas, plus one stand from 4,100 m on Volcdn Chimborazo). All three communities were broadly similar in composition, but varied sufficiently to be separated by TWINSPAN.
Bamboo Pâramo
On the eastern slopes of the eastern range of the Ecuadorian Andes, with very
humid conditions and over 5 m of rainfall per year, places which would be dominated
by Calamagrostis sp. [251] tussocks elsewhere in the country are instead covered by
tussocks of bamboo grass. This type of vegetation was only encountered in one study
site, between 3,800 and 4,000 m on the eastern slopes of El Altar. Thus, only one
community of bamboo paramo has been described (Neurolepis elata Bamboo Pâramo). However, similar vegetation was observed in many other sites on the outer
slopes of the eastern Cordillera (Bromley, 1971; Black, 1982). It would appear, therefore, that this community occupies an equivalent niche to the bamboo-bunchgrass
paramo (the community of Chusquea and Oreobolus obtusangulus ssp. rubrovaginitus)
described by Cleef (1981) for the Colombian Cordillera Oriental.
It has been suggested that the presence of Neurolepis elata indicates a lack of disturbance (Bromley, 1971). While this may be true in some cases, an undisturbed
habitat does not appear to be a prerequisite for the presence of Neurolepis, evidenced by the relatively high burning scores for the samples of this type in the paramo of eastern El Altar. However, human habitation and disturbance is much lower
in these paramos, most likely the consequence of high rainfall and unpleasant living
conditions. This same rainfall and humidity is undoubtedly a major factor in determining the success of Neurolepis. In the same way, bamboo species in Colombian
paramos are confined to similar areas, and their presence has also been linked with
high annual precipitation (Gradstein, Cleef & Fulford, 1977; Cleef, 1978) and with
the higher night temperatures associated with permanent atmospheric humidity
(Cleef, 1981).
2. Community Ecology of the Ecuadorian Pâramos
71
Espeletia P6ramo
In the northern part of the country, and also in one isolated region in central Ecuador, the Calamagrostis sp. [251] tussock vegetation is replaced by paramo communities similar to those of Colombia. They are dominated by Espeletia pycnophylla
rosette plants in conjunction with Calamagrostis sp. [251] tussocks. In this study, vegetation samples from Volcan Chiles provided quantitative descriptions of such communities. This area is connected directly to the expansive paramo of El Angel, which
is perhaps the most famous area of Espeletia Paramo in Ecuador.
The lower Espeletia Pdramo community on Volcdn Chiles occurred at 3,600 and
3,700 m (Calamagrostis sp. [251] and Espeletia pycnophylla Tussock Grassland with
Paspalum tuberosum). It corresponds well with the community of Calamagrostis effusa and Espeletia cf. pycnophylla described by Sturm & Rangel (1985) for the CumbalChiles region. Calamagrostis and Espeletia are codominant, with Blechnum loxense
and Paspalum bonplandianum key elements of the flora. Sturm & Rangel (1985) report that this community was present on "very humid sites" and associated with the
greater soil moisture and the protection offered by hollows. Franco, Rangel & Lozano (1986) described a similar community (Castratello-Calamagrostietum effusae) in
the Colombian Cordillera Oriental and linked its distribution to soil moisture conditions. It is also similar in many respects to the Calamagrostis effusa and Espeletia
hartwegiana ssp. centroandina community described for two neighbouring Colombian paramos by Rangel & Franco (1985) and Rangel & Lozano (1986).
Miller & Silander (1991) report a community from 3,415 m in the nearby Paramo
del Angel dominated by tussocks and giant rosettes. Espeletia hartwegiana (=E.
pycnophylla) covered approximately 15% of the area, whereas Puya hamata was responsible for roughly 30% cover. Tussocks of Calamagrostis intermedium (c. 30%)
and Carex pichinchensis (c. 15%) were codominants. This community appears similar
in many ways to that just described for Volcdn Chiles and may represent a lower altitude equivalent of the Espeletia and Calamagrostis tussock grassland. It is notable,
however, that Paspalum is absent.
At 3,800 and 3,900 m on Volcdn Chiles, a community of Calamagrostis sp. [251]
and Espeletia pycnophylla Tussock Grassland with V iola sp. [192] was described. Tussocks of A grostis sp. [240] displaced those of Calamagrostis sp. [251] at the highest altitude. This vegetation may correspond in some way with Sturm & Rangel's (1985)
Loricaria cf. colombiana and A grostis foliata community. They describe a shift at
higher altitudes: a decrease in Calamagrostis and Espeletia in favour of Loricaria and
A grostis. Although, Loricaria was not present in the Ecuadorian plots, there was certainly a change from Calamagrosti.s. to A grostis with increasing altitude.
Cushion Paramo
At the highest elevations, tussock vegetation gives way to cushion paramo. Often,
the transition is very rapid, occurring over very short distances. The physiognomic effect of replacing tussocks with cushions results in a much more open vegetation, with
increased species diversity. The cushion paramo communities, by virtue of their
residence on the tops of mountains, have a rather localised and isolated distribution.
2. Community Ecology of the Ecuadorian Paramos
72
The stratified sampling technique employed in this study was prone to overlooking
some examples of this vegetation. However, six representative communities were described.
In most regions, Calamagrostis sp. [251] tussocks begin to decrease in cover between 3,950 and 4,000 m, and their dominant position in the vegetation is assumed
by Plantago rigida cushions. At 4,200 m on the western slopes of El Altar, three
stands were described which were typical of the vegetation which develops above the
tussock paramos. In this Plantago rigida Cushion Paramo, the dominant species' cushions covered 50-75% of the surface area. Some Calamagrostis sp. [251] tussocks
persisted, with a cover of 25-50%. This type of vegetation was reported from Volcan
Pichincha, Ecuador, by Benoist (1935).
At the same altitude on the more humid eastern slopes of the same mountain,
Plantago rigida cushions were co-dominant with mats ofA ciachne flagelhfera and
V aleriana aretioides. One hundred metres lower on this slope and Plantago rigida was
completely absent. In its place, were cushions and mats of A zorella corymbosa, Distichia muscoides and W erneria humilis.
At 4,300 m, vegetation cover was thinner. Although cushion plants were present
(W errzeria humilis, A zorella corymbosa) they were not dominant. In fact, no species
was able to assume a dominant position in the community.
On Volcan Cotacachi at 4,200 m, cushions of W errzeria humilis and Plantago rigida
were co-dominant with Lycopodium sp. [289]. A number of other cushion forming
species were also present (A ciachne flagelhfera, V aleriana aretioides) along with some
dwarf shrubs (Diplostephium rupestre and Loricaria sp. [334]).
At 4,400 m in the Paramo de Guamani a similar cushion paramo was sampled. Werneria humilis was again a significant element in the community, though this time codominant with A zorella corymbosa.
Although Harling (1979) noted the widespread existence of cushion paramos in
Ecuador, there are few descriptions of them in the literature. In Colombia and Venezuela, cushion communities occur almost exclusively in wet azonal bogs and mires,
and therefore Cleef (1981) did not find the type of zonal communities described
here. 011gaard & Balslev (1979) describe several cushion communities in general
terms. Communities close to the W erneria humilis & Plantago rigida Cushion Paramo
described here were found at Locations 15, 72, 78 and 79.
Rainshadow Desert Paramo
Two forms of desert paramo were present in the study sites: high altitude desert
paramo (the result of consistently low temperatures) and rainshadow desert paramo
(occupying a wide altitudinal range in certain areas). General classifications of Ecuadorian paramo vegetation have sometimes confused these two kinds of desert (for
example, Harling, 1979), though the differences are pronounced and are reflected in
the floristic composition of the vegetation.
2. Community Ecology of the Ecuadorian Pâramos
73
A regional desert paramo was sampled on Volcan Chimborazo. The 'Grande Arenal' or Great Beach of Chimborazo was described by Acosta-Solis (1985) as an
example of puna vegetation in Ecuador, very different from the paramo vegetation
elsewhere in the country, but comparable to the puna of Peru, Bolivia and Argentina.
This was confirmed by the analyses in this study: the stands from Chimborazo were
consistently separated from the remainder.
Below 4,000 m on Volcan Chimborazo, the vegetation was dominated by Stipa sp.
[253] and Senecio teretifolius with Plantago sp. [301]. The presence of Stipa tussocks
symbolises the link between the vegetation of the Arenal and that of the puna, where
Stipa ichu is predominant.
The arid nature of the north-western part of Volcan Chimborazo can be explained
by two factors. Firstly, the area is in rainshadow, both from the volcano itself, and
also from the eastern range of the Andes upon which most of the rain falls. Secondly,
the sandy substrate found here does not support surface waters. Acosta-Solis (1985)
reports that meltwater from the snow-cap of the mountain flows below ground only
to emerge at an altitude of around 4,000 m.
There are a number of other sites in Ecuador which are subject to similar conditions. Acosta-Solis (1984, 1985) gives a number of examples including the slopes of
Volcan Carihuayrazo and Volcan Iliniza, and the paramos of Palmira and Moyocancha. By virtue of the free-draining substrate, a number of other regions may support
similar vegetation locally. For example, Miller & Silander (1991) describe communities dominated by Stipa ichu tussocks on Volcan Cotopaxi, though they were
more diverse than those described here.
High Altitude Desert Paramo
At the limits of plant survival at high altitudes, a sparse vegetation cover of plants
may exist amongst the rocks and scree which provide shelter from the harsh conditions. This is equivalent to the super-paramo defined by Cuatrecasas (1954, 1958,
1968). The super-paramo reaches its greatest extent the south-west of Ecuador
where it is driest (Furrer & Graf, 1978).
The vegetation of the upper reaches of Volcan Chimborazo provided a good
example of this kind of desert paramo, though the rainshadow effects mentioned
above may have contributed to its formation. The majority of the region is dominated by three species: Chuquiraga jussieui, Calamagrostis sp. [251] and Geranium
sp. [157]. One of 011gaard & Balslev's (1979) collecting locations was in a similar
area on Volcan Cotopaxi (Location 74), with Chuquiraga jussieui a conspicuous element of the vegetation, but lacking Calamagrostis tussocks and Geranium cushions.
At the highest altitudes, the vegetation was heavily dependent upon sheltered sites
amongst the rocky substrate. Pfitsch (1988) demonstrated that significant thermal advantages (and thus benefits to survival) were experienced beside large rocks. On
Chimborazo, this resulted in a very variable floristic composition, which is reflected
by the complex community patterns produced by TWINSPAN at these altitudes. All of
2. Community Ecology of the Ecuadorian P gramos
74
the communities were dominated by Chuquiraga jussieui, Calamagrostis sp. [251] and
cushions of Geranium sp. [157], with an associated range of other species.
On the humid slopes of Volcan Tungurahua, another desert paramo community
was recorded at altitudes of 4,000 m and above. Tussock grasses were absent, the
vegetation consisting of a low carpet of plants. Many of the species in this community
may also be found in the cushion paramos which have already been described.
The distributions of the pdramo communities which have been described above
are determined by a number of environmental factors. Many of these factors are interdependent, and altitude, in particular, proved to be correlated with almost all of
the other environmental variables measured in this study. It was also strongly related
to species and community distributions. Thus, altitude provides a useful overview of
many of the elements which control the presence and extent of paramo vegetation.
Baruch (1984), in a study of Venezuelan paramo vegetation, also found that altitude
was the principal component of his first ordination axis, and that most of the environmental parameters he measured were statistically related to it.
Cleef (1981) proposed that the zonation found in the Colombian Cordillera Oriental was related to the incidence and frequency of sub-zero temperatures. The mechanisms involved may be related to tissue damage and early morning water stress,
illustrated by studies such as that by Goldstein & Meinzer (1983) and Goldstein,
Meinzer & Monasterio (1985). Unfortunately, it was not possible to collect representative climatic information during the field visits and no data was available from
other sources.
Although the altitude of an area would indicate the general level of minimum temperatures, they are buffered by humidity. Paramos in humid, cloudy regions tend to
be subject to fewer, less intense frosts than their counterparts in drier areas (Sarmiento, 1986). In a similar way, soil moisture may buffer temperatures close to the
ground and around the roots. Thus, the drainage and water holding capacities of
paramo soils may also be relevant to plant distribution.
It is not clear whether plant communities develop in response to soil pH, or vice
versa. For example, the paramo soils of Sumapaz in Colombia are acidic despite overlying limestone (Fosberg, 1944). However, certain patterns of soil pH are evident
from this study, despite the loss of some samples. Whilst cushion and tussock Oramos occupy areas of relatively average pH (around 4.5-5.5), Espeletia paramos appear to inhabit more acidic areas (pH of about 4.5) and desert paramos are found
largely on soils with higher pH values (5.5). Cleef (1981) also linked soil depth to
plant distribution in the Colombian paramo. This is particularly important at higher
altitudes where soil formation is slow.
In high altitude paramos, vegetation was restricted to sheltered sites next to rocks
or other plants. Exposure was important in determining the presence or absence of
species throughout the paramo vegetation studied. Smith (1978) clearly demonstrated that Polylepis sericea seedlings were unable to establish in open paramo in
2. Community Ecology of the Ecuadorian Paramos
75
Venezuela, and many other species may be similarly restricted in the Ecuadorian
Andes.
Overall disturbance was rated for each sample on a six-point scale. This index
proved a reliable estimator for three other indices used in this study: burning, trampling and grazing. These three variables are strongly linked with each other. Usually,
paramo areas are burned because a farmer wishes to improve the nutritional value of
the land for livestock. Implicitly, the same area must be subject to grazing pressure.
Furthermore, cattle and horses with freedom to roam the paramo appear to favour
recently burned areas (personal observation; Verweij & Kok, 1992). This will concentrate grazing activity in burned areas. Of course, wherever there is grazing pressure,
there must also be trampling. On steep slopes throughout the Ecuadorian Andes, the
effects of trampling can be clearly seen: the hillsides are minutely terraced . along the
contours by the movements of livestock. It might appear from the aforementioned
that agricultural use of the paramos is intensive. In fact, the reverse is usually the
case, cattle and horses are grazed over wide areas and at low density. However, in an
environment such as the paramo, recovery from disturbance is slow. This leads to
widespread indications of disturbance, though the rate of disturbance may be low.
Schmidt & Verweij (1992) observed that cattle on the paramo graze for long hours
and over large areas to meet their nutritional requirements. It was estimated that
cattle walked an extra 5 km per day, with an estimated ascent of 50 m, when compared to the foraging behaviour of lowland animals.
Disturbance (both natural and human) was considered to be an important factor
determining the distribution of Colombian and Venezuelan 'Aram° species by Lozano & Schnetter (1976) and Baruch (1984), respectively. In Ecuador, disturbance was
clearly correlated to species and community distributions. This will be discussed further in the final chapter.
Species diversity was found to rise with increasing altitude. Baruch (1984) observed the same phenomenon in a Venezuela paramo. This follows general observations that diversity tends to increase in places with relatively high environmental
stress (Peet, 1978). In such places, stress tolerance is the main strategy and competitive exclusion is less influential on community composition (Grime, 1979). In the
Ecuadorian paramos, the diminishing dominance of Calamagrostis tussocks and its
eventual disappearance at the highest altitudes results in a much more even distribution of species abundance, which increases diversity. The cushions, which tend to
dominate the higher altitudinal zones, are less able to exclude other species which
grow upon them.
This study has described a number of paramo communities throughout Ecuador
and covering a wide altitudinal range. The distributions of the plant species making
up these communities were statistically related to environmental variables such as altitude, exposure and disturbance. Other factors seem to be involved which were not
measured, like minimum temperatures and humidity. These factors have been linked
to plant distributions in the paramos of Venezuela (Faritias & Monasterio, 1980; Baruch, 1984) and Colombia (Lozano & Schnetter, 1976; Cleef, 1981).
2. Community Ecology of the Ecuadorian Pâramos
76
Three areas of research, in particular, warrant further study. First of all, more
quantitative descriptions of paramo areas of Ecuador are needed, to determine the
representativeness of the current study and to give a better picture of variability in
composition. Paramos to the south and east were under-represented in this study
and the understanding of Ecuadorian mountain vegetation would benefit from further attention to these regions.
Secondly, the mechanisms which limit plant species distribution require attention.
This requires detailed study over a long period of time so that due consideration is
given to climatic factors. A useful starting point for such research would be the transition zone between tussock and cushion paramo found so often at around 4,000 in,
which is described in more detail in Chapter 4. Here, there are pronounced changes
in physiognomy over a very short distance (both laterally and altitudinally) and an opportunity to conduct a detailed study of the processes involved.
Finally, a third avenue of study concerns the distribution and composition of azonal vegetation in the Ecuadorian paramos. Bogs, woodlands and thickets are frequently discovered amidst the more extensive zonal pâramo dealt with here.
Consideration of the adaptations evolved by plants in response to such environments
may reward investigation.
Chapter 3
Plant Form In The
Ecuadorian Paramos
77
3. Plant Form in the Ecuadorian Paramos
78
I. A Growth Form Classification for
the Ecuadorian Nramo
The form of plants has long been recognized as an important descriptive feature of
both individual plants themselves and of the vegetation they comprise. Terms such
as tree and herb for plants, and forest and grassland for plant communities are in
general use and relate to particular forms of plant life. There are other, more specialized terms like xerophyte, which suggest particular adaptations to specific environmental conditions.
Not surprisingly, there exists a large body of literature on plant forms with varied
applications. Throughout, there has been confusion over the use of the terms growth
form and life form. In this work, the definitions of Clements (1920) in Schulze (1982)
have been adopted — the growth form is the direct, quantifiable response made by a
plant to different habitats and conditions, while the life form is a morphological feature of a species which is insensitive to environmental changes. This distinction is a
useful one, and any temptation to use the terms synonymously should be resisted.
However, it is difficult to establish which of these terms applies to a particular feature without first subjecting the plant to artificial environments. It is often impractical to determine whether particular plant forms are fixed regardless of the
environment, and for this reason all plant forms will be referred to here as growth
forms. Future research may yield the information necessary to distinguish between
those features that have a plastic response to the environment and those which do
not.
Von Humboldt (1806) offered the first widely recognized classification of 'Hauptformen' (physiognomic types) following his travels in the Andes. His system grouped
plants according to physiognomic similarities, rather than by conventional taxonomic
comparison. He described nineteen distinctive plant forms, named and characterized
by a genus or family in which that form was clearly represented (for example, palms,
banana, lianas, lily, fern, grass). Later in the century, other authors advocated similar
physiognomic systems of plant classification. For example, Grisebach (1872) described sixty vegetative forms and attempted to show their correlation with the climate in which they were found.
Warming (1884), Schimper (1898) and others, stressing the ecological significance
of plant form, classified plants partly by an assessment of their response to water supply and transpiration.
A simpler system was proposed by Raunkiaer (1907, 1908, 1934): his life form divisions were based on the position of the buds or organs from which new shoots or foliage developed after an unfavourable season. Plant behaviour during the growing
season was largely ignored. Ellenberg & Miiller-Dombois (1967) found it necessary
to modify the system to include more emphasis on structure and seasonality of the
crown, foliage and shoot systems. Despite difficulties, the Raunkiaer system and its
3. Plant Form in the Ecuadorian Paramos
79
derivatives have enjoyed wide usage for comparing different vegetation types and
their relationship to the environment.
Criticism of the Raunkiaer approach led to the development of other systems. Du
Rietz (1931) concluded that it was impossible to formulate one, all-encompassing
classification; rather, more was to be gained from the use of several parallel systems,
emphasizing different features. He proposed six classifications including main life
forms (based upon the general physiognomy of the plant in the growing season), bud
height life forms (as in the Raunkiaer system) and leaf life forms (based upon the
character of the leaves).
Most of the classification approaches have been devised in the temperate zone,
with only a secondary incorporation of tropical regions. Hedberg (1964) points out
the hopeless task of creating a system to incorporate all plants on earth for all climates. It is not surprising, therefore, that the applicability of general-use systems for
classifying growth forms in the peculiar environments of the tropical alpine regions is
limited. Many problems arise from the lack of climatic seasonality, which results in
the absence of a growing season (and of "resting buds"). The "unfavourable season"
required by a host of classifications occurs every night, invalidating such systems. In
addition, it is impossible to distinguish annuals from perennials, monocarps from
polycarps, or half-shrubs from herbs. Many plants are frutescent, but the younger
shoots are herbaceous, only acquiring woodiness with age (irrespective of the time of
year).
For these reasons, Hedberg (1964) found it necessary to establish his own system
for grouping plant forms in the Afroalpine environment. His system consisted of five
classes: giant rosette plants, tussock grasses, acaulescent rosette plants, cushion
plants and sclerophyllous shrubs. Some 45% of the flora fitted these classes, the rest
being ignored (showing "less conspicuous adaptations to this environment"). Vareschi (1970) referred to nine `biotypes' of the Venezuelan Andes. Stem rosettes,
cushion plants, 'trellis' plants, dense bunch plants, plants with clustered flowers,
dwarf shrubs, rosette plants, geophytes and therophytes comprised his list. Troll
(1975 — cited by Lauer, 1979) associated nine life forms with the paramo: paramo
grasses, stem rosettes, basal rosettes, evergreen shrubs with dense scales or involuted
leaves, macrophyllous evergreen shrubs, shrubs with pubescent leaves, cushion
plants, dwarf rosette plants and dwarf semi-woody shrubs forming cushions.
Hedberg (1964) and Hedberg & Hedberg (1979) used their five classes of growth
form to examine the adaptive significance of these plants to the environment. However, it is also of interest to compare regions on the basis of growth form and to describe communities by their growth form composition. For these to be accomplished,
it is desirable to attempt to account for the 55% of the flora left out of Hedberg's system, though Mena & Balslev (1986) compared the paramo of El Angel with Afroalpine vegetation using only Hedberg's five growth forms. The categories proposed by
Vareschi (1970) and Troll (1975) go some way to accounting for the omitted flora,
though their classifications do produce further problems of undue complication, particularly with regard to shrubs.
3. Plant Form in the Ecuadorian Paramos
80
Perhaps ideally, a classification system should indicate relationships between the
various groups in a hierarchical manner. However, in practice, this is very hard to
achieve with growth forms. Von Humboldt (1806) recognised that life forms are "by
their nature not capable of strict classification". Warming (1909) considered it "an intricate task to arrange the life-forms of plants in a genetic system, because they exhibit an overwhelming diversity of forms, ... also because it is difficult to discover
guiding principles that are really natural". Only small modifications to a growth pattern are required to change the growth form (the fact that certain taxa may have representatives in a number of growth form categories is evidence for this). A tussock
can thus be viewed as an exploded cushion (Hodge, 1946), which in turn may be seen
as a contracted shrub and so on. Therefore, an hierarchical classification has not
been attempted since the end groups are all that is required.
Tansley & Chipp (1926) state that "the independent student of evolution will do
well, however, to make his own classification of life-forms of the communities he actually studies". It is my intention to add to Hedberg's system in an attempt to include
a larger portion of the flora than the five classes currently accommodate, and to
apply it to the pâramo flora of Ecuador.
Hedberg's classes are dealt with first, with examples from the Ecuadorian paramo
flora and further afield. Then, additional types of growth form are defined and examples given.
Stem Rosettes
(Figure 3.1)
Hedberg's classification (1964) included both stem rosettes and basal rosettes in a
group termed 'Giant Rosettes'. These forms are differentiated in the present system.
Stem Rosettes are characterized by thick and unbranched stems covered by dense,
dry leaves that remain attached to the plant when they die (marcescent). The single
aerial meristem produces lateral inflorescences. It is polycarpic and growth is not
determinate (that is, flowering does not halt the development of the axis). Cuatrecasas (1979) terms this growth form as a polycarpic, more or less tall, monocaul caulirosula. Vareschi (1970) and Troll (1975) both include this form in their systems.
A widely used concept of plant form is the architectural model (Halle & Oldeman,
1970): an inherent growth strategy which defines both the manner in which a plant
elaborates its form and the resulting architecture (Barthelemy, Edelin & Halle,
1989). The architectural model of any plant is based on observations of the type of
growth, branching pattern, morphological differentiation of axes and the position of
sexuality (Barthelemy et al., 1989). Therefore, the architectural model is entirely independent of taxonomic boundaries between plants. It may express both the phenotypic plasticity of plants (including branch ageing, die-back, release of dormant
meristems and re-iteration) and the genetic control of metameric growth and iteration (Halle, Oldeman & Tomlinson, 1978).
3.
Plant Form in the Ecuadorian Paramos
81
.
111;0
4
fi
P
a1
1>1;44•;:*•.„
g
•
.'-‘5T:!•.''
• r^-,r
•
;
Figure 3.1.
The stem rosette growth form (for example, Espeletia pycnophylla).
The stem rosette form corresponds, in a wide sense, to Corner's model of tree
architecture (Halle et al., 1978). However, there are differences — differentiated reproductive branches, spiral phyllotaxis in the stem but decussate phyllotaxis in the inflorescence —which strictly requires another architectural model (Cuatrecasas,
1986). In the Raunkiaer system (Ellenberg & Miiller-Dombois, 1967) these plants
would be classified as phanerophytes.
It is exemplified by Espeletia pycnophylla ssp. angelensis in the paramos of northern Ecuador (illustrated in Figure 3.1). Tree ferns also conform to this model, and so
Blechnum loxense has been incorporated into this class.
Elsewhere in the northern Andes, Espeletia species are a more important element
of the paramos (Cuatrecasas, 1979). Monasterio (1986) estimated that Espeletia timotensis in the Venezuelan Andes has a life-span of 170 years. In Venezuela, Planta-
3. Plant Form in the Ecuadorian Paramos
82
go penymondii and Bulbostylis sp. (Cyperaceae) also exhibit this form (Vareschi,
1970). In the puna vegetation of the Southern Andes, stem rosettes are not present.
On a wider scale, stem rosettes are common in the afroalpine region: Senecio, Dendrosenecio and Carduus species (Mabberley, 1986). Cyathea, Dicksonia and Cibotium
tree ferns occur in subalpine habitats in the grasslands of Malaysian mountains (Van
Royen, 1967).
Basal Rosettes
(Figure 3.2)
Holttum's model (Halle et al., 1978) is defined as a plant with a unique axis provided
by a single aerial apical meristem which always remains unbranched. After a phase of
Figure 3.2.
The basal rosette growth form (for example, Puya hamata).
3. Plant Form in the Ecuadorian Paramos
83
stem building, the terminal meristem differentiates completely into an inflorescence.
It is, therefore, monocarpic. Sometimes they are referred to as 'candle plants'. These
plants are rosulate phanerophytes according to the modified Raunkiaer system of Ellenberg & Miiller-Dombois (1967).
Hedberg (1964) failed to differentiate this group from the stem rosettes, though
Troll (1975) did make the distinction. The basal rosette growth form is more commonly represented in the Ecuadorian paramos than the stem rosettes. Several
species of Puya (for example, Phamata, shown in Figure 3.2) conform to this growth
model. Several other species in Ecuador have a similar overall appearance, and belong to this category but, unlike Puya, are neither woody nor monocarpic: Rumex tolimensis, V aleriana plantaginea, Lupinus alopecuroides and Culcitium sp. A number of
fern species can also demonstrate this form, for example, Thelypteris sp.
Basal rosettes are present, but are not as important in Colombia and Venezuela—
representatives of this growth form, such as Draba spp. and Senecio spp., were reported in several studies of the Colombian paramos (for example, Sturm & Rangel,
1985); and Vareschi (1970) fails to mention them at all in his study of Venezuelan
paramo vegetation. In the puna of Pert, however, they are much more conspicuous
(for example, the impressive Puya raimondii). In East Africa, Lobelia spp. (Campanulaceae) are very significant basal rosettes. Rheum nobile (Polygonaceae), Lobelia
(Campanulaceae), Eremurus himalaicus (Liliaceae), A nrebia (Boraginaceae) and
Saussurea (Compositae) in the Himalayas, Silverswords (A rgyroxiphium spp., Haleakala spp. —Compositae) in Hawaii and Echium spp. (Boraginaceae) in the Canary Islands are all examples of 'candle plants' in mountain areas.
Tussock Plants
(Figure 3.3)
In tussock or bunch grasses, erect tillers are produced from tightly packed culms at
the soil surface (though often growth of the tussock raises this region of dense culmbases above ground level). Dead leaves are retained and decay while still attached to
the plant. The scleromorphic leaves tend to be filiform, either tightly folded or inrolled. This growth form would classify as a caespitose hernicryptophyte in Raunkiaer's system. It is often difficult to distinguish between true tussock-formers and
those plants which are only loosely tufted (Hedberg, 1964).
In the Ecuadorian paramos, tussock grasses most commonly belong to the genera
Calamagrostis, Cortaderia, Festuca and Stipa. Non-graminoids also belong to this
group, in particular a number of sedges (Carex, Rhynchospora, Uncinia) and Sisyrinchium spp.
Tussock grasses are common elsewhere in the pdramos of the northern Andes,
though their prominence declines northwards as the giant rosettes become increasingly significant. Vareschi (1970) notes that bunch grasses (particularly A grostis haenkeana and Helleria fragilis) occur at the highest altitudes in Venezuela. Yet further
north, in the Mexican zacatal, large stands of tussock grasses occur, the genera Briza,
Bromus, Calamagrostis, Festuca, Muhlenbergia and Stipa well-represented (Breed-
3. Plant Form in the Ecuadorian Paramos
84
Figure 3.3.
The tussock growth form (for example, Calamagrostis effuse).
love, 1973). In the puna of the Altiplano, tussock grasses (Stipa, Calamagrostis and
Festuca) represent an important growth form element (Ruthsatz, 1977; Seibert,
1983). In the afroalpine zone of East Africa tussock grasses are equally significant in
community physiognomy. Hedberg (1964) cites Festuca, Poa, A grostis, Pentaschistils
and Carex as the main tussock-forming genera. Van Royen (1967) cites the main tussock grasses of New Guinea as Danthonia, A grostis, A nthoxanthum and Festuca. This
growth form is also highly characteristic of Sub-Antarctic regions, exemplified by Poa
and Stipa (Sewell, 1954), and Chionochloa, Poa, Festuca and Notodanthonia for New
Zealand (Mark & Adams, 1973).
Acaulescent Rosettes
(Figure 3.4)
The basal rosette of these forms is initiated at or below ground level, the leaves being
attached at virtually the same level. An overground stem is absent (though in some
members the flowers are not sessile, but borne on more or less leafless flowering
stems). These plants commonly possess a large tap root. They are usually small (up to
30cm across, but generally smaller), though some of the larger species may be viewed
as small versions of the giant basal rosette form. Rosette or semi-rosette hemicryptophyte approximates to this growth form in the Raunkiaer system (Ellenberg & Maller-Dombois, 1967).
85
3. Plant Form in the Ecuadorian Paramos
•••nI
Figure 3.4.
The acaulescent rosette growth form, for example, Eryngium humile (left) and Hypochaeris sessiliflora (right).
The pâramo flora of Ecuador contains many representatives of this growth form.
The commonest include Hypochaeris sessiliflora, Oritrophium peruvianum, Hieracium
frigidum, Senecio pimpinellifolius, Senecio repens and Plantago major W erneria nubigena is a variation on the rosette form, the leaves stacked on top of each other. Luzula
spp. and a number of sedges are included here. Less well-defined acaulescent rosettes are Lachemilla hispidula (and other similar members of the same genus), Oreomyrrhis andicola and Niphogeton dissecta. Isoetes is also placed in this category,
though perhaps it merits a group of its own.
Acaulescent rosettes are common elsewhere in the pdramo; for example, Rhizocephalum candollei Wedd. in the Venezuelan paramo (Hedberg & Hedberg, 1979).
Cleef (1978) includes A caena cylindristachya, V aleriana spp., Castratella spp. and Oritrophium spp. among the Colombian paramo acaulescent rosettes. The form is wellrepresented in the puna (Hedberg & Hedberg, 1979). Wilcox et al. (1986) list a
number of acaulescent rosette species in their bofedal communities in the highlands
of Central Perti. Cabrera (1968) lists a number of species occurring in the puna, including Chaptalia similis, Trechonaetes lanigera, Plantago monticola and Northoscordurn sessile.
In East Africa, there are several notable examples, including Ranunculus cryptanthus, Oreophyton falcatum, Haploccupha rueppellii and Carduus chamaecephalus
(Hedberg, 1964). In New Zealand, there are a large number of species belonging to
this growth form, notably A ciphylla congesta, Celmisia spp. (C.major var. brevis appears remarkably similar to the Andean Oritrophium peruvianum — Solbrig (1960)
considered Oritrophium a section of the genus Celmisia but Cuatrecasas (1968)
treated them as separate genera) and Craspedia spp. (Mark & Adams, 1973). Himalayan rosette species (Polunin & Stainton, 1987) include Pycnoplinthopsis bhutanica
(Cruciferae) and A stragalus rhizanthus (Leguminosae).
86
3. Plant Form in the Ecuadorian Paramos
z
F
.7•44.
1,2;71/4;14
•
‘("
4C:1
:t
•
Figure 3.5.
The cushion and mat growth form, for example, Plantago rigida (right) and
Azorella pedunculata (left).
Cushion Plants
(Figure 3.5)
The term cushion is used here in its widest sense, to include soft mat-cushions and
hard, compact bolster plants.
The plants in this class show a variety of shapes, ranging from semi-spherical
through hummock to flat mat. Rauh (1939) distinguished a number of morphological
types of which rosette cushions, creeping cushions and ball cushions are commonest
in the Ecuadorian paramos. All have profusely branched stems with short internodes, the branches terminated by imbricate leaves in more or less evident rosettes,
forming a dense layer covering the peaty interior (formed by the decay of the remains of leaves, branches and roots). As the older branches die and decay, the
younger parts become isolated from the original plant (Heilborn, 1926; Hedberg,
1964). It is difficult to establish whether cushions or mats are composed of one individual or more; the complete fusion of cushions of different species has been observed (Heilborn, 1926). In the Raunkiaer system, cushion plants would be classified
as semi-woody dwarf shrubs, more specifically as "suffrescent" pulvinate chamaephytes. The classification of mat or hummock plants is less precise.
Cushion plants are very conspicuous in the Ecuadorian paramos, particularly
above 4,000m. Harling (1979) recognized an entire vegetation type as "cushion paramo". Notable examples of cushion-formers are Plantago rigida, V aleriana rigida, Nototriche jamesonii, Geranium sp., A zorella diapensioides, A rcytophyllum spp., W erneria
humilis, V iola sp. and Distichia muscoides. Mat-formers include Oreobolus obtusangulus, Distichia muscoides, Plantago rigida, A ciachne flagelltfera and A zorella
pedunculata.
3. Plant Form in the Ecuadorian Paramos
87
According to Cleef (1978) six taxa constitute common vegetation types in the Colombian paramos: Plantago rigida, Distichia muscoides, Distichia tolimensis, Oreobolus sp., A zorella multifida and A ciachne pulvinata. In the Venezuelan paramo around
Mucubaji, Hedberg & Hedberg (1979) classified A ciachne pulvinata, Paepalanthus
karstenii, Calandrinia acaulis, A renaria venezuelana f. caespitosa, A zorella crenata and
Plantago rigida as cushion plants.
The cushion growth form reaches its height of supremacy in the puna vegetation
where desert conditions produce the so-called puna mat vegetation (Hodge, 1946).
Notable puna species with a cushion or mat form include A zorella compacta, A desmia erinacea, A desmia patacana, Pycnophyllum spp., A nthobgum tetragonum and
Opuntia atacamensis (Cabrera, 1968). Hodge (1946, 1960) lists a number of cushion
plants for the Peruvian puna and emphasises the use of 'llareta' (A zorella spp.) for
fuel. Ruthsatz (1978) identified some thirty species of cushion plants in North-west
Argentina including several which occur in Ecuador (for example, Distichia muscoides, W emeria pygmaea). In Northern Chile, Quintanilla (1983) reports Pycnophyllum bryoides, A zorella sp. and A desmia sp., whilst Armesto, Arroyo & Villagran
(1980) studied cushions of Laretia acaulis, A zorella monantha and A zorella madreporica in Central Chile.
In East Africa only five species were found by Hedberg (1964) to belong to this
class: A grostis sclerophylla, Sagina afroalpina, Swertia subnivalis, Myosotis keniensis
and Haplocatpha ruppellii. In Malaysia cushions are also very rare (Van Steenis, 1935
p.346, 1939 p.448). Van Royen (1967) describes mats of Euga brasii and Oreobolus
sp. in Papua New Guinea. In the alpine zone of New Zealand mats and cushions are
a very conspicuous group, with Colobanthus canaliculatus (and other species),
Drapetes lyallii, Dracophyllum muscoides, Celmisia sessiliflora, Haastia pulvinaris,
Raouli a spp., Phyllachne spp. and Pygmaea spp. good examples (Mark & Adams,
1973; Godley, 1978). Reference to Polunin & Stainton (1987) shows that cushion and
mat plants are represented in the Himalayas, notably by Thylacospennum caespitosum (Caryophyllaceae), A renaria bryophylla and A .densissima (Caryophyllaceae),
Sarifraga saginoides and S.pulvinaria (Saxifragaceae) and A ndrosace delarayis and
A .tapete (Primulaceae).
Upright Shrubs
(Figure 3.6)
The sclerophyllous (tough-leaved) and dwarf shrubs of Hedberg's (1964) classificatory system are characterized by thin and distinctly woody branches with thin bark.
The leaves are rigid, more or less coriaceous (leathery), often small, folded or revolute. The leaf surface is sometimes covered by dense, white pubescence, often mixed
with gland hairs. The shrubs vary in height from 50cm to 2m or more. Not all shrubs
in Hedberg's classification were sclerophyllous and likewise, there are a number of
different leaf forms in the paramo species forming this group. Troll (1975) divided
upright shrubs into three classes: those with dense scales or involuted leaves (like
Loricaria), those with large leaves (such as Befaria) and shrubs with pubescent leaves
(for example, Helichrysum).
3. Plant Form in the Ecuadorian Paramos
88
Figure 3.6.
The upright shrub growth form (for example, V aleriana microphylla).
The group is continuous with suffrescent herbs and borderline cases are difficult
to distinguish. The absence of a resting season makes the classification of this growth
form according to the system of Raunkiaer impossible. Perhaps evergreen nanophanerophytes without bud protection is close, but so too is suffruticose/frutescent chamaephytes (Ellenberg & Miiller-Dombois, 1967). Sturm (1978) found this to be a
problem too —many of his dwarf shrubs were denoted as "chamaephyte or nanophanerophyte".
Dwarf shrubs occur quite frequently in the Ecuadorian paramos, particularly in its
lowest reaches. V aleriana, Gynoxys, Diplostephium, Pentacalia (= Senecio), Chuquiraga, Berberis, Hypericum, Gnaphalium, Lupinus, Loricaria, Calceolaria and Hesperomeles are all genera with representatives of this growth form.
Cleef (1981) referred to a dwarfshrub pdramo and a number of azonal dwarfshrub
comunities in the Colombian Cordillera Oreintal. Sclerophyllous shrubs constitute
3. Plant Form in the Ecuadorian Paramos
89
Figure 3.7.
The prostrate shrub growth form (for example, Pemettya prostrata).
one of the most frequent growth forms of the Venezuelan paramo, including Hesperomeles pernettyoides, Lachemilla verticillata, Polylepis sericea, Draba cf. funckiana, Hypericum laricifolium, V aleriana parvi flora and Baccharis prumfolia (Hedberg &
Hedberg, 1979).
In the remainder of paramo regions, dwarf shrubs are common in the lower portions, the subparamo of Cuatrecasas (1958, 1968). In the puna of the Central Andes,
tola (Parastrephia lepidophylla) plays a major role in the plant community, in conjunction with other composite shrubs (Seibert, 1983). Quintanilla (1983) describes Chuquiraga kuschelli and Chuquiraga spinosa for the Chilean puna. `Sclerophyllous'
shrubs are a common element of the afroalpine flora: according to Hedberg (1964)
about fifty taxa belong to this group, including Protea kilimandscharica, Thesium kilimandscharicum, A khemilla spp., A denocatpus mannii, Hypericum spp., all afroalpine
Ericaceae, Bartsia spp., Helichrysum spp., Senecio spp. and Eugops spp. J.M.B. Smith
(1975) notes the importance of shrubs in the tropicalpine zone of New Guinea (for
example, Hypericum). In New Zealand, a number of species belonging to the genera
Drapetes, Hebe and Helichrysum are similar in general appearance to Loricaria spp.
in the paramo. Some Hebe spp. are also like the paramo shrubs of V aleriana (Mark &
Adams, 1973).
Prostrate Shrubs
(Figure 3.7)
Woody plants which form a dense covering on the ground have already been dealt
with in the cushion and mat section. However, certain creeping dwarf shrubs have a
more open cover and these are the plants which belong to this class. Vareschi (1970)
termed them 'trellis' plants.
It is possible to view these prostrate shrubs as dwarf shrubs or as mats in a loose
sense. For example, when dealing with the tropicalpine species Eutya brasii, J.M.B.
3. Plant Form in the Ecuadorian Paramos
—I/
90
n Oit
4/ lot
0/
fr
l
7
- teel
Figure 3.8.
The erect herb growth form, for example, Bartsia laticrenata (left) and Jamesonia
alstonii (right).
Smith (1975) talks of a shrub, while Van Royen (1967) terms it a mat. These plants
would classify as frutescent chamaephytes under the Raunkiaer system.
Pemettya prostrata, Disterigma empetnfolium, Baccharis alpina and A rcytophyllum
aristatum are common representatives of this growth form in the Ecuadorian paramo. These and similar species occur elsewhere in the paramos of the northern Andes
(Cleef, 1978, 1981). Vareschi (1970) offers Pernettya prostrata, Cyrilla racemiflora,
Eugenia triquetra and Hesperomeles pemettyoides as Venezuelan representatives. A desmia horrida, A .atamensis and Nardophylum sp. are cited by Quintanilla (1983) as
prostrate shrubs of the Chilean puna. In New Zealand examples of prostrate shrubs
are found in the alpine zone, such as Pernettya nancz, Palpina, Gaultheria depressa
and Coprosma petriei (Mark & Adams, 1973). Salix cayculata (Salicaceae) is a good
example of a Himalayan prostrate shrub (Polunin & Stainton, 1987).
Erect Herbs
(Figure 3.8)
Since the 'herbaceous stems' grade continuous into small shrubs through varying degrees of woodiness, it is difficult to distinguish the larger members of this class from
the smaller ones of the 'dwarf shrubs' group.
In the Raunkiaer system, plants from this group can be included in the phanerophytes, particularly phanerophytic grasses (maybe lignified as in the bamboos) and
3. Plant Form in the Ecuadorian Pâramos
Figure 3.9.
The prostrate herb growth form (for example, Geranium multipartitum).
Figure 3.10.
The trailing Herb growth form (for example, Vicia setifolia).
91
3. Plant Form in the Ecuadorian Paramos
92
herbaceous phanerophytes. Several hemicryptophytes, which do not form tussocks,
are also included here. Some members may even be assigned to the geophytes. A
number of forms included here were termed 'plants with clustered flowers' by Vareschi (1970).
Common elements of the Ecuadorian paramo flora in this group include: Bartsia,
Castilleja, Draba, Jamesonia, Gentianella, Lobelia, Lycopodium, Bomarea, Lupinus
and a variety of genera from the Grarnineae; similarly for the paramo vegetation
elsewhere in the northern Andes.
This growth form is prominent in other high altitude tropical vegetation types,
along with most plant communities throughout the world.
Reptant and Prostrate Herbs
(Figure 3.9)
These plants lack erect, leafy stems and possess stolons or other means of spreading
vegetatively, along the soil surface or just underneath it. Reptant hemicryptophytes or
reptant herbaceous chamaephytes defined in Raunkiaer system belong to this growth
form and are common throughout the world.
Lachernilla, Geranium, Satureja, Bidens, Gentiana, Gentianella, Halenia and
Ranunculus are all extremely common genera with a reptant herb growth form in the
Ecuadorian paramos.
Trailing Herbs
(Figure 3.10)
With weak ascending stems, some with tendrils or minutely toothed stems aiding
their support among other plants, this growth form is relatively common in the grassy
paramo. The tussock grasses provide an ideal framework for these plants. The Raunkiaer system would classify Ecuadorian paramo climbers into various categories:
spreading, climbing, hemicryptophytic lianas (Galium); tendril-climbing, hemicryptophytic lianas (V icia); or spreading, climbing, geophytic lianas (Oxalis). Other climbers
include Stellaria (on tussocks) and Bomarea (on shrubs). Here they are all grouped
together.
Cryptogams
Clothing the surface of soil or rocks in a close mantle, mosses and lichens are particularly common in undisturbed humid paramos. Though not of the same character as
the mats of the earlier section, bryophyte mats can be the dominant growth form, especially at high altitudes on ashy soils. In the Raunkiaer system, these plants are
termed adnate thallophytes. This category includes leafy and thalloid cryptogams
(mosses, leafy and thallose liverworts, filmy ferns) and foliose, fruticose and thallose
3. Plant Form in the Ecuadorian Paramos
93
lichen mats. One lichen species, Thamnolia sp., is a solifluction floater (according to
the definition of Lind & Morrison, 1974), which merely lies, unattached to the substrate, on the surface of the ground.
Owing to the loss of voucher specimens for a number of sites, it was not possible
to include this growth form category in later analyses.
Other Growth Forms
are a number of species not strictly classifiable into the above growth forms,
T here
though dealt with by Raunkiaer's system. These include trees (for example,
Polylei* Gynoxys) which are occasionally found in the paramo, and several geophytes such as Stenomesson aurantiacum (Amaryllidaceae).
None of these plants was present in the following studies described in Sections II
and III of this chapter.
3. Plant Form in the Ecuadorian Paramos
94
..57.42:::4:•7M:WM:M57.40•AV.Q7
II. The Growth Form Composition of
the Ecuadorian Paramos
Introduction
In Chapter 2, a number of vegetation samples from the paramo vegetation of Ecuador were classified according to their species composition. Many of the resulting
communities were assigned to various physiognomic groups (for example, cushion
paramo, tussock paramo), with the same dominant growth forms but not necessarily
the same species present.
Given that all characters of an organism are likely to reflect the action of selection
(Fisher, 1930), differences in plant form can be thought of as visible indications of
niche differentiation. Thus, selection gives rise to organisms that are ecologically
equivalent, a process known as convergent evolution (Johnson, 1973). It is often the
case that where the dominant growth form is the same, there is a similar degree of
convergence in the subordinate units of vegetation (Mooney, 1974).
The vegetation of high tropical mountains has often been cited as support for the
concept of convergent evolution (for example, Hedberg, 1964; Troll, 1968; Hedberg
& Hedberg, 1979; Halloy, 1983). Such reports are based on qualitative observations
of physiognomic parallelism between, for example, the mountains of East Africa and
the Andes. To test the extent of convergent evolution, more quantitative approaches
are required. However, before different mountain systems are compared, it is necessary to determine the degree of variability of the vegetation within single mountain regions.
Mena & Balslev (1986) used Hedberg's (1964) growth form classification and applied it to three 10 x 10 m plots in the paramo of El Angel in northern Ecuador,
where they estimated the percentage cover of each of five growth forms. However,
the majority of their work related to a floristic comparison with the Afroalpine vegetation.
In this study, using ten of the eleven growth forms defined in the first section of
this chapter (bryophytes and lichens were not used here, because a large number of
voucher specimens were stolen in Ecuador), the growth form composition of Ecuadorian paramo vegetation is described in detail. Comparisons between the samples are
made and growth forms are related to environmental variables which may be responsible for variations in their distribution.
3. Plant Form in the Ecuadorian Paramos
95
Methods
In Chapter 2, the vascular plant composition of 192 pdramo quadrats was described. Using these same data, collected from twelve sites, each vascular plant
species was assigned to one of the ten growth form categories described in the previous section. Appendix 1 indicates the growth form category of each of the 348 taxa.
Where several species in a quadrat belonged to the same growth form category,
their Braun-Blanquet abundance scores were summed (by converting the individual
scores to their mid-point percentage cover, summing and converting the resultant
cover value back to a Braun-Blanquet score). In this way, the growth form composition of the 192 quadrats was determined.
Environmental data for each quadrat were also collected (described in Chapter 2).
Altitude, aspect and slope were measured, and exposure, burning intensity, grazing
intensity, trampling intensity and overall disturbance were estimated using subjective, semi-quantitative scales. The coverage (as a percentage) of bare ground and
rock cover (including scree) were also noted for each sample.
The 192 paramo stands and the ten growth forms were classified using the TWINSPAN algorithm (Hill, 1979), part of the VESPAN-II package (Malloch, 1988).
In order to explore the relationship between the growth form composition of the
stands and the environmental variables measured in the study, direct gradient analysis was performed using the CANOCO programme (ter Braak, 1988). The TWINSPAN classes were later superimposed as centroids upon the resulting ordination.
The first axis of the ordination and the trace statistic (the sum of all axes) were tested
for statistical significance by means of a Monte Carlo permutation test (Hope, 1968),
also part of the CANOCO package.
Results
The summary statistics on the growth form composition of the sample stands in
the twelve regions is presented in Table 3.1. In total, the growth form composition of
192 samples was recorded. The only site to have all ten growth forms was Volcdn Tungurahua, but the remaining sites all had eight or nine of the ten present, with the exception of the pdramo near Ona which had only seven.
Overall, the mean number of growth forms in a 25 m 2 sample was just under
seven. Notably different from this value was the mean for the samples from Volcdn
Chimborazo, at 4.38.
Combining the three plots at each altitude level resulted in a mean growth form
complement of 7.25 for all of the sites. Again, Volcdn Chimborazo had fewer growth
3. Plant Form in the Ecuadorian Paramos
No of
Altitude
Levels
Site
VolcAn Chiles
VolcAn Cotacachi
PAramo de Guamanf
VolcAn Tungurahua
El Altar (west)
El Mar (east)
Daldal
Volcin Chimborazo
Paramo de Zapote Naida
Cajas
Cumbe
Dia
Overall
No of
Quadrats
No of
Forms in
all Quadrats
Mean No
Forms per
Quadrat
96
Mean No
Forms per
Altitude
(3 Quadrats)
7
7
7
5
5
6
6
7
3
7
3
1
21
21
21
15
15
18
18
21
9
21
9
3
9
8
8
10
8
8
9
8
8
8
9
7
7.67
7.38
6.33
6.87
6.73
6.06
7.67
4.38
7.33
6.95
7.67
6.33
8.14
7.57
6.86
7.40
7.20
6.33
8.00
5.43
8.00
7.57
8.33
7.00
64
192
10
6.73
7.25
Table 3.1.
The location and summarised growth form composition of 192 paramo quadrats.
For each locality the number of altitude levels sampled and the number of quadrats used are stated. The total number of growth forms found in the stands at each
locality are given. The mean number of growth forms found in each stand and at
each altitude level (three quadrats combined) are shown.
forms on average (6.33). The altitude levels in the paramo of Cumbe had the highest
mean with over eight growth forms on average.
Figure 3.11 indicates the frequency of occurrence of the ten growth forms in the
sample quadrats. Erect herbs, acaulescent rosettes and prostrate herbs were recorded in over 90% of the stands. Seven of the ten forms occurred in over threequarters of the vegetation samples (tussocks, acaulescent rosettes, cushions, upright
shrubs, prostrate shrubs, erect herbs, prostrate herbs). The remaining three forms
were much less frequent: trailing herbs and basal rosettes were present in approximately one-quarter of the stands, and stem rosettes were observed in just over 10%
of the samples.
The four most dominant growth forms were tussocks, acaulescent rosettes, erect
herbs and prostrate herbs, each of which accounted for more than 5% of the area of
about three-quarters of the plots. Only tussocks and acaulescent rosettes achieved a
cover of greater than 50% (in 65 and 8 plots, respectively).
The CANOCO analysis produced only two useful axes, due in part to the overall
similarity of the samples in terms of growth form composition. Subsequent axes were
unable to explain sufficient variation in the dataset to merit interpretation, demonstrated by low eigenvalues (A < 0.2). Another possible cause for such low eigenvalues
could be that the environmental variables recorded in this study were not sufficient
to ordinate the growth forms beyond the first two axes. However, the relationship between the distribution of the growth forms and the ordination axes was tested using a
Monte Carlo permutation test and was found to be statistically significant
(p < 0 .0 01) .
97
3. Plant Form in the Ecuadorian Paramos
Frequency
200
150
100
50
A_Pnmr
a•
SR BR
>0%
>5%
A=.J
rA
L 3
VA
I
>25%
>50%
>75%
T AR C/M US PS EH PH TH
Figure 3.11.
The growth form composition of in 192 samples of Ecuadorian paramo vegetation. The frequency of each growth form is shown cumulatively for five cover
values. [SR, Stem Rosettes; BR, Basal Rosettes; T, Tussocks; AR, Acaulescent Rosettes; C/M, Cushions; US, Upright Shrubs; PS, Prostrate Shrubs; EH, Erect
Herbs; PH, Prostrate Herbs; TH, Trailing Herbs]
Figure 3.12 shows the CANOCO biplot for the ten growth forms with the environmental variables superimposed on the ordination. The growth forms are well separated in the ordination space along a number of planes, which demonstrates a
complex relationship with the environmental variables.
Stem rosettes and basal rosettes have their optima in the lower right portion of the
ordination. This area can be partially characterised by relatively high exposure scores
and low bare ground coverage. Disturbance, altitude and rock cover appear to be
poorly correlated to their presence. However, none of the environmental variables
3. Plant Form in the Ecuadorian Paramos
98
2
• Cush ons
& Mats
Bare Ground
1
_
A t tude
/4
i
Acaulescent
Rosettes
Up( ght
Shrubs
I
-3
Trail ng
Herbs RP
I
-2
1
•
TLISSOCkS
/(
i R°ck i
Erect
1
Herbs
-1
Prostrate .n
Herbs w
D sturbance
Exposure
IP- I
2
Axis 1
Prostrate
Shrubs
-1 _
[
3 ^
•
Basa
Rosettes
-4 -
Ax 2
5 -
• Stem
Rosettes
Figure 3.12.
biplot ordination of the ten growth forms recorded in 192 samples of 'Aram° vegetation. Environmental variables are depicted as sol id arrows. The axes
are divided into standard deviation units (Ai =0.057, A2 = 0.026). To the left-hand
side of the ordination, trailing herbs and tussocks are located, characterised by
lower altitudes, higher disturbance levels, less rock cover and lower exposure
scores. By contrast, cushions, and to a lesser extent acaulescent rosettes have
their optima at higher altitudes, with higher exposure levels, more rock cover and
less disturbance.
CANOCO
measured in this study explains the distribution of stem and basal rosettes satisfactorily.
Tussocks and trailing herbs are found on the left-hand side of the ordinaton and
are associated with plots of low exposure and relatively high disturbance. High altitude with comparitively high rock cover characterises the plots containing acaulesent
rosettes, and cushions and mats.
Erect herbs, prostrate herbs, prostrate shrubs and upright shrubs are located
around the origin of the ordination and are not, therefore, associated with extremes
of any of the studied environmental variables.
99
3. Plant Form in the Ecuadorian Péramos
Growth Forms
10
7
2
0.178
0.103
5
3
Tussocks,
Trailing Herbs
3
0.113
Acaulescent
Rosettes,
Erect Herbs,
Prostrate Herbs
2
Upright Shrubs,
Prostrate
Shrubs
1
Stem Rosettes,
Basal Rosettes,
Cushions/Mats
Figure 3.13.
classification of the ten growth forms recorded in 192 paramo vegetation samples. Those growth forms which show similar distributions in these samples are grouped together. Four groups resulted from the classification. The
eigenvalue, which gives an indication of the importance of each division, is shown
directly beneath each division. The number of growth forms in each group is provided either side of a division.
TWINSPAN
Using TWINSPAN, the growth forms were grouped together according to their
similarity of distribution throughout the 192 paramo samples. The results of this analysis are depicted in Figure 3.13.
The first division of the ten growth forms separates a group containing the stem rosettes, basal rosettes and cushions from the remainder. Of the remaining seven
forms, tussocks and trailing herbs showed a linkage in distribution, as did the upright
shrubs and prostrate shrubs. The final association consisted of the acaulescent rosettes, erect herbs and prostrate herbs.
The paramo samples were analysed by virtue of their growth form composition,
both with TWINSPAN and CANOCO. The TWINSPAN classification of the 192
paramo quadrats in terms of growth form is shown in Figure 3.14. Twelve groups resulted from the TWINSPAN classification. Each class contained at least six of the ten
growth forms, while two had the full complement of ten. The usual number (and
mean) was eight growth forms. It is clear from this that the differences in composi-
100
3. Plant Form in the Ecuadorian Pâramos
Stands
192
65
43
27
21
0.179
0.143
2
Class
A
I
22
0.160
16
0.168
8 0.226
0.148
8
0.139
0.113
8
Classl 1Classi IClassl I Class' [Class' !Class' Class
G
H
BCDEF
15
0.138
10
4 0.143
14
10
127
0.115
Class
I
6
Class
J
Figure 3.14.
classification of 192 samples of pãramo vegetation according to growth
form composition. The composition of each of the twelve end groups is described
in detail in the text. The eigenvalue, which gives an indication of the importance of
each division, is shown directly beneath each division. The number of stands in
each group is provided either side of a division.
TWINSPAN
Class
L
101
3. Plant Form in the Ecuadorian Paramos
Class B
Class A
Frequency
Frequency
ez
2.6"
7
2-
6
1 .6 -
4
3
2
0 .6 4MINI
nIII.nnn••n••1%
0
SR BR
T
SR BR
AR C/61 US PS EH P14 TN
Class D
Class C
Frequency
10/
4
2
SR BR T
- SR BR T AR C/1.1 US PS EH PH TH
AR C/M US PS EH PH TN
Class F
Class E
Frequency
Frequency
10 /
•
4
2
SR BR T AR C/N US PS EH PH TH
SR BR
.0"
.14 W INMEd ZAMMIAMMI
AR C/N US PS EH PH TH
AR C/61 US PS EH PH TN
Figure 3.15.
The growth form composition of the twelve groups defined by TWINSPAN from 192
samples of Ora= vegetation. KEY: SR Stem Rosettes; BR Basal Rosettes; T Tussocks; AR Acaulescent Rosettes; CM Cushions/Mats; US Upright Shrubs; PS Prostrate
Shrubs; EH Erect Herbs; PH Prostrate Herbs; TH Trailing Herbs.
0%
65
3. Plant Form in the Ecuadorian Paramos
102
Class H
Class G
requenc-y
requency
•„
0%
%
n 2 16
SR BR T
SR BR T AR C/IA US PS EH PH TH
AR C/I4 US PS EH PH TH
Class I
Class J
Frequency
Frequency
,y
4
3
2
1
SR BR T AR C/14 US PS EH PH TH
SR BR
AR CAI US PS EH PH
Class L
Class K
requency
SR BR
AR C/M US PS EH PH TH
SR BR T AR C/M US PS EH PH TN
Figure 3.15. (Continued)
The growth form composition of the twelve groups defined by
samples of Oramo vegetation.
TWINSPAN
from 192
TH
3. Plant Form in the Ecuadorian Paramos
103
tion were largely in terms of the relative abundances of the growth forms rather than
their presence or absence from the stands.
The growth form composition of each of the twelve TVVINSPAN classes is shown
in detail in Figure 3.15, and their distribution throughout the study areas indicated in
Figure 3.16.
Class A (7 stands)
The plots belonging to this group were all located on Volcan Chimborazo (two at
4,600 m, one at 4,500 m, two at 4,400 m, one at 4,300 m and another at 4,000 m). All
of the plots had a sparse cover of vegetation, consisting of tussocks (5-25% cover)
with cushions, and upright shrubs (both growth forms with up to 25% cover). Acaulescent rosettes were present in six stands ( < 5% cover), erect herbs in five ( < 25%
cover), with a sparse cover of prostrate shrubs and prostrate herbs (each <5%
cover).
Class B (2 stands)
One stand at 4,200 m on Volcan Chiles and another at 4,400 m on Volcan Chimborazo were grouped by the TWINSPAN analysis. Cushions, and erect herbs were recorded in both plots with a cover of more than 5%. Accounting for less than 5%
cover, tussocks, acaulescent rosettes, upright shrubs and prostrate shrubs were found
in both stands, whilst prostrate herbs and trailing herbs were only present in one
stand each.
Class C (8 stands)
This group comprised three plots from 4,200 m on Volcan Cotacachi, one from
4,200 m and another from 4,100 m on Volcan Chiles, and one plot from 4,200 m and
another two from 4,100 m on the eastern slopes of El Altar. Acaulescent rosettes
were dominant, with a cover of greater than 25% in most cases. Prostrate herbs and
erect herbs occurred in all of the stands, sometimes with more than 5% cover. Prostrate shrubs and upright shrubs were also present in all of the stands, though responsible for less than 5% of the ground cover in most cases. Cushions frequently
accounted for 5-25% cover. Tussocks were found in the majority of the samples, with
a cover of up to 25%.
Class D (4 stands)
One plot at 3,400 m in the paramo of Cumbe, one at 4,200 m in the paramo of
Guamani, one at 4,200 m and another at 4,100 m on Volcan Chiles made up Class D.
All of the plots had a cover of 5-25% tussocks, and in one quadrat, tussocks were responsible for 25-50% cover. All four stands were vegetated by acaulescent rosettes,
cushions, and erect herbs (5-25% cover each), with lesser coverage by upright shrubs
and prostrate shrubs. Prostrate herbs were present in three of the samples and
reached more than 25% cover in one. Basal rosettes were also recorded in three of
the plots, but with less than 5% cover.
3. Plant Form in the Ecuadorian Paramos
104
Class E (6 stands)
Three plots at 4,200 m on the western side of El Altar, two plots at 4,000 m in the
paramo of Daldal and one quadrat at 4,300 m in the pdramo of Guamanf were dominated by acaulescent rosettes. In all of the stands, they covered over 50% of the
ground, and in two-thirds of cases more than 75%. Erect herbs were present in all of
the stands and covered 5-25% of the area in three plots. Prostrate herbs, prostrate
shrubs and upright shrubs were also found in all of the stands, each with a cover of
up to 25%. The majority of the plots had a cover of tussock plants, sometimes in excess of 25%.
Class F (8 stands)
Eight stands were grouped together by TWINSPAN in this growth form class
(three at 4,400 m in the paramo of Guamanf, three at 4,300 m and two at 4,200 m on
the eastern flank of El Altar). Cushions were present in all of the stands, and in over
half of them covered 5-25% of the ground surface and in a quarter covered more
than 25%. Acaulescent rosettes (5-50% cover), erect herbs (up to 25% cover) and
prostrate herbs ( <5% cover) formed the remainder of the vegetation cover. Prostrate shrubs were recorded in three samples with up to 25% cover. Tussocks were
present in only one plot, with a cover of less than 5%. Upright shrubs were completely absent.
Class G (8 stands)
On Volcan Tungurahua, in all three quadrats at 4,000 and at 4,100 m, one at 4,200
and another at 4,300 m, the vegetation was similar in many ways to that described for
Class F: acaulescent rosettes covered 5-25% of the ground, erect herbs up to 25%
and prostrate herbs less than 5%. Cushions were less conspicuous, responsible for
less than 5% cover. Prostrate shrubs were present in all of the stands and accounted
for more than 5% cover in all but one. Prostrate herbs were also present in three of
the quadrats. Basal rosette plants were recorded in seven of the eight plots, but
covered less than 5% in every case. Upright shrubs, contributing up to 25% cover,
were recorded in half the samples belonging to this class. Again, tussocks were rare
members of the community.
Class H (7 stands)
Four stands on Volcan Tungurahua (three at 3,900 m and one at 4,200 m) and
three on Volcan Chiles (all at 3,900 m) had a similar growth form composition. Stem
rosettes and erect herbs were present in all of the samples, sometimes covering more
than 25% each of the plot's area. The other growth forms, present in the majority of
samples, were basal rosettes ( <5% cover), tussocks ( <5% cover), acaulescent rosettes ( < 25% cover), cushions ( < 25% cover), upright shrubs ( < 25% cover), prostrate shrubs ( <5% cover) and prostrate herbs ( < 25% cover). Trailing herbs were
recorded in two of the plots with minimal cover.
•
3. Plant Form in the Ecuadorian Priramos
gI
1101
IL
105
ca
a)
i
:E
C.)
-E
c..)
co
0
co
I
..5
0
'E
CO
E
as
c
i
0
_
co co co co
0000
to
c
_c
:2
=
cp
c
II
-0
(1)
_0 0)
_J
CO co CO CM
0) CA L A CA
C.
Q3
•
E
cr)
1:=10
•
ca
cc)
•
0
1:3 'CT'S
12
1
.0 w
— -6'
•Tc °,.,
LT,
CD
^
LLI
(_/
CL ) CL )
U)
CO
23 "
CD
CO •-
01
81)
to u)
(c3
-C
O
0000
(13 -a
n
<
•
5 "a"i
TC (°
w
-0
co
LTI —
CD
0
0 -6
•
cr)
CO
‘-
CD
CD 0
CO
-o
CO
0
4—
< 03 0 0
U) C O L A CA
cn CA
C/7
CO CO CO CO
0-C3-0C)
o
N
E
cp
13
E
:E
o
1101111011
C;)
_
C
a)
;s
cr)
'0 a)
2 C) L—
c:5) o
=
Ll
•
(C
a_ a")
Cl)
z
112
c;)
a) 03
E
cc)
02
G. CO
•
Z
NJ
a)
1:11.
cc)
CO
C
0
c
CO
0
•
•
_C
co
0
CD CD CD 0 CD
CD CD CD CD CD
CD If) szr 0) N
cr
CD 0 CD CD CD
CD CD CD CD 0
•-• 0 0) CO hCr
C,)
C7
CD
CD
CD
CO
CD CD 0
0 0 0
CD
07 CI C.)
C1)
"0
n
CD CD
N
CO C)
Z)
-C)
.0
3. Plant Form in the Ecuadorian Pdramos
106
1.5
Bare Ground
Altitude
10
111
II Rock
Exposure
Figure 3.17.
CANOCO biplot ordination of the twelve growth form classes defined by TWINSPAN
from 192 samples of paramo vegetation. The axes are divided into standard
deviation units (A =0.057, A.2=0.026).
Class I (5 stands)
Four samples from Volcan Chiles (two from 3,700 m and two from 3,800 m) and a
fifth from 3,400 m in the paramo of Zapote Naida, were notable for the presence of
stem rosettes (with a cover up to 50%). Co-dominant were tussocks (25-50% cover),
and the other growth forms recorded were prostrate shrubs (up to 25% cover), cushions (up to 25% cover), erect herbs (5-25% cover), prostrate herbs (5-25% cover)
and upright shrubs ( < 5% cover). Acaulescent rosettes (5-25% cover) and basal rosettes ( <5% cover) were found in four and two of the five stands, respectively.
Class J (4 stands)
Two plots on Volcan Tungurahua (one at 4,300 m, the other at 4,200 m), one plot
at 4,300 m in the paramo of Guamani, and a fourth plot at 4,100 m on Volcan Chiles
comprised another TWINSPAN group. Prostrate herbs, erect herbs and basal rosettes each covered more than 5% of the area of all plots. Acaulescent rosettes were
also recorded in all of the samples, but with a cover of less than 5%. In addition, cu-
3. Plant Form in the Ecuadorian Paramos
107
shions, upright shrubs and prostrate shrubs were all present in at least half of the
plots, with a cover of less than 5%.
Class K (6 stands)
Six plots were dominated by tussocks (25-50% cover). They were located at 4,000
m, 3,900 m (two) and 3,800 m in the paramo of Cajas, and at 3,400 m and 3,200 m in
the paramo of Cumbe. Basal rosettes ( <5% cover), prostrate shrubs ( <5% cover),
erect herbs ( < 25% cover) and prostrate herbs ( < 25% cover) were present in all
samples, and acaulescent rosettes, cushions and upright shrubs were recorded in the
majority of the plots, each with up to 25% cover each.
Class L (127 stands)
The remaining 127 stands were included in the largest TWINSPAN group, representing two-thirds of the samples. All ten growth forms were present in this class,
though stem rosettes, with up to 50% cover, were found in only nine plots. The most
notable feature of this group was the dominance of tussocks, occurring in all but two
of the plots and covering over 75% in some samples. Acaulescent rosettes, erect
herbs and prostrate herbs were also strongly represented, each accounting for up to
50% cover in the majority of plots. Responsible for up to 25% cover in most samples
in this group, upright shrubs, prostrate shrubs and cushions formed a significant part
of the growth form community. Basal rosettes were only present in 25 stands, exceeding 5% cover in some of these. Finally, trailing herbs were recorded in 53 plots (only
three plots outside this class were found to have this growth form). In the majority of
cases, the coverage of trailing herbs was less than 5%, though this was surpassed in
three plots.
The twelve classes described above were super-imposed on the CANOCO biplot
(Figure 3.17). Most of the classes were clustered around the origin of the ordination,
confirming their overall similarity of composition. However, Classes I and H were located away from the main cluster, towards the position occupied by the stem rosettes
and basal rosettes. This indicates the importance of these forms in Classes I and H.
In the same way, Classes F, C, B and E were situated in the upper right sector of the
plot, indicating the predominance of cushions & mats and acaulescent rosettes. Class
L was sited close to the origin (as expected for the group representing two-thirds of
the samples), but towards the left-hand side, dominated by tussocks and trailing
herbs.
Discussion
n Hedberg's (1964) growth form classification, 55% of the flora were not classified.
IHedberg
(1992) maintains that although many paramo plants cannot be assigned to
one of the five forms in his system, all of the dominant plants can. Mena & Balslev
(1986), using the Hedberg classification, concluded that those species which did not
fit amounted to very little cover. It is interesting to find, therefore, that the additional growth form categories used in this study do account for a significant part of
3. Plant Form in the Ecuadorian Paramos
108
the vegetation cover. In particular, erect herbs and prostrate herbs covered 5-25% of
the sample area in the majority of plots. By comparison, stem rosettes and basal rosettes were much less frequent and abundant members of the Ecuadorian paramo
communities.
It may be that this shift of emphasis in growth forms accompanies the change from
stem rosette-dominated paramos in the north to the grassy paramos of Ecuador and
northern Peril in the south. However, erect and prostrate herbs were significant elements of even the stem rosette communities (Classes H & I) of Volcan Chiles.
Clearly, for a full and illuminating comparison of paramo regions within the Andes
(and more so for inter-continental comparisons), the full growth form spectrum
should be investigated. Merely because certain growth forms do not appear to be
adapted to the high altitude environment does not mean that they are not so. In fact,
their very presence must mean that they are successfully tackling the problems of
such environments. Furthermore, by omitting a large part of the flora from a growth
form classification system, important differences may be overlooked.
According to Mooney (1974), there is an optimal dominant growth form for a
given climatic-substrate-successional combination. Further, he suggests a similar degree of convergence in the subordinate units of vegetation. On this assumption, one
would expect to find a high degree of growth form similarity in the 192 päramo
stands sampled in this study. This is indeed the case, as evidenced in Table 3.1. In
most of the study areas, the average number of growth forms per plot was found to
be between 6 and 7, with the overall average 6.73. Examination of Figure 3.11 reveals that stem rosettes, basal rosettes and trailing herbs were generally absent. The
dominant growth form was the tussock. It was responsible for more than 50% of the
vegetative cover in one-third of the samples, and more than 25% in two-thirds.
The similarity of composition of the samples was also demonstrated by the TWINSPAN and CANOCO analyses, with low eigenvalues indicating little variation between plots.
However, some differences were evident. The TWINSPAN classification resulted
in twelve groups of stands according to growth form composition. One group, Class
L, contained about two-thirds of the samples and represented the most widespread
growth form composition. Tussock grasses were clearly dominant and associated with
them were trailing herbs present in 53 plots. Acaulescent rosettes, erect herbs, prostrate herbs, upright shrubs, prostrate shrubs and cushion plants were all strongly represented in this group.
A group of six stands, all found in the southern paramo regions (Class K), was very
similar to the large group just described. It differed in that tussocks, though still
dominant, were less abundant and trailing herbs were absent. Basal rosettes were
present in all of the plots but accounted for less than 5% of the cover.
In four samples from the humid northern 'Aram° of Volcdn Chiles and one from
Zapote Naida in the south (Class I), tussocks were co-dominant with stem rosettes.
3. Plant Form in the Ecuadorian Pâramos
109
In other parts of Volcan Chiles and on Volcan Tungurahua, stem rosettes were dominant in their own right, with tussocks accounting for less than 5% cover (Class H).
The remaining eight groups of samples were typically from plots with extreme conditions where tussock grasses were less abundant, generally at higher altitudes.
In places where such extremes restricted plants to microhabitats which permitted
establishment and growth, dominance of a single growth form was rare. In the arid
conditions of the Grande Arenal de Chimborazo, a number of samples showed codominance of tussocks, cushions and upright shrubs (Class A). Another plot from
Volcan Chimborazo was similar to a sample of vegetation from the highest altitude
sampled on Volcan Chiles (Class B). Although eight growth forms were present, only
erect herbs and cushions were able to exceed 5% cover.
Where a more continuous vegetative cover existed at high altitudes, the dominant
growth form was often the acaulescent rosette. In Class D, which consisted of plots
from the paramos of Cumbe, Guamani and Volcan Chiles, acaulescent rosettes were
co-dominant with tussocks, cushions and erect herbs. Class C was more clearly dominated by acaulescent rosettes with tussocks, cushions and erect herbs subordinate
members of the community. In the six samples of Class E from the paramos of El
Altar (west), Daldal and Guamani, acaulescent rosettes accounted for over 50% of
the vegetative cover (more than 75% in half of these plots).
In some paramo regions, above the vegetation dominated by acaulescent rosettes,
cushion plants became co-dominant (Class F). Upright shrubs which were present in
the lower altitude vegetation were not found and tussocks were rare. This vegetation
consisted of only six growth forms, the fewest of all the TWINSPAN groups.
On Volcan Tungurahua, much of the vegetation (Class G) showed a similar composition to that of higher altitudes elsewhere. Acaulescent rosettes were once again codominant, this time with erect herbs and prostrate herbs. Basal rosette plants were
present in almost all samples in this group.
The co-dominance of basal rosettes, erect herbs and prostrate herbs characterised
the vegetation of plots from Volcan Tungurahua, Volcan Chiles and the paramo of
Guamani (Class J).
From the CANOCO analysis, it is evident that altitude, and environmental variables closely correlated to it (for example, rock cover and disturbance), were related
to the distribution of acaulescent rosettes and cushions. These low stature plants are
probably confined to high-altitude, rocky sites by competition from other species for
light. Plantago rigida has been grown in a more favourable greenhouse environment
in the absence of competition from other species (personal observation). A later
chapter will examine the nature of the relationship between this species and Calamagrostis tussocks.
Tussocks and trailing herbs (which relied heavily on the tussocks for physical support) tended to occur more frequently in stands at lower altitudes, subjected to
higher disturbance.
3. Plant Form in the Ecuadorian Paramos
110
According to Bliss (1971), cushions and mats increase and tussocks and acaulescent rosettes decrease as the environment becomes more severe. This supports the
results of the current study.
To summarise, the growth form composition of the Ecuadorian paramos can be
described as follows:
• The majority of paramo vegetation is dominated by tussocks. The accompanying growth forms are mostly acaulescent rosettes, cushions, upright shrubs, prostrate shrubs,
erect herbs and prostrate herbs, sometimes with stem rosettes, basal rosettes or trailing herbs.
• At higher altitudes, the dominance of tussocks is reduced.
At first, acaulescent rosettes become dominant, but at yet
higher altitudes their dominance is shared with cushions.
At the highest altitudes of all, where plant cover is thin, no
single growth form is dominant.
• In other locations where plant cover is sparse, once again
no single growth form is dominant.
• In humid paramos, stem rosettes may be co-dominant with
tussocks or erect herbs. Basal rosettes, erect herbs and
prostrate herbs may be locally co-dominant at higher altitudes.
In terms of percentage cover, prostrate shrubs, erect herbs and prostrate herbs are
as important as upright shrubs and giant rosettes, but less conspicuously so. They
also account for a large proportion of the species present in the paramo. Their persistence in the paramo environment implies that strategies other than those of Hedberg's five forms are successful and deserve attention. Therefore, it is suggested that
all growth forms are considered in future paramo studies, at least in Ecuador.
Both basal and stem rosettes were present in stands with the least bare ground.
Miller (1987b) reports that establishment of Puya clava-herculis (a basal rosette
species) in the Ecuadorian paramo was reduced on bare, exposed soil, and even
more so in vegetation dominated by cushions and mats.
However, the environmental variables which were measured did not explain the
distribution of stem and basal rosettes satisfactorily. Miller & Sillander (1991) suggest that the upper elevational limit of Puya clava-herculis (a basal rosette species) in
the paramo of Virgen, Ecuador, is due to the combined effects of physiological
drought and low temperatures. Similar explanations have been offered to explain the
distribution of Espeletia species (stem rosettes) in other paramo regions of the Northern Andes (Faririas & Monasterio, 1980; Perez, 1987).
Billings (1973) linked the local distribution of growth forms in the equatorial alpine region to the availability of soil moisture: graminoids in the wetter sites, acaules-
3. Plant Form in the Ecuadorian Paramos
111
cent rosettes and cushions in drier sites (and ridges, rocky places and disturbed sites)
and prostrate shrubs along the whole moisture gradient.
According to Barkman (1988), plant forms which grow together can be expected
to have some similarity in their physiology. In the Ecuadorian pdramo, four groups of
such forms were determined (Figure 3.13), such as that containing acaulescent rosettes, erect herbs and prostrate herbs. The CANOCO analysis did not place members
of the same group close together in the ordination. This provides more evidence that
additional environmental factors are important in determining the distribution of
growth forms in the paramo.
In particular, soil characteristics, climatic features (especially measures of temperature and atmospheric humidity) and studies of plant water balance may yield interesting relationships with plant forms in the Andes. The next section in this chapter
looks at morphological adaptations to temperature in the paramos of Ecuador.
3. Plant Form in the Ecuadorian Paramos
MW44ZZZ•VX44.4.:ZW.:4Z44.4.4.4::44.4M4W4.4.4WKWWW
4.4.4WWW 4ZWW:4WZWWZWX4W
112
ZW44,M Z.:WW:WW
III. Temperature Characteristics of
Major Growth Forms in the
Ecuadorian Paramos
Introduction
High elevation tropical grasslands have often been used as an example of convergent
evolution (for example, Monasterio, 1986). Accepting Fisher's (1930) view that "no
character is likely to remain immune from selection for very long", the structure and
form of tropicalpine plants can be considered to be adapted to the prevailing environment. Similar selective agents in East Africa and the Andes, for example, will give
rise to plants that are ecologically equivalent and therefore alike in form and function.
One striking aspect of the tropicalpine environment which differentiates it from
other alpine and arctic regions is the diurnal temperature climate. "Summer every
day and winter every night" (Hedberg, 1964) present unique problems to tropicalpine plants. Temperature and its effects on water balance have been repeatedly used
to explain convergent evolution (Walter, 1973; Carlquist, 1974; Hedberg, 1964; Hedberg & Hedberg, 1979; Monasterio, 1986).
Wind induced cooling and water stress has been implicated in delayed flowering in
Hypericum laricifolium (Smith, 1972). Differences in the responses of plants to wind
in the paramo are closely correlated with differences in their growth form (Smith,
1972).
It is usually the reproductive organs of a plant that are most sensitive to chilling
and frost (Larcher & Bauer, 1981). Miller (1987a) carried out detailed observations
of the temperature relations of Puya inflorescences.
Hedberg & Hedberg (1979) presented temperature records for five species, each
representing one of Hedberg's (1964) growth forms, in the Venezuelan paramo in
Mucubaji, Merida. The evidence was seen to support the hypothesis that the various
growth forms represent different strategies to maintain the water balance in the
tropicalpine environment. Pfitsch (1988) stated that of Hedberg's five growth forms
that characterise the paramos, only sclerophyllous shrubs have no morphological
means of moderating the temperature extremes experienced by growing plant tissues.
Similar temperature measurements to those collected by Hedberg & Hedberg
(1979) were carried out in an Ecuadorian paramo over a 24 hour period.
3. Plant Form in the Ecuadorian Paramos
113
Methods
Study Site
This study was carried out in the paramo on the slopes of Volcan Chiles, about
381cm from Tulcan (0°47'N 77°57'W), near or in a boggy depression just below
4,000m. 011gaard & Balslev (1979) visited the site during the third Danish botanical
expedition to Ecuador in 1976 (Location 23) and described it floristically.
The area was not far from the three quadrats used to sample the vegetation at
4,000m in the phytosociological study of Volcan Chiles (one of the sites used in Chapter 2 and the previous section of this chapter).
Temperature Measurements
A Comark 2007 digital thermometer equipped with thermocouples (wire and
probe attachments) were used to record temperatures of the plants at various positions within their structure. Measurements were also carried out to provide contemporaneous records of air and soil temperature near the plants involved in the study.
The measurements taken were as follows (a wire thermocouple was used unless
stated otherwise):
Ambient Air
Measured at 1.5 m above ground surface, shielded
from the sun.
Ambient Soil
Measured at 100 mm below ground surface using
a probe thermocouple.
Giant Stem Rosettes
Espeletia pycnophylla ssp. angelensis
Centre of flowers
Surface of stem beneath marcescent leaves
Surface of stem lacking marcescent leaves
Surface of living leaf
Tussock Plants
Calamagrostis sp. [251]
Air between leaves at base of tussock
Air between leaves in upper part of tussock
Cortaderia sericantha
Air between leaves at base of tussock
Surface of inflorescence
3. Plant Form in the Ecuadorian Paramos
Acaulescent Rosettes
114
Valeriana bracteata
Surface of basal leaf
Surface of flower
Senecio sp
[voucher no. 847 in Ramsay & Merrow-Smith 1987 collection, corresponding to the "pretty Senecio with large solitary nodding heads"
(no. 8450) in 011gaard & Balslev, 1979.1
Surface of basal leaf
Surface of flower
Oritrophium peruvianum
Surface of basal leaf
Surface of flower
Cushion Plants
Wemeria humilis
Surface of rosette
Cushion at 100 mm depth (using probe)
Oreobolus obtusangulus
Surface of mat
Plantago rigida
Surface of rosette
Cushion at 100 mm depth (using probe)
Upright Shrubs
Loricaria ilinissae
Tip of branch
Pentacalia stuebellii
Tip of branch
Hypericum sp. [coil no. 9151
Tip of branch
Centre of flower
Pentacalia andicola
Tip of branch
Air within interior of shrub
Erect Herbs
Jamesonia sp [call no 861]
Apex of stalk
Lycopodium sp [coil no. 859]
Apex of stalk
Perezia pun gens
Surface of stem
Centre of flower
Centre of unopened flower bud
Castilleja sp [coil no. 946]
Centre of flower
Culcitium ovatum
Surface of stem
Centre of flower
The measurements were taken on the 20th and 21st of October 1987, with five records over the 24 hour period: on the first day at 14.30 and just after sunset at 18.30,
3. Plant Form in the Ecuadorian Paramos
115
then on the second day at 01.30, 05.30 (just before sunrise) and finally at 13.30. It was
clearly impossible to measure all plants simultaneously and so these times mark the
start of the temperature recording sessions. These sessions followed a precise sequence from plant to plant.
Over the course of this study, the sky was overcast during the day and for most of
the night, with intermittent drizzle, though occasional patches of clear sky appeared
during darkness hours.
Results
The ambient air temperature reached a maximum of 8.9°C at 14.30 hrs on the first
day. This temperature fell quickly after sunset (approximately 18.00 hrs) to 5.0°C and
reached a minimum of 3.7°C at 05.30 hrs on the following day (approximately half an
hour before sunrise). Air temperature rose quickly after sunrise, and by 13.30 hrs
had reached 7.3°C.
The temperature of the soil 100 mm beneath the surface showed little variation.
The maximum temperature was 6.9°C in the early hours of the second day, and the
minimum temperature was 6.4°C at 13.30 hrs later the same day. Clearly, there is a
considerable delay in warming up and cooling down at this depth.
The temperature measurements for the plants are described below.
Giant Stem Rosette
Espeletia pycnophylla spp. angelensis was the only species examined which belonged to this growth form. Unlike some of the other species of Espeletia, E.pycnophylla spp. angelensis does not appear to exhibit nyctinasty (the closure of the leaves
around the leaf buds at night). The thermocouple measurements relating to this
plant are presented in Figure 3.15.
In general plant parts were found to follow the ambient air temperature closely at
night. During the day, however, their temperatures were at times more than 10°C
above the air temperature.
Flower temperatures were high during the day (about 14°C) but dropped considerably at night, to below the ambient air temperature at 01.30 hrs. The flower temperatures showed oscillations of up to 11.2°C over the 24-hour period.
The leaves of Espeletia remain fixed to the stem after death (marcescence). The
majority of specimens in this area lacked marcescent leaves on the lower portion of
their stems as a direct consequence of burning. The insulating effect of marcescent
leaves was demonstrated by the reduced amplitude of the stem temperature beneath
the mantle of dead leaves (9.9°C) compared to the temperature of part of the stem
which lacked them (18.8°C). This was largely the result of the higher daytime tern-
3. Plant Form in the Ecuadorian Paramos
116
peratures of the bare stem (which was black because of charring). The marcescence
did maintain the stem temperature slightly above the ambient air temperature during the night.
Examination of the temperatures for the living leaves revealed a similar pattern to
that exhibited by the stem clothed by dead leaves, namely, a reduction of extreme,
high temperatures in the daytime and the maintenance of a slightly higher temperature than the air at night.
Tussock Plants
Two species of tussock grasses were represented in this study, Calamagrostis effusa
and Cortaderia sericantha (Figure 3.16). The former species is found as the co-dominant over most of the area, the latter is a common element of the flora in boggy areas.
The upper portion of the Calamagrostis tussock was found to maintain a high temperature (close to 15°C) in the trapped air between the leaves. The amplitude of the
measurements taken was 10.6°C. Lower down the tussock in the dense base, the amplitude was half this range (5.3°C). The air between the basal parts of the leaves
cooled more slowly after darkness fell and was not subjected to temperatures above
10°C over the entire 24-hour period.
Even the extremes of temperature of the ground surface between the tussocks
were reduced, remaining slightly above the ambient air temperature throughout the
night.
Cortaderia sericantha has a more open tussock structure. In some respects, the
basal portion of the Cortaderia tussock is most similar to the upper (rather than the
lower) portion of the Calamagrostis tussock; sunlight penetration and air circulation
are greater. Thus, daytime temperatures are in excess of 15°C in the bases of the Cortaderia tussock. At night the leaf bases were found to be approximately 1.5-2.0°C
higher than the corresponding air temperature. The inflorescence temperatures follow closely the air temperature.
Acaulescent Rosettes
Three species inhabiting the boggy area were used as representatives of the acaulescent rosette form: Ori trophium peruvianum, V aleriana bracteata and Senecio sp.
[847]. The flower temperatures of the former two species were observed to be approximately 0.5-1.0°C higher than the air temperature at night (Figure 3.17). The taller flower of Senecio sp. followed the air temperature closely for most of the study
period, falling below it at one time during the night.
The basal rosettes of the three species were higher than the ambient air temperature at all times over the course of the 24-hour study. In particular, the basal rosette
of Senecio sp. was strikingly higher than the air temperature at night, by some 2-3°C.
3. Plant Form in the Ecuadorian Paramos
Temperature (°C)
20 -
Time (hours)
• Leaf
--0 -- Soil (10cm)
—F— Flower
x Marc Stem
* Air at 1.5m
0 Bare Stem
Figure 3.18.
Temperature variation over 24 hours at a number of measuring points on Espeletia
pycnophylla ssp. angelensis at 4,050m in the paramo of Volcan Chiles. The points
measured were the flower disc, upper leaf surface, the surface of the stem clothed in
dead leaves ('Marc Stem') and the surface of the bare stem. Air and soil temperatures
are also shown.
117
3. Plant Form in the Ecuadorian Pâramos
118
Cushion/Mat Plants
Two species of cushion plant (Werneria humilis and Plantago rigida) and one mat
species (Oreobolus obtusangulus) formed the basis for the examination of these
growth forms (Figure 3.18).
The surfaces of all the cushion and mat species follow roughly the same pattern.
During the day, temperatures were high, in some cases in excess of 20°C. At night,
temperatures were higher than the surrounding air temperature by about 0.5-3.0°C.
At 10cm depth in the two cushion species, temperatures were conspicuously constant at about 9°C, less variable than that of the waterlogged soil surrounding them.
Upright Shrubs
Figure 3.19 presents the temperature data collected for four species of shrub: Loricaria ihnissae, Pentacalia stuebellii, Hypericum ? strictum and Pentacalia andicola. A
similar pattern of temperature variation was observed for all four species. At night,
the tips of the branches were almost always slightly above the air temperature, while
by day they were often 5-10°C higher than the ambient temperature.
The dense branches of Pentacalia andicola formed an effective screen against light
and air circulation, but the temperature within this space did not demonstrate an
amelioration of the temperature extremes; in fact, it deviated little from the ambient
air temperature.
The flowers of Hypericum ? strictum showed similar temperature patterns over the
course of the study period to this species' branch tips.
Erect Herbs
This growth form was found to experience similar fluctuations in diurnal temperature to those observed in other plants in this study: temperatures close to air temperatures at night complemented by daytime temperatures well in excess of the
ambient conditions (Figure 3.20). Stem, bud and flower temperatures did not vary
markedly from this pattern in the five species studied: Lycopodium ? crassum, Castilleja sp., Perezia pungens, Cukitium ovatum and Jamesonia ? goudotii.
The most interesting observation in this group concerns the higher night-time temperatures of the hairy stalk apex of Jamesonia, which did not reflect the corresponding ambient air temperatures as faithfully as the remaining erect herbs.
Discussion
Unfortunately, the night-time temperatures experienced on the study dates were
not particularly low. In 1976, 011gaard & Balslev (1979) had measured a night-time
3. Plant Form in the Ecuadorian Paramos
119
Calamagrostis
Temperature (C)
14.30
18.30
5.30
1.30
13.30
Time (hours)
— Base
0-
—I— Leaves
* Air (1.5m)
Soli (10cm) —)4— Ground Surface
Cotaderia sericantha
Temperature (C)
14 30
18.30
1.30
5.30
13.30
Time (hours)
— Base
* Air at 1.6m
—f— Flower
- 0- Soil at 10cm depth
Figure 3.19.
Temperature variation over 24 hours at a number of measuring points on two species
with a tussock growthform (Calamagrostis sp. and Cortaderia sericantha) at 4,050m in
the paramo of VoleAn Chiles. The temperature was measured at the base of the tussock, the upper leaf region (Calamagrostis only), the surface of the inflorescence (Cortaderia only) and the ground surface at the edge of the tussock. Air and soil
temperatures are also shown.
120
3. Plant Form in the Ecuadorian Paramos
Senecio sp.
V aleriana bracteata
Temperature 1°C)
Temperature (CC)
18.30
14 30
1.30
5.30
3.30
14.30
18.30
Time (hours)
—
Rosette
Alt at 1.6m
t30
5.30
Time (hours)
—
—1— Flower
Roll at 10cm depth
Rosette
—I— Flower
Air at tern
Coll at 10em depth
Oritrophium peruvianum
Temperature (°C)
18.30
1.30
5.30
13.30
Time (hours)
— Rosette
* Air at 1.5m
Flower
- 0- Soil at 10cm depth
Figure 3.20.
Temperature variation over 24 hours at a number of measuring points on three species
with an acaulescent rosette growthform (Oritrophium peruvianum, Valeriana bra cteata
and Senecio sp.) at 4,050m in the paramo of Volcan Chiles. Basal leaf surface and
flower temperature were measured. Air and soil temperatures are also shown.
3.30
3. Plant Form in the Ecuadorian Paramos
121
minimum of about +10C a few hundred metres away from the location of this present study (where the minimum temperature was 3.7°C). Examination of Hedberg &
Hedberg's (1979) figures reveals that much of the evidence used to support their hypotheses was derived from one of the three nights for which they had recorded data,
when temperatures dropped to around + 1°C. On the remaining two nights, minimum temperatures were about 7-9°C, and the thermoregulatory properties of their
study plants were not so pronounced, if apparent at all. Clearly, better insights into
plant strategies can be gained when the nights are cold, preferably with a frost.
Figures 3.16-3.20 give the impression that a sharp temperature decline takes place
between 1430 and 1830, as a consequence of the intervals between measuring times.
It should be noted, however, that the majority of the temperatures decline occurred
between 1800 and 1830. After sunrise, however, air temperatures increased more
steadily, taking several hours to achieve values similar to those indicated at 1330;
though as a result of insolation, some surface temperatures may have risen considerably immediately after sunrise.
Radiation frost —the loss of radiated heat from surfaces —is an important consideration in interpreting these results. Surfaces of vegetation or ground cool down several degrees more than air at 2m. Usually, minimum plant temperatures on clear
nights are 1-3°C below the minimum air temperature (Larcher & Bauer, 1981). According to Grace (1988), short vegetation would be expected to be cooler than tall
vegetation because mixing of air is reduced closer to the ground and therefore radiated losses are more important.
Stem Rosettes
The stems of giant rosette plants in Africa and South America contain voluminous,
parenchymatous pith that acts as a water source during periods of low water availability (Hedberg, 1964; Goldstein, Meinzer & Monasterio, 1984). Many of these giant
rosette plants exhibit nyctinasty: the leaves close around the single apical bud at
night and open during the day (Smith, 1974), damping diurnal temperature fluctuations. Thus they avoid freezing stress on cold nights and overheating (and resultant
water stress in young leaves) early in the morning. Smith (1974) has demonstrated
that leaf wilting and death results from the prevention of nyctinastic leaf movements
in such species. Mabberley (1986) attributed damped heating and cooling of stem rosettes to their massive construction, a view that is supported by the findings of Smith
(1980) that Espeletia schultzii plants were larger at higher altitudes. Coespeletia lutescens was found to modify the microclimate beneath the plant (air temperatures 4.77.0°C higher than in the open; soil at 20 cm depth 2.4-4.2°C higher), and was linked
to better seedling survival and greater water uptake.
The species observed in this study, Espeletia pycnophylla ssp. angelensis, did not exhibit nyctinasty. However, the living leaves making up the apical rosette were densely pubescent (Acosta-Solis, 1984, refers to them as "donkey's ears"). This fur-like
covering may explain why these leaves cooled down more slowly and remained slightly warmer than the other parts of the plant throughout the night. Meinzer & Goldstein (1985) found that the thickness of pubescence in a Venezuelan species of
Espeletia increased by 1.5 mm along a 1,600 m gradient of increasing altitude. Hed-
122
3. Plant Form in the Ecuadorian Paramos
Oreobolus obtusangulus
W emeria hum&
Temperature CC)
20
Temperature CC)
26
16
20
16
10
10
14.30
1.30
18.30
5.30
Time (hours)
0
14 30
5.30
1.30
18 30
13.30
Eturfaco of Mat
Time (hours)
— Cushion Surface
* Air at Lent
* Air at 1.61n
C. Soil at 10em depth
—4— Cushion flOcal depth)
-0 Soil at lOcia depth
Plantag0
Te m p erature (QC)
14.30
1.30
18.30
Time (hours)
— Cushion Surface
* Air at 1.5m
-4— Cushion (10cm depth)
'El' SOH at 10cm depth
Figure 3.21.
Temperature variation over 24 hours at a number of measuring points on three species
with a cushion or mat growthform (Wemeria humilis, Plantago rigida and Oreobolus
goeppingen) at 4,050m in the paramo of Volcan Chiles. Cushion or mat surface temperature and that 10cm below the surface were measured. Air and soil temperatures
are also shown.
13.30
3. Plant Form in the Ecuadorian Paramos
Pentacalia stuebellii
Loricaria ilinissae
Temperature (C)
Temperature (C)
14 30
18.30
5.30
1.30
13.30
—1— Branch Tip
-9-
g Soil at 10cm depth
5.30
1.30
13.30
18.30
14 30
—4— Branch Tip
- a-
Soil at 10cm depth
1.30
5.30
Time (hours)
Time (hours)
* Air at 1.5m
Air at 1.5m
Temperature (°C)
Temperature (C)
— Flower
4'
Soil at 10cm depth
Pentacalia and/cola
Hyper/cum
18.30
13.30
Time (hours)
*- Air at 1.5m
—I— Branch Tip
6.30
1.30
18.30
14 30
Time (hours)
14 30
123
--- inside Shrub
* Air at 1.5m
—I— Branch Tip
- a-
Soil at 10cm depth
Figure 3.22.
Temperature variation over 24 hours at a number of measuring points on four species
with a shrubby growthform (Loricaria ilinissae, Pentacalia stuebellii, Hypericum sp.
and Pentacalia and/cola) at 4,050m in the paramo of Vo'can Chiles. The temperature
of the branch tip, flower and the inside of the shrub were measured. Air and soil temperatures are also shown.
13.30
3. Plant Form in the Ecuadorian Paramos
Lyco podium
James onia
Temperature fC)
Temperature CC)
18.30
1.30
6.30
13.30
14.30
18.30
Time (hours)
I
—
Slant Apex
124
1.30
6.30
13.30
Time (hours)
A Al, al Lan 0- $ae Mem/
Stalk Apex Air
A
al t6111
.0- loll (10are)
Perezia pun gens
Temperature eCt
15
10
18.30
14.30
1.30
6.30
13.30
Time (hours)
--
Al, al ilea
-0 led (lawn) -"-
Culcitium ova turn
Castilleja
Temper•ture CC)
Tempersture rC)
15
10
•
14 30
e
14 30
18.30
1.30
6.30
18.30
130
6.30
13.30
Time (hours)
13:30
Time (hours)
Flower
— Flower
Air at L5er -
0 - Bell (Vera)
Son (10cm)
A Air al 1.11,
-A- Sem
Figure 3.23.
Temperature variation over 24 hours at a number of measuring points on five species
with an erect herb growthform (Jamesonia sp., Lycopodium sp., Castilleja sp., Perezia
pungens and Culcitium ovatum) at 4,050m in the paramo of Vo'can Chiles. The temperature of the branch tip, flower, bud and halfway up the stem were measured. Air
and soil temperatures are also shown.
3. Plant Form in the Ecuadorian P6ramos
125
berg (1964) and Baruch & Smith (1979) hypothesised that the adaptive significance
of leaf pubescence in tropical alpine giant rosette species lay in reduced radiation absorption, leading to reduced leaf temperature and lower rates of transpiration. Meinzer & Goldstein (1985), however, suggest that prevailing air temperatures indicate
that latent and convective heat loss are more critical in determining the thermal balance of the leaf. They predict that leaf pubescence could result in up to 5°C higher
leaf temperature. Similar pubescence on the inflorescence of Puya hamata was
shown to increase tissue temperature significantly and thus increase seed production
(Miller, 1987a).
A clear effect of the marcescent leaves on the surface temperature of the stem was
observed. The diurnal range of temperature was reduced from 18.8°C to 9.9°C by this
covering of dead leaves. Minimum surface temperatures on the stem were approximately the same regardless of the presence of marcescent leaves. The buffering effect of the marcescent layer was, therefore, largely in the prevention of extreme high
temperatures. Hedberg & Hedberg (1979) demonstrated the good insulating capacity of the mantle of marcescent leaves in Espeletia schultzii, which remained remarkably constant at around 7.5°C, regardless of the temperature outside the mantle. The
mean temperature of the stem beneath the dead leaves of E.pycnophylla ssp. angelensis over the course of the present study was also 7.5°C.
The marcescent leaf mantle was incomplete —the lower portion having been destroyed by fire — and this may have resulted in some loss of insulatory protection.
Goldstein & Meinzer (1983) removed the dead leaf layer of Espeletia timotensis and
showed that stem temperature was altered, resulting in transient and permanent effects on water balance. Smith (1979) and Goldstein, Meinzer & Monasterio (1984)
report similar conclusions. The mechanism attributed to this effect by Goldstein &
Meinzer (1983) was considered to be one or more of the following: the inhibition of
pith recharge by subfreezing stem temperatures, the formation of embolisms in the
stem xylem and freezing injury to pith tissue.
Flower temperatures were close to ambient air temperatures for most of the 24hour period. This indicates that the inflorescences possess little ability to modify temperature from that of the surrounding air and rely on tolerance of low temperatures
rather than avoidance. Smith (1974) reported that the parabolic form of Espeletia
schultzii leaves concentrated the sun's rays, raising the temperature of the bud. Based
on the evidence of this study, E.pycnophylla ssp. angelensis does not function in the
same way, since flower temperatures were not found to be greater than the air. As
mentioned earlier, high inflorescence temperature was linked to increased seed production in Puya hamata (Miller, 1987a). Fewer numbers of flowers were found on
the windward sides of the Espeletia plants by Smith (1974). These are clear illustrations of how temperature stress can effect reproductive potential.
Tussock Grasses
Tussocks provide a well-defined boundary layer of dead air (Geiger, 1966; Jones,
1983). The outer leaves of the Calamagrostis tussock are subject to greater temperature variability than the basal leaves, but the trapping of air within the tussock allows
the temperature to rise more than 5°C above that of the ambient air during daylight
3. Plant Form in the Ecuadorian Paramos
126
hours. At night the temperature does not deviate greatly from the ambient air temperature, but since these leaves are old and hardy, low temperatures may not be damaging.
The dense bases of Calamagrostis sp. are well insulated against extremes of temperature, cooling slowly after dark and not exceeding 10°C during the day. It would
appear likely that during severe frosts the developing tillers of Calamagrostis are protected by the surrounding leaves. The retention of dead leaves within the tussock
structure may enhance this shielding effect. Hedberg (1964) observed that the dense
base of a tussock of Festuca pilgeri ssp. pilgeri on Mount Kenya, East Africa, was
7.5°C warmer than the -5°C temperature in the outermost leaves of the tussock. Coe
(1967) presented similar findings for the same species. In the Venezuelan Andes,
Hedberg & Hedberg (1979) showed a similar phenomenon in Stipa sp.
The insulatory properties are similar with respect to factors other than climate.
During a fire, for example, this portion of the tussock is shielded against radiated
heat in much the same way as it is protected from intense cold (Chapter 4).
The hairy basal leaves of Cortaderia sericantha serve a similar function to the pubescence on the marcescent Espeletia leaves, with the same result. The protection from
frosts afforded by these hairs allows newer leaves to develop undamaged.
The flowers of C.sericantha project beyond the vegetative leaves of the tussock. Although this results in lower night-time temperatures and thus lower seed production,
the flowerheads serve the function of pollen and seed dispersal which requires good
air circulation to be effective. The benefits of increased pollination and dispersal
might outweigh the disadvantages of low seed production.
According to Nishikawa (1990), tussock formation provides stable growth conditions against fluctuations in water level, air temperature and other factors. Tussock
formation changes a plant from a competitor in an unformed tussock to a stress tolerator in maturity.
Acaulescent Rosettes
The higher night rosette temperatures of the three acaulescent rosette species
compared with the air temperature corroborate the findings of Hedberg & Hedberg
(1979) with Hypochaeris sessihflora in Venezuela. In particular, Senecio sp. showed
the same degree of difference between these temperatures.
Hedberg & Hedberg (1979) suggest that the position of these plants at the air/soil
interface enables them to buffer temperature variation, but they do not offer a mechanism for this, nor explain their results, which show the rosette temperature above
both air and soil surface temperature over the three day period. Possible explanations include the protective properties of the outermost leaves and the beneficial
heat output of groundwater during cold nights (Carlquist, 1974). Hedberg (1964)
noted that water is more viscous at low temperatures and that the short internodes of
acaulescent rosettes mitigate this problem, and perhaps explains their success.
3. Plant Form in the Ecuadorian Paramos
127
The three species covered in the present study are found on soil which is heavily
waterlogged. The temperature of this wet soil is several degrees higher than that of
neighbouring areas. These higher soil temperatures may help plants considerably in
buffering extreme cold. Heat transfer from lower in the soil profile would clearly be
advantageous. In Hedberg & Hedberg's (1979) study, the soil temperature 10cm
below ground is 10-14°C, several degrees higher than the soil temperature found in
the boggy areas of the Volcan Chiles study location. It is likely that the soil temperature will remain several degrees above the night-time air minimum throughout the
paramo, and acaulescent rosettes can therefore, exploit the soil/air interface over a
wide range of temperatures.
The flower temperatures of V aleriana braeteata and Oritrophium peruvianum were
observed to be 0.5-1.0°C higher than the air temperature at night. It is difficult to establish an external morphological explanation for this. One possible explanation may
be that these structures are able to exploit the heat release associated with condensation of water vapour on the flower surface. By encouraging condensation, the flowers
may sustain a higher temperature than the surrounding air through the night.
Cushions
The surfaces of the cushions followed a similar diurnal pattern to that found by
Hedberg & Hedberg (1979) for Plantago rigida in Venezuela. By day, temperatures
reached in excess of 20°C, while at night these surfaces fell to within a few degrees of
the air temperature. Ruthsatz (1978) observed the diurnal temperature regimes of
five cushion species in the puna of North-west Argentina and reported similarly wide
thermal fluctuations just beneath the cushion surface.
Hedberg & Hedberg (1979) point out that cushions merely represent an aggregate
of acaulescent rosettes and they may be viewed as adopting a similar approach to
thermoregulation, that is, taking advantage of the soil/air interface (Rauh, 1939; Hedberg, 1964; Billings & Mooney, 1968; Billings, 1973; Armesto, Arroyo & Villagran,
1980).
The inside of the cushions (10cm below the surface) remained markedly constant — more so than the soil at the same depth —at around 9°C in both Plantago rigida and W emeria humilis. In support of these observations, Ruthsatz (1978) found
that temperature measurements 10 cm deep within five cushion species in Argentina
had much smaller oscillations than the ambient conditions.
Therefore, the cushion form may enjoy the advantages of an enhanced soil/air interface situation while the increased height which the domed shape provides for
some species may reduce waterlogging and increase the competitive ability of the
plant with regard to light. In addition, the grouping of rosettes may provide mutual
protection against strong winds and desiccation. Therefore, it does seem plausible
that the cushions can effectively raise the soil/air interface to their rosettes by means
of the cushion structure. Alliende & Hoffmann (1985) demonstrated that for some
puna species cushions provide an ideal substrate for germination; indeed, some
species were found almost exclusively on cushion plants. This indicates that the physical characteristics of cushions ameliorate the extremes of environment in such cases.
3. Plant Form in the Ecuadorian Paramos
128
Shrubs
These plants showed little adaptation towards temperature regulation, relying
heavily on low temperature tolerance. By day they were warmed by insolation and by
night they cooled with the air temperature. Hedberg & Hedberg (1979) proposed
that these plants do not possess morphological features to avoid low temperatures;
instead, their morphology enables these plants to withstand them. Thus the scale-like
leaves of Loricaria ilinissiae, the needle-like leaves of Hypericum sp., the waxy leaves
of Pentacalia stuebellii, and the leathery leaves of Pentacalia andicola all serve to reduce transpiration during low temperatures, and by these means prevent water
stress. If this were so, then one would expect to see increasing xeromorphy as conditions become more severe: Hedberg (1957) found this to be the case in East Africa.
Carlquist (1974) affirmed the frost resistant function of the `cupressoid' habit of
Loricaria and added the functions of minimising transpiration and withstanding the
effects of alpine light conditions. He also pointed out that Loricaria has ultraspecialised wood with an abundance of vascular tracheids which is related to cold tissue
temperatures.
Erect Herbs
Like the shrubs, it would appear that four of the five erect herbs in this study do
not possess morphological features to ameliorate their temperatures. Lycopodium
sp. appears to rely on low temperature tolerance and was found to show significant
altitudinal trends in leaf and plant size for Central Ecuador (Buckland & Ramsay, in
press), which may be a response to temperature and water stress.
Culcitium ovatum has leaves covered with downy hairs, but does not appear to gain
thermal benefit from this pubescence at the temperatures encountered in this study.
These hairs may instead serve to reduce transpiration during periods of water stress.
Unlike Culcitium, Jamesonia goudotii was found to stay approximately 0.5-1.0°C
above the air temperature overnight. Dense pubescence around the developing
frond tip and along the midrib characterises this species, and may explain the slightly
higher temperatures.
E
ssentially, there are three major problems associated with low temperatures in the
paramo:
• direct damage to tissues by low temperatures.
• reduction in rates of growth and development
• water stress caused by transpiration demand when cold
temperatures restrict the rate of water uptake.
Minor thermal differences can have a significant effect on plant water balance
(Goldstein & Meinzer, 1983) and survival (Smith, 1979). These problems are particularly pronounced when temperatures fall below zero. In a cold environment, there is
3. Plant Form in the Ecuadorian Paramos
129
strong selective pressure for the evolution of freezing avoidance and/or tolerance
mechanisms (Azocar, Rada & Goldstein, 1988). In habitats where temperatures at
night do not fall far below zero and remain there only for short periods of time, the
main resistance mechanism should be freezing avoidance (Larcher, 1981; Sakai &
Larcher, 1987). On the other hand, if temperatures drop well below freezing at night
and stay there for several hours, tolerance should be the selected resistance mechanism (Larcher, 1981; Rada eta!., 1985; Sakai & Larcher, 1987). In the study area, a
combination of both avoidance and tolerance would be expected, since the plants
there must endure both short and more long-lasting periods of freezing stress, according to the season (Sarmiento, 1986).
In the case of tolerance, physiological adaptations are most important permitting
tropical alpine plants to recover their full photosynthetic capacity after a night frost
(Schulze et al., 1985). AzOcar et al. (1988) studied Draba chionophila in the Venezuelan paramo. This rosette plant was not insulated from low night-time temperatures
and leaves, pith and roots were observed to freeze without causing injury to the
plant.
Morphological features may be important in reducing transpiration (for example,
by means of xeromorphy or pubescence) or in maintaining the water balance in some
other way (such as the water-storing pith of Espeletia spp.— Goldstein et aL, 1984).
A number of avoidance strategies have been adopted by paramo plant species.
One approach is the shielding of delicate parts with dead, hardy or expendable parts:
as in the case of the marcescent leaf mantle clothing the Espeletia stem, or the protection of developing tillers and leaves by the outer leaves in tussock grasses. Many rosette plants protect their inner developing leaves with outer ones (for example, Puya
hamata, W erneria nubigena). Trees of the genus Polylepis buffer temperatures by
means of many thin layers of exfoliating bark (Simpson, 1979) — a significant reduction in the extremes of high and low temperatures beneath the bark was measured by
Liley (1986). As mentioned earlier, insulatory functions of a plant can increase its
survival rate after a fire by shielding part of the plant from intense radiated heat.
Pubescence is another common strategy for low temperature avoidance. Meinzer
& Goldstein (1985) demonstrated by model simulation that leaf pubescence works
by increasing the thickness of the boundary layer of still air and reducing convective
heat transfer from leaf to air. This is particularly pronounced when many pubescent
layers lie together (as in a developing bud).
In this study, Espeletia pycnophylla, Cortaderia sericantha and Jamesonia goudotii
maintained higher temperatures than that of the air by means of hairiness. Miller
(1987a) reported an increase in inflorescence pubescence for various species of
Puya, and with a combination of temperature measurements and pubescence removal demonstrated that the layer of hairs was responsible for up to 80% of the difference between flower and air temperature. He then linked this higher thermal
regime with increased success in seed production.
Finally, by inhabiting the boundary between soil and air, some smaller plants are
able to benefit from the warmer soil temperatures at night just below the surface.
3. Plant Form in the Ecuadorian Paramos
130
The acaulescent rosette growth form adopts this strategy, as do mat-forming species.
Taken one stage further, cushion plants are able to artificially raise the soil surface,
perhaps increasing their competitive abilities or reducing the effects of waterlogging.
In addition, by retaining a smooth surface, the individuals of a cushion or mat are
able to offer mutual protection from desiccation and wind action.
In some cases, both avoidance and tolerance strategies occur in combination. For
instance, nyctinasty was found to enhance the avoidance of low temperatures in
young leaves at night in Espeletia semiglobulata, whilst the outer leaves undergo regular freezing and appear to be undamaged (Larcher, 1975). The acaulescent rosette,
Senecio sp., and the tussock, Cortaderia sericantha, both employ avoidance in their
vegetative parts (by means of the soil/air interface and pubescence/mutual shelter, respectively) and tolerance in the floral parts.
According to Dobzhansky (1950), any organism that lives in a temperate or cold
climate is exposed at different periods of its life cycle or in different generations to
sharply different climates. To survive and reproduce, any species must be at least
tolerably well adapted to every one of the environments which it regularly meets.
Changeable environments put the highest premium on versatility rather than perfection in adaptation. This view is supported by Tomlinson (1987) who suggests that
plasticity is more significant in adaptive terms than initial architecture.
The thermal regime is just one of the elements of the environment addressed by
growth form. It has already been mentioned that resistance to fire, protection from
solar radiation, transpiration, reproduction and competition are rival considerations
for inclusion in the overall form of a paramo plant. Therefore, the form of a plant
represents the outcome of many selection pressures, some weightier than others.
The form is a structural and functional compromise which allows for the optimization of cost-benefit relationships (Baruch, 1982). As long as a plant gains more carbon than it pays for its architecture and physiology it may survive (Kiippers, 1989).
So, the inflorescence may suffer reduced seed production in an exposed position,
but the fewer seeds that are produced may be dispersed more efficiently. Such forces
are not necessarily antagonistic: it has already been cited that morphological features
which insulate sensitive tissues from extremes of climatic temperature can also serve
to protect against the high temperatures experienced during a paramo fire. Givnish,
McDiarrnid & Buck (1986) suggest that the evolution of a stem rosette species in the
Venezuelan tepuis has been driven by fire rather than low temperatures and were
able to demonstrate that fire survival was correlated with rosette height. Beck,
Scheibe & Schulze (1986) found that tussock grasses were increased after an East African alpine fire, suggesting that fire favours the tussock form.
Despite competition for morphological adaptations from other considerations, it is
clear from the results of this study that resistance to low night-time temperatures has
been evolved by a number of plants and the growth form plays a major role in this. A
study of this kind, performed on a very cold night, would provide further and possibly
more conclusive evidence for thermoregulation by growth form and other morphological features.
Chapter 4
Aspects of Plant
Community Dynamics in
the Ecuadorian Paramos
131
4. Dynamics of Ecuadorian Paramo V egetation
132
Introduction
p aramo plant species are not randomly scattered throughout the vegetation; they
A' exist in repeated patterns of particular species —as plant communities. Chapter 2
provided a descriptive treatment of paramo communities and correlated species composition with a number of environmental variables.
This century has seen a productive and well-documented difference of opinion between those who believed, like Clements (1916), that communities were 'super-organisms' (and succession an entirely deterministic process) and those who saw the
vegetation as merely the resultant of two factors, the fluctuating and fortuitous immigration of plants and an equally fortuitous and fluctuating environment (Gleason,
1917, 1976). The debate between the Clementsian holists and the Gleasonian individualists resulted in the rejection of the holist approach and the application of Darwinian reductionist thinking to the development of plant communities. Plant
communities are therefore viewed as the result of three influences:
• the response of plants to variation in external factors in
their environment.
• the response of plants to each other through competitive
interactions between individuals.
• historical chance events, reflecting both colonisation and
extinction.
Emphasis has shifted away from the abstract of plant communities to the components of the vegetation, the individual plants themselves. Although particular communities may appear static in composition, these communities are longer-lived than
their component parts (the plants) and they are maintained by a dynamic process of
death and replacement of individual plants. Two main approaches have been used to
examine the dynamic nature of plant communities. One involves piecing together a
picture of the processes involved by comparing contemporaneous plots at different
stages of development. This means of investigation is somewhat subjective, but as
Watt (1947) points out: "the formulation of laws and their expression in mathematical terms will be facilitated if an acceptable qualitative statement of the nature of relations between the components of the vegetation is first presented."
The second, and perhaps more satisfying approach examines the actual changes in
plant communities over time, looking at individual plants or plants falling into categories (patches involving dominant growth forms or smaller associations within the
community). The advantage of this quantitative approach is that it is open to statistical interpretation. In its simplest form, each individual or category has a certain probability of being replaced by another of its kind or by an individual or category of
another kind. This approach has been used widely and has led to the development of
Markov modelling where these probabilities of replacement are used to predict suc-
4. Dynamics of Ecuadorian Paramo Vegetation
133
cession in vegetation (Horn, 1975; Usher, 1979, 1981; Hobbs & Legg, 1983; Lough et
al., 1987).
These approaches can be used to investigate not only the dynamic nature of the
maintenance of community composition in climax vegetation, but also to examine
succession or recovery from disturbance. It has been proposed by a number of authors that much of today's grassy pdramos (and other Andean high elevation grasslands) are secondary vegetation types, maintained artificially by man via burning
(Ellenberg, 1979; Laegaard, 1992), though other authors disagree (for example, Simpson, 1979). In Chapter 2, burning (an element of disturbance) was found to be at
the very least correlated with species distributions.
Fire is commonplace in the pdramo, a tool used by farmers to improve their pastures. If fire is a very rare (catastrophic) event then it is unlikely to exert a selective
influence on the vegetation, but if (as is the case in the paramo) burning is frequent,
the vegetation might be expected to show some kind of fire adaption. Although the
paramos are very humid, this does not preclude the occurrence of natural fires — Givnish, McDiarmid & Buck (1986) describe a fire started by lightning in an exceedingly
rainy tepui in Venezuela. From the arguments presented above, burning should affect the composition of paramo communities through its influence on plant population dynamics. Smith & Young (1987) noted the apparent cyclic succession induced
by fire in the paramos of Colombia and Ecuador, and pointed to the lack of data on
such phenomena.
With this in mind, a number of experiments were set up to investigate the dynamics of paramo tussock grass communities. The physiognomy of pdramo grassland is
not suited to a comprehensive strategy of sampling, particularly at the individual
plant level, because it consists of both large and small plants and may consist of several layers. In addition, the study period was very short. For both these reasons, a
number of different approaches, both descriptive and probabilistic, were employed
to examine mechanisms involved in the maintenance of the community and its recovery from burning.
Three approaches were used: the measurement of fire temperatures, general observations of changes in plant communities and monitoring the fate of individual
plants. Each of these approaches will be described and the results presented. Finally,
the discussion will draw upon the results from all three studies.
Methods
Study Sites
of the data were collected in the valley of Daldal, Chimborazo Province, on
M ost
the Cordillera Oriental about 40km south-east of Riobamba (one of the areas
surveyed in Chapter 2). This paramo begins around 3,500m, the lower reaches main-
4. Dynamics of Ecuadorian Paramo Vegetation
134
tamed by regular burning. The grasslands extend thence to more than 4,200m before
the Andes begin their descent to the tropical lowlands of the Amazon Basin.
In general. the vegetation is grassy, dominated by Calamagrostis sp. [251] tussocks.
The lower 'Aram° also contains shrubs (including members of the genera Lupinus,
Brachyotum, Chuquiraga, Baccharis, Pent acalia, Gynoxys and Pemettya), a number of
grasses (notably Paspalum sp.) and a large plant of the Cyperaceae, Uncinia
phleoides. These plants extend up to around 3,750m. At around 4,000m the tussock
grass is largely displaced by large cushions of Plantago rigida.
Throughout the Paramo de Daldal, burning is a major feature of land management
by local farmers. Bolivar Coronel, the owner of the land on which the study was carried out, burns areas of paramo at least every three years, but the practice appears to
be somewhat erratic: burning is carried out according to the appearance of the vegetation and only if the weather conditions are suitable. Usually, only one match is
needed to start a blaze (if the fire base needs to be widened, pieces of tussock are
used to carry the flames from one spot to another). The fire is left to extinguish itself.
The experiments to measure fire temperatures were carried out near Laguna
Luspa in El Area Nacional de RecreaciOn Cajas on the Cordillera Occidental (near
Cuenca), and above Laguna de Hoyas in the Paramo de Guamani on the Cordillera
Oriental (not far from the road between Quito and Baeza). Tussocks of Calamagrostis sp. [251] dominate the vegetation in both areas, in much the same way they do in
Daldal. In Cajas, agriculture is restricted to certain valleys within the national recreation area, but many areas are subject to acts of vandalism by tourists and fishermen
(Ramsay, 1988). The Paramo de Guamani, a much wetter area than Daldal or Cajas,
still appears prone to agricultural burning below 4,000m.
Temperatures during 13 6ramo Fires
Two experimental burns were conducted to determine the fire temperature in the
vegetation structure during a fire. The first burn was carried out beside Laguna
Luspa in Cajas in September 1985, the second above Laguna de Hoyas in the Paramo de Guamani in November 1987.
In both of these experiments, THERMOCHROM ® crayons by A.W. Faber-Castell
were used. Each crayon contains a pigment which changes colour at a set temperature: by using a number of crayons containing different pigments, a range of temperatures were encompassed. A set of 18 such crayons covering a temperature range
from 65°C to 670°C was used (Table 4.1).
135
4. Dynamics of Ecuadorian Pdramo Vegetation
GLASS
PLATE
NA
STEEL
PLATE
(C)
CRAYON
SECTIONS
ALUMINIUM
CORRUGATE
Figure 4.1.
crayon pyrometer construction. For the experiment conducted in
the P6ramo of Cajas, thin slices of crayon were sandwiched between a steel baseplate and an upper plate of thermal glass (A), secured with wire (B). In the P6ramo
of Guamani, crayon slices were held between two corrugated sheets of aluminium
(C).
THERMOCHROM ®
—.
....................,
—.
-
100 cm
1
1
I
-
75
I
I
2
50
I
I
3
-
4
I
1
S
6
1
1
1
7
Figure 4.2.
The positioning of the pyrometers in the tussock structure. Pyrometers were
placed within tussocks at approximately 1 m (position 1), 650 mm (position 2) and
350 mm (position 3) above ground. Pyrometers were also placed within the dense
tussock bases approximately 50 mm above surrounding ground level (position 4),
and at the edge of tussock bases (position 7). In the spaces between tussocks, pyrometers were placed on the ground surface (position 5) and 20 mm below
ground (position 6). It should be noted that pyrometers at position 6 were not
placed directly beneath pyrometers at position 5 —it is shown this way for diagrammatic purposes only.
-
25
Ground
Level
4. Dynamics of Ecuadorian Paramo Vegetation
Colour No.
2815/65
2815/75
2815/100
2815/120
2815/150
2815/175
2815/200
2815/220
2815/280
2815/300
2815/320
2815/350
2815/375
2815/420
2815/450
2815/500
2815/600
2815/670
Original
Colour
Pink
Pink
Pink
Light Green
Green
Violet
Blue
White
Green
Green
Green
Yellow
Pink
White
Pink
Brown
Blue
Green
Changed
Colour
Blue
Blue-Green
Blue
Blue
violet
Blue
Black
Yellow
Black
Brown
White
Red-Brown
Black
Brown
Black
Black
White
White
136
Temp. at which
Colour Changes
65°C
75°C
100°C
120°C
150°C
175°C
200°C
220°C
280°C
300°C
320°C
350°C
375°C
420°C
450°C
500°C
600°C
670°C
Table 4.1.
THERMOCHROM ®
crayon information, calibrated for an exposure time of 30 minutes
(data from manufacturer).
These crayons were developed for use in industry to detect the temperatures of
pre-heated hot bodies within about two seconds. The pigments they contain are the
same as those used in thermocolour paints but are mixed with various waxes and extruded in strand form.
In Cajas, thin slices of the crayons were sandwiched between a stainless steel plate
and a thermal glass plate, held in place firmly with wire (Figure 4.1 a, b). This
allowed the pigments to be viewed, whilst protecting them from direct flames and
smoke. In the Paramo de Guamanf aluminium corrugates (cut from the sheets used
to ventilate plant presses) replaced the steel and glass plates (Figure 4.1 c). These pyrometers were positioned in six parts of the vegetation structure (Figure 4.2). The pyrometers in positions 1, 2 and 3 were suspended by wire in the leaves of the grass
tussocks, those in position 4 were placed in amongst the dense bases of the tussocks.
They correspond to 1000 mm, 650 mm, 350 mm and 50 mm above ground level respectively. Positions 5, 6 and 7 correspond to the intertussock region, 20 mm below
ground in the intertussock region and the tussock/intertussock boundary, respectively. In Cajas, three replicates were used in each of positions 2 to 7. The base of one of
the tussocks was then lit and the fire's progress recorded. In the Paramo de Guamanf, nine replicates were used in position 1 with three replicates in each of positions 46.
4. Dynamics of Ecuadorian Paramo Vegetation
137
Qualitative Observations on the Recovery of
Grass Paramo from Fire
The dominant species in the Paramo de Daldal is Calamagrostis sp. [251] and is
crucial to the functioning of the community. The dense, 1m-tall tussocks that this
grass forms make its detailed study difficult. The short time available for the fieldwork ruled out the possibility of all but the most basic of investigations. A sample of
tillers at 3,750m provided a figure for the average number of leaves per tiller. The
ratio of live to dead leaves is of interest and was examined by random samples of a
number of tussocks at 3,750m and 3,950m. A recently burned area provided an excellent opportunity to accurately map tussock bases in a 25m 2 area, with subsequent
measurements of basal area of the species involved.
Near to this sampling area, two recently burned tussocks were randomly selected
and 40 tillers tagged using small plastic rings. The survivorship of these tillers was
monitored. In the same area, general notes were made in June and July 1987 on the
recolonization of burned areas and some quantitative data collected on the number
of colonists on newly burned tussocks and older, established ones. In addition, the relationship between Calamagrostis tussocks and Paspalum sp. was noted in this
burned area and in adjacent recovered vegetation.
Transition Matrix Experiments
The basic unit of a plant community is the individual plant, and to study dynamics
at this level requires a sampling technique with sufficient resolution to differentiate
between individuals. It must also be capable of recognizing gaps which might be important in the regeneration process (Grubb, 1977). Empirically, 100 mm2 was found
to be adequate, containing one plant only with reasonable consistency. Therefore,
this area was used as the basis for this part of the investigation. Single plant modules
(tillers, etc.) were treated as individuals for vascular plants, but grouped together for
mosses.
To record changes in occupancy in the intertussock areas, 1 m x 10 mm belt transects were used, each containing one hundred 100 mm microquadrats in which the
presence of individuals was noted, similar to the approach adopted by ThOrhallsdOttir (1990).
For each sampling plot, three times the number of transects ultimately required
were mapped and the transects to be used were randomly selected from them.
In order to identify the positioning of the transects with the degree of accuracy required, two wooden pegs were securely fixed in the ground, 50 mm from the beginning and end of the transect (primary and secondary pegs, respectively). A tight
string from peg to peg marked the exact line of the transect and the starting point,
measured from the primary peg. The 10 mm wide transect lay to the left of the string
138
4. Dynamics of Ecuadorian Paramo Vegetation
Primary
..40.......••• Post
Secondary
Post
--\..........„_6...
St ring \\1/4
...::::...,:::::::....:;.:;;;:..;•:-..;::::.:::.:::::::::•:.--;;;S•fi.::::•.Y..-:
•:.':••••••••••%."...:::::•::::•.i.::.:::::'..*-:.:-::::::•::::•:::;!....-....::::.:::::::.-..
•• ......./...
• .**.:.:::•':•*.i;.::::............
Start of
/ Transect
•
''
::•:.•:11.:::::: .. '':''?..*;:::::..::.
-1Vlicroquadrat"
End
Figure 4.3.
The layout of the transects used for the transition matrix studies.
running from primary to secondary post (Figure 4.3). The pegs were located with the
aid of sketch maps of the tussock bases in the immediate vicinity of the sampling
plots.
The species occupying each microquadrat were noted. Only rooted individuals
were recorded and sampling units in which nothing was rooted were defined as gaps.
Both mosses and vascular plant species were registered, though in the case of the
bryophytes, their small size called for a different approach: the presence of a
bryophyte species in a microquadrat was counted as one individual, regardless of the
actual number of individuals there. If one of the larger plant species was rooted
across several microquadrats, it was recorded for each of those quadrats (for
example, the large tap root of Hypochaeris sessiliflora can span 30-40 mm). No special treatment was made for clonal individuals: if rooted in the sampling units they
were recorded as individuals. This is an important point since many paramo plants reproduce vegetatively (for example, A zorella pedunculata, Paspalum sp and Geranium
multipartitum). Where more than one species occurred within a unit, the frequency
was recorded as a fraction of 1. Thus, two individuals in the same microquadrat each
received a score of v2. Using this method, the total frequency for each transect always
added up to 100. Frequency measures for each species were obtained by summing
their frequencies in each transect.
The pattern of replacement of species within the sampling plots was analysed by
constructing a matrix such that the rows represented the species recorded at the start
of the study period, the columns the species present at the end. The j th column of the
.th
i row represents the number of microquadrats where species i has been replaced by
4. Dynamics of Ecuadorian Paramo Vegetation
139
species j. It is then possible to test the individual cells in the matrix, using Chi-Square
(x 2 ) analysis, to determine whether the pattern of replacement is random or not.
The analysis is complicated by an assumption inherent in the x2 test. Consider the
case where a microquadrat is occupied by a certain species both at the start and the
end of the experiment. The x 2 test assumes that the last individual has replaced a
member of the same species over the course of the experiment. Since the experiment
was conducted within the lifetime of many plants, this is probably not the case: the
same individual has probably persisted during the time interval. This would result in
an over-estimation of the frequency with which a species replaces one of its own
kind, and may disguise the actual changes taking place elsewhere. This situation is
undesirable and therefore the princi3a1 diagonals of the matrix (representing "no
change") were eliminated from the x test. Of course, this hides any replacement of a
species by another individual of the same species and no probabilities are available
for such transitions. Another assumption of the x 2 test is that every change of occupancy is a single transition. Bearing in mind the brevity of the experiment this is a reasonable assumption: the case of an individual replacing another then being itself
replaced is unlikely.
Simply stated, the x2 test will determine the probability that the pattern of replacement observed is completely random. If for each species pair, the species present at
time 1 is called the itil species and the species present at time 2 the jth species, then
the null hypothesis states that "species i will be replaced by species j in that proportion which the total replacements made by species j contribute to the overall number
of changes" or:
Eg =
E (nil-- nil) x (n4— nii)
E (n4— nil)
where Y represents all species other than i or j, 'nil.' the total number of times
species i is followed by all other species, 'nu' the total number of quadrats occupied
by species i at time 1 and time 2, `nd' the total number of times species j follows all
other species, and 'nil' the total number of quadrats occupied by species j at time 1
and time 2 (ThOrhallsdOttir, 1990). Put another way, the expected value is:
Eij =
Total number of quadrats x Total number of quadrats
h
invaded byj species
vacated by ith species
Grand Total of All Changes
provided the diagonal terms (the species replacing themselves) in the matrix are subtracted before making the calculation. An example calculation is provided in Appendix 2.
Most of the species involved in the data were rare and to avoid bias in the x 2 2
values those species with an expected value less than 5 were not subjected to a x
test. The rarer species were treated as a group to overcome this problem. Yates' correction for continuity was applied (Zar, 1984).
4. Dynamics of Ecuadorian Paramo Vegetation
140
Experiments were set up at 3,750 m and 3,950 m the Paramo de Daldal. For each
altitude, observations were made between tussocks in three vegetation types:
• Control areas where burning had not taken place for a
number of years. Ten transects were recorded at 3,750 m
and six at 3,950 m.
• Recently burned areas. Five transects were recorded at
3,750 m and three at 3,950 m.
• Artificially bared ground (prepared by removing the top
few centimetres of the ground surface, exposing the bare
soil beneath). Five transects were recorded at 3,750 m and
three at 3,950 m.
Data were collected at the beginning of July 1987 and again at the end of October
of the same year.
Results
Temperatures during Pâramo Fires
Table 4.2 presents the maximum temperatures reached within the typical vegetation structure during experimental burns in two different locations. The chromatic
thermometer crayons used in the construction of the pyrometers are said to be accurate to 5°C, but since the colour changes are a function of time, and because the casing of the crayon slices may shield them slightly from radiated heat, it is suggested
that 10°C is more appropriate in this instance. Fire temperatures were maximum in
the upper leaves of the tussocks, with temperatures over 500°C. Temperatures
greater than 420°C occurred just 350 mm above the ground surface, but 250 mm
lower in the dense tussock bases, the temperature was much lower: often less than
65°C and with a maximum of 100°C. The edges of the tussock bases at ground level,
however, were subjected to much higher temperatures similar to those midway up
the tussock (375-420°C).
4. Dynamics of Ecuadorian Paramo Vegetation
Position
141
Maximum Temperature (°C)
Cajas
Guamanf
1. Top of Tussock (1000 mm above ground) 420-450
420-450
420-450
420-450
420-450
450-500
450-500
500-600
500-600
2. Inside Tussocks (750 mm above ground) 350-375
420-450
450-500
3. Inside Tussocks (350 mm above ground) 420-450
420-450
420-450
4. Tussock Bases (50-100 mm above ground) <65
<65
<65
<65
75-100
100-120
5. Intertussock (ground level)
350-375
75-100
350-375
65-75
220-280
100-120
6. Buried (20 mm below ground)
<65
<65
<65
<65
<65
<65
7 Intertussock/Tussock Interface (ground level) 375-420
375-420
375-420
Table 4.2.
Maximum fire temperatures, obtained from THERMOCHROM ® crayon-based pyrometers, in the paramo grasslands of El Area Nacional de RecreaciOn Cajas and
the Paramo de Guamanf. The positions of the pyrometers in the tussock grass
structure is shown in more detail in Figure 4.2.
Qualitative Observations on the Recovery of
Grass Paramo from Fire
Figure 4.4 presents a detailed map of tussock bases in a 5 x 5 m plot at 3,750 m. Calamagrostis species tussocks dominate the area with a basal area of approximately
0.2871 m 2 IT1-2 or about 29% of the total. Uncinia phleoides (Cyper,ae2) is much
less important tussock-former with a basal area of around 0.0327 m ni or around
3% of the total area. In all, the bases of these two tussock species account for roughly
32% of the ground surface in the sample plot. It should be remembered that this is
basal area and not the area shaded by the plants' leaves: this often exceeds 75% at
this altitude.
From a random sample of tillers taken from a single mature tussock, each tiller
has on average 2.86 leaves per tiller (range 2-4; sd 0.7827). This species does not
shed its dead leaves but retains them amongst the living ones. At 3,750m, just over
half of the standing leaves are dead (54.2%) and decay within the tussock itself. This
reduces the photosynthetic potential of the plant, with dead leaves shading out living
4. Dynamics of Ecuadorian Pâramo Vegetation
142
Figure 4.4.
Detailed map of tussock bases in a recently burned 5 m x 5 m plot at 3,750 m in
the ['Aram° de Daldal. The enclosed areas represent the tussock bases, mostly
belonging to Calamagrostis sp. (clear) with some Uncinia phleoides (dotted). The
positions of the five transects used to sample the small-scale changes in the community are also shown.
tissue, but may serve as a defence against predation by herbivores by decreasing the
overall nutritional value of the leaves (Schmidt & Verweij, 1992).
Continual burning of the tussocks and the destruction-renewal cycle that results
can produce cyclical patterns of species dynamics. One example of this is the interaction between Calamagrostis tussocks and Paspalum sp. Areas of paramo at 3,750m
4. Dynamics of Ecuadorian Nramo Vegetation
143
which have not been burned for a number of years do not possess much Thspalum.
This plant has a growth form which is highly suited to opportunistic vegetative spread
after a fire (Figure 4.5). The addition of each new leaf moves the growing point
along the ground and invades new territory.
Although Paspalum does not survive burning well, some individuals remain after a
fire and grow rapidly, utilizing the abundant nutrients released by the fire. Owing to
this response to burning, Paspalum has been noted as a characteristic plant of burned
areas (Cleef, 1979; Ramsay, 1988). The species favours drier ground and so grows
onto the tussock bases. Once there, it rarely descends back into the intertussock
spaces (Figure 4.6). After a time, Paspalum `stolons' come to cover much of the tussock, suppressing the recovery growth of the Calamagrostis tillers. Other species are
then able to colonise the tussocks, among them Rumex acetosella, Disterigma empetrifolium, Geranium multipartitum, Oxalis sp. and Lachemilla sp.
At this point two possibilities exist. Burning may occur again, soon after the first
fire. Since Paspalum does not survive fires well, this may be to the advantage of the
Calamagrostis particularly since the temperatures produced by a fire at this stage are
not so great. However, the tussock base can be damaged by repeated burning and
may start to crumble. If Calamagrostis has become so weakened by repeated burning
and competition from Paspalum and the others, the tussock may die. New tussocks
are formed in the intertussock zone by the multiplication of any surviving fragment
of the original tussock or by seed. The hummock left behind is gradually broken up
as it dries and as the old culm bases decay. Paspalum becomes less important as the
other species suppress its growth.
The other possibility is that the Calamagrostis is sufficiently resilient to resprout
successfully over much of the tussock base and force Paspalum towards the sides of
the tussock, by blocking the light to its leaves. In this way, the cycle is completed.
A number of observations were made on the recovery of Calamagrostis tussocks
after a burning episode. Immediately after a fire, new leaves begin to sprout from the
charred tussock base. Many of these first leaves are damaged towards their tips and
soon wither. However, re-growth continues with the appearance of many more
leaves, borne from tillers produced after the fire.
Two weeks after a burning episode, the tussock gives the appearance of relatively
straightforward recovery: new leaves rapidly replacing those lost to the flames. Tiller
ringing at this stage revealed that mortality is extremely high, with 37.5% of ringed
tillers dead five weeks later, 40% dead after ten weeks and 72.5% dead some fifteen
weeks after ringing (Figure 4.7). In fact, twenty weeks after the fire, the tussocks
were very similar in appearance to that only two weeks after the event, such is the effect of this mortality.
Clearly, this long-term exposure of the tussock base to light makes colonisation attempts by other species possible. Table 4.3 presents data collected at 3,750m in the
Pdramo de Daldal. Burning allows species to colonise the tussock by removing the
leaves that block light. Lupinus sp. shrubs were commonly observed growing in established tussocks. Following a fire, germination of seeds already present in the tussock
144
4. Dynamics of Ecuadorian Pâramo Vegetation
0
Direction
of Growth
Figure 4.5.
The habit of Paspalum sp.
Ca amagrosns Tussock
Paspa urn
Raised Tussock Base
RECOVERY
+
RECOVERY
BURNING
New Tussock from Fragment
of Old Tussock
N
New Tussock from Seedling
Ca arnagrostis Regrowth
\
i
Paspalum Invades Raised
Tussock Base
i
H
___P11411aL
'
L_
Crumbling Tussock Base
f
NO BURNING
INITIAL RECOVERY
Ilp
REPEATED BURNING
4/
Ca/amagrostis Regrowth
Figure 4.6.
The dynamic relationship between Calamagrostis sp. and Paspalum sp, mediated
by fire at 3,750 m in the Pâramo de Daldal.
145
4. Dynamics of Ecuadorian Paramo Vegetation
bases takes place: an average of around 37 seedlings per square metre of tussock was
found. Stellaria leptopetala shows similar behaviour. Mature individuals of both these
species were found exclusively in and around the tussocks and seedlings were not
present elsewhere.
Burned Tussocks
Tussock dimensions (cm)
Tussock area (cm 2)
Unburned Tussocks
91x56
5096
77x52
4004
60x60
3600
40x40
1600
40x40
1600
47
28
6
1
B
B
B
B
15
15
3
5
8
2
B
2
B
2
Rumex acetosella
Lupinus sp-.
Stellaria leptopetala
Dryopteris sp. [1066]
Lachemilla rupestris
Hydrocotyle bonplandii
Lachemilla orbiculata
Stachys elliptica
Vicia sp. [144]
Geranium sibbaldioides
Pentacalia arbutifolius
Paspalum sp. [103]
Relbunium croceum
4+B
6+B
B
87x35
3045
1
1
B
2
4
B
1
1
1
1
1
Table 4.3.
Tussock colonisation immediately after a less severe burn in relation to plants occupying mature tussocks. Numbers shown are the number of individual seedlings
of each species colonising the tussocks. 'B indicates species colonising the outer
portion around the sides of the tussock base.
Percentage Survivorship
100
10
I
o
5
10
15
20
Weeks after Fire
Figure 4.7.
Survivorship of Calamagrostis sp. tillers following a paramo fire. The survivorship
axis is on a logarithmic scale.
1
4. Dynamics of Ecuadorian Paramo Vegetation
146
The long period of tussock recovery after a fire gives the seedlings time to reach a
size where they can compete effectively for light once the tussock begins to grow
again.
The burned tussock is also open to opportunistic colonization. The behaviour of
Rumex acetosella is a good example of this. Within days of a fire, seeds of this species
germinate all over the tussocks. One tussock was seen to have 47 seedlings of
R.acetosella present. Unlike the other species mentioned, Rumex is not found exclusively on tussocks; in fact, it is quite rare in established tussocks of Calamagrostis and
much commoner between them.
Species such as Lachemilla rupestris, Lachemilla orbiculata and Stachys elliptica are
able to take advantage of the bare ground around the tussock bases which is suddenly opened to sunlight after burning (Table 4.3). They were also present on the tussocks but are eventually killed as the tussock leaves overshadow them.
Some 250 m higher in the 'Ammo, at around 4,000m, tussocks and mats are codominant in the community and burning is less frequent. As altitude increases, the
tussocks of Calamagrostis are gradually displaced by Plantago rigida cushions until
eventually the tussocks are well spaced out. At 4,000m, a co-dominance exists, with
the cover tussocks and mats (covering the intertussock region) more or less equal. It
is here that some insight into the processes in operation might be gained. From observations of vegetation showing different stages of development, the following dynamic process is proposed. First of all, large tussocks of Calamagrostis are invaded by
a mat of Plantago rosettes, initially just a 'dent' in the tussock (Figure 4.8). This may
be the result of opportunistic growth following a fire, but since such events are rare
in this location, it was not possible to verify this by direct observation. Another possibility is that the tussock base develops to the extent that water may become less accessible and the plant's growth is weakened. Having gained a foothold, the mat
spreads across the top of the tussock base (which is raised above the surface of the
ground by up to 500 mm), splitting the original tussock into smaller ones around the
periphery of the mound. By this stage, a Plantago rigida cushion has developed. Dissection of ten large cushions of this species revealed that they were all overlying former grass tussock bases.
At this point, other species are able to colonise the cushion covering the centre of
the hummock: Lachemilla orbiculata, Cotula ? mexicana, Oreomyrrhis andicola and
Lachemilla andina. These species are common in the intertussock region. Some
species, rarely encountered in the low, intertussock depressions are relatively common on the Plantago hummocks, namely MIsterigma empetnfolium, Pernettya prostrata and Lachemilla hispidula. These species almost certainly benefit from the better
drainage afforded on the mound, but which was previously shaded by the tussock
leaves.
With time, the hummock becomes colonised by more species. Plantago rigida
cover drops from close to 100% to around 40%. In addition to those species
already mentioned, Hypochaeris sessiliflora, Rumex acetosella and Festuca sp. are
all later colonists. As more species invade the hummock, it begins to dry up. At first
the sloping surfaces become uneven and crumble. Finally, the P rigida rosettes
147
4. Dynamics of Ecuadorian Paramo Vegetation
die and only moss species are able to survive on the crown of the hummock. Eventually, this too dies and the mound disintegrates.
While this process is going on, the fragments of the original tussock survive nearby
and, given the right conditions, are able to reproduce and gain in size. One or several
of these patches of Calamagrostis may attain full size and begin building a new
mound by repeated tillering on top of dead culm bases. After some time, P rigida
may invade once more to begin another cycle.
PLAN VIEW
Plantago rigida Rosettes
Calamagrostis Tussock
Dense Base of Tussock
SIDE VIEW
Plantago Cushion
Remnants of
Former Tussock
C2:3
11114,
.440
Vel
A**
oe•
PLAN VIEW
New Tussock
Develops
from Remnant
Prostrate Shrubs (Pemettya
prostrata, Disterigma empetrifolium)
Plantago Cushion
SIDE VIEW
Cushion Disintegrates and Dies
Figure 4.8.
The dynamic relationship between Calamagrostis sp. and Plantago rigida at
4,000 m in the Paramo de Daldal.
4. Dynamics of Ecuadorian Paramo Vegetation
148
Transition Matrix Experiments
We have already seen that the intertussock vegetation accounted for approximately 68% of the ground cover at 3,750 m. The contributions of the species present in
the microquadrat transects were calculated for both sampling times using frequency
measures (Figure 4.9). In July, 32 species were recorded in the 1000 microquadrats,
rising to 35 species in October.
Much of the intertussock region at 3,750m was bare ground, with 51.0% of microquadrats unoccupied in July, and 40.1% in October. A zorella pedunculata was by far
the most important plant species beneath the Calamagrostis tussocks, accounting for
25.7% and 27.8% of the quadrats in July and October respectively. This species was
even more influential in open areas around 3,200-3,500m in this valley, where montane forest had been cleared (see Chapter 5). The other species present were much
less frequent, the most abundant being Paspalum sp. (4.2 and 6.5% in July and October, respectively).
The dynamic interactions between these species is most interesting. Table 4.4
shows the number of transitions occurring between the main species in the intertussock vegetation. During the study period (115 days), a remarkably high 36.9% of the
microquadrats showed a change of occupancy. 61.2% of all unoccupied 100 mm2
areas remained so throughout. x 2 analysis showed that most of the transitions were
explained by random replacements of one species by another. However, some transitions were found to depart from randomness (Figure 4.10).
42.7% of all changes involved gaps being replaced by A zorella pedunculata or vice
versa. This pattern is characteristic of a short-lived ephemeral species, an opportunist
which invades bare ground quickly and vacates it after its short lifetime ends. However, A. pedunculata is not an ephemeral. In fact, it is a `k-selected' species forming
mats of tough rosettes borne on thick, woody rhizomes (Figure 4.11). The A . pedunculata plant can be viewed as a 'raft' of rosettes on the soil surface, rather like corks
on water. As old rosettes die, the raft is rearranged to fill the gaps. Similarly as a new
rosette grows, the neighbouring rosettes are "reshuffled" to accommodate it by the
turgor pressure of the new growth. Thus, the plant is able to make use of its surface
area extremely efficiently. It is not clear whether this process can continue indefinitely: senescence may occur, the mobility of the rosettes lessened as the layer immediately beneath them becomes cluttered with decaying rhizomes.
Five 1 m transects were used to sample burned intertussock areas not far from
those samples just described. Figure 4.12 shows the principal species in the intertussock zone, notably A zorella pedunculata, Lachemilla orbiculata, Hydrocotyle bonplandii and V iola humboldtii. Of the five hundred 100 mm2 samples, 59.0% were devoid
of plants. In total, nineteen species were present in the intertussock community samples one week after a fire (double this sample size revealed 32 species in nearby unburned vegetation at the same time). Moss species were noticeably infrequent at
1.2%, about one-fifth of that observed in the unburned vegetation.
149
4. Dynamics of Ecuadorian Paramo Vegetation
Species at
Time 1
1 Alch orbi
3 Azor pedu
5 Pasp sp.
9 Cala sp.
14 Relb croc
15 Care tris
24 Gnap pens
29 Dist empe
42 Moss 2
43 Moss 3
44 Moss 4
R REST
G GAPS
T TOTALS
T-D
1
2
12
/
-
3
Species at Time 2
5
9 14 15
-
-
169,4/4 3
3
31
8
5
11/4
1
101/4
24
1/4
29
46
1
1
3
5
R
G
1
21/4 65
6
9
13
5
8
111/4
121/2
1/2
1
1
31/2 2
4
921/2 22
71/2 278
65
51/2 108 1/4 34
42
1
2
12
3
-
1
21/4
17
4
9
4414 151/2 25
311/2 41/4 1516
1
514
.
-
314 3
91/2
61/2
3
11/2
8
3
1
54 14 3
314
2
4
141/2 27 316
9
25
181/2 99 401
31/2 1516 1714 441/4 89
T T-D
3
2561/2
411/2
9
341/2
121/2
131/2
51/2
161/2
101/2
151/2
70 1/2
514
1000
1
861/4
101/2
0
211/2
11/4
1
0
7
101/2
151/2
1514
198
36814
Table 4.4.
Transition matrix for "unburned" vegetation at 3,750m in the Paramo de Daldal.
Codes: 1 Alchemilla orbiculata; 3 Azorella pedunculata; 5 Paspalum sp.; 9 Calamagrostis sp.; 14 Relbunium croceum; 15 Carex tristicha; 24 Gnaphalium aft. pensylvanicum; 29 Disterigma empetrifolium; 42 Moss 2; 43 Moss 3; 44 Moss 4; 46
Moss 6. REST: Hypochaeris sessiliflora, Halenia weddelliana, Rumex acetosella,
Trifolium repens, Gentiana sedifolia, Alchemilla andina, Hydrocotyle bonplandii,
Geranium multipart/turn, Festuca sp., Poa sp., Equisetum bogotense, Holcus lanatus, Plantago major, Oreomyrrhis andicola, Ranunculus pilosus, Aphanactis
jamesonia, Nertera granadensis, Bidens andicola, Geranium reptans, Viola humboldtii, Moss 1, Moss 7, Moss 8, Moss 11. G Gaps; T Totals;
T-D Totals—Diagonals.
125 days later, the same samples in the burned area had 24 species, of which eight
were new to the transects. Bare ground had fallen to 45.8%. The most frequent
species was Hydrocotyle bonplandii, with A zorella pedunculata abundant too. Lachemilla orbiculata, Rumex acetosella and V iola humboldtii were also important elements in the intertussock community at this time.
A high proportion of the samples showed a change of occupancy (47.3%) from the
first sampling one week after the fire to the second one, 107 days later (Table 4.5). A
number of transitions within the matrix were significantly different from that expected by chance (Figure 4.13).
150
4. Dynamics of Ecuadorian P6ramo Vegetation
Gaps
•Aw....w.N~.....+.1E1
Azor pedu
13
Pasp sp.
Cala sp.
Relb croc
Care tris
Gnap pens
Moss 2
Moss 3
Moss 4
Moss 6
Rest
1=11
500 400 300 200 100 0
Frequency (N=1000)
lel
July 1987
,
100 200 300 400 500
Frequency (N=1000)
October 1987
Figure 4.9.
The composition of unburned intertussock vegetation at 3,750 m in the Paramo de
Daldal. The frequency of occurrence in one thousand 100 mm 2 microquadrats
was recorded in July 1987 and 115 days later in October 1987. Full species names
are given in Table 4.4.
Paspalum
sp.
Azorella
pedunculata
Gaps
/..
'Rest'
Figure 4.10.
Constellation diagram showing significant deviations from random species replacements in recently unburned vegetation at 3,750m in the Paramo de Daldal.
The species comprising the 'Rest' are given in Table 4.4. Solid arrows = more
than expected. Dashed arrows = less than expected. * p < 0.05; ** p < 0.01; ***
p<0.001.
4. Dynamics of Ecuadorian Pâramo Vegetation
A
Figure 4.11.
The habit of A zorella pedunculata. A. Cross-section. B. Details of a portion of a
branch. C. Diagrammatic representation of an area of A zorella pedunculata mat.
Rosettes belonging to the same parent plant are the underlying branch system are
indicated. The arrows show the forces applied by new rosettes and the movement
of existing rosettes into spaces left by the death of other rosettes.
151
4. Dynamics of Ecuadorian Paramo Vegetation
Species at
Time 1
1 Lachemilla orbiculata
2 Rumex acetosella
3 Azorella pedunculata
12 Hydrocotyle bonplandii
28 Viola humboldtii
R REST
G GAPS
T TOTALS
T-D TOTALS-DIAGONALS
1
2
3
71/2
-
4
2
2
14
-
12
8
18
16
1014 12
Species at Time 2
12 28
R
G
13
1
451/4 5
2214
4
21/4 314 1
24
51
9
731.4 95
15
281/4 7214 11
31/2 11
314
71/4 24
1
6
-
-
4 1
13 1/2
167
25
531.4229
40 1/4 62
152
T T-D
38
3014
71/2 31/2
8314 381.4
30
714
4
0
274/4
41
296
129
500
23614
Table 4.5.
Transition matrix of replacements at 3,750m in the burned paramo of the Daldal
valley. REST: Trifolium repens, Paspalum sp., Bidens and/cola, Geranium multipartitum, Relbunium croceum, Carex tristicha, Cotula ? mexicana, Plantago linearis,
Gnaphalium aff. pensylvanicum, Holcus lanatus, Disterigma empetrifolium, A phanactis jamesonia, Hypochaeris sessiliflora, Gentiana sedifolia, Geranium reptans,
Plantago major, V aleriana microphylla, Moss 2.
Once again, the reshuffling behaviour of A zorella pedunculata was evident from its
higher than expected replacements of, and by, bare ground (but with a net loss overall). Bare ground was replaced by Hydrocotyle bonplandii and Rumex acetosella more
often than random. As a group, the rarer species showed a very significant mortality
rate ('being replaced by gaps'). This was probably the result of delayed fire damage.
These species tended not to replace Hydrocotyle bonplandii, which in turn replaced
A zorella pedunculata less than expected.
For comparison, Table 4.6 shows the species present in five sample transects located on ground that had been artificially cleared of vegetation. After 15 weeks, 10%
of the microquadrats had been colonized — about two-fifths of these by Rumex acetosella, again demonstrating its opportunistic abilities. V iola humboldtii and Cotula ?
mexicana between them accounted for a further quarter of the occupied quadrats.
A zorella pedunculata was present in only two of the 500 microquadrats; probably regenerated from underground fragments of previous individuals.
153
4. Dynamics of Ecuadorian P6ramo Vegetation
Gaps
Alch orbi
Rume acet
Azor pedu
Hydr bonp
111
Viol humb
Rest
I
I
1
300 250 200 150 100 50 0 50 100 150 200 250 300
Frequency (N = 500)
Frequency (N=500)
1:31 July 1987
October 1987
Figure 4.12.
The composition of burned intertussock vegetation at 3,750 m in the Paramo de
Daldal. The frequency of occurrence in five hundred 100 mm 2 microquadrats was
recorded in July 1987 and 125 days later in October 1987. Full species names are
given in Table 4.5.
Rumex
acetosella
'Rest'
***
Gaps
A zorella
pedunculata
• Hydrocotyle
bonplandii
Figure 4.13.
Costellation diagram showing significant departures from random replacements in
burned vegetation at 3,750m in the Paramo de Daldal. The species comprising the
'Rest' are given in Table 4.5.Solid arrows = more than expected.Dashed arrows
= less than expected. * p <0.05; ** p < 0.01; *** p < 0.001.
4. Dynamics of Ecuadorian Paramo Vegetation
154
Freq. %Freq.
Species
Fiumex acetosella
Viola humboldtii
Cotula ? mexicana
Hydrocotyle bonplandii
Gnaphalium aff. pensylvanicum
Aphanactis jamesonia
Hypochaeris sessilillora
Azorella pedunculata
Carex tristicha
Stachys elliptica
Stellaria recurvata
19
6
6
4
3
3
2
2
2
1
1
1
3.8
1.2
1.2
0.8
0.6
0.6
0.4
0.4
0.4
0.2
0.2
0.2
Gaps
450
90.0
Moss 2
Table 4.6.
Composition of bared ground transects (500 microquadrats) at 3,750m after 108
days.
Similar experiments were carried out at 3,950 m. Figure 4.14 shows the major
plant species found in this vegetation. There were fewer species in this area (27 at
both times) than at 3,750m, though the sample area was smaller (0.06 m2 rather than
0.10 m2). Bare ground accounted for 34.8% of the microquadrats at the start of the
study and 26.5%, 107 days later—much less than that observed 200m lower. A .pedunculata, so important at 3,750m, was not so frequent, occupying less than 5% of the
microquadrats. Mosses were better represented, with a cover around 25% at both
sampling times. Geranium multipartitum, Lachemilla andina and Festuca sp. were all
important members of the intertussock community.
The transition matrix in Table 4.7 shows that 27.7% of the microquadrats were occupied by a different species at the end of the study from the one resident at the
start; this is a much slower turnover than the 36.9% observed 200m lower. In fact,
omitting the unchanged quadrats, only one replacement value departed significantly
from random expectation. As a whole, the category in which all the rarer species
were grouped ('Rest') did not invade gaps as frequently as chance alone would predict (p < 0.001). It is also worth noting that 64.1% of those quadrats which were unoccupied in July remained in this state throughout the course of the study period.
4. Dynamics of Ecuadorian Pg ramo Vegetation
Gaps
Azor pedu
Pasp sp.
Alch andi
Gera mult
Care tris
Fest sp.
Moss 2
CZ=
Moss 6
Moss 9
.1111MINWIt
Rest
01111111MIM1111
250 200 150 100 50
Frequency (N=600)
0
MCI July 1987
50 100 150 200 250
Frequency (N=600)
October 1987
Figure 4.14.
The composition of unburned intertussock vegetation at 3,950 m in the Paramo de
Daldal. The frequency of occurrence in six hundred 100 mm 2 microquadrats was
recorded in July 1987 and 107 days later in October 1987. Full species names are
given in Table 4.7.
155
156
4. Dynamics of Ecuadorian P6ramo Vegetation
Species at
Time 1
3
15
3 Azorella pedunculata
5 Paspalum sp.
11 Lachernilla andina
13 Geranium multipart/turn 15 Carex tristicha
16 Festuca sp.
42 Moss 2
46 Moss 6
49 Moss 9
R REST
3
G GAPS
18
T TOTALS
3
T-D Totals-Diagonals
5
11
13
Species at Time 2
15 16 42 46
1
1
1
171/2
331/2
1/2
2
41/2
1
411/2
1
181/2
1
21/2
Vz
2
1/2
301/2
4
1
2
411/2
2
1
1
1
21/2
2
2
1
2
1
6
2
2
14
13
6
671/2
541/2 551/2 251/2 40
24
61/2 21
14
7
91/2 26
49
R
1
11/2
25
2
36
1
1
12
3
421/2 40
17 1/2 4
1
G
5
1/2
4
3
2
1
2
1
3
4
1
3
5
1
7
41
14
134
731/2 159
321/2 25
T T-D
23
19
38
55
28
37
5514
3314
46
56
209
600
8
11/2
41/2
1314
91/2
61/2
14
814
10
15
75
166
Table 4.7.
Transition matrix for "unburned" vegetation at 3,950m in the Paramo de Daldal.
REST: Lachemilla orbiculata, Calamagrostis sp., Hypochaeris sessiliflora, Rumex
acetosella, Re/bun/urn croceum, Geranium reptans, Pemettya prostrata, Azorella
aretoides, Gentiana sedifolia, Hydrocotyle bonplandii, Trifolium repens, Oreomyrrhis andicola, Ranunculus pilosus, Gnaphalium aff. pensylvanicum, Ste//aria
recurvata, Moss 1, Moss 3, Moss 5, Moss 7, Moss 10.
At 3,950m, bryophytes are responsible for most of the vegetation cover one week
after a fire (Figure 4.15). The 300 microquadrats contained just eight species and
65.3% of the 100 mm 2 units were unoccupied. Fifteen weeks later, bare ground had
dropped to 59.3% and the vegetation comprised 14 species with bryophytes again
dominant. The transition matrix (Table 4.8) indicates the changes of occupancy of
the microquadrats over the 108 days between these sampling times. In all, 45.2% of
the microqtTdrats showed a change, nearly double that of the unburned vegetation
nearby. By x analysis, the rarer species taken as a group were shown to be significant
invaders of gaps (Figure 4.16). The relationship between Moss 6 and gaps was significant in both directions, with the moss both vacating and invading bare ground. This
may be the result of delayed burning-related mortality and opportunistic behaviour
in colonizing gaps.
Moss 6 was shown to be capable of invading bare ground in the study transects located in the artificially bared areas at 3,950m (Table 4.9). 93.6% of the area remained uncolonized after 15 weeks. Hypochaeris sessihflora regenerated from its
thick tap roots.
157
4. Dynamics of Ecuadorian Päramo Vegetation
Gaps
Moss 2
Moss 6
Moss 8
Rest
200
150
100
50
Frequency (N=300)
MI
0
July 1987
50
100
150
Frequency (N=300)
October 1987
200
Figure 4.15.
The composition of burned intertussock vegetation at 3,950 m in the Paramo de
Da!dal. The frequency of occurrence in three hundred 100 mm 2 microquadrats
was recorded in July 1987 and 107 days later in October 1987. The species comprising the 'Rest' are given in Table 4.8.
*
Gaps
4 **,
--7--
Moss 6
1.
'Rest'
Figure 4.16.
Constellation diagram showing significant departures from random replacements
in burned areas at 3,950m in the Paramo de Daldal. Solid arrows = more than expected. Dashed arrows = less than expected. * p <0.05; ** p < 0.01; ***
p<0.001.
4. Dynamics of Ecuadorian Paramo Vegetation
Species at
Time 1
42
8
6
42 Moss 2
46 Moss 6
48 Moss 8
R REST
G GAPS
T TOTALS
T-D TOTALS-DIAGONALS
15
29
21
46
Species at Time 2
G
48
R
8
121/2
3
1
14
1
5
8
32
5
1
132
178
46
7
1
5
261/2 414 18
50
14
29
371/2 7
24
158
T T-D
26
56
15
7
196
300
18
431/2
8
2
64
13514
Table 4.8.
Transition matrix of replacements at 3,950m in the burned paramo of the Daldal
valley. REST: Lachemilla orbiculata, Rumex acetosella, Azorella pedunculata, Hydrocotyle bonplandii, Geranium multipartitum, Cotula ? mexicana, Geranium reptans, Plantago linearis, Azorella crenata, Viola humboldtii, Sibthorpia repens,
Moss 5.
Freq. %Freq.
Species
Hypochaeris sessiliflora
Moss 2
Moss 6
Hydrocotyle bonplandii
Lachemilla andina
Aphanactis jamesonia
7
5
4
1
1
1
2.3
1.7
1.3
0.3
0.3
0.3
Gaps
281
93.9
Table 4.9.
Composition of bared ground transects (300 microquadrats) at 3,950m after 107
days.
Discussion
The pdramo sites used for these studies were subjected to burning practices similar to those reported throughout the paramos (Smith & Young, 1987b; Laegaard,
1992; Verweij & Budde, 1992) and in other tropical alpine grasslands (Smith, 1975;
Beck, Scheibe & Schulze, 1986; Veldzquez, 1992). Generally, in a single paramo
area, there exists a mosaic of vegetation representing different periods of recovery
from burning which is clearly visible in the colouration of the vegetation.
The temperature distribution within the vegetation structure during a paramo fire
has great implications for plant survival. Firstly, all plant material forming the upper
part of the tussock is destroyed during a fire by high temperatures, sometimes approaching 600°C. Clearly, growth forms with unprotected meristems in or above this
zone, such as upright shrubs, may be disadvantaged by burning (Hedberg, 1964). A
number of species are able to regenerate from their roots. In the Paramo de Daldal,
such species included: Monnina crassifolia, Brachyotum ledifolium, Pentacalia andicola, Hypericum laricifolium, V aleriana microphylki, Chuquiraga jussieui, V accinium ?
4. Dynamics of Ecuadorian Paramo Vegetation
159
floribundum and Lupinus ramosissimus. Laegaard (1992) reports similar findings for
some of these species. Beck et al. (1986) found that shrubs have a high capacity for
regeneration in the East African tropical alpine grasslands: even after a severe burn
nearly all shrubs regenerated.
The middle part of the tussock also reaches lethal temperatures (350-450°C), and
this is of relevance to species living within the tussock. Such species include climbing
herbs such as Lobelia tenera, Draba sp., V icia ? set ifolia, A geratina azangoroensis and
Oxalis sp., as well as erect herbs like Festuca sp. and Trisetum spicatum.
However, for these species and for Calamagrostis itself, the dense tussock bases
offer protection from the intense heat generated in the canopy above. In both burning experiments, the dense base was subjected to relatively low temperatures, often
below 65°C because the dense leaves of the tussock shielded the inner parts from the
heat. Therefore, those plants with apical or axillary buds or rootstock capable of suckering within this region may be able to regenerate even though parts of the plant
have been lost by fire.
Despite these relatively low temperatures during a fire, the subsequent exposure
of new shoots to the harsh 'Aram° environment can slow regeneration of the tussock. This allows opportunistic species, lying within the tussock as seed, to establish
on the tussock. Some of these young plants will survive only until the tussock canopy
closes (such as Rumex acetosella). Others can survive for longer periods of time and
become established within the mature paramo community, though this may be only a
small proportion of those seedlings which germinated. A good example of a species
of this nature is Lupinus cf. pubescens.
Certain species appear to favour the edge of tussocks, demonstrated by a number
of species in Table 4.3. The distribution of Uncinia pheleoides in Figure 4.4 shows the
pattern well. It is not clear whether these distributions reflect a degree of protection
from fire afforded by the tussock base or are the result of favourable sites for seedling establishment. The latter explanation was proposed by Miller & Silander (1991)
to explain why Puya clava-herculis was frequently found inhabiting the tussock edge
in Ecuadorian paramos.
The intertussock spaces and the sides of the tussock bases may reach temperatures
on the ground of up to 375°C or 420°C if there is a good deal of dry matter in the tussocks. Temperatures of around 100°C are probably commoner in Daldal, where burning is a regular practice. This is also true of more humid paramos such as the Para=
de Guamani, which produced temperatures between 65-120°C in the experimental
burn.
Though temperatures on the ground may be lethal, temperatures remained low 20
mm under the surface (in all cases less than 65°C, the minimum limit for the pyrometers). Therefore, plants occupying the spaces between tussocks may have their aboveground parts destroyed but the subterranean organs are unaffected by the fire. Many
plants can `recolonise' a burned area simply by regenerating from rhizomes or roots
(for example, A zorella pedunculata and Hypochaeris sessiliflora). Other species, including those just mentioned, shield delicate buds within less sensitive plant parts.
4. Dynamics of Ecuadorian Paramo Vegetation
160
According to Laegaard (1992), the apical buds of acaulescent rosettes and cushion
plants are often situated 10-20 mm below the surface. From the results of the experiments described here, this would afford such plants adequate protection from lethal
temperatures.
At 3,750 m in the Paramo de Daldal, the species surviving a fire could be seen to
benefit from the above factors. The commonest survivors following a fire were those
able to resprout from rhizomes or rootstock: Lachemilla rupestris, Lachemilla orbiculata, Hydrocotyle bonplandii, Rumex acetosella, Halenia weddelliana, V iola humboldtii, Sonchus oleraceus, Plantago major, Uncinia pheleoides and Cotula ? mexicana. It is
interesting to note that those species shielding the apical buds with plant parts did
not survive the most intense fire, but were present following a less severe burn. Such
species included A zorella pedunculata and Hypochaeris sessiliflora. It was also evident
that particularly after a less intense fire, small patches of intertussock vegetation
were commonly left alone by the fire. Within these patches, plants which did not appear to survive by means of one of the above strategies were observed. Such fortuitous survival enabled species such as Pczspalum sp. to capitalise on the abundance of
space and nutrients after a fire, as described earlier. Laegaard (1992) confirms these
strategies for the majority of the above species.
The transition experiments corroborated these findings. In addition to the species
mentioned above, the burned transects at 3,750 m also contained Tnfolium repens,
Bidens andicola, Geranium multipartitum, Relbunium croceum, Holcus lanatus, A phanactis jamesonia and one species of moss.
Following this initial survival, remaining bare ground was colonised by opportunistic species from seed. In particular, Rumex acetosella and Hydrocotyle bonplandii
were shown to increase their presence more than chance alone would predict. Hydrocotyle bonplandii was shown to be resistant to invasions by the rarer species, but was
itself less likely to replace A zorella pedunculata rosettes.
By contrast, a highly significant mortality of A zorella pedunculata rosettes and a
number of the rarer species was observed in the transects. This indicates that a number of individuals which survived the fire initially did not persist, perhaps because
they sustained critical damage which could not be repaired.
Although colonisation was taking place, this was a slow process: the proportion of
gaps in the transects decreased from 59% to 46%. However, nearby unburned vegetation showed a similar proportion of gaps (51% and 40%). The data from the control transects suggest that A zorella pedunculata is likely to proliferate at the expense
of many of the other species which have colonised the burned area, and that a further reduction in bare ground is unlikely.
Of the twelve species found in the cleared transects at this altitude, all but two
rare species were present in the burned plots. This indicates that these colonists do
not rely on burning for seed germination, but merely take advantage of disturbance,
whatever form it may take. Interestingly, 90% of the bared transects were unoccupied, compared with 40% and 46% in the associated control and burned plots. Removal of the upper 20 mm of soil has much more serious implications for the
4. Dynamics of Ecuadorian Paramo Vegetation
161
regeneration of intertussock vegetation than a fire. The importance of regrowth from
plant parts just beneath the surface is confirmed.
At 3,950 m, survival following burning was much lower than in the transects just
described. About 65% of the intertussock spaces were bare, and only eight species
survived, with half of these being bryophytes. The diversity of these transects increased in the following 108 days to fourteen species, largely by the invasion of bare
ground (which decreased to 59%). There was a high turnover of species in these
plots, however, mostly the result of delayed mortality in Moss 6. This species also colonised bare ground, and replaced a number of other bryophyte species.
Unburned vegetation at 3,950 m was found to contain almost twice as many
species (in twice as many sample units). The commonest species from the burned
plots were also present in the controls. Bare ground was very low: 35% at the start of
the experimental period and 27% at the end. In this crowded situation, it was found
that the rarer species did not colonise gaps as frequently as chance would predict,
perhaps because competition for resources was high.
When this vegetation was cleared by removing the upper 20 mm of soil, recolonisation was slow (similar to that 200 m lower in the paramo). After 107 days, only 6% of
the ground had become occupied, more than half of this by regeneration by Hypochaeris sessihflora from tap roots. Bryophytes were less successful colonists than in
the burned plots, but the two principal species in the burned transects were important in bared ground. In experimental studies in the Venezuelan paramo, Pfitsch,
Smith & Rodriguez Poveda (unpublished — cited by Smith & Young, 1987b) also
found that recolonisation of bared plots at high altitudes was a slow process. These
plots were colonised by a gradual accumulation of species from the mature community, without early specialist species. In the Paramo of Daldal at 3,950 m, there appears to be some evidence to support these observations (the establishment of
Hypochaeris sessihflora) but other colonists were more opportunist species which
characterised the invasion of disturbed ground at lower altitudes (Hydrocotyle bonplandii and A phanactis jamesonia). However, the climatic conditions at this altitude
were not so severe as those reported in the Venezuelan experiment.
From these studies in the Ecuadorian paramos, a number of generalisations can be
made:
• The temperature during a fire is determined mainly by the
structure of the vegetation. The highest temperatures are
produced in the tussock canopy, the lowest within the tussock base and just beneath the surface.
• Regeneration from below ground plant parts (including
those within the tussock bases) is the main form of recovery from fire.
• The severity of the fire (largely a function of the interval
since the last burn) determines the degree of survival of intertussock species on the ground surface.
4. Dynamics of Ecuadorian Nramo Vegetation
162
• Initial survival does not guarantee persistence in the community. Both tussock grasses and intertussock species show
significant mortality rates in subsequent months.
• Recovery is slower at higher altitudes.
• Burning may induce cyclical patterns of community development, illustrated by the interactions of Calamagrostis sp.
and Paspalum sp. in the NI-am° de Daldal.
• Certain species quickly colonise bare ground by seed.
These may persist to maturity, but most will be killed by
competition from neighbours as the vegetation matures.
• Some species (such as Lupinus cf. pubescens) appear to
rely heavily on burning for establishment within tussocks
(where mature individuals are found).
It is clear that paramo vegetation in Ecuador is able to regenerate relatively rapidly after burning. Similar rates of renewal are inferred from studies in Colombian
pdramos (Pels & Verweij, 1992; Verweij & Budde, 1992). In Chirrip6 National Park,
Costa Rica, recovery was well underway a few months after a huge paramo fire
(Boza, 1978). This contrasts with observations by Janzen (1973), again in Costa Rica,
where regeneration was very slow, with large patches of bare ground still present
three years after the fire.
Transition matrix studies have often been associated with predictions of succession
(Horn, 1975; Usher, 1979, 1981; Noble & Slatyer, 1981; Hobbs & Legg, 1983; Lough
et al., 1987). However, in the present study this approach was considered inappropriate, mainly because of the short timescale over which the observations were made.
The early changes in specific composition during recovery from disturbance are
usually faster than the later changes (Shugart & Hett, 1973), and several of the pressures acting upon individuals are not uniform over time (for example, grazing influence is especially common immediately after a fire —Verweij & Kok, 1992).
Therefore, the fixed transition probabilities demanded by Markov modelling do not
apply to 'Aram° vegetation after a fire (Pels & Verweij, 1992).
Further to this argument, in East African tropical alpine grasslands, the studies of
Beck et al. (1986) indicated that linear succession after burning did not take place.
Rather, burning began a series of complete or incomplete (if burning was repeated
too soon) successional cycles. In order to predict the patterns of replacement during
such cycles, observations are required over the entire simulation period. It is not
clear how long such recovery cycles might be in the Ecuadorian 'Aram°, and complete recovery could involve several decades or more.
Apart from the dynamics directly relating to paramo burning, these studies have
brought to light interesting aspects of the small-scale changes associated with unburned vegetation. Notably, the majority of changes of occupancy in the transition
4. Dynamics of Ecuadorian Paramo Vegetation
163
studies involved gaps and appeared to be random replacements. Similar findings
were made by ThOrhallsdOttir (1983, 1990) in a grassland community in North Wales.
Gaps are known to be important in the dynamics of many plant species, and even
small ones can influence local conditions (Silvertown, 1981). The frequency of gaps
was high bearing in mind the dense appearance of the vegetation (51.0% and 40.1%
at 3,750 m, and 34.8% and 26.5% at 3,950 m, in July and October 1987 respectively).
ThOrhallsdOttir (1983, 1990) reported gap frequencies of 17-60% (mostly around
30%) in North Wales. She also noted that clonal species tended to replace and be replaced by gaps more often than chance would predict. This was found to be true in
the paramo, exemplified by the behaviour of A zorella pedunculata.
The mobility of mats of A zorella pedunculata in many ways resembles the floating
rafts of the Water Hyacinth (Eichhornia sp.—Watson & Cook, 1982). Clearly this
mobile collection of rosettes and underlying rhizomes presents several problems to
other plant species. To compete with A . pedunculata a plant must be capable of resisting the movement of the mat or must itself be flexible to move with the rosettes. Certainly, Paspalum sp. and the rarer species (grouped together) were found to be
significantly less likely to replace anA. pedunculata rosette than chance would predict.
The fact that the main means of spread for Calamagrostis sp. is by vegetative reproduction from large, established tussocks may explain why the mat is unable to exclude the tussock grass. While the A . pedunculata mat may be an able competitor for
space, it may not be so adept at competing for light. The leaves of the grass shade out
the A zorella rosettes, allowing new tillers to develop at the tussock edges.
Grubb (1977) noted the lack of information on Calamagrostis sp. regeneration in
the Andes. Although some seedlings of Calamagrostis sp. were encountered in the
transition matrix studies, these were relatively uncommon and vegetative spread in
kaleidoscopic pattern is the principal means by which tussocks are maintained within
the community.
Some 250 m higher in altitude, and the dominance of Calamagrostis tussocks over
cushion and mat species is lost. The cyclical processes in operation in the boundary
zone between grass and cushion paramos was described earlier and is similar to that
described by Lough et al. (1987) for a New Zealand alpine cushion community.
It is not clear what forces drive this process. However, the existence of an apparently persistent and stably cyclic dynamic relationship implies either an extrinsic environmental cycle (unlikely in the paramo) or a cycle in one of the dominant species
(Horn, 1974). Since both species can exist both above and below this altitude (personal observation), it would appear that some environmental factor or factors result
in a spatial change in relative competitive abilities of the two dominant species. The
lower limit of extensive cushion vegetation elsewhere in the Andes seems to be the
result of interspecific competition (Armesto, Arroyo & Villagran, 1980; Alliende &
Hoffmann, 1985).
4. Dynamics of Ecuadorian Paramo Vegetation
164
It may be, as suggested by Laegaard (1992), that tussocks become overmature and
can no longer supply water and nutrients to satisfy the needs of the plant (water requirements become more difficult to meet at higher altitudes —Meinzer & Goldstein, 1986). The same factor may also explain the reduced vitality of the cushions
and their subsequent decay, also observed in the Colombian paramos by Cleef
(1981). With the competitive ability of the tussock reduced, the cushion invades. Alternatively, it could be that occasional burning at this altitude exerts a significant
stress on the tussocks (fire can spread between the widely-spaced tussocks if it is
windy). Some other factors, relating to climatic or edaphic features of the local environment may also be involved.
The invasive behaviour of Plantago rigida, overgrowing other plants to form cushions is mirrored elsewhere. Nathaniel (1985) describes a similar mechanism in the
formation of Plantago rigida cushions on the slopes of Volcan Cotopaxi, Ecuador,
and W erneria humilis also appears to share this behaviour (personal observation).
Lough et al. (1987) observed an analogous process with different species in New Zealand.
The widespread existence of this transition from grass paramo dominated by Calamagrostis sp. tussocks to cushion pâramo dominated by Plantago rigida implies that
the controlling factor or factors are also widespread. The explanation of this fundamental physiognomic change would provide a valuable insight into the mechanisms operating throughout the pdramo ecosystem.
Clearly, fundamental research into the small-scale dynamics of paramo vegetation
is a rewarding undertaking. Even very brief studies, such as those described here, can
facilitate the interpretation of the large-scale community in terms of the agents that
maintain them.
Chapter 5
An Assessment of Net
Aboveground Primary
Productivity in the
Andean Grasslands of
Central Ecuador
165
5. Productivity of Andean Grasslands in Ecuador
166
Introduction
severity of the high altitude tropical environment has often lead to compariT he
sons with arctic and temperate alpine ecosystems (Bliss, 1971; Tieszen & Detling,
1983). Comparisons between temperate alpine and tropical alpine areas have shown
that, although temperate alpine regions experience more favourable conditions during the growing season, the tropical montane environment experiences a greater
number of degree hours per year (Billings, 1973). From this evidence, Smith &
Young (1987b) suggest that tropical alpine communities may be more productive on
a yearly basis than their temperate counterparts.
The paramos throughout the Northern Andes are used for extensive grazing of
cattle, sheep, horses and mules. In the early 1950s, it was estimated that over half of
Ecuador's cattle and around 85% of its sheep were grazed on the paramos (AcostaSolis, 1960). With the colonisation of Amazonia in recent times, the paramos no
longer contribute such high proportions of Ecuador's cattle production, but they are
nevertheless a critical element of the rural economy in highland regions. In fact, it
seems likely that increasing population pressure, diminishing isolation and the loss of
self sufficiency in many rural communities has led to a recent increase in the head of
domestic livestock on the paramos of Ecuador. A particularly important feature of
livestock is that they represent an investment immune from inflation— a "living bank
account on the hoof" (Brush, 1976).
The previous chapter looked at the dynamic processes associated with burning to
improve forage quality in the paramo. Grazing and trampling pressures are also important influences on plant community composition (Verweij & Budde, 1992). Unfortunately, such pressures have resulted in the degredation of some paramo
ecosystems (Grubb, 1970; Smith, 1981; Acosta-Solis, 1984; Ramsay, 1988; Grubb,
Lloyd & Pennington, unpublished) and the neighbouring ceja andina forests (Brandbyge & Holm-Nielsen, 1986; Laegaard, 1992; Verweij & Beukema, 1992).
Up to this date, there have been few published studies of the productivity of natural communities in any tropical alpine region of the world (Smith, 1987). One study
carried out by Hnatiuk (1978) reported net aerial productivity rates of 1.28-4.42 t ha1 yr-1ini the grasslands of New Guinea. Acosta-Solis (1984) estimated a yield of 4.35 t
of dry matter per hectare per year in the Pdramo de Chiquicagua, Ecuador, though
the methodology described indicates that the value given is in fact an estimate of
standing biomass rather than productivityper se. A table of data for plots in the Oramo of Volcan Antisana, Ecuador, appears in Black (1982). Values range from 2.6021.40 t and plots were subjected to burning, cutting and fertilizer treatments.
However, the reported details of the experimental design (exact method of data collection, sample sizes, whether harvested material was dried before weighing, etc.) are
not sufficient to allow any interpretation of the values presented.
This chapter reports preliminary studies of grassland productivity over an altitudinal gradient in the Andes of central Ecuador. Apart from providing data on standing
crops, comparisons are made between the aboveground net primary productivity of
5. Productivity of Andean Grasslands in Ecuador
Figure 5.1.
The location of the productivity experiments. A. The village of Alao is situated to
the south-east of Riobamba. One exclosure (site A) was located at the Sangay National Park guardpost in Alao. B. The general location of this area is shown in (A)
as a box with dashed lines. Four exclosures (sites B-E) are shown in the upper
Daldal valley.
167
168
5. Productivity of Andean Grasslands in Ecuador
4200 m
Cushion
p.a.ramo
4000 m
Grassy
pa.rarno
3800 m
Shrubby pa.ramo
Montane cloud
3600 m
forest zone
with recent
clearings
3400 m
Old pastures in
cloud forest zone
Cultivated
valley floor
3200 m
......'''---------.-4-Site A
3000 m
Figure 5.2.
Diagrammatic representation of the five exclosure sites used in the productivity
studies in the valleys of Alao (lower valley system with site A) and Daldal (upper
valley system with sites B-E). The slopes have been greatly exaggerated.
5. Productivity of Andean Grasslands in Ecuador
169
areas covering an altitudinal range of nearly 1,000m, from improved pastures in the
valley bottoms to the upper reaches of agricultural use in the grass paramo at nearly
4,000m. Applications of fertilizer were used to assess the potential improvement of
natural grasslands and cutting regimes were applied to simulate grazing and burning.
Methods
Study Sites
were carried out in the highland valley systems of Alao and DalThedal,experiments
about 40-50Icm south-east of Riobamba (Figure 5.1). Five pastures were
chosen to cover the altitudinal gradient and the transition from improved grasslands
to unimproved ones (Figure 5.2).
The Alao Valley, Site A: 3,100m
The valley of Alao has been used intensively for agriculture for many years. Cereals,
potatoes and other crops are grown on the steep slopes of the valley, and animals are
Mean Precipitation (mm)
120
100
80 -
60 -
40 -
20 H
Jan
Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month 1981-83
Figure 5.3.
.Monthly precipitation for Alao, with means calculated from data for the period
1981-1983 (from the lnstituto Nacional de Meteorologia e Hidrologia, Quito, for
station M396 Alao). Temperature records for this period were incomplete.
170
5. Productivity of Andean Grasslands in Ecuador
reared on small pastures and amongst the crops. Alao is also the gateway to Sangay
National Park and a guardpost has been built on the flat valley plain to monitor visitors. The experimental site was located within the grounds of this guardpost. The area
was flat and sheltered to some extent by trees and buildings. The soil showed good
drainage.
The meteorological station at Alao recorded 100 and 147 days of rain in 1981 and
1983, amounting to precipitation totals of 737.7 and 975.3 mm yr -1 respectively. Over
these years, a maximum of 33.8 mm fell during any 24-hour period. The pattern of
rainfall for the years 1981-83 is presented in Figure 5.3. Around 61% of the annual
precipitation falls between March and August, with a peak in March to May. No
other meteorological data was available.
The vegetation consisted of short-cropped grasses (about 5cm tall), which were intensively grazed from time to time by horses and cattle. There were no signs of cultivation, though nearby, vegetable crops were growing in tilled soil.
Site
Species
Dactyhs glomerata
Trifohum repens
Holcus lanatus
Anthoxanthum odoratum
Lolium sp.
Alopecurus sp.
Azorella pedunculata
Carex trishcha
La c hemilla orbiculata
Ranunculus sp.
B/dens and/cola
Gentiana sedifolia
Geranium sibbaldioides
Paspalum sp.
Hydrocotyle bonplandii
Taraxacum officinale
Bromus sp.
Agrostis sp.
Ste//aria leptopetala
Gnaphalium aft. pens ylvanicum
Geranium multipartitum
Cerastium sp.
Bromus lanatus
Festuca sp.
Trisetum spicatum
Festuca sp.
Calamagrostis sp.
Paspalum tub erosum
Cotula ? mexicana
Halenia weddelliana
Senecio pimpinellifolia
Valeriana microphylla
Hypochaeris sessiliflora
A
B
d
+
+
+
+
+
+
C
D
+
+
+
+
+
+
d
d
+
-i-
+
+
+
+
+
+
+
+
+
+
1-
+
+
+
+
+
+
+
i-
+
4
4
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
4-
+
+
d
+
+
+
+
+
+
Table 5.1 (Part 1).
Species list for the exclosures at the five study sites. Site codes: A, 3, 100m; B,
3,250m; C, 3,450m; D, 3,750m; E, 4,000m. "d" represents the dominant species.
1
-14
+
+
4
+
5. Productivity of Andean Grasslands in Ecuador
171
Site
Species
A
B
C
D
+
+
+
+
+
+
+
Rumex acetosella
Hieracium lrigidum
Lachemilla andina
Plantago major
Ranunculus peruvianus
Relbunium sp.
Poa sp.
Agrostis sp.
Gramineae
Gramineae
Sisyrinchium jamesoni
Vaccinium sp.
Equisetum bogotense
Vicia sp.
Lycopodium sp.
Ericaceae
Oreomyrrhis andicola
Pemetlya prostrata
Sibthorpia repens
Disterigma empetrifolium
Oritrophium peruvianum
Cerastium sp.
Azorella aretoides
Mphogeton dissecta
Viola humboltii
E
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Table 5.1 (Part 2).
Species list for the exclosures at the five study sites. Site codes: A, 3, 100m; B,
3,250m; C, 3,450m; D, 3,750m; E, 4,000m.
The stand was characterized by grasses such as Dactylis glomerata, Lolium sp., A nthoxanthum odoratum, Holcus lanatus and A lopecurus sp. Other herbaceous elements included Trifolium repens (Table 5.1).
The Daldal Valley
The remaining four sites were located in the valley of the Rio Daldal (Figure 5.1, b).
A small farming community occupied the lower end of the valley, but the huts above
3,400m were used solely for temporary accommodation. The lower part of the valley
above the settlement of Daldal was once forested, perhaps within the last century.
Some small patches of forest still remained but they were highly modified. Clearance
of these patches was taking place and during the course of the study, a significant area
was cleared for firewood. Forest clearance was carried out more for local fuel requirements than for agricultural purposes. As a result of these activities, this part of
the valley showed a mosaic of forest and clearings of varying ages, some of the older
pastures having been colonized by species characteristic of 'improved' grasslands. In
this zone, some tributaries of the Rio Daldal showed signs of re-routing by man and
irrigation ditches had been skilfully constructed along the contours to channel water
for agricultural and domestic use. Above 3,650m, the forest was replaced by shrubby
grassland (sub-paramo). The upper limit of the forest was gradually being pushed
lower by burning from above. In some parts of the valley only a narrow belt of forest
5. Productivity of Andean Grasslands in Ecuador
172
existed between 3,400-3,450m, a consequence of burning from above and felling from
below. The higher portions of the Daldal valley were covered with paramo grassland,
typical of the highest reaches of the northern Andes. From 3,750-4,000m, the paramo
was dominated by tussock grasses, but above this level, cushion plants became codominant. The vegetation of the Daldal paramo has been discussed in earlier chapters. No meteorological data exists for this area.
The experiments were carried out on the land of Bolivar Coronel, a local farmer,
who owned a stretch of pastures and forest from 3,250m to 4,200m.
Site B: 3,250m
The chosen area was sited on the valley floor, beside the Rio Daldal at 3,250m. The
vegetation at one time would have been naturally forested, but at the time of study it
consisted of short grasses (no more than 3cm high) and other herbs with a substantial
cover of a mat-forming species. The field was used to graze 10 cattle throughout the
year. There was little evidence for significant grazing by rabbits. At one time, a small
rivulet flowed through the pasture to join the R. Daldal, but had since been rerouted. The ground sloped 5° from horizontal with a north-westerly aspect, and no
visible signs of previous cultivation. The pasture had been colonized by some species
characteristic of improved grassland. The dominant species, however, was A zorella
pedunculata, which formed a dense mat over the surface of the ground. Small patches
and individual plants of other species grew through this mat, notably Paspalum sp.,
Hokus lanatus, Dactylis glomerata, Tnfolium repens and Bidens andicola (Table 5.1).
Site C: 3,450m
Amongst the remnants of secondary ceja andina cloud forest in the Daldal valley
were cleared patches of land up to half a hectare in area (but usually much less than
this). Often, the stumps of the once-dominant trees and shrubs were still very much
in evidence, some resprouting to form small bushes. The exclosure was erected in an
area free from such bushes, at 3,450m on sloping ground (8° from horizontal) with a
westerly aspect. Cattle freely grazed, dividing their time between these forest clearings and the field lower down (site B), this latter site being favoured more often.
Rabbit grazing was more important here than at site B.
Like the exclosure at Site B, these plots were dominated by the presence of a mat
of A zorella pedunculata. Other important floristic elements included Paspalum sp.,
Holcus lanatus, Tnfolium repens and Bidens andicola (Table 5.1).
A small area of land not far from the study plot was used to cultivate potatoes,
beans, oca and carrots.
Site D: 3,750m
The fourth exclosure was situated at 3,750m, above the forest patches, in paramo
grassland. The slope of the ground was 18° from horizontal with a northerly aspect.
The plots were exposed to strong up-slope winds. Tussock grasses of Calamagrostis
sp. dominated the vegetation with other grasses and herbs performing a secondary
5. Productivity of Andean Grasslands in Ecuador
173
role within the dense tussocks or in the intertussock spaces. Small, woody plants of
V aleriana microphylla and Lupinus sp. were frequently found amongst the tussocks,
and larger shrubs (Brachyotum ledtfolium, in particular) were locally common, though
absent in the experimental plots (Table 5.1). This type of vegetation is often described as sub-paramo (Cuatrecasas, 1958, 1968). Cattle were allowed to graze freely
in this area, the vegetation being regularly burned (once every 2-4 years) to remove
dead leaves and to stimulate the production of nutritious, young shoots. Rabbit grazing may have been as important as that of livestock, but there was no data available to
support this.
Site E: 4,000m
The highest exclosure was located at 3,900m in grassy paramo or paramo proper
(Cuatrecasas, 1958, 1968). The plot sloped 12° from horizontal towards the southwest, and was more sheltered than the exclosure at 3,750m. Tussock grasses of Calamagrostis sp. were dominant and the only noteworthy physiognomic difference
between this and the paramo 250m lower was the absence of large shrubs and lupins,
though V aleriana microphylla was still common (Table 5.1). This area was visited by
cattle and horses, though less often than the lower vegetation, partly due to old, often
ineffective, man-made earthen walls and ditches (to limit herd movements) and partly because of the unfavourable climate.
Above 4,000m, the vegetation became increasingly dominated by cushions of Plantago rigida, and, as a consequence, unsuitable for grazing livestock. The unfavourable
climate also deterred cattle grazing above this altitude.
Experimental Design
At each of the above sites, 10.5m x 10.5m fenced exclosures were set up (Figure 5.4). For each exclosure, eight 1.5m tall eucalyptus fence posts were driven into the
ground, and joined by wire mesh and barbedwire strand. Sixteen 2 x 2m plots were located within the exclosure, separated by buffer zones of 0.5m between each of the
plots and between the plots and the fence. This design excluded cattle but did not
prevent small mammals such as rabbits and rodents from entering the areas.
The experiment at each of the five sites consisted of an unreplicated 2 4 factorial
plot design, incorporating two fertilizer treatments and two cutting regimes in all
combinations. The sixteen treatments were assigned to plots at random within each
location.
The treatments were a combination of:
5. Productivity of Andean Grasslands in Ecuador
OD,
POST
lx 2M
PLOT
0.5M
BUFFER
ZONE
FENCE
OUARTERS
FOR
HARVESTING
_II— -)-_-_•
AfklAINk;" NOW .
GROH('
LEVEL
Figure 5.4.
The design of the exclosures used in the productivity studies. Each exclosure consists of sixteen 2 m x 2 m plots, harvested by quarter, and separated by 0.5 m buffer zones. The 1 m high perimeter fences of wire mesh topped with barbed wire
strand were supported by eight eucalypt posts.
174
5. Productivity of Andean Grasslands in Ecuador
175
a. Fertilizer treatments (applied at start of experiment)
• No fertilizer
• 174g per plot of nitrogen (46:0:0) in the form of Urea,
corresponding to 200 kg ha -1 of nitrogen.
• 87g per plot of phosphorus (0:46:0), corresponding to 100
Kg ha-1 , and 133g per plot of potassium (0:0:60), corresponding to 200 Kg ha-1.
• Application of nitrogen, phosphorus and potassium in the
above rates (that is, both treatments ii. and iii.).
b. Cutting treatments
• No cuts.
• Early cut at start of experiment only.
• Late cut only after 70 days.
• Cut both at start of experiment and after 70 days.
The experiments were started between the 13-20 t1 July 1987. The initial cutting
treatments (ii. and iv.) enabled the weight of the standing crop to be estimated. Harvesting was carried out by plot quarter (that is, a different quarter of each plot was
harvested at each time) at approximately 35 day intervals (see Figure 5.5). The plots
were cut using small shears. The vegetation was cut back to ground level (or to just
above the hummocky bases of the tussocks in the 'Aram° sites, and to the level of
the A zorella pedunculata mats in sites B and C). In the paramo sites (D and E), much
of the harvested material was dead. However, this was not separated from the live
material. "Biomass" and "phytomass" are used here to incorporate these dead leaves
still attached to the tussocks, but not to include litter. This is only important in previously uncut plots, since the dead leaves take time to accumulate to a significant
level.
Assessment of the net primary production of underground plant parts by biomass
methods are fraught with difficulties. To begin with, it is a destructive process and
cannot be used for repeated measurements of the same area. Also, there is great difficulty in separating recently produced rootstock from older, perhaps even dead, material in the soil. For these reasons, and bearing in mind the limitations on resources,
the assessment of belowground productivity was not attempted.
The harvested plant material was then taken to the Sangay National Park guardpost at Alao for initial drying at 65°C in a homemade oven (to prevent rotting), before being transported to the Politecnico Nacional in Quito for final drying at 105°C
for 24 hours. Weighing was carried out using an accurate balance to the nearest lg.
176
5. Productivity of Andean Grasslands in Ecuador
No cuts
Aw =
a
a
Early cut only
Aw =
a
Pw =
a+b
Late cut only
Early and late cuts
AIN= a+b
START &
EARLY CUT
SECOND
HARVEST&
FIRST
HARVEST
END &
THIRD HARVEST
LATE CUT
TIME NE•100Figure 5.5.
Diagrammatic representation of the cutting and harvesting regimes. At the start of
the experiments, half of the plots were cut. The dry weight at the start of the experiment is shown by a small circle. Harvesting was carried out twice during the
course of the experiments and once at the end of the study period. The overall
change in weight (SW) was calculated according to the formulae shown.
177
5. Productivity of Andean Grasslands in Ecuador
Results
Standing Crops
Estimates of the dry weights for the initial aboveground biomass (standing crop)
of the five exclosure sites at the start of the experiments are given in Table 5.2. The
most striking feature of these dlta is the very highvalues for standing crop in the
paramo exclosures: at 794 g rn - and 837 g m 2, some 15-27 times larger than the
values for the other three sites. The upper four exclosures have roughly the same degree of variability about their means, about half that of the lowest site.
Site Altitude
A
3,100m
B
3,250m
C
Pasture Type
Improved/
Agricultural valley
Standing 95% Confidence
Crop (g m -2)
Limits
CV%
54
40a
87.9
Semi-improved/
Cleared ceja
42
12a
33.3
3,450m
Semi-improved/
Cleared ceja
31
8a
33•5
D
3750m
Unimproved Lower
paramo grassland
837
161 b
23.0
E
4,000m
Unimproved Upper
piramo grassland
794
219b
33.1
Table 5.2.
Standing above-ground biomass estimates collected 13-20 th July 1987 from five
exclosure sites at five altitudes in the Andes of Central Ecuador. Based on 8 observations per site of (a) 4m 2 and (b) 1m2.
The increments of dry weight from harvest to harvest were calculated and an analysis of variance (ANovA) was performed on all of these data. Since there is no replication of treatments, the usual procedure in such cases was followed: using some of
the higher level interactions as the error term (in this case, "Harvest x Sites x Fertilizer", "Harvest x Sites x Cuts", "Harvest x Fertilizer x Cuts", "Sites x Fertilizer
x Cuts" and "Harvest x Sites x Fertilizer x Cuts"). Despite significant differences
between altitude levels and an interaction between the time of harvest and the site
position, the exceptionally high coefficient of variation (CV), at 344%, did not allow
much confidence to be placed on these results. The data were examined for correlation between variance and sample dry weights: logarithmic transformation of the
data did not reduce the variability of any of the data (determined by eye from residual versus fitted value plots) and so, throughout these analyses, it has not been used.
•
•
5. Productivity of Andean Grasslands in Ecuador
178
===p---1
0
0
o3
0
0
0
0
0
0 0 0 0 0 0 0 0 0
0 0 0 0 0 0
00
co 0 o3 co •n:1- N
N
I
I
a
C
a)
CCS
0
Li
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0
0 0 0 0 0 0 0 0
000
0(0(0
ED
N
N
(0 0 CO
N •ct cf)
N
I
I
I
I
I
I
(z_w 6) wawai o uilL161oM Ala
(0
2
179
5. Productivity of Andean Grasslands in Ecuador
Owing to the very high variation of the harvest to harvest increment data, an analysis was performed on the overall dry weight changes over the entire experimental
period. The method of calculation of these increments is shown diagrammatically in
Figure 5.5 and the data is presented in Figure 5.6. Some of the values are negative,
showing an overall decrease in weight from the start to the end of the trial period.
Only those plots which were uncut at the start of the experiment show this and the
paramo sites (D and E) exhibit the greatest tendency to high negative values.
The NPK fertilizer treatment produced higher yields than the other fertilizer treatments. Site A shows the greatest increase, within each fertilizer treatment, and for
the majority of cutting treatments. Sites B and C exhibit the smallest changes in
aboveground biomass.
The ANOVA results for this data are given in Table 5.3. The coefficient of variation is 101%, still extremely high, but much less than that for the harvest to harvest
data.
Source of Variation
df
SS
SS%
MS
Sites
Fertilizers
Cuts
Sites X Fertilizers
Sites X Cuts
Fertilizers X Cuts
Error (Sites X Fertilizers X Cuts)
Total
4
3
3
12
12
9
36
79
1847027
401373
242323
398846
739158
141235
539844
4309805
42.86
9.31
5.62
9.25
17.15
3.28
12.53
100.00
461757
133791
80774
33237
61596
15693
14996
54554
F
30.79
8.92
5.3
2.22
4.11
1.05
p
Significance
0.000
0.000
0.004
0.033
0.000
0.424
***
***
**
*
***
NS
Table 5.3.
Analysis of Variance table for the dry weight increments from the start to the end
of the trial period. The highest level interaction term was used as the error, since
there was no replication. The level of significance is denoted by the conventional
symbols: NS, not significant; *, significant at the 95% level; **, significant at the
99% level; ***, significant at the 99.9% level.
A
Mean Dry W3ight (g nf 2)
B
Site
C
D
E
424.06a 45.69 b 27.38 b 73.38 b 38.00b
Table 5.4.
Mean increments of dry weight (g m -2) from the start to the end of the experiment
(103-110 days) for each of the sites. Means sharing a letter have not been separated by the Student-Newman-Keuls test.
5. Productivity of Andean Grasslands in Ecuador
(b)
(a)
Dry weight increment (gm )
Dry weight increment (gm )
No PK
180
PK
Figure 5.7.
Mean dry weight gains for all five sites. a. Fertilizer treatments,
b. Cutting treatments. Means sharing a letter were not separated by a StudentNewman-Keuls test.
Site
A
No Fertilizer
N
PI(
NPK
330 25`
515.00b
195.50d
655 50°
12.75°
62.50°
44.75°
62.75°
6.75°
32.75°
31.25°
38 75°
50.50°
12.50°
550°
236.00d
80.25de
-25.25°
-80.50ef
170.50d
Table 5.5.
Mean increments of dry weight (g m -2) from the start to the end of the experiment
(103-110 days) for each of the site x fertilizer interaction terms. Means sharing a
letter have not been separated by the Student-Newman-Keuls test.
The analysis shows a highly significant effect of site position. A Student-NewmanKeuls (SNK) test, a multiple range test for comparing means (Zar, 1984), identifies
this difference: the increase in weight at Site A being much higher than that for the
other sites (Table 5.4). The fertilizer treatments give a high probability of being different, the NPK treatment being significantly higher than the other fertilizer treatments (Figure 5.7a). Cutting regimes also result in important differences, the no cut
treatment being much lower than the other treatments (Figure 5.7b). Significant interactions are shown between sites and fertilizers (site A plots responding more vigorously to the fertilizer applications, and the PK treatment in the highest exclosure
showing the poorest response: Table 5.5). Another important interaction was that between sites and cutting treatments. The SNK test revealed seven groups of "sites x
5. Productivity of Andean Grasslands in Ecuador
181
cuts" means (Table 5.6). The uncut plots at site A produced the greatest increase in
biomass (551.75 g rri-2 on average) and was distinct enough to form a separate group
from the rest. The other three cutting regimes at site A constituted the next heaviest
group (343.25-408.75 g 111-2 ). The remaining groups showed some overlap, but roughly speaking, the uncut plots from the paramo sites (D and E) formed the bottom
group (with a reduction in weight of 261.25 and 126.5 g m -2), the montane forest
plots (B and C) formed the next lightest groups (showing increases in weight from
5.25-84.25 g m -2), and the otherzroups consisted of the remaining plots from the
paramo sites (141.5-206.75 g m -`). The only exception was the cut 2 plot from the
most elevated site, which, with a reduction in dry weight of 46 g m 2, was placed in
the group with the plots from sites B and C.
Although at 101% the coefficient of variation of these data is much less than that
for the harvest to harvest increments, it is still too high to allow firm conclusions to
be drawn. There are two further approaches available to reduce the variability of the
data, both requiring subdivision of the data. The first is to analyse the sites separately, the other to look at the cut plots only.
The complex interactions with site position (Table 5.3) merit analyses of the sites
separately. In fact, the interactions, particularly those shown in Tables 5.5 and 5.6,
warrant the use of the following groups: site A, sites B + C and sites D + E.
The analysis of variance for the data from Alao (site A) is shown in Table 5.7. The
coefficient of variation is 39%. A significant effect of fertilizer is demonstrated by
the analysis of variance (p = 0.015). The result of the SNK test on the fertilizer
means is shown in Figure 5.8a: NPK and possibly N fertilizer treatments are higher
than the others. There appears to be little interaction between N and PK fertilizers,
the increased growth resulting from the addition of NPK is roughly the sum of that
produced by applications of N and PK alone.
No significant differences were found between sites B and C (Table 5.8). Both fertilizer and cutting treatments were found to exhibit pronounced differences. The
coefficient of variation for the analysis was 70%. The SNK test of the fertilizer
means differentiated the lowest, unfertilized plot from the other fertilizer plots (Figure 5.8b), although the PK treatment was not clearly defined. The NPK plots yielded
less than would be expected if the relationship between N and PK was a straightforward additive effect.
Site
A
No Cuts
Early Cu/
Late Cut
Early & Late Cut
551.75a
392.50b
343.25b
408.75b
5•25f
42.25°
51.00ef
84.25def
7•75f -261.25g -126.50g
34.00af 141.50` d ° 146.75`de
-46.001
30.00 8f 206.50'
37.75af 206.75a 177.75cd
Table 5.6.
Mean increments of dry weight (g m -2) from the start to the end of the experiment
(103-110 days) for each of the site x cuts interaction terms. Means sharing a letter
have not been separated by the Student-Newman-Keuls test.
5. Productivity of Andean Grasslands in Ecuador
Site A
000
703
NPK
I
100
a
.
Sites B & C
GNPK
30
r.r.
cr,
10
Figure 5.8.
Mean dry weight increments for the four fertilizer treatments.
a: Site A; b. Sites B & C; c. Sites D & E. Means sharing a letter were not separated
by a Student-Newman-Keuls test. The bar to the left of each figure represents the
Least Significant Difference (LSD). Key to fertilizer treatments: C, Control; N, Nitrogen only; PK, Phosphorus and Potassium only; NPK, Nitrogen, Phosphorus
and Potassium.
182
5. Productivity of Andean Grasslands in Ecuador
Source of Variation
Fertilizers
Cuts
Error (Fertilizer X Cuts)
Total
df
3
3
9
15
SS SS%
491498
96262
244093
831853
19.69
11.57
29.34
100.00
MS
163833
32087
27121
55457
183
F
p
6.04
1.18
0.015 *
0.370 NS
Significance
Table 5.7.
Analysis of Variance table for the dry weight increments for site A (Alao) from the
start to the end of the experiment. The fertilizer x cuts interaction term is used as
the error since there is no replication of treatments.
Source of Variation
df
SS
SS%
MS
Sites
Fertilizers
Cuts
Sites X Fertilizers
Sites X Cuts
Fertilizers X Cuts
Error (Sites X Fertilizers X Cuts)
Total
1
3
3
3
3
9
9
31
2682.8
8357.1
12151.1
675.8
2672.3
4287.3
5823.5
36650.0
7.32
22.80
33.15
1.84
7.29
11.70
15.89
100.00
2682.8
2785.7
4050.4
225.3
890.8
476.4
647.1
1182.3
F
p
Significance
4.15
4.31
6.26
0.35
1.38
0.74
0.072
0.038
0.014
0.792
0.311
0.672
NS
*
*
NS
NS
NS
Table 5.8.
Analysis of Variance table for the dry weight increments for sites B and C (Daldal)
from the start to the end of the trial period. The highest level interaction term was
used as the error, since there was no replication.
Table 5.10 shows no difference between sites D and E, though differences were
demonstrated for the fertilizer and cutting treatments. However, the coefficient of
variation was 221% - too high to allow confident interpretation of these results. Figure 5.8c illustrates the high variability of these data-both the N and PK treatments
resulted in a decrease in yield, while the NPK treatment showed a strong positive effect.
The comparison of cutting treatment means indicated that the uncut plots were,
on average, significantly lower than those plots which were cut (Table 5.9).
Another approach to reducing the variability of the data (shown in Table 5.3) is to
exclude those plots which were left uncut at the start of the experiment from the analysis. Since half of the original data (the uncut and late cut only plots) depend on the
initial estimates of standing crop (which have CV values of values between 23% and
88%), the variability of the data is increased. By removing these data from the analysis, the overall CV of the increments from the start of the experiment to the end
was reduced from 101% to 44%. The excluded plots had a CV of 208%. The
ANOVA of the early cut plots is given in Table 5.11.
5. Productivity of Andean Grasslands in Ecuador
184
The two cutting treatments (early cut only, early and late cut) were not differentiated and all interactions were insignificant. Once again, the sites contributed a
great deal to the overall variance of the data and proved to be highly significant. The
SNK test determined three populations from the five means, site A yielding more
than the others, with the dry weight increments from sites B and C being significantly
lower than the rest (Table 5.12). The fertilizer treatments were significantly different,
the plots treated with NPK growing fastest (SNK test: Table 5.13).
In a similar way to the subdivision of the whole dataset previously, it may be
possible to eliminate some of the variability of these cut data by looking at some sites
independently. It is clear from Table 5.12 that site A, sites B & C, and sites D & E
represent meaningful groups to be treated in this way.
The analysis of the Alao (site A) data is presented in Table 5.14, the coefficient of
variation being 36%. Despite the percentage sums of squares showing values similar
to those of previous analyses, the fewer degrees of freedom in this ANOVA do not
allow any significant differences between treatments to be expressed. Figure 5.9a
shows the growth increments of the fertilizer treatments, NPK again yielding the highest, with PK depressing the growth rate. The two cutting treatments produced very
similar yields of around 400 g m-2 over the trial period.
The coefficient of variation for the cut plots of sites B and C was 28%, and the analysis of variance of these data is shown in Table 5.15. The difference between the
mean yield of site B (63 g m -2) and site C (36 g m -2) was found to be significant. The
plots which were cut twice yielded significantly more dry weight than those which
were only cut once, at the start of the experiment (61 g m 2 average as opposed to
38 g m - ). There were no apparent differences between the fertilizer treatments,
though the percentage sum of squares was high. The effects of N and PK appear to
be independent (Figure 5.9b).
Table 5.16 presents the analysis of variance of the cut plots from the paramo sites
(D & E). The coefficient of variation was 34%. Again, there were no significant effects of the treatments, despite the high sums of squares (particularly for the fertilizer treatments). NPK application increased growth dramatically, despite N and
PK treatments showing little gain independently (Figure 5.9c).
Cutting Treatments
Uncut Early Late Early & Late
Cut Cut Cuts
Mean Dry Weight (g m-2)
6.50°
44150 61.00b
b
38.13b
Table 5.9.
Mean increments of dry weight from the start to the end of the experiment (107108 days) for the cutting treatments at sites B and C (Daldal). Means sharing a letter have not been separated by the Student-Newman-Keuls test.
5. Productivity of Andean Grasslands in Ecuador
S00
702
S . tes B & C
70
ONPK
Sles 0 & E
no
100
120
Figure 5.9.
Mean dry weight increments for the four fertilizer treatments in plots which were
subjected to an initial cut. Analysis of variance did not indicate any significant difference between the means in all three cases.The bar to the left of each figure represents the Least Significant Difference (LSD). a. Site A; b. Sites B & C; c. Sites D
& E. Key to fertilizer treatments: C, Control; N, Nitrogen only; PK, Phosphorus and
Potassium only; NPK, Nitrogen, Phosphorus and Potassium.
185
5. Productivity of Andean Grasslands in Ecuador
Source of Variation
Sites
Fertilizers
Cuts
Sites X Fertilizers
Sites X Cuts
Fertilizers X Cuts
Error (Sites X Fertilizers X Cuts)
Total
SS
dl
1
3
3
3
3
9
9
31
10011
284317
714842
15371
155554
291034
135841
1606969
SS%
0.62
17.69
44.48
0.96
9.68
18.11
8.45
100.00
10011
94772
238281
5124
51851
32337
15093
51838
186
MS
F
0.66
6.28
15.79
0.34
3.44
2.14
0.436
0.014
0.001
0.797
0.066
0.136
p
NS
*
***
NS
NS
NS
Table 5.10.
Analysis of Variance table for the dry weight increments for sites D and E (Daldal)
from the start to the end of the trial period. The highest level interaction term was
used as the error, since there was no replication.
Source of Variation
dl
Sites
Fertilizers
Cuts
Sites X Fertilizers
Sites X Cuts
Fertilizers X Cuts
Error (Sites X Fertilizers )< Cuts)
Total
4
3
1
12
4
3
12
39
SS
SS%
660892
78117
10017
77102
4570
13709
63757
908099
MS
72.78
8.60
1.10
849
0.50
1.51
7.02
100.00
165223
26039
10017
6425
1126
4570
5313
23285
F
31.10
4.90
1.89
1.21
0.21
0.86
Significance
p
0.000
0.019
0.195
0.374
0.927
0.488
***
*
NS
NS
NS
NS
Table 5.11.
Analysis of Variance table for the dry weight increments from the start to the end
of the trial period. Only plots with an early cut treatment are included.The highest
level interaction term was used as the error, since there was no replication.
A
Mean dry weight increment (g m-2)
B
Sites
C
D
E
400.63a 63•25b 35.88b 174.12c 162.25C
Table 5.12.
Mean increments of dry weight (g rT1-2) from the start to the end of the experiment
(103-110 days), excluding initially uncut plots, for each of the sites (A, 3,100m; B,
3,250m; C, 3,450m; D, 3,750m; E, 4,000m). Means sharing a letter have not been
separated by the Student-Newman-Keuls test.
5. Productivity of Andean Grasslands in Ecuador
No fertilizer N
Mean dry weight Increment (g m-2)
143.5a
187
NPK
PK
238.2b
a
167.0120.2
a
Table 5.13.
Mean increments of dry weight (g m -2) from the start to the end of the experiment
(103-110 days), excluding initially uncut plots, for each of the fertilizer treatments.
Means sharing a letter have not been separated by the Student-Newman-Keuls
test.
Source of Variation
Fertilizers
Cuts
Error (Fertilizers X Cuts)
Total
df
3
1
3
7
SS
122718
528
63367
186614
SS%
65.76
0.28
33.96
100.00
MS
40906.1
528.1
21122.5
26659.1
F
p
1.94
0.03
0.300 NS
0.884 NS
Significance
Table 5.14.
Analysis of Variance table for the dry weight increments for site A (Alao) from the
start to the end of the experiment. Only plots with an early cut treatment are included. The fertilizer x cuts interaction term is used as the error since there is no
replication of treatments.
Source of Variation
dl
SS
SS%
MS
F
Sites
Fertilizers
Cuts
Sites X Fertilizers
Sites X Cuts
Fertilizers X Cuts
Error
Total
1
3
1
3
1
3
3
15
2997.6
4380.2
2093.1
354.7
1463.1
972.2
573.2
2833.9
23.36
34.13
16 31
2.76
11.40
7.58
4.47
100.00
2997.6
1460.1
2093.1
118.2
1463.1
324.1
191.1
855.6
15.69
7.64
10.95
0.62
7.66
1.70
p
Significance
0.029
0.064
0.045
0.649
0.070
0.337
*
NS
*
NS
NS
NS
Table 5.15.
Analysis of Variance table for the dry weight increments for sites B and C (Daldal)
from the start to the end of the trial period. Only plots with an early cut treatment
are included. The highest level interaction term was used as the error, since there
was no replication.
5. Productivity of Andean Grasslands in Ecuador
Source of Variation
df
SS
SS%
Sites
Fertilizers
Cuts
Sites X Fertilizers
Sites 5 Cuts
Fertilizers X Cuts
Error
Total
1
3
1
3
1
3
3
15
564.1
23611.2
9264.1
4154.7
1173.1
2557.7
9995.7
51320.4
1.10
46.01
18.05
8.10
2.29
4.98
19.48
100.00
MS
564.1
7870.4
9264.1
1384.9
1173.1
852.6
3331.9
3421.4
188
F
p
Significance
0.17
2.36
2.78
0.42
0.35
0.26
0.708
0.249
0.194
0.755
0.595
0.854
NS
NS
NS
NS
NS
NS
Table 5.16.
Analysis of Variance table for the dry weight increments for sites D and E (Daldal)
from the start to the end of the trial period. Only plots with an early cut treatment
are included. The highest level interaction term was used as the error, since there
was no replication.
Length
of expt.
(days)
Net Annual Aboveground Production (g nf2 yr-1)
Overall
Fertilized
16
12
4
1,503
154
93
243
128
1,614
192
117
269
73
1,170
43
23
168
295
Unfertilized
Altitude
n
Site A
S te B
S te C
S te D
Site E
3100m
3250m
3450m
3750m
4,000m
103
108
107
110
108
=
Initially
Initially
cutuncut
a 8
1,420
214
122
578
548
1,586
95
64
91
292
Unfertilized
Cut
2
1,359
110
70
430
512
Table 5.17.
Estimated net annual above-ground productivity for the five exclosure sites. At
each altitude sixteen productivity estimates were made (including all fertilizer and
cutting treatments) and the mean of these values is shown in the "Overall" column. The mean values for fertilized and unfertilized sites are presented, along with
those plots which were initially cut and those which were not. The most reliable estimate of natural productivity is given by the final column: those plots which were
unfertilized and cut at the start of the experiment.
Table 5.17 shows the increments in dry weight over the experimental period extrapolated to a yearly base. Overall, the lowest site (Mao) was much more productive
than the rest. The productivity estimates of the sites in the Daldal valley were an
order of magnitude lower.
The addition of fertilizer stimulated production in the lower four plots, particularly at Mao and in the montane forest sites. The highest paramo site showed depressed
growth on addition of fertilizer (73 compared to 295 g M-2 yr" ).
An initial cut at the start of the experiment produced higher yields in the montane
forest and paramo plots. This represented roughly a doubling of productivity in three
of these sites, and a five-fold increase in the lowest paramo plots. The dry matter production in Mao was not increased by this early cut.
5. Productivity of Andean Grasslands in Ecuador
189
Owing to the extreme variability of the data in plots which were not cut at the start
of the experiment, the best estimates of natural productivity rates are given by those
plots which were initiallyicut and not fertilized. The lowest plots in Alao produced an
estimated 1,359,g m- yr- (13.59 t ha- yr- ). The two montane forest zone plots at
110 and 70 g m-` yr-1 produced ten to twenty time less than the Alao lots. The paramo plots yielded substantially more dry weight (430 and 512 g M -2 yr- ).
Discussion
Patchiness within the vegetation was responsible for the high variability of the harvest data. Such heterogeneity is unavoidable in natural vegetation, and is caused by a
number of factors. One of these is the distribution of the dominant species and occurs, for example, in the 'Aram° sites (with the patchiness inherent in the tussock
grass physiognomy) and in the montane forest sites (with a very high mat cover, restricting growth). The pattern of distribution of other species may contribute to the
dry weight variability. For example, in the Alao site, Dactylis glomerata forms coarse,
spreading culms at maturity, which are heavier but do not possess such a plastic response to the environment as younger D. glomerata —the distribution of such mature
patches is reflected in the dry weight values. Another possible cause of patchiness at
Alao is grazing. The usual practice of tethering animals to a stake would produce patches of different grazing intensity and recovery ages.
Such variability in the vegetation could have been reduced by replication of treatments or by increasing the plot size to a level which was influenced less by pattern.
However, neither of these approaches was possible for logistical reasons.
Half of the data are estimated increments (based on the standing crop values,
Table 5.1) while the other 50% are exact increments, the plots having been cut back
to a known level, namely zero, during the early cut. These latter data are much more
dependable than the uncut plots, reflected in the lower CV values (44% as opposed
to 208%). Throughout, the CV values are very high, but it should be noted that these
data are rates of growth, and rates are susceptible to high variability.
Logistics determined the timespan of the experiment: only slightly more than 100
days. Clearly this is by no means ideal and seasonal trends within vegetation processes may well confound the picture presented here. It may explain, for example,
the fact that four of the five control plots decreased in standing crop weights over the
experimental period. In addition, plant nutrient deficiency may act by limiting the
amount of photosynthetically active organs. Such a mechanism would lead to a lag
between fertilizer application and the vegetation's response (Tamm, 1975); thus, a
complete response to fertilizer application would require a much longer study of perhaps several years. Both the small plot size and the short duration of the experiment
complicated the analyses by over-emphasizing interaction effects, mostly attributable
to the high variability of the data (evidenced by the high coefficients of variation).
5. Productivity of Andean Grasslands in Ecuador
190
The harvest approach to plant productivity is a well-established one, but recent
studies of productivity have used CO2 exchange to estimate the rates of energy flow
in and out of producers. Such research has shown that traditional harvest approaches
seriously underestimate dry matter production. This underestimation is a result of
events taking place between shoot harvests and changes of plant material from one
category to another. In particular, the role of decomposers has been neglected —
Clark & Paul (1970) reported a living biomass of decomposers and consumers in excess of 50% that of the primary producers. Herbivory is also important and while the
activities of large herbivores have been intensively studied, the biomass consumed by
small herbivores (rabbits, rodents, grasshoppers, etc.) has been largely ignored. Coupland (1972 — cited by Coupland, 1975) has shown that invertebrates may ingest and
drop as much as 80% that consumed by cattle.
A major difference between the paramo sites and the lower grassland pastures was
demonstrated by their higher standing crop biomass estimates at 837 and 794 g mfor the plots at 3,750 m and 3,950 m respectively. There was an order of magnitude
of difference between these sites and those lower down. In fact, the values are extremely high compared with grasslands throughout the world (Rodin et al., 1975).
However, these estimates are consistent with reports from other parr° and tropical alpine vegetation. Acosta-Solis (1984) recorded a value of 435 g m 2 the PAramo de Chiquicagua in Ecuador. Further north in Colombia, Tol & Cleef (1992)
measured the aboveground standing biomass of mainly dwarf bamboos and
bryophytes in a Colombian bamboo paramo at 2,282.5 g M-2 (63% dead material),
much higher than that in Daldal. In a study of Venezuelan paramo (physiognomically
very different from the one in this study), Smith & Klinger (1985) recorded values
ranging from 130 to 601 g m -2. In a wider context, Hnatiuk (1978) reported aboveground phytomass estimates of 436-628 g M -2 in tropical alpine tussock grassland in
New Guinea.
The explanation for such high standing crops is straightforward: the tussock
grasses which dominate the paramo vegetation retain their dead leaves, which decay
whilst still attached to the tussock. A very large proportion of the aboveground standing crop biomass is therefore dead material. It would also be expected that decomposition would take place at a slower rate at higher altitudes because of the colder
temperatures. It is suggested that there is adaptive significance attached to the retention of dead leaves —they may insulate the delicate, developing tillers from climatic
extremes and from lethal temperatures during grass fires (as discussed in earlier
chapters). It is also proposed that this habit may prevent colonization by other
species and reduce the overall nutritional value of the tussock, reducing levels of predation by herbivores. Schmidt & Verweij (1992) showed that dead matter in Calamagrostis effusa tussocks is of much lower digestibility and that tussock grasses are
preferred least by cows (making up just 30% of the diet) despite its availability. Tol
& Cleef (1992) suggested that by tying up nutrients in dead material still attached to
the living plant, they are not available to other plants and this gives the tussock a
competitive advantage.
The lower three sites did not share this tendency to retain dead leaves and their
values for aboveground standing crop biomass were much lower as a result. The
5. Productivity of Andean Grasslands in Ecuador
191
values were very low, perhaps reflecting grazing pressure at the lowest site and, at
the montane forest sites, the additional effect of the dominance of the A zorella mats.
From the initial analysis of the harvested material (Table 5.2), the five sites separate into three groups. The lowest site (Alao) was very significantly different from the
others, with a much higher production of dry matter. The two montane forest sites
made up the second group, with the two paramo sites comprising the third category.
However, it is the analysis of the initially cut plots that shows this distinction most
clearly (Table 5.13).
Before the start of the experiment, the primary limitation on growth at Alao was
grazing. Removal of herbivory (at least by cattle), allowed a massive increase in phytomass to take place. At the other sites, however, grazing was not so limiting on plant
growth and, as a result, exclosing the plots did not have such a large effect. The estimated annual productivity of the Alao site, at 1359 g m - yr-1, is very large, comparable to boreal mountain forest and semi-arid savanna (Rodin et al., 1975).
However, because the experiment was carried out during the wettest part of the year
for Mao, coupled with the removal of grazing, the burst of growth recorded is perhaps not sustainable, leading to an over-estimation of annual production.
That productivity is linked to water availability is well-documented (Tieszen & Detling, 1983). Weischet (1969), Lauer (1976) and Sarrniento (1986) report that the
zone of maximum precipitation in the Andes occurs at middle altitudes, corresponding to the position of montane forest vegetation. Therefore, it would be expected
that the sites in this zone (B and C) would be the most productive of those studied.
In fact, they are the least productive. The steep slopes of the Andes at these elevations may result in high rainwater run-off, and indeed it is common practice for the
pastures at site C to be periodically waterlogged with water diverted from a nearby
watercourse to promote growth (this suggests water-limitation is a factor in the productivity of this site). The lowest Daldal site (B) is more or less flat and the rate of
growth was always higher than site C, 200m above (Table 5.16), though it was only
found to be significantly different from the steeper slope on one occasion (Table 5.14). Water limitation may, therefore, be a constraint on plant growth in these pastures, which may lead to an interaction between the treatments and site position.
The main restriction on growth in these sites is the mat-forming habit of A zorella.
The mat suppresses the vigour of the species growing within it, allowing only a few
shoots through— and these are quickly cropped by herbivores. The annual productiv- and 70 g rn- is roughly equivalent to estimates for deity for these sites, at 110 g m2
sert vegetation (Rodin et al., 1975) and semi-desert (Lieth, 1975). However, it must
be pointed out that this study does not take into account the productivity of the A zorella mat itself, which may increase the overall annual production very significantly
indeed. It was impossible to harvest this species without disrupting the very nature of
the vegetation structure, leaving bare earth, and without destroying the other species
in the process.
A zorella pedunculata dominance is strongly associated with heavy grazing pressure. Grubb, Lloyd & Pennington (unpublished) reported a similar mat covering at
4,050m on the intensively grazed pastures of Volcan Antisana, Ecuador — although
5. Productivity of Andean Grasslands in Ecuador
192
they suggest that some of the mats were probably quite large before the practices of
burning and grazing were introduced. This latter point is debatable, since A zorella pedunculata is only found in large mats where grazing intensity is high; elsewhere it is
found in very small patches and contributes little to the vegetation structure. It
would seem likely that light competition, with Calamagrostis and other grasses, is a
major factor in determining its distribution—where grazing and trampling alters the
competitive balance, A zorella is able to dominate.
However, it is clear that once an extensive A zorella mat has formed, it becomes
very difficult for other species to compete. Grubb et al. (unpublished) describe the
suppression of shoots of other species as the A zorella mat comes to surround them.
One of the reasons for this could be the mobile nature of the A zorella mats, consisting of rafts of rosettes, borne on rhizomes, which are constantly changing position to
accommodate the growth of new rosettes. Any shoots of other species must straddle
this mat (roots below and leaves above) and therefore must be resistant to the lateral
movement of A zorella. Such plants would include species with short-lived, easily replaced shoots (for example, grasses) and plants which can physically resist the movement (such as woody plants).
It is difficult to assess the potential for recovery of these mat-dominated pastures,
but without the continual cropping of shoots as they appear through the mat (i.e., in
the absence of grazing), other species may establish and become locally more dominant. This may eventually lead to the elimination of the mat through competition for
resources. It is most likely, however, that in the Daldal montane forest site, these colonists would be woody species and the pastures would revert to forest once more.
If, after the clearance of forest, production is to be maintained at a reasonable
rate, grazing pressure must not reach the threshold for the formation of A zorella
mats. This is an area of research which demands more attention.
The upper two sites, in the paramo zone, were also influenced strongly by biotic
factors. Grazing is not so heavy here and conditions are often harsh. The accumulation of dead leaves by the tussock grasses is of major importance, both in terms of
"adaptation" to the environment and with regard to the potential for dry matter production. As dead leaves build up within the tussock, the plant loses vigour and
becomes less palatable to herbivores. For this reason, burning has become a well-established management technique used by peasant farmers to stimulate available production for livestock. The two paramo sites examined in this study were nearing a
condition when burning is applied (though its application is somewhat erratic) and
were perhaps at, or near to, a steady state — a condition of equilibrium where decomposition of dead material and respiration proceed at the same rate as production,
leading to an absence, in terms of weight, of tussock growth (a common state in vegetation, described by Horn, 1974).
This balance may explain why many of the uncut plots in the paramo did not show
weight gains over the course of the experiment (Figure 5.7). In fact, all of these uncut
plots showed a loss in weight, indicating, perhaps, that some inhibition of growth was
taking place. This phenomenon is also reflected in the annual net productivity estimates for the initially uncut plots in the paramo (Table 5.16), which are much lower
5. Productivity of Andean Grasslands in Ecuador
193
than the estimates for the initially cut plots. It would seem, therefore, that the local
agricultural practice of burning the grassland to increase production is well-founded,
though it is likely that fire would inflict greater damage on the tussocks than mere
clipping. The fact that grazing animals tend to concentrate their foraging in recently
burned areas further complicates matters.
Physiological water limitation may be a major factor in inhibiting growth in the
paramo. Low soil temperatures, particularly during the clear early mornings when
photosynthesis and transpiration are occurring at rapid rates, may contribute to overall water stress in the higher altitude sites by reducing water uptake. This has been
demonstrated for Dendrosenecio (Smith & Young, 1987a) and Lobelia keniensis
(Young & Van Orden Robe, 1986) for East Africa, and for Espeletia schultzii in the
Venezuelan paramo (Smith, 1972). Water shortage would inhibit leaf elongation
(Wardlaw, 1969) and translocation of photosynthate may be directed to belowground
structures as a result. This growth of the root systems would not be apparent in this
study.
It is worth noting at this point that, in accordance with the data currently available,
tropical alpine grasslands have the highest ratio of aboveground to belowground phytomass of all vegetation (Smith & Klinger, 1985). Root phytomass ranged from 66.7386.4 g M -2 in a Venezuelan paramo (Smith & Klinger, 1985) and 1,084-1,598 g M-2
in a Colombian paramo (Rossenaar & Hofstede, 1992). It has been proposed (Smith
& Klinger, 1985) that this high ratio is the result of a stressful environment (vegetation is of short stature and therefore extensive root systems are not required for support) with a year-round growing season (less need for storage in roots).
The net annual productivity estimates for these paramo sites (168 g M -2 and
295 g M -2 for sites D and E respectively) is quite low, similar to that for tundra ecosystems and about one-quarter that of mountain steppe (Lieth, 1975; Rodin et aL,
1975). They are in accordance with estimates of North American mountain grasslands (Sims & Singh, 1978) and values recorded by Hnatiuk (1978) in the tropical alpine grasslands of New Guinea (128-442 g m -2). Annual net above ground
1986)
productivity was estimated for Espeletia timotensis at 700 g m 2
(Sturm
&
Abouchaar,
1981)
in
physiogand for Espeletia grandiflora at 1,500 g M-2
nornically different paramos in Venezuela and Colombia, respectively. At higher altitudes, desert paramo productivity was much lower at 140 g m 2 (Lamotte, Garay
& Monasterio, 1989).
This rate of growth would be required for at least 3-5 years to accumulate the
standing material shown in Table 5.1, which fits in very well with the observed practice of burning the tussocks every 2-4 years to renew the plants' vigour. This rate of
growth and renewal is much higher than that estimated by Mann (1966) for Peruvian
puna vegetation: the gross annual primary productivity of dry and humid punas
- dry weight (with standing crop weights of
being, respectively, 0.3 m-` and 8 g m2
200 g 111 -2 and 700 g m- )*
It is likely that the environment could support a higher rate of carbon assimilation,
but this would be exploited with some risk to the plant. In a global survey of mineral
nitrogen content in high altitude plant tissues, tropical alpine plants had the lowest
5. Productivity of Andean Grasslands in Ecuador
194
content of those latitudes studied (KOrner, 1989). A tentative hypothesis for this situation was proposed: that tropical alpine plants keep their growth rates under control
(KOrner, 1989).
In temperate regions, there is a period of climatic cooling in the autumn, allowing
a step-wise 'hardening' of plant tissues to take place before winter sets in. In the
tropical mountains, however, with "summer every day, winter every night", the plants
must retain the hardy state throughout the year to prevent night-time damage by
cold temperatures and desiccation. The induction of frost hardiness may lead to a decrease in photosynthetic capacity and, in any case, only 60% of the optimal CO2 uptake is achieved by chilling tolerant plants at temperatures between +5 and + 10°C,
and is completely blocked when ice forms in the assimilatory organs (Larcher &
Bauer, 1981). Kaufmann (1977) discussed feedback inhibition via carbohydrate accumulation and stomatal closure induced by water stress in the leaves caused by low
temperatures (especially in the soil). Prolonged exposure to low temperatures can
further depress the photosynthetic capacity of plants (Larcher & Bauer, 1981). For
example, the altitudinal limit of successful potato cultivation is set by the intensity of
episodic night frosts (Li & Palta, 1978). There are also structural responses to low
temperatures (such as growth form) which may be at the expense of optimum photosynthetic performance.
The analysis of variance for the data from the start to the end of the experimental
period for all the sites shows a very significant effect of fertilizers (Table 5.2). The
NPK treatment gave significantly better results than the other treatments (Figure 5.8, a). There was a site-specific effect of fertilizer application, but this was not a
straightforward relationship (Table 5.4). It may be linked to soil moisture phenomena. Drier areas are less likely to be limited by nutrients—certainly the paramo sites
would be subjected to a relatively high degree of water stress and as a result, the addition of nutrients would not be expected to boost production by any great amount.
Furthermore, (KOrner, 1989) noted that the nitrogen content of high alpine vegetation globally is naturally much higher than in other areas. Whether plants can respond to further additions of nutrients by increased growth is still uncertain.
Another factor which may lead to a fertilizer-site interaction is the slope of the
plots. Generally, Andean slopes are very steep and rainfall quite high. In some of the
plots, this may have led to the lateral translocation of nutrients across the buffer
zones. The waterlogging practices carried out at site C may have contributed to this
phenomenon. It is worth mentioning that the very low value of the PK treatment
(Figure 5.8,a) is the result of some heavy weight losses in the paramo sites in the
uncut plots (Figure 5.7). Analysis of the initially cut plots only resolves this anomaly,
for although the PK treatment is still the lowest, it is much closer to the other treatments in its group (Table 5.12).
The main analysis gives a high significance to the cutting treatments, the uncut
plots proving to be much less productive than the cut ones (Table 5.2). Since limited
burning is used to increase production in the paramo, it is not surprising that cutting
has a similar effect. In fact, clipping would be expected to damage the plants less
than burning (which not only removes material but may inflict lethal temperatures
on parts untouched by flames). It should also be considered that a certain intensity of
5. Productivity of Andean Grasslands in Ecuador
195
grazing may actually stimulate production. The early and late cut treatment increased the yield over the value obtained by either the early or late cut treatments on
their own (Figure 5.8b). This was true not only in the paramo, but also in the montane forest sites (Table 5.8). It would appear that the frequency of the cutting in this
treatment did not adversely affect production. Further research to determine the optimum cutting/burning/grazing regimes in these environments is required.
Longer-term studies are required to provide more accurate estimates of annual
production in the Andes, preferably with replication to reduce the variability of the
harvest data. In particular, emphasis should be placed on the overgrazed clearings of
montane forest.
Luteyn (1992) highlights the increasing demand for agricultural land in the Andean highlands and the threat this may pose for the fragile paramo ecosystem. This
study indicates that better management (to avoid the low productivity associated
with overgrazing), the production of the lower pastures in the highlands could be
raised to a level that eliminates the need to exploit new areas of land (many farmers
exploiting montane forest and/or paramo have pastures at lower altitudes). This area
of research must be explored in some detail in the future.
Chapter 6
A Greenhouse Study of
Competition Between
Three Andean Grasses
at Two Regimes of
Water Availability
6. Greenhouse Competition Between Paramo Grasses
197
Introduction
'Waffler chapters have looked at the zonation of paramo communities and their de'velopment with time at both small-scale and large-scale levels. The interactions
between species are clearly of major significance in the evolution of a plant community, and are of vital importance to the understanding of community dynamics. Interactions between immediately adjacent plants may largely determine plant
performance (Weiner, 1982; Silander & Pacala, 1985). In the ecocline of the Andean
altitudinal gradient, each individual species will show a distribution "according to its
own genetic, physiological and life-cycle characteristics and its way of relating to
both physical environment and interactions with other species" (Whittaker, 1973). In
this latter respect, a plant's distribution on the gradient will have been modified by
niche differentiation and restriction.
Dramatic, sharp discontinuities in vegetation are usually associated with strong enironmental
discontinuities or disturbance, most often in the paramo by fire. In
'
general, individual plant species, and the communities of which they are a part, intergrade continuously along the altitudinal gradient, rather than forming distinct zones.
The altitudinal ecocline is far from simple, however. An increase in altitude is associated with several factors, including a reduction in air temperature (the lapse rate
for the Andes is around 0.6°C per 100m of altitude), an increase in wind speed and
increased cloud cover (Sarmiento, 1986, 1987). These effects are in turn modified by
factors such as topography, regional climatic patterns, grazing and disturbance by
man. These major processes are then involved in smaller patterns, within the soil for
example. To study the whole host of variables relating to altitude would be impractical; so t o %%ould be the study of such factors at the community level. A reductionist
stance must be applied, dealing with only a few species and varying an artificial environment in clear, measurable ways. By this process, the effects of the measured environmental variables on the growth and interactions of the species can be observed.
In spite of the approximate nature of such observations, they can be used to assess
the effects that are likely to be important in the field (Williams, 1962).
One of the most important factors governing plant distributions in the paramos
was found to be water availability (Ramsay, 1988). It has been postulated by several
authors that with increasing altitude physiological drought becomes important (Walter & Medina, 1969; Perez, 1987; Smith & Young, 1987b). It would be expected,
therefore, that plants from higher altitudes might respond better to drought conditions than those from lower elevations, since they are regularly exposed to earlymorning droughts in their natural habitat.
To investigate the hypothesis that drought tolerance increases with altitude, three
grass species from the Ecuadorian paramo were collected for greenhouse experiments. One of these species was obtained from a high altitude of 4,150m, another
from a lower altitude of 3,750m, and the remaining one from an intermediate elevation of 4,000m. Simple diallel tests of competitive interactions between these
species were carried out, growing each species with each of the others and with itself.
6. Greenhouse Competition Between Paramo Grasses
198
Materials and Methods
Plant Material
All of the plants used in this study were collected on the 12 th November 1987 from
the Pâramo de Guamani (Papallacta), between Quito and Baeza (Figure 6.1).
This pâramo is quite wet, located on the Cordillera Occidental and represents a relatively undisturbed paramo, though burning and grazing do occur in some places, including the areas where grass material was collected.
At 4,150m the vegetation was dominated by Plantago rigida, interspersed with low
tussocks of Calamagrostis sp. A [724] (Plate 6.1). Other components of the vegetation included Culcitium ovatum, Oritrophium peruvianum, Oritrophium hieracioides,
W erneria nubigena, Senecio repens, Senecio chionageton, Lycopodium sp. [3731, Pernettya prostrata, Disterigma empetnfolium, Gentianella cernua, Gentiana sedifolia, Halenia weddelliana, Hypericum lancioides, A grostis nigritella, Poa sp. [723], Carex
lemanniana, Bartsia laticrenata, Castilleja sp. [2221, Geranium sibbaldioides, V aleriana
bonplandiana, Satureja nubigena, Oreomyrrhis andicola, Dyngium humile and A phanactis jamesonia. Thamnolia vermicularis, a small white tubular lichen was conspicuous, lying on the ground surface. Tillers of Calamagrostis A were collected.
Lower down, at 4,000m, the area was covered with tussocks of Calamagrostis sp. B
(voucher specimen of cultivated plant deposited at the Royal Botanic Gardens,
Kew), with Poa sp. (voucher specimen of cultivated plant deposited at the Royal Botanic Gardens, Kew) growing within them and in the intertussock spaces (Plate 6.2).
The vegetation was burned at intervals, probably once every four years or so, though
it was estimated that approximately 2-3 years had elapsed since the last burning of
the vegetation at the collection site. Other prominent species in the plant community
were Puya clava-herculis, Oritrophium hieracioides, Senecio repens, Senecio chionageton, Senecio pimpinelhfolia, Bidens andicola, Hypochaeris sonchoides, Gnaphalium
aff. pensylvanicum, Diplostephium glutinosum, Pernettya prostrata, Disterigma empetrifolium, Gentianella nummalarzfolia, Hypericum lancioides, A grostis sp. [8221 Poa sp.
[944], Care.x tristicha, Bartsia laticrenata, Geranium sibbaldioides, Sisyrinchium jamesoni, Lachemilla andina, Lachemilla orbiculata, Ranunculus sp. [339], A zorella pedunculata, A zorella aretoides, Oreomyn-his andicola, Ezyngium humile and A phanactis
jamesonia. Tussock material of Calamagrostis B was collected.
The lowest collection site (3,750m) was dominated by tussocks mainly of Calamagrosti.s C (voucher specimen of cultivated plant deposited at the Royal Botanic Gardens, Kew) but also of Festuca sp. [742] (Plate 6.3). Important elements of the flora
at this elevation included Oritrophium hieracioides, Bidens andicola, Gnaphalium gnaphaloides, Gnaphalium dysodes, Cukitium adscendens, Gnaphalium aff. pensylvanicum, Son chus ? oleraceus, V icia andicola, Relbunium hypocarpium, Satureja nubigena,
Stachys eliptica, Rumex acetosella, Hydrocotyle bonplandiana, A grostis sp. [822], Poa
sp. [356], Uncinia pheleoides, Bartsia laticrenata, Geranium reptarzs, Geranium multipartitum, Sisyrinchium jamesoni, Lachemilla orbiculata, Ranunculus sp. [339], A zorel-
6. Greenhouse Competition Between Paramo Grasses
199
dunculata, Oreomyrrhis andicola, Dyngium humile and A phanactis jamesonia. Tussock material of Calamagrostis C was collected.
Sample tillers were collected by excising sections of tussocks of the species concerned. The plant material was immediately placed in polythene bags, the roots
covered with damp newspaper, and the leaves emerging out of the top of the bags.
The plants were then transported in this state to the greenhouses at the University of
Wales, Bangor: a journey which took nearly two weeks. During this time, many of the
outer tillers of the tussocks died, but the protected inner ones survived.
The plant material was then divided into individual tillers (though some tussocks
were retained to provide stock material for further experiments). The tillers were
then placed in deep boxes of soil and allowed to grow: half in a warm house (1823°C) and the other half in a cold house (5-15°C). Unfortunately, tiller mortality was
extremely high. In the warm house, Calamagrostis C showed the best survival, with
18.8% of the tillers still alive after six months. Half as many tillers of Calamagrostis
A survived and CalamagrostLs B had a survival rate of 7.8% (with 100% of the survivors flowering). In the cold house, lower survival rates were observed of 7.8, 4.7 and
1.5% for Calamagrostis spp. A, B and C respectively. None of the cold house plants
flowered. The great majority of tiller deaths occurred during the first week after separation from the mother plant. The tillers that survived showed some growth, occasionally prolific. Some tillers from each species had formed small tussocks of over
100 tillers after six months.
Figure 6.1.
Map of the Paramo de Guamani, Ecuador, showing the collection localities of the
three grass species used in the greenhouse study (s ). Based on the Institut°
Geografico Militar (Quito) map for Oyacachi. The scale is 1:1,000.
6. Greenhouse Competition Between Paramo Grasses
200
Experimental Design
The experiment was constructed along the lines of the classic diallel design (see
The
1962; Norrington-Davies, 1967). In such an experimental design, plants
of two species are grown together in the same pot such that each species is grown
with each of the others and with itself once in each replication.
On 28 th June 1988, individual tillers of similar size and condition were arranged in
pots containing two tillers, one target species and one neighbour species, such that
each replication consisted of six pots:
Neighbour Species
Calamagrostis
A
Target
Species
Calamagrostis A
Calamagrostis B
Calamagrostis C
B
C
•
•
•
Square pots (15cm x 15cm) were used, filled with John limes' "Humax" compost.
Twenty replications were planted, amounting to a total of 120 pots and 240 tillers.
Reserve tillers were also planted at the same time.
The pots were well-watered in warm conditions (18-23°C) to allow for the establishment of the tillers. An unusually long establishment period of six months was decided upon (following the slow growth rates and high mortality of the species over
previous months). During this time, any dead tillers were replaced with living ones of
the same age from the reserve stock. However, mortality was again very high and the
number of replicates was reduced.
Following this establishment period the pots were randomly assigned bench positions in the warm house. One set of five replicates was placed on capillary matting,
allowing water uptake as necessary for growth: the other treatment, again of five replicates, consisted of watering with a fine rose from above twice weekly, creating
periods of water shortage. These regimes were maintained for six months.
Harvesting was carried out on 3 rd July 1989, some 270 days after planting. The
plant material was washed clean of soil and divided into above and below ground sections. Each of these portions were oven dried at 105°C for 24 hours and weighed.
Analysis of Results
he differences between the competitive abilities of the three grass species and the
effects of the two moisture regimes were assessed by way of an analysis of variance and multiple range tests for comparing means (Student-Newman-Keuls tests;
Zar, 1984).
T
6. Greenhouse Competition Between Paramo Grasses
201
Three indices of competitive abilities were also calculated. Relational effects (Harper, 1977), the competitive advantage of species a over species b, were calculated according to the formula:
1
1
11
Rab = — (Yab - — Yaa) + — (- Ybb - Yba)
2
2
2 2
where Yaa is the yield of the pure species a, Ybb the yield of the pure species b, Yab
the yield of species a when grown with species b, and Yba the yield of species b when
grown with species a.
Summational effects (Harper, 1977), a measure of how the yield of a mixture compares with that predicted from the pure stands, were calculated following the
formula:
11
1
Sab = — ( - Yaa + —Ybb) - • 1(Yab + Yba)
2 2
2
using the same notation as above.
An alternative measure of how the mixture yield performs relative to the pure
stands, the Relative Yield Total (RYT— de Wit, 1960), was calculated according to
the formula:
RYT = Yab
Yba
+
Yaa
Ybb
again using the same notation as previously.
Results
he mortality of tillers during the establishment phase of the experiment is of some
T interest
(Figure 6.2). It can be seen that over twice as many tillers of Calamagrostis B survived as Calamagrostis A, with Calamagrostis C higher still. In examining
how each species survives when grown with each of the other species it is evident
that the survival of tillers of Calamagrostis spp. B and C is largely unaffected by the
species with which they are grown. Calamagrostis A, however, shows much greater
mortality when grown with Calamagrosttls B.
A summary of the dry weight yields of the experiment is displayed in Figure 6.3.
The analysis of variance for the harvested aboveground plant material is presented
in Table 6.1. The coefficient of variation (CV) is high for a greenhouse experiment at
69.25%. No important differences were found between the two watering treatments.
202
6. Greenhouse Competition Between Paramo Grasses
The target species performances were found to be considerably disparate (p <
0.001), while the neighbour species effects were considered insignificant. The interaction terms proved of little consequence.
df
Source
Water Availability
Targets
Neighbours
Water Availability X Targets
Water Availability X Neighbours
Targets X Neighbours
Water Availability X Targets X Neighbours
Error
Total
SS
1
0.97
2 1181.55
2
46.72
2
76.77
2
53.84
61.01
4
4
83.06
72 1205.12
89 2709.04
CV = 69.25%
MS
F
0.969 0.06
590.776 35.30
23.360 1.40
38.386 2.29
26.918 1.61
15.253 0.91
20.766 1.24
16.738
30.439
p
Significance
0.811
0.000
0.254
0.108
0.207
0.462
0.301
NS
*
NS
NS
NS
NS
NS
Table 6.1
Three-way analysis of variance performed on aboveground (shoot) dry weight
yields of three species of Calamagrostis growing at high and low water availability.
The yields of each species are referred to as 'targets' and the yields of plants
grown with each species as 'neighbours'.
Number of Surviving Tillers
50
40 -
3o
20
10
0
Calamagrostis A
Calamogrostis B
Calamagrostis C
Associate Species
Colomagrostis A
Calomogrosfis B
Colamogroslis C
I=
Figure 6.2.
Number of surviving tillers after the establishment phase of six months. The total
survivorship for each species is subdivided into the number remaining when
grown with itself and each of the other two species. The values for pure stands
represent mean half-pot yields. The maximum possible survivorship for each
species is thus sixty tillers (twenty with each neighbour species).
o
•••
•
•
6. Greenhouse Competition Between Paramo Grasses
203
-o
coo°
c _c 2
20
co
0) Cl) Cn
•a. _ a) .ca
O
_a
0 a)
cti 0.
, QS
r 4-a)
E
>
a
Cl)
co
0)
8
(ts
w
(I)
O
En
°
0 C)
CU
0
1:3
c
o
>
2 13
t,>' •-:.
4E. a)
_c ce
-5 a)
H _c
.0
• io
t(1))
:E" a_
a)
ea
2
-to
2 >
cts
C) ca
7: a)
.
0)
u)
2-
•
In
0 .0 0 .0 0 .0 0
CNI 041 -
v
0
E
E
.1
U) 0
ca ;12 a,)
3 0.)
In
• .c
• .(12,
N.-
)7( rt
(a
Ws
•E'
"U)
p
C
,
E
>
_E
o
CC
• a) _ca) a)
ce)
c.)
.0
0
0
_C
L.
o
5
.c
• _c ••-•
CC) 1— L.•
<
-W 3
a) -..Sw
0) F.) 75 I c P_
v)
E
c
i)
C
-c)
cti a tr,
ct
IT) 6
51
O
o
Ca
0
C
CI)
E
a)
2 u„',
cc; 0_ u,
a) 3
7.5
caa)
>
Cn .5_
a)
E-
a,)
,,
- 5, 0 as
.=
0
,.
a) _c -=
s•-•
CC ..-C
C
r_
(. T)-
:
C.CI)
CC -0
a) a) I
E c Cn
CC
c
,,,
.e.L5
1:5
Ta
0 .9
0. En
'cf) 2
r
a)
a) t
co•>
CC o
C)
1630
V)
a,
co
13• C.)
-a3 0_5
-C
or) C-)
0
0. 0
•
0 cts
_c
•
-
Cf)
2
204
6. Greenhouse Competition Between Paramo Grasses
Table 6.2 presents the analysis of variance for the belowground harvested plant material. A highly significant difference between the watering regimes was exposed (p
< 0.001), despite a coefficient of variation of almost 90%. Meaningful differences
between target species yields and between neighbour effects were also discovered (p
< 0.001 and p = 0.005, respectively). An effect of watering treatment on target
species yields was shown to be highly significant (p < 0.001).
Source
df
Water Availability
Targets
Neighbours
Water Availability X Targets
Water Availability X Neighbours
Targets X Neighbours
Water Availability X Targets X Neighbours
Error
Total
1
2
2
2
2
4
4
72
89
SS
296.84
439.47
218.95
401.62
30.11
48.39
50.08
1403.05
2888.51
MS
296.84
219.73
109.48
200.81
15.06
12.10
12.52
19.49
32.46
F
15.23
11.28
5.62
10.30
an
0.62
0.64
p
Significance
0.000
0.000
0.005
0.000
0.466
0.649
0.634
QQQ
QQQ
QQ
000
NS
NS
NS
CV = 89.67%
Table 6.2
Three-way analysis of variance performed on belowground (root) dry weight
yields of three species of Calamagrostis growing at high and low water availability.
The yields of each species are referred to as 'targets' and the yields of plants
grown with each species as 'neighbours'.
The combined root and shoot analysis is presented in Table 6.3. A significant difference was found between watering regimes (p = 0.031). Target species yields (p <
0.001) and the effects of neighbour species on the yields (p = 0.031) were considered noteworthy. Finally, a watering-dependent effect on target species growth
was demonstrated as important (p = 0.001).
Source
dl
SS
Water Availabi ty
Targets
Neighbours
Water Availability X Targets
Water Availability X Neighbours
Targets X Neighbours
Water Availability X Targets X Neighbours
Error
Total
1
2
2
2
2
4
4
72
89
263.9
2977.1
396.3
822.3
134.2
164.2
226.6
3919.9
8904.4
MS
F
4.85
263.89
1488.54 27.34
198.15
3.64
411.14
7.55
67.11
1.23
41.04
0.75
56.65
1.04
54.44
100.05
Significance
p
0.031
0.000
0.031
0.001
0.298
0.559
0.392
000
QQQ
NS
NS
NS
CV = 68.12%
Table 6.3
Three-way analysis of variance performed on whole plant dry weight yields of
three species of Calamagrostis growing at high and low water availability. The
yields of each species are referred to as 'targets' and the yields of plants grown
with each species as 'neighbours'.
6. Greenhouse Competition Between Paramo Grasses
205
A series of figures summarize the results of the Student-Newman-Keuls (SNK)
tests. Figure 6.4 presents the mean yields for shoots, roots and whole plants for the
two watering treatments. Differences in watering regimes were found to be unimportant in aboveground yields. Roots were found to yield significantly more in drier
conditions, which was sufficient to influence the importance of the effect of watering
on the whole plants.
Calamagrostis A and B were inseparable in terms of mean aboveground dry weight
yields, though CalamagrostLy C was significantly higher than the others (Figure 6.5).
The belowground yields of each species were substantially different from the others:
Calamagrostis C yielding more than Calamagrostis B which in turn produced a
greater dry weight than Calamagrostis A. A similar situation was seen for the whole
plant yields.
Figure 6.6 shows the mean yields of the three species when grown with each neighbour species, that is, how well the other species grew when mixed with a particular
species. Thus, the mean dry weight yield for shoots of Calamagrostis spp. A, B and C
when grown with Calamagrostis A was 6.3g. No significant difference was found between neighbour species means for the aboveground plant portions. Belowground, a
greater mean yield was obtained for plants grown with Calamagrostis A than for
those grown with the other two species. For whole plant yields, Calamagrostis A
Dry We • ght (g)
Whole Plant
Roots
Shoots
Figure 6.4.
Mean yields of roots, shoots and entire plant (roots and shoots combined) for the two water
availability treatments in terms of dry weight. Separate comparisons between the two watering regimes were made for each of the roots, shoots and entire plant datasets. Means sharing a letter were not separated by a Student-Newman-Keuls test (n = 45).
6. Greenhouse Competition Between Par-am° Grasses
206
Dry Weight (g)
Whole Plant
Roots
Shoots
Calamagrostis
Figure 6.5.
Mean dry weight yields of roots, shoots and entire plant (roots and shoots combined) for
each target species regardless of the water availability regime and neighbour species. Separate comparisons between the three species were made for each of the roots, shoots and entire plant datasets. Means sharing a letter were not separated by a Student-Newman-Keuls
test (n = 30) .
Dry We • ght (g)
10
I
A
B
Whole Plant
Roots
// Shoots
C
Calamagrostis
Figure 6.6.
Mean dry weight yields of plants (regardless of species or watering regime) when grown with
each of the neighbour species. Separate comparisons between the three species were made
for each of the roots, shoots and entire plant datasets. Means sharing a letter were not separated by a Student-Newman-Keuls test (n = 30).
6. Greenhouse Competition Between Paramo Grasses
207
allowed plants growing with it to perform better than they did when grown with Calamagrostis C. The mean yield for plants mixed with Calamagrostis B was intermediate
between the other two species and could not be separated from either of them by a
SNK test.
No interaction effects were seen to be significant in aboveground yields. Figure 6.7
presents the mean target species yields for the two watering treatments for roots and
the entire plant. For the roots, no difference was found between these target species
x watering regime interaction means, with the exception of the Calamagrostis C
yield in the drier regime, which was significantly higher than the rest. For entire
plant production, once again the Calamagrostis C yield in the dry conditions was significantly higher than the others. This species also performed well in the wetter
watering treatment. Calamagrostis A in the drier conditions yielded the lowest dry
weight, with the remainder of the means intermediate.
The competitive balances between the species (relational effects) are shown in
Table 6.4. The most striking feature of these results is that the competitive differences between the species is much smaller than, sometimes even reversed, in the wetter conditions compared with the drier regime. Calamagrostis A is the weakest
competitor performing significantly worse than the other two species. The relationship between Calamagrostis B and C is interesting. In the dry regime, Calamagrostis
C yields much more than Calamagrostis B. However, in the wet regime, these species
yield almost the same overall (with Calamagrostis B yielding more aboveground and
Calamagrostis. C more belowground).
The yield of a mixture, when taken as a whole, may differ from that predicted from
the yields of the pure stands alone. These elements of interaction, the summational
effects of Harper (1977), are given in Table 6.5. Mixtures containing Calamagrostis C
consistently yield more than predicted from the performance of the pure stands, in
both wet and dry conditions. In the wet, the mixture of Calamagrostis A and B produced a greater yield than expected, whereas in the drier regime the yield of the
same mixture was depressed.
A different approach to comparing the yields in mixture with those of the pure
stands is presented in Table 6.6. The Relative Yield Total (RYT) equals one when
the yield of the mixture reflects the corresponding yields of the species grown in isolation. A value greater than one indicates that mutual stimulation of yield occurs,
while a value less than one indicates suppression. In general, the performance of mixtures in the wetter regime was enhanced compared with the pure stands. In the dry
treatment, mixtures containing Calamagrostis A were depressed. The mixture of Calamagrostis B and C followed expectation from the pure stands for the whole plant
(with the shoots yielding more and the roots less than predicted).
Figure 6.8 illustrates this situation for whole plants in the dry treatment. The joint
yield diagram for Calamagrostis B with C is almost a straight line linking the pure
stand yields, which indicates that the mixture yield can be predicted from the yields
of both species in isolation. The two other mixtures (A with B and A with C) lie
below this hypothetical straight line: the mixture yields are less than predicted. The
replacement series diagrams (on the right-hand side of Figure 6.8) illustrate the de-
6. Greenhouse Competition Between Pal-am° Grasses
(a) Roots
Dry Weight (g)
B
Calamagrostis
(b) Whole Plant
Dry Weight (g)
30
25
20
15
10
5
B
Calamagrostis
Figure 6.7.
Mean dry weight yields of species regardless of neighbour for the two moisture
regimes for (a) roots and (b) whole plant. Yields sharing a letter were not separated by a Student-Newman-Keuls test (n = 15).
208
6. Greenhouse Competition Between Paramo Grasses
209
tail of these findings. Compared with the pure stands (intraspecific effects), one
species influences the other more than predicted while itself being affected less.
In the wet regime (Figure 6.9), all mixtures show enhanced performance compared with the pure stands, shown by the convex line in the joint yield diagrams. The
replacement series diagrams also share the same form: both species show a convex
line, indicating that they both perform better than expected from the pure stand
yields.
Mixture
A versus B
A versus C
B versus C
Dry
Wet
Shoots Roots Whole Shoots Roots Whole
-0 93
-2.89
-5.12
-4.50
-5.01
-2.93
-4.44
-7.90
-8.05
0.48
-0.19
0.66
-1.52
-0.75
-0.61
-1.05
-0.94
0.05
Table 6.4.
The competitive relationships between the three Calamagrostis species (Relational effects of Harper, 1977). If the value is positive then the first species holds a
competitive advantage over the second species; if negative then vice versa.
Values are in terms of dry weight (g) per species.
Mixture
Dry
Wet
Shoots Roots Whole Shoots Roots Whole
A with B 1 22
A with C -059
B with C -2 41
-0.09
-1.19
-0.10
2.13
-1.78
-2.51
-2.65
-2.58
-1.75
-3.40
-1.97
-0.03
-6.05
-4.55
-1.78
Table 6.5.
The relationships between the yields of mixtures and pure stands of the three Calamagrostis species (Summational effects of Harper, 1977). Values are in terms of
dry we'ght (g) per species. If the yield of the mixture exceeds that predicted from
the yield of the pure stands then the value is negative; if below this predicted value
then the figure will be positive.
6. Greenhouse Competition Between Paramo Grasses
A with B
A with B
Dry Weight of A (g) Per Pet
e
Dry Weight (p) per pot
16
io
Dry We ght of B (g) per pot
20
25
IA eture
A with C
A•B
Pu
eB
A with C
Dry We el t A g per p I
2
210
Dry We ght g per pot
15
30
20
5
10
20
10
30
40
Dry We ght of C (g) per p t
0
Pure A
0 y We gni
Pure C
B with C
B with C
25
IA xture A • C
IS g ow pot
40
Dry We pet p per pot
30
16
20
10
6/-
0
0
10
ao
30
Dry We ght of C (g) per pot
40
o
Pure B
IA Aunt B•C
Figure 6.8.
Joint yield diagrams (left) and replacement series diagrams (right) for the three
Calamagrostis mixtures grown in the dry regime. In the replacement series
diagrams, the two solid lines represent the yield of each Calamagrostis species
and the dashed line the joint yield.
Pure C
6. Greenhouse Competition Between Paramo Grasses
A
with B
A
Dry Weight of A (g) per pot
25
211
with B
Dry Weight (g) per pot
20
16
10
2
4
6
a
Dry Weight of B (g) per pot
12
14
Pure A
A with C
Mixture MB
A
Dry We ght of A (g) per pot
Pure B
with C
Dry Weight (9) per pot
26
------------------------20
10
6
10
16
20
26
Dry Weight of C (g) per pot
0
Pure A
B with C
Mixture MC
Purl C
B with C
Dry We ght of B ftit per pot
26 Dry Weight (g) per pot
20
10
10
16
Dry Weight of C (g) per pot
20
26
Pure B
Mixture B•C
Figure 6.9.
Joint yield diagrams (left) and replacement series diagrams (right) for the three
Calamagrostis mixtures grown in the wet regime. In the replacement series
diagrams, the two solid lines represent the yield of each Calamagrostis species
and the dashed line the joint yield.
Pure C
6. Greenhouse Competition Between Paramo Grasses
Mixture
212
Wet
Dry
Shoots Roots Whole Shoots Roots Whole
A with B 0.63
A with C 0.84
B with C 1.15
1.00
0.96
0.90
0.72
0.90
1.02
2.20
1.80
1.47
2.75
2.36
1.08
2.46
1.89
1.24
Table 6.6.
The relationships between the yields of mixtures and pure stands of the three Ca/amagrostis species (Relative Yield Totals, RYTs, of de Wit & van den Burgh, 1965).
The values have no units. If the RYT >1, then a yield advantage is obtained in mixture; if RYT =1, no advantage is obtained; and if RYT <1 a disadvantage is indicated.
Discussion
The Cultivation of the Calamagrostis Species
efore dealing with the experimental results per se, it would be helpful to look at
B some of the problems encountered when trying to grow these species in the greenhouse.
It is standard procedure to begin an experiment of this nature with tillers taken
from mother plants. In this case, with large mother plants, it was easy to select suitable tillers of the same status. However, these tillers demonstrated remarkably high
mortality after separation and planting. The overall mortality was close to 52% (with
76%, 48% and 32% for Calamagrostis A, B and C respectively). It seems unlikely
that tillers would have been able to exert competitive effects on neighbours at this
stage. The discrepancy between the species is probably a result of the performance
of the plants themselves rather than the effects of neighbour tillers, despite the high
mortality of Calamagrostis A when grown with Calamagrostis B.
The cause of such mortality is unknown. The cultivation of tropical alpine plants is
notoriously difficult. At the Royal Botanic Gardens, Kew, in an attempt to simulate
the tropical alpine environment, artificial 12-hour daylength and marked diurnal temperature regimes (21°C during the day and 5°C at night) have been introduced to the
greenhouses by Tony Hall at the Alpine Department at the Royal Botanic Gardens,
Kew (personal communication). Even under this regime, some plants show very different growth (for example, Plantago rigida adopts a very different habit from the
cushion growth form seen in the paramo). Such environmental control was not
possible in this experiment and may have resulted in the death of a high proportion
of tillers. However, the survival of new individual tillers of recently burned Calamagrostis tussocks in the field (Pdramo de Daldal) was also low, at 85% and 60% for
two tussocks studied (Chapter 4). Perhaps low tiller longevity is a natural feature of
these plants. Clearly more research is needed to establish the expectations of tiller
mortality in these grass species.
6. Greenhouse Competition Between Paramo Grasses
213
Whatever the causes, high tiller mortality makes it difficult to design satisfactory
experiments. Further complications to experimental design followed the establishment phase. A number of tillers were found to remain 'dormant' after planting,
that is to say, they stayed alive without producing any new growth. The experimental
protocol of replacing dead tillers did not allow for the substitution of dormant ones;
besides, it would not have been possible to determine the potential of tillers for dormancy during the establishment period.
The presence of tillers more or less unchanged since planting at the start of the experiment contributed greatly to the variability of the data. In all, fifteen tillers were
classed as dormant out of the ninety that comprised the dataset: if these are removed
then the coefficient of variation is reduced from around 68% to below 28% for the
whole plant data. This would probably reduce further if it were possible to compensate for the extra yields obtained by the neighbours of dormant tillers.
The growth of these plants taken as a whole was much slower than that of local
Welsh upland grasses grown in similar conditions (Ramsay, unpublished data). The
stature of Calamagrostis A was always small, even in stock tussocks not involved in
the experiment. It would appear that some internal constraint on growth is present,
in addition to those limitations imposed on the plants by the greenhouse conditions
discussed earlier.
Finally, it is worth noting that a difference between species was observed in root
formation. Plants from the highest elevation (Calamagrostis A) developed much
finer roots than the other species, with the roots of Calamagrosth B appearing finer
than Calamagrostis C. This purely passing observation is supported by the work of
KOrner & Renhardt (1987) who found that plants from higher altitudes developed
about 4.5 times more fine root length per unit leaf area than low altitude plants. In
the Calamagrostis species used in this study, there appears to be a genetic basis for
these differences.
DiaIlel Yields
results of the diallel experiment do not support the hypothesis of better perforT he
mance in drought conditions by species from high altitude compared with low altitude ones. In fact, the results are the reverse of those expected: the species from the
highest altitude was the only one to yield less in the drier regime than in the wetter
regime. Furthermore, the relative competitive abilities of the plants used in these experiments were more pronounced in the drier conditions.
Overall, Calamagrostis C performed best, followed by Calamagrostis B with the
species from the highest altitude, Calamagrostis A, doing worst of all. This pattern
was reciprocated, in that the species which yielded most depressed the yield of its
neighbours most, and so on. These relative competitive abilities of the species were
sometimes influenced by the watering regime. For example, Calamagrostis C yielded
much more than Calamagrostis B in the drier regime, whereas in the wetter treatment they showed almost equal competitive ability. However, in general it appears
6. Greenhouse Competition Between Paramo Grasses
214
that Calamagrostis grasses from higher altitudes are less able competitors than those
from lower altitudes, irrespective of the watering regime. This supports the expectation of Grime (1979) that competition is less important at higher altitudes (where
stress is high) and that species from lower altitudes should be fast-growing, capturing
resources as quickly as possible. The upper distributional limit of several pdramo
plants and communities have been attributed to stress tolerance factors (particularly
frost), whereas the lower limit has most frequently been ascribed to competitive effects (inter alia Armesto, Arroyo & Villagran, 1980; Faritias & Monasterio, 1980; Alliende & Hoffmann, 1985; Perez, 1987; Miller & Silander, 1991).
In productivity terms, greater yields were obtained from plants subjected to the
drier treatment, mainly due to better root development (aboveground production
was little influenced by the watering regime), an effect that was particularly pronounced in Calamagrostis C. It seems unlikely that capillary matting would result in
root suppression in these species: more likely is that root growth was stimulated in
the dry regime.
The relationships between the yields of mixtures and those of their component
species in pure stands was also investigated. The RYT has been widely used to investigate the extent to which species in a mixture compete for common limiting resources (Snaydon, 1991). The term 'resource complementarity' was coined by Snaydon &
Satorre (1989) to describe such an index. The lower the RYT, the greater is the competition for limited resources. The summational effects, with minor differences, followed the same pattern as the RYTs described below.
In the drier treatment, RYTs were approximately equal to one, or less than one.
This implies that the species are making demands on the same limiting resources of
the environment (Harper, 1977), and in some cases this results in a reduced yield
compared with the pure stand yields.
In the wetter conditions, mixtures exceeded the performance of pure stands, with
RYTs greater than 1 and convex lines in the species replacement diagrams. Mixtures
containing Calamagrostis A resulted in very high RYT values: close to 2 or even
higher. This result is very rare in plant mixtures (Snaydon, 1991) and indicates complete or nearly complete resource complementarity (that is, the species avoid competition, making different demands on the environment). The most plausible
explanation for these very high RYT values with Calamagrostis A in the wet regime is
that, owing to the inherently slow growth of this species, competition for resources
between the species in the mixture did not occur at all, or was delayed until towards
the end of the experimental period.
The lack of phenotypic plasticity in Calamagrostis A clearly had an important bearing on the overall result of the experiment, decreasing the potential yield response of
this species. This may be advantageous in the High Andes by restraining opportunistic growth (and utilization of precious resources) in response to transient climatic
conditions: Grime's (1979) stress tolerant strategy. In the greenhouse, however, it
makes experiments of this nature difficult to interpret.
6. Greenhouse Competition Between P6ramo Grasses
215
One of the principal drawbacks of replacement experiments, such as that described here, is that the results obtained only apply to the particular conditions under
which the experiment was carried out (Jolliffe, Minjas & Ruenckles, 1984; Connolly,
1986). Firstly, the experiments were carried out at fixed densities, a situation which
does not reflect the field situation (Inouye & Schaffer, 1981). Secondly, the climatic
conditions did not echo those prevalent in the Andes. The limiting factors on growth
in such changed conditions may have been different and the competitive balance between the grass species may have been altered in favour of the lower altitude species.
For example, the warm temperatures throughout the 24-hour period may have
allowed the lower elevation species to capitalize on the potentially higher metabolic
activity at the expense of Calamagrostis A.
A second possibility is that the drought treatment was not severe enough to exert
sufficient influence on the outcome of competition between species. In fact, the
grasses did significantly better in the drier regime, yielding over one-third more than
in the wetter conditions, largely the result of lower root growth in the wet. Calamagrostis A is an exception in this, though, performing better in the wet —this is the result of greater tiller dormancy in the drier conditions. In the design of the
experiment the severity of the drier treatment was deliberately restricted because of
the high mortality rate experienced with these grasses. More stringent deprival of
water would perhaps have produced a different effect.
Therefore, before dismissing the original hypothesis outright, it is worth considering these arguments. Nonetheless, attention should be given to the rejection of the
hypothesis in hand, and to alternatives to explain the distribution of plant species
over the altitudinal gradient in the Ecuadorian paramos. Temperature certainly
shows a marked trend across the elevation gradient with a lapse rate of 0.6°C per
100m. The number of frost days also increases sharply over the altitudinal range of
the paramo (Sarmiento, 1986). This factor might influence plant distributions
through plant competition, or via flower and seed development and germination requirements. Frost could be particularly important with regard to the latter.
Future research in this direction should combine precise environmental measurements in the field with distributional data for the plant species. Greenhouse studies
should aim to simulate the paramo environment as closely as possible, since the complex relationship between environmental factors is perhaps more important than any
one factor taken alone. Field experiments to determine the outcome of competition
between species, in situ in the Andes, may reveal more useful information than the
artificial greenhouse environment.
Chapter 7
Overall Discussion
216
7. Overall Discussion
217
he paramo environment is unique to tropical alpine regions. It presents particular
T difficulties for the flora, described in earlier chapters. The harshness of these conditions confers a special fragility on the plant communities of the paramos. The evolutionary history of the Andes (principally the uplift of the Andes, periodic volcanic
activity and the periods of cooler weather during glaciations alternating with warmer
interglacials) has further contributed to this precarious existence.
Geographically, the isolated nature of paramo regions has resulted in distinct floras. During the warmer interglacial episodes, paramo vegetation contracted as the
vegetation zones were pushed higher up the mountains. High rates of speciation and
extinction are thought to have occurred during these periods of isolation, and accounts for the high level of endemism in pdramo taxa. At cooler times, the paramos
covered a larger area in the Andes, as the vegetation zones were lower down the
mountains. At such times, new taxa, evolved in isolation, were free to migrate
through the extensive paramo belt. The vegetation of the paramo has undergone a
series of about 15-20 contraction-expansion cycles with accompanying speciation and
radiation episodes, each cycle lasting approximately 100,000 years (Van der Hammen
& Cleef, 1987).
In a sea of warm tropical vegetation and subject to the above climatic and catastrophic changes, the paramos can be thought of as typically insular and short-lived
(Smith & Cleef, 1988). They rely on long-distance dispersal rather than local adaptation from the lowland tropical flora as the primary source of recruitment to the plant
communities. Cleef (1979) and Van der Hammen & Cleef (1986) found that about
half of the paramo flora of the Colombian Cordillera Oriental were of temperate
origin, the remainder of tropical (mostly neotropical) origin. In Ecuador, temperate
taxa dominate the zonal paramos, making up about two-thirds of the flora (Chapter
2). In all studies so far, paramo endemics account for just under 10% of the total
flora (9% in this work).
As a consequence of the above factors, the vascular plants of the paramo represent
the richest mountain flora in the world (Smith & Cleef, 1988). Luteyn, Cleef & Rangel (1992) estimate that it consists of 112 families, 479 genera and between 3,000 and
4,000 species. However, these estimates include all types of paramo vegetation: azonal bogs, woodlands and thickets as well as the zonal vegetation types from shrubby
sub-paramo to high altitude desert super-paramo.
In Chapter 2, the zonal paramo vegetation of Ecuador was sampled by means of
192 quadrats, amounting to a total area of 0.48 ha. In these samples, 348 vascular
plant taxa were recorded (9-12% of the total paramo flora of the Andes). These belonged to 117 genera and 46 families (24% and 41% of the total Andean paramo estimates, respectively). In view of the objective to sample only the typical zonal
vegetation and that the study areas represent only a small fraction of the whole paramo province from Central America to Peril, these numbers are unexpectedly high.
Basing his work on a limited sample of species, Balslev (1988) estimated that 60%
of the highland flora of Ecuador were found outside Ecuador, and 72% were transAndean (that is, present on both the eastern and western ranges of the Andes). In
the present study, 143 trans-Andean species were found ( or 41% of all species re-
218
7. Overall Discussion
corded). Approximately 33%, or 114 species, were restricted to samples from the
Eastern Cordillera sites (Paramo de Guamanf, Volcan Tungurahua, El Altar west, El
Altar east, Paramo de Daldal, Pdramo de Zapote Naida, Paramo de Cumbe and
Pdramo de Ona). The remaining 91 species (26%) were recorded only in quadrats
from study sites on the Western Cordillera (Volcan Chiles, Volcan Cotacachi, Volcan
Chimborazo and Päramo de Cajas).
Table 7.1 shows the number of species found in each study site. Less than onethird of the 348 species found in all of the sites was present at any one study area.
The three samples from the Paramo de Oria contained just 24 species, while 21 samples from the Paramo de Cajas were comprised of 117 species. These figures represent 0.6-0.8% and 2.9-3.9% of the total paramo flora, respectively.
Paramo Region
No of
Samples
Total Area
(ha)
No of
Species
% of Total
Paramo Species
Volcin Chiles
Volcan Cotacachi
Paramo de Guamanf
Volcin Tungurahua
El Altar (west)
El Altar (east)
Paramo de Da!dal
Vo can Ch mborazo
Piramo de Zapote Naida
Paramo de Cajas
Paramo de Cumbe
Paramo de Ona
21
21
21
15
15
18
18
21
9
21
9
3
0.05
0.05
0.05
0.04
0.04
0.05
0.05
0.05
0.02
0.05
0.02
0.01
94
89
97
52
92
71
91
37
47
117
71
24
2.4-3.1
2.2-3.0
2.4-3.2
1.3-1.7
2.3-3.1
1.8-2.4
2.3-3.0
0.9-1.2
1.2-1.6
2.9-3.9
1.8-2.4
0.6-0.8
Total
192
0.48
348
8.7-11.6
Table 7.1
Distribution of vascular plant species in the twelve study sites in Ecuador. The percentage of paramo species is based on an overall estimate of 3000-4000 species
(Luteyn, Cleef & Rangel, 1992).
Almost half of the species were recorded in just one study area and three-quarters
were present in fewer than four sites (Table 7.2). Only four species were present in
ten or more sites: Pernettya sp. [185], Hypochaeris sessiliflora, Calamagrostis sp. [251]
and Etyngium humile.
These differences in species composition between the study areas were reflected
in the plant communities. The principal patterns were Espeletia paramo in the north,
Neurolepis bamboo 'Aram° in the east and Rainshadow desert paramo in the west.
Furthermore, the grassy paramos, present all over the country, revealed similar regional patterns.
219
7. Overall Discussion
Frequency
Species recorded in only 1 Site
Species recorded in 2 Sites
Species recorded in 3 Sites
Species recorded in 4 Sites
Species recorded in 5 Sites
Species recorded in 6 Sites
Species recorded in 7 Sites
Species recorded in 8 Sites
Species recorded in 9 Sites
Species recorded in 10 Sites
Species recorded in 11 Sites
Species recorded in all 12 Sites
Total
%
168
66
26
26
24
16
10
3
5
2
1
1
48.3
19.0
7.5
7.5
6.9
4.6
2.9
0.9
1.4
0.6
0.3
0.3
348
100.0
Table 7.2.
Frequency distribution of vascular plant species in the twelve study sites
in Ecuador.
Despite these regional differences, altitudinal zonation of plant communities was
pronounced. The mid-altitude zones (tussock paramos and lower cushion paramos)
were extensive in many of the study areas. By contrast, other communities were restricted to much smaller areas, for a number of reasons. The high altitude desert
paramos and cushion paramos were confined by the limited extent of the land at
such elevations. The high altitude dwarfshrub paramo communities were restricted
to a very narrow attitudinal band for reasons that are not yet clear, but which may be
related to humidity (Cleef, 1981). Finally, at the lower end of the paramo range, the
shrubby sub-paramo was largely absent because of burning and agriculture.
There is very little information on the degree of intraspecific variation in paramo
species. Altitudinal ecotypes may exist in the paramos (Smith, 1980). Buckland &
Ramsay (in press) measured morphological parameters for several species along two
altitudinal gradients. Some of these species showed distinct morphological responses
to altitude. For example, Lycopodium sp. demonstrated a decrease in leaf size and
plant height with increasing altitude. Other species showed little or no correlation
with altitude.
In the greenhouse study described in Chapter 6, Calamagrostis sp. tussocks from
the highest altitude in the Paramo de Guamani had a fixed response regardless of
the prevailing environmental conditions. Growth was slow compared to Calamagrostis spp. from lower altitudes and tussock stature was also small. This implies that, for
some species at least, rate of growth, plant form and other features are under genetic
control. In other species, however, the plant form is partially controlled by the environment. For example, cultivated Plantago rigida at Kew has a very different habit to
the dense cushion observed in nature.
Thus, the zonal Ecuadorian paramos are relatively species-rich (possibly with significant intra-specific variation) and exhibit strong regional variations in composition.
Many of the communities are restricted to small areas, while their sensitivity to disturbance is high because of extreme environmental conditions.
7. Overall Discussion
220
The main threat to this diversity comes from agriculture. Although the high Andes
have been populated for thousands of years (Eckholm, 1975), there are two to three
times as many people living in the highlands now than were there immediately before the arrival of the first Europeans (Baker, 1978). In Ecuador, approximately 50%
of the population live in the Andean Highlands (Luteyn, 1992) and population
growth is among the highest in South America.
Traditionally, highland peoples of the Andes have achieved sufficiency by their
ability to exploit several distinct life zones simultaneously. Murra (1972) termed this
system "vertical control". A highland community would farm a number of geographically separated areas, deriving different products from each and at different times of
the year.
The arrival of the Spanish modified this lifestyle, through the introduction of new
crops and livestock and the widespread resettlement of peoples (Brush, 1976). However, despite these changes, Murra's model of vertical control remains the basis for
the subsistence economies of many Andean communities.
The valley of Daldal, surveyed in Chapter 2 and the main site for Chapter 5's productivity studies, is a good example of how such a system operates even at a local
scale. Farmers living in the settlement of Daldal (3,100 m) have a number of fields
nearby in which they grow crops such as maize, quinoa (Chenopodium), peas and a
range of other produce for domestic use and for sale at the local market. These farmers may also possess land at higher altitudes. Many farmers have land at 3,400 m,
which is used to cultivate tubers (potatoes, oca) and beans, and another more extensive holding in the paramo zone used for the rearing of livestock (mostly cattle and
horses). In the nearby settlement of Alao, some farmers occasionally visit the montane forest on the eastern slopes of the Andes to supplement their produce by hunting and fishing.
Since the land-use of extensive highland haciendas was changed by the 1964 Law
of Agrarian Reform and Colonisation, smaller, more intensive farms have become
common (Cabarle et al., 1989). As the population grows and as communities become
less isolated and part of the wider economy, many farmers are unable to operate vertical control. Instead, they are forced to survive with smaller plots or land of lower
quality (usually at higher altitudes). Thus, in the valley of Daldal, there were farmers
living at 3,400 m with only nearby plots on which to grow food. This requires the intensive use of this land. It is also less resilient than the traditional lifestyle. Cropping
at this altitude is less dependable and requires more land for the same crop. Since
these people are no longer able to meet all of their food demands, they are forced to
sell part of their harvest and trade for other necessities (including fuel, now that
most of the montane forest has gone). As a result of these pressures, conversion of
natural or semi-natural habitats to agriculture has proceeded very rapidly in recent
times, and the sub-paramo has completely disappeared in the Daldal valley.
This trend is mirrored in most other highland settlements and the sub-paramo has
been destroyed in many areas by conversion to arable land. Much of the remainder
of the paramo has also been affected by its widespread utilisation as grazing pasture.
As reported in Chapters 2 and 4, the poor herbage quality of mature tussocks leads
7. Overall Discussion
221
farmers to burn tussock paramos every two to four years. The effect of this practice
on the Andean environment has led to speculation that the Ecuadorian paramos
(covering an area of approximately 20,000 km2 —Bonifaz, 1981; Encalada, 1986)
may not be the true climax vegetation, but a secondary type maintained by burning.
Scattered throughout the paramo zone, there are woodlands of varying sizes, mostly
consisting of the genus Polylepis (Rosaceae) but often in association with Gynoxys
(Compositae). It has been argued that high-altitude forests, similar to these woodlands, once covered much of today's paramos, but have been destroyed over many
years by man-induced fires and replaced by the grasslands present today. Evidence
for this view includes:
• the regular practice of burning in many paramo grasslands
as a management tool for improving pasture quality (Ellenberg, 1958; Laegaard, 1992);
• the presence of woodlands and small patches of trees growing in areas unlikely to sustain fires, for example, scree
slopes and beside rivers (Laegaard, 1992);
• the ability of certain tree species to survive in the grassland
zone, some 400-500 m above the present forest limit (Ellenberg, 1958; Brandbyge & Holm-Nielsen, 1986);
• observations that paramo fires erode the edges of adjacent
forests (Laegaard, 1992);
• the lack of a transition zone from trees to grassland at the
present treeline and observations of such transitions at
higher elevations (Laegaard, 1992); and
• biogeographical information relating to birds endemic to
high-altitude woodlands (Fields, 1992).
Counter to these arguments, other authors believe the paramos to be a largely
natural phenomenon, representing the true climax vegetation of high altitudes in the
northern Andes. Observations used to support such a view include:
• the apparent restriction of high altitude forests to specific
microclimates in sites such as rocky slopes, river courses,
valley bottoms, etc. (Troll, 1959);
• the common occurrence of trees characteristic of the high
altitude forests around dwellings which replicate the microhabitats described above (Simpson, 1979), supported by
evidence from Brandbyge (1992) that Polylepis incana
growth near walls is twice that in the open;
7. Overall Discussion
222
• experimental investigations carried out by Smith (1978) in
Venezuela showed convincingly that Polylepis sericea was
restricted to microsites by poor establishment elsewhere.
• biogeographical evidence relating to the tree species themselves (Simpson, 1979) and fauna associated with them
(Simpson, 1979); and
• palaeohistorical data on Polylepis suggesting wide expansions and contractions with climatic changes without man
(Simpson, 1979).
Clearly, both sides of the debate have drawn upon powerful support for their case.
Without doubt, the current extent of paramo grassland is considerably greater than
would naturally exist in the absence of man. However, as Balslev & Luteyn (1992)
put it, "the question is not whether man has cut the Andean forest and continues to
maintain and increase paramo area, but rather whether he alone has been responsible for wiping out the high-elevation forests up to the presently observed timberline or if natural forces have controlled this."
The biogeographical evidence is confused, and has been used to support both
views. Owing to the complex climatic history of this region, with periods of speciation and radiation, it is difficult to assess whether species patterns observed today
reflect current forest limits or are the consequence of previous isolations and expansions. Similarly, palaeological data have not yet provided sufficient information to
determine the former extents of forest cover in this regard.
The ability of some tree species to survive at higher elevations than they are found
in nature is undisputed. However, this does not mean that they could colonise grassland if burning were halted. Simpson (1979) suggests Polylepis seeds are dispersed by
birds and germination rates of P incana are low (Brandbyge, 1992). Brandbyge &
Holm-Nielsen (1986) suggested that natural regeneration of Polylepis might depend
on favourable microclimatic conditions not offered by tussock paramo, and Smith
(1978) found Polylepis sericea unable to establish in Venezuelan 'Aram° vegetation
or on bare soil. The author has not observed Polylepis seedlings growing in open
grass paramo in Ecuador, only within woodlands or at their edges, on scree or beside
water near existing woods.
From this evidence, even without burning, Polylepis would invade paramo grassland at a very slow rate, gradually encroaching into the paramo from its current
woodland edges. In view of the relatively recent climatic changes (with the last warm
period ending about 3,000 years BP—Van der Hammen & Cleef, 1986), in order for
Polylepis to cover most of today's paramos, it would have been necessary for it to colonise new areas at a relatively fast rate. The means by which such a rate of spread
would have been accomplished has not been demonstrated.
Furthermore, the lack of relict traces of former forests in the grassland demands
further explanation. Particularly at the highest altitudes (Laegaard, 1992, suggests a
true timberline between 4,100-4,350 m in Ecuador) decomposition of charred trees
7. Overall Discussion
223
would be slow, and remnants might be expected in some places at least. There are no
reports of such finds.
Laegaard (1992) noted that transition zones from forest to grassland are largely absent from the present treeline, and that they have been observed at the upper limit of
the high altitude woodlands. However, Polylepis woodlands usually have sharp upper
boundaries and transitions are not found in the majority of cases. In addition, the
present treeline does occasionally present a transition zone, where human disturbance is absent. Therefore, these observations do not favour a higher or lower
treeline, but merely confirm the effects of burning on the boundary between forest
and grassland, especially in the most accessible areas.
Laegaard (1992) states that "all grass paramos are more or less regularly burned".
This seems unlikely. Although great areas of paramo are frequently burned, many
remote areas are not. The question then becomes, how frequently must fires occur to
prevent woodland establishment? When tussock grasses have not been burned for a
long time, fires can be both intense and far-reaching. Natural fires could occur under
certain circumstances and have been recorded in similar situations elsewhere (Givnish, McDiarmid & Buck, 1986). Such fires could cover very large areas if climatic
and topographic conditions allow. Therefore, if very occasional burning could prevent woodland establishment, natural explanations may be forthcoming.
If only the forest edges are destroyed by a neighbouring grass fire, very many fires
would be required to isolate pockets of woodland in the manner that has been proposed. It is unlikely that the frequency of natural fires could account for such a number.
Patches of paramo may be found which are isolated from other grasslands by efficient fire breaks (islands in lakes, land isolated by cliffs and watercourses, etc.) and
are almost certainly not burned by man. If Polylepis woodland was the natural climax
vegetation, then these areas would be forested. In many cases, they are not. According to Siltanen, Thurland & Casanova (1987), Polylepis trees can grow on a wide
range of soil types and depths. It exists on wet and dry soils and even on rocky scree
slopes. Therefore, an inappropriate substrate is not a viable explanation for the
trees' absence in these sites.
It is generally accepted that human populations only became a significant influence on the vegetation of the high Andes in the last few thousand years (Eckholm,
1975). The apparent capacity of certain plant species to survive fires demands explanation of how such ability evolved if fires are such a recent phenomenon in the paramos. Laegaard (1992) suggests that the ability to survive burning is a fortuitous
side-effect of selection for other traits, such as drought tolerance and resistance to
UV radiation. Although these features may confer a degree of fire resistance, it is
perhaps cold temperature avoidance that offers a more plausible explanation, since
insulation from cold temperatures is equally effective against heat.
One possible explanation for the current distribution of Polylepis woodlands is that
they are relict populations from the warm period some 3,000 years BP that are able
7. Overall Discussion
224
to regenerate within the self-perpetuating environment of the forest interior, but are
unable to expand much beyond the limits of the current woodlands.
Exactly where the natural timberline lies is difficult to judge on current information. However, it seems likely that it will vary considerably from region to region in a
similar way to the vegetation zones described in Chapter 2. Furthermore, local topographic features will alter its distribution at a finer scale.
To answer the fundamental question of high altitude forests requires a great deal
of further effort. However, even if the paramos are secondary vegetation types, this
does not alter the fact that they now cover large areas of the Andes. They are economically valuable and their sustainable management depends upon an understanding
of how the grassland functions.
Paramo fires are important determinants of plant communities. In Chapter 4, the
survival of paramo plants following burning was related to fire temperatures. Radiated heat and flames were responsible for the loss of aboveground leaves and stems
and the majority of regrowth was from belowground parts. However, high rates of
leaf and shoot mortality were observed in surviving plants in subsequent weeks and
recovery was a slow process, especially at high altitudes.
Despite these slow growth rates, the productivity experiments in the paramo
(Chapter 5) suggested that tussock productivity reaches equilibrium in 3-5 years,
which corresponds well to the observed practice of burning every 2-4 years. Furthermore, cutting the plots (to simulate burning) resulted in a higher yield in the paramo
plots. In agricultural terms, therefore, the practice of burning appears well founded.
In areas subjected to regular fires, plants which are poorly adapted to burning
might be displaced. Other species, able to survive fires or to colonise bare ground
after burning, might increase. It is likely, then, that paramo fires might be responsible for the loss of biodiversity. However, regularly burned paramo grassland is not
subject to the very high temperatures reached by unburned vegetation. Infrequent
fires, therefore, may result in more losses than frequent fires because of their higher
intensity.
In some paramos, burning can cause long term damage to the soil. Over-frequent
burning appears to be most prevalent at higher altitudes, where the process of recovery takes longer. Thus, burning every 2-4 years does not allow enough time for
complete recovery. Over-burning does occur at lower altitudes too, if the frequency
of fires is high. Where over-burning is practised, erosion can result (Portsch & Hicks,
1980; Ponce, 1984). The humus content of the soil can be lost during the fire, and the
lack of vegetation to bind the remaining soil material can lead to its loss during the
runoff events associated with heavy rainfall. Eventually, a sparse plant community remains, growing on a largely mineral substrate. The productivity of the land is lost, in
complete contrast to the original aim of the burning.
Grazing and trampling also affect the paramo vegetation. They are strongly associated with burning, since recently burned areas are preferred for foraging (Verweij &
Kok, 1992). To meet their nutritional requirements in the paramo, cattle must travel
7. Overall Discussion
225
long-distances and forage for long hours (Schmidt & Verweij, 1992). Diet selection
does occur (short grasses and sedges are preferred rather than tussock grasses —
Schmidt & Verweij, 1992), but the effect of this on the composition of plant communities is still unclear. Widespread trampling effects include the creation of
micro-terracing (an intimate series of cattle paths following the hillside contours)
and the compaction of wet ground by poaching of livestock hooves. Intertussock
plants are damaged by trampling, which may favour tussock species and tough rosette plants capable of surviving trampling.
Occasionally, small patches of highly modified paramo are found, which represent
the sheltering sites for livestock. Usually, they are dominated by short herbs (particularly, Lachemilla orbiculata) often with thistles (Sonchus oleraceus). None of the
paramos in this study were grazed by sheep and none showed the extreme modifications in response to heavy grazing described by Grubb, Lloyd & Pennington (unpublished) for the paramo of Volcan Antisana. In addition to the usual tussock grassland,
they found large areas of short turf and other areas dominated by tough mats of A zorella pedunculata.
However, at lower altitudes in the Daldal valley, in the ceja andina zone, heavy
grazing and trampling had resulted in the dominance of A zorella pedunculata mats
(Chapter 5). These mats reduced the herbage production to levels comparable to desert conditions. In such circumstances, the addition of fertilizers produced minimal
effect.
Population pressure is not only reflected in increased agricultural impacts. Fuel requirements have resulted in the widespread destruction of montane forests. In some
areas, shrubs from the sub-paramo are now used as the principal source of fuel (personal observation). In other areas, particularly around Volcdn Chimborazo, shrubs of
Chuquiraga jussieui are collected from altitudes of 4,300 m and above. This corresponds to the use of Llareta cushions (A zorella spp.) for fuel in Peru (Hodge, 1946,
1960). Efforts are underway to resolve the 'poor man's energy crisis' by planting the
highlands with native species (Brandbyge & Holm-Nielsen, 1986; Brandbyge, 1992).
The paramo zone between 3,600-3,700 m has been proposed for aforestation with
Pinus radiata (Miller, 1976).
A number of Ecuadorian National Parks and other protected areas include large
areas of paramo (Sangay, Podocarpus, Cotopaxi, Cajas, Cayambe-Coca, CotacachiCayapas). One of the fundamental roles of these areas is to foster contact with the
natural environment for urban populations (Ponce, 1984). Of course, this has lead to
conflicts with one of the other roles of these areas —to promote nature conservation.
For example, in Cajas, with more than 25,000 visitors per year, fishermen are responsible for a high proportion of paramo fires, spread from campfires. In Sangay National Park, hunters of deer and tapir set fires to flush out their prey. Other threats
come from trampling and the collection of shrubs and trees for fuel. As the number
of visitors to paramos increases, it is important to protect the ecosystems from a
corresponding rise in environmental degradation.
Clearly, the pdramo ecosystem is under threat from many sources, the main one
being an increase in agriculture within the pdramo zone. The paramo is a valuable re-
7. Overall Discussion
226
source in biological terms (with high biodiversity and significant regional and altitudinal variation), but is also an important resource for major centres of human population. Most highland towns and cities depend upon paramo regions for their water
supplies and increasingly for leisure activities and tourism. Furthermore, there is a
long history of medicinal uses for paramo species. The best-known pdramo plant is
Chuquiragua (Chuquiraga jussieui) which, among a wide range of uses, has been advocated for the treatment of malaria (Paredes, 1962). Concern over these issues has
led to public pressure to conserve and manage these regions in a sustainable manner.
Practices of burning may be sustainable within certain limits, though an increase in
the frequency and extent of burning should be considered cautiously. Wherever
possible, the conversion of paramo into arable land and pasture should be avoided. It
appears from the results of the productivity studies that lower altitude pastures may
be substantially under-productive and attention here may negate the requirement for
more land in the paramo zone.
Although a strong political commitment and improved social standards are important elements in the long-term conservation of the paramos and the maintenance of
their biological diversity, the understanding and participation of the rural poor in
the planning, design and management of such strategies are essential. Before this can
occur, one major difficulty needs to be resolved. The precarious existence of highland families makes the adoption of new methods and management practices a high
risk enterprise. One cannot expect people to venture their very lives on new management models, even if they have been demonstrated scientifically, without adequate
financial and social backing.
A great deal of research is currently being carried out in Colombia and Venezuela
on the issues of paramo biodiversity, management and community dynamics. While
much of this work will be directly relevant to Ecuador, it is clear from this study that
the Ecuadorian paramos differ from those further north and more research is
needed in Ecuador itself. Researchers who follow this path will find it both challenging and rewarding.
References
227
References
228
Acosta-Solfs, M. (1937). Excursion botanica al paramo del Angel. Flora (Quito), 1:103-118.
Acosta-Solfs, M. (1960). Los pastizales naturales del Ecuador: conservaciOn y approvechamiento
de los paramos y sabanas. Revista Geografica (Rio de Janeiro), 53: 87-99.
Acosta-Solfs, M. (1966). Las divisiones fitogeograficas y las formaciones geobotanicas del Ecuador. Revista de la Acadamia Colombiana de Ciencias Exactas, Fisicas y Naturales, 12: 401-47.
Acosta-Solfs, M. (1984). Los Paramos Andinos del Ecuador. Quito: Publicaciones Cientfficas
M.A.S.
Acosta-Solis, M. (1985). El Arenal del Chimborazo, ejemplo de puna en el Ecuador. Revista Geogreica (Quito), 22: 115-122.
Alliende, M.C. & Hoffmann, A.J. (1985). Plants intruding Laretia acaulis (Umbelliferae), a high Andean cushion plant. Vegetatio, 60:151-156.
Armesto, J.J., Arroyo, M.T.K. & Villagran, C. (1980). Attitudinal distribution, cover and size structure of umbelliferous cushion plants in the high Andes of Central Chile. Oecologia Generum, 1:
327-332.
AzOcar, A. & Monasterio, M. (1979). Variabilidad ambiental en el 'Aram° de Mucubajf. In, M.L.
Salgado-Labouriau (ed.), El Medio Ambiente Pdramo. Merida (Venezuela): Ediciones de Estudios Avanzados.
AzOcar, A. & Monasterio, M. (1980). Estudio de la variabilidad meso y microclimatica en el Paramo de Mucubajf. In, M. Monasterio (ed.), Estudios EcolOgicos en los Nramos Andinos. Merida
(Venezuela): Ediciones de la Universidad de los Andes.
AzOcar, A., Rada, F. & Goldstein, G. (1988). Freezing tolerance in Draba chionophila, a 'miniature' caulescent rosette species. Oecologia, 75: 156-160.
Baker, P. (1978). The Biology of High Altitude Peoples. Cambridge: Cambridge University Press.
Baldock, J.W. (1982). Geology of Ecuador. Explanatory Bulletin of the National Geological Map of
the Republic of Ecuador, 1:1,000,000 Scale. Joint publication. London: Natural Environmental
Research Council. Quito: Ministerio de Recursos Naturales y Energeticos.
Balslev, H. (1988). Distribution patterns of Ecuadorean plant species. Taxon, 37: 567-577.
Balslev, H. & De Vries, T. (1982). Diversidad de la vegetaciOn en cuatro cuadrantes en el paramo
arbustivo del Cotopaxi, Ecuador. Publicaciones del Museo Ecatoriano de Ciencias Naturales,
Arlo 3, 3: 20-32.
Balslev, H. & Luteyn, J.L. (1992). Paramo: An Andean Ecosystem under Human Influence. London: Academic Press.
Banco Central del Ecuador (1990). Invierta en el Ecuador. Quito: Banco Central del Ecuador.
Barkman, J.J. (1988). New systems of plant growth forms and phenological plant types. In: M.J.A.
Werger, RJ.M. van der Aart, H.J. During & J.T.A. Verhoeven (eds.), Plant Form and Vegetation
Structure. The Hague: SPB Academic Publishing.
Barnes, P.W., Flint, S.D. & Caldwell, M.M. (1987). Photosynthesis damage and protective pigments in plants from a latitudinal arctic/alpine gradient exposed to supplemental UV-B radiation
In the field. Arctic and Alpine Research, 19: 21-27.
Barnett, A. & Gordon, C. (1985). Mammal Report. In, A. Barnett (ed.), U.C.N. VV. Exploration Club
Expedition to CAJAS National Park, Ecuador, 1984: Preliminary Report. Privately published.
References
229
Barry, R.G. & Van Wie, C.C. (1974). Topo- and microclimatology in alpine areas. In, J.D. Ives &
R.G. Barry (eds.), Arctic and Alpine Environments. London: Methuen.
Barthelemy, D., Edelin, C. & Halle, F. (1989). Architectural concepts for tropical trees. In, L.B.
Holm-Nielsen, I.C. Nielsen & H. Balslev (eds.), Tropical Forests: Botanical Dynamics, Speciation and Diversity. London: Academic Press.
Baruch, Z. (1979). Elevational differentiation in Espeletia schultzii (Compositae), a giant rosette
plant, from the Venezuelan paramos. Ecology, 60: 85-98.
Baruch, Z. (1982). Patterns of energy content in plants from the Venezuelan paramos. Oecologia,
55: 47-52.
Baruch, Z. (1984). Ordination and classification of vegetation along an altitudinal gradient in the
Venezuelan paramos. Vegetatio, 55: 115-126.
Baruch, Z. & Smith, A.P. (1979). Morphological and physiological correlates of niche breadth in
two species of Espeletia (Compositae) in the Venezuelan Andes. Oecologia, 38: 71-82.
Beck, E., Scheibe, R. & Schulze, E.D. (1986). Recovery from fire: observations in the alpine vegetation of western Mt. Kilimanjaro (Tanzania). Phytocoenologia, 14: 55-77.
Beek, K.J. & Bramao, D.L. (1968). Soil map of South America. In, E.J. Fittkau, J. lilies, H. Klinge,
G.H. Schwabe & H. Sioli (eds.), Biogeography and Ecology in South America, Vol. 1. The
Hague: Dr. W. Junk.
Benoist, R. (1935). Le Plantago rigida H.B.K., sa structure, sa biologie. Bulletin de la Societe Botanique Francais, 82: 462-466; and 82: 604-609.
Billings, W.D. (1973). Arctic and alpine vegetations: similarities, differences and susceptibility to
disturbance. Bioscience, 23: 697-704.
Billings, W.D. & Mooney, H.A. (1968). The ecology of arctic and alpine plants. Biological Reviews, 43: 481-529.
Black, J. (1982). Los paramos del Antisana. Revista Geogràfica (Quito), 17: 25-52.
Bliss, L.C. (1971). Arctic and alpine plant life cycles. Annual Review of Ecology and Systematics,
2: 405-438.
Bonifaz, E. (1981). Agricultura y población de los Andes. Revista Geografica (Quito), 14: 31-42.
Bon, M.A. (1978). Parque Nacional ChirripO. In, M.A. Boza (ed.), Los Parques Nacionales de
Costa Rica. Madrid: Instututo de la Caza Fotografica y Ciencias de Naturaleza.
Brandbyge, J. (1992). Planting of local woody species in the paramo. In, H. Balslev & J.L. Luteyn
(eds.), Paramo: An Andean Ecosystem under Human Influence. London: Academic Press.
Brandbyge, J. & Holm-Nielsen, L.B. (1986). Reforestation of the High Andes with Local Species.
Reports from the Botanical Institute, University of Aarhus, No. 13.
Breedlove, D.E. (1973). The phytogeography and vegetation of Chiapas (Mexico). In, A. Graham
(ed.), Vegetation and Vegetational History of Northern Latin America. Amsterdam: Elsevier.
Bromley, R.J. (1971). Neurolepis as a sensitive indicator of human activity in the High Andes.
American Antiquity, 36: 475.
Brush, S.B. (1976). Man's use of an Andean ecosystem. Human Ecology, 4:147-166.
References
230
Buckland, S.M. & Ramsay, RM. (in press). An investigation of intra-specific variation in plant morphology along altitudinal gradients in the Ecuadorian paramos. In, PM. Ramsay, Ecological
Studies in Parque Nacional Sangay, Ecuador: a Report on the Sangay '89 Expedition.
Cabarle, B.J., Crespi, M., Dodson, C.H., Luzuriaga, C., Rose, D. & Shores, J.N. (1989). An assessment of biological diversity and tropical forests for Ecuador. Washington: The Center for International Development and Environment of the World Resources Institute.
Cabrera, A.L. (1957). La vegetaciOn de la Puna Argentina. Revista de las lnvestigaciones Agricolas de Buenos Aires, 11:317-412.
Cabrera, A.L. (1968). Ecologfa vegetal de la puna. In, Geo-ecology of the Mountainous Regions
of the Tropical Americas. Proceedings of the UNESCO Mexico symposium, Aug. 1966. Colloquium Geographicum, 9: 91-116.
Cambridge Llanganati Expedition 1969 (1970). Cambridge Uanganati Expedition 1969. Cambridge Expeditions Journal, 1969-70.
Carlquist, S. (1974). Island Biology. New York: Columbia University Press.
Carlquist, S. (1987). Anatomy of tropical alpine plants. In, P Rundel, F. Meinzer & A.P. Smith
(eds.), Tropical Alpine Environments: Plant Form and Function. Berlin: Springer-Verlag.
Cer6n, C.E. (1985). Los paramos de Pisayambo. Revista Geogrdfica (Quito), 22: 7-24.
Clapperton, C.M. & McEwan, C. (1985). Late Quaternary moraines in the Chimborazo area,
Ecuador. Arctic and Alpine Research, 17:135-142.
Clark, F.E. & Paul, E.A. (1970). The microflora of grassland. Advances in Agronomy, 22: 375-435.
Cleef, A.M. (1978). Characteristics of neotropical paramo vegetation and its subantarctic relations. In, C. Troll & W. Lauer (eds.), Geoecological Relations Between the Southern Temperate
Zone and the Tropical Mountains, Erdwissenschaftliche Forschung, 11: 365-390.
Cleef, A.M. (1979). The phytogeographical position of the neotropical vascular paramo flora with
special reference to the Colombian Cordillera Oriental. In, K. Larsen & L.B. Holm-Nielsen (eds.),
Tropical Botany. New York: Academic Press.
Cleef, A.M. (1981). The Vegetation of the Colombian Cordillera Oriental. University of Utrecht:
Proefschrift.
Cleef, A.M. (1983). Fitogeograffa y composiciOn de la flora vascular de los paramos de la Cordillera Oriental Colombiana (estudio comparativo con otras altas montafias del trOpico). Revista
de la Academia Colombiana de Ciencias Exactas, Fisicas y Naturales, 58: 23-29.
Clements, F.E. (1916). Plant Succession: An Analysis of the Development of Vegetation. Washington (DC): Carnegie Institution.
Clements, F.E. (1920). Plant indicators: the relation of plant communities to process and practice.
lbid, 290.
Coe, M.J. (1967). The Ecology of the Alpine Zone of Mount Kenya. Monogr. Biol. 17. The Hague:
Junk.
Connolly, J. (1986). On difficulties with replacement series methodology in mixture experiments.
Journal of Applied Ecology, 23:125-137.
References
231
Coupland, R.T. (1975). Productivity of grassland ecosystems. In, Productivity of World Ecosystems: Proceedings of a Symposium. International Biological Program. Washington, D.C.: National Academy of Sciences.
Crawford, R.M.M., Wishart, D. & Campbell, R.M. (1970). A numerical analysis of high altitude
scrub vegetation in relation to soil erosion in the eastern Cordillera of Peru. Journal of Ecology,
58: 173-191.
Cuatrecasas, J. (1934). Observaciones geobotanicas en Colombia. Trabajos del Museo Nacional
de Ciencias Naturales Madrid, Serie Botanica, 27.
Cuatrecasas, J. (1958). Aspectos de la vegetaciOn natural de Colombia. Revista de la Academia
Colombiana de Ciencias Exactas, Fisicas y Naturales, 10: 221-264.
Cuatrecasas, J. (1968). Paramo vegetation and its life forms. Colloquium Geographicum, 9: 163186.
Cuatrecasas, J. (1979). Growth forms of the Espeletiinae and their correlation to vegetation types
of the high tropical Andes. In, K. Larsen & L.B. Holm-Nielsen (eds.), Tropical Botany. New York:
Academic Press.
Cuatrecasas, J. (1986). Speciation and radiation of the Espeletiinae in the Andes. In, F. Vuilleumier & M. Monasterio (eds.), High Altitude Tropical Biogeography. Oxford: Oxford University
Press.
Descimon, H. (1986). Origins of the Lepidopteran faunas of the high tropical Andes. In, F. Vuilleumier & M. Monasterio (eds.), High Altitude Tropical Biogeography. Oxford: Oxford University
Press
Diels, L. (1934). Die paramo der Aquatorial Hoch-Anden. Sitzungsberichte der Preussischen
Akademie der Wissenschaften, 1934: 57-68.
Dobzhansky, T. (1950). Evolution in the tropics. American Scientist, 38: 209-221.
Drew, W.B. (1944). Algunas observaciones hechas en mi segunda excursion por las cercanias del
Cayambe. Flora (Quito), 4: 77-80.
Du Rietz, G.E. (1931). Life Forms of Terrestrial Flowering Plants. I. Acta Phytogeographic Suecica,
3:1-95.
Eckholm, E. (1975). The deterioration of mountain environments. Science, 189: 764-770.
Eidt, R.C. (1968). The climatology of South America. In, E.J. Fittkau, J. lilies, H. Klinge, G.H.
Schwabe & H. Sioli (eds.), Biogeography and Ecology in South America, Vol. 2. The Hague:
Junk.
Ellenberg, H. (1958). Wald oder Steppe? Die natOrliche Pflanzendecke der Anden Perus. I. Umschau Heft, 21: 645-648.
Ellenberg, H. (1979). The Tansley Lecture: Man's Influence on tropical mountain ecosystems in
South America. Journal of Ecology, 67: 401-416.
Ellenberg, H. & Milller-Dombois, D. (1967). Tentative physiognomic-ecological classification of
plant formations of the earth. Bericht des Geobotanischen lnstituts der E.TH., Stiftung RiThel in
Zurich, 37: 21-55.
Encalada, M. (1986). Evidencias del Deterioro Ambiental en el Ecuador. Quito: Gangotena & Ruiz
Editores S.A.
References
232
Farifias, M. & Monasterio, M. (1980). La vegetaciOn del ['Aram° de Mucubajf. Analisis de ordan amiento y su interpretaciOn ecolOgica. In, M. Monasterio (ed.), Estudios EcolOgicos en los Pdramos Andinos. Merida (Venezuela): Ediciones de la Universidad de los Andes.
Field* J. (1992). Biogeography of the birds of the Polylepis woodlands of the Andes. In, H. Balslev & J.L. Luteyn (eds.), Pdramo: An Andean Ecosystem under Human Influence. London: Academic Press.
Fisher, R.A. (1930). The Genetical Theory of Natural Selection. Oxford: Clarendon Press.
Flohn, H. (1974). Contribution to a comparative meteorology of mountain areas. In, J.D. Ives &
R.G. Berry (eds.), Arctic and Alpine Environments. London: Methuen.
Forster, N.R. (1989). Minifundistas in Tungurahua, Ecuador: survival on the agricultural ladder. In,
W.C. Thiesenheusen, Searching for Agrarian Reform in Latin America. London: Unwin Hyman.
Fosberg, F.R. (1944). El Paramo de Sumapaz, Colombia. Journal of the New York Botanical Garden„ 45: 226-234.
Fosberg, F.R. (1967). A classification of vegetation for general purposes. In, G.F. Peterken (ed.),
Guide to the Check Sheet for MP Areas, IBP Handbook No. 4. Oxford: Blackwell Scientific Publications.
Franco, R, Rangel, 0. & Lozano, G. (1986). Estudios ecolOgicos en la Cordillera Oriental. II. Las
comunidades vegetates de los alrededores de la Laguna de Chingaza (Cundinamarca). Caldasia, 15: 219-248.
Frei, E. (1958). Eine Studie Ober den Zusammenhang zwischen Bodentyp, Klima und Vegetation
in Ecuador. Plant and Soil, 9: 215-236.
Furrer, G. & Graf, K. (1978). Die subnivale HOhenstrufe am Kilimandjaro und in den Anden Boliviens und Ecuadors. Erdwissenschaftliche Forschung, 11: 441-457.
Garnett, A. (1937). Insolation and Relief. Transactions (Institute of British Geographers), 5.
Geiger, R. (1966). The Climate Near the Ground. Cambridge (Massachussets): Harvard University
Press.
Geiger, R. (1969). Topoclimates. In, H. Flohn (ed.), General Climatology. Amsterdam: Elsevier.
Givnish, T.J., McDiarmid, R.W. & Buck, W.R. (1986). Fire adaptation in Neblinaria celiae (Theaceae), a high elevation rosette shrub endemic to a wet equatorial tepui. Oecologia, 70: 481-483.
Gleason, H.A. (1917). The structure and development of the plant association. Bulletin of the Torrey Botanical Club, 44: 463-481.
Gleason, H.A. (1976). Delving into the history of American ecology. Bulletin of the Ecological Society of America, 56: 7-10.
Godley, E.J. (1978). Cushion Bogs. In: C. Troll & W. Lauer (eds.), Geoecological Relations Between the Southern Temperate Zone and the Tropical Mountains. Wiesbaden: Franz Steiner.
Goldstein, G. & Meinzer, F. (1983). Influence of insulating dead leaves and low temperatures on
water balance In an Andean giant rosette plant. Plant Cell and Environment, 6: 649-656.
Goldstein, G., Meinzer, F. & Monasterio, M. (1984). The role of capacitance in the water balance
of Andean giant rosette species. Plant Cell and Environment, 7: 179-186.
References
233
Goldstein, G., Meinzer, F. & Monasterio, M. (1985). Physiological and mechanical factors in relation to size-dependent mortality in an Andean giant rosette species. Acta Oecologia/Oecologia
Plantarum, 6: 263-275.
Gemez-Pampa, A. (1973). Ecology of the vegetation of Veracruz. In, A. Graham (ed.), Vegetation
and Vegetational History of Northern Latin America. Amsterdam: Elsevier.
Gondard, R (1988). Land use in the Andean region of Ecuador: from inventory to analysis. Land
Use Policy, 5: 341-348.
Grace, J. (1988). The functional significance of short stature in montane vegetation. In: M.J.A. Werger, RJ.M. van der Aart, H.J. During & J.T.A. Verhoeven (eds.), Plant Form and Vegetation Structure. The Hague: SPB Academic Publishing.
Gradstein, S.R., Cleef, A.M. & Fulford, M.H. (1977). Studies on Colombian cryptogams. II. Hepaticae— oilbody structureand ecological distribution of selected species of tropical Andean
Jungermanniales. Proceedings, Koninklijke Nederlandse Akademie van Wetenschappen,
Series C, 80: 377-420.
Griesebach, A. (1872). Die Vegetation der Erde nach ihrer klimatischen Anordnung. Ein Abriss
der vergleichende Geographie der Pflanzen, Leipzig.
Grime, J.P. (1979). Plant Strategies and Vegetation Processes. Chichester: Wiley.
Grubb, P.J. (1970). The impact of man on the paramo of Cerro Antisana, Ecuador. Journal of Applied Ecology, 58: 7-8.
Grubb, P.J. (1977). The maintenance of species-richness in plant communities: the importance of
the regeneration niche. Biological Reviews, 52:107-145.
Grubb, P.J., Lloyd, J.R. & Pennington, T.D. (unpublished). The Paramo of Cerro Antisana in the
Ecuadorian Andes: the impact of grazing and burning.
Halle, F. & Oldeman, R.A.A. (1970). Essai sur l'Architecture et la Dynamique de Croissance des
Arbres Tropicaux. Paris: Masson.
Halle, F., Oldeman, R.A.A. & Tomlinson, P.B. (1978). Tropical Trees and Forests: An Architectural Analysis. Berlin: Springer-Verlag.
Halloy, S. (1983). Some ecological data on Nototriche caesia, the high Andean Malvaceae in the
Cumbres Calchaquies Tucuman, Argentina. Lilloa, 36: 85-104.
Haney, E.B. & Haney, W.G. (1989). The agrarian transition in highland Ecuador. In, W.C. Thiesenheusen, Searching for Agrarian Reform in Latin America. London: Unwin Hyman.
Hanselman, D.R (1975). Species Diversity in Tundra Environments along a Latitudinal Gradient
from the Andes to the Arctic. M.Sc. thesis, Duke University, Durham, USA.
Harling, G. (1979). The vegetation types of Ecuador—a brief survey. In, K. Larsen & L.B. HolmNielsen (eds.), Tropical Botany. Academic Press.
Harling, G. & Sparre, B. (1973-). Flora of Ecuador. Department of Systematic Botany, University
of Goteborg, and the Section of Botany, Riksmuseum, Stockholm.
Harper, J.L. (1977). Population Biology of Plants. London: Academic Press
Hedberg, 0. (1957). Afroalpine Vascular Plants: a Taxonomic Revision. Symbolae Botanicae Upsaliensis, 15:1-411.
References
234
Hedberg, 0. (1964). Features of Afroalpine Plant Ecology. Acta Phytogeographic Suecica, 49: 0-0.
Hedberg, 0. (1992). Afroalpine vegetation compared to paramo: convergent adaptations and
divergent differentiation. In, H. Balslev & J.L. Luteyn (eds.), Pdramo: An Andean Ecosystem
under Human Influence. London: Academic Press.
Hedberg, I. & Hedberg, 0. (1979). Tropical-alpine life-forms of vascular plants. Oikos, 33: 297-307.
Heilborn, 0. (1925). Contributions to the ecology of the Ecuadorian paramos with special reference to the cushion plants and osmotic pressure. Svensk Botanisk Tidskrift, 19: 153-170.
Herrera, N. (1987). La racionalidad campesina andina y la alimentatciOn: el caso de la comuna de
Yanaturo en la Sierra Central del Ecuador. Agricultura y Sociedad, 45: 183-227.
Hill, M.O. (1979). TW INSPA N: a FORTRA N Program for Arranging Multivariate Data in an Ordered Twoway Table by Classification of the Individuals and Attributes. Cornell Ecology Programs Series.
Ithaca, (N.Y.): Cornell University.
Hnatiuk, R.J. (1978). The growth of tussock grasses on an equatorial high mountain and on two
Sub-Antarctic islands. In: C. Troll & W. Lauer (eds.) Geoecological Relations Between the
Southern Temperate Zone and the Tropical Mountains. Wiesbaden: Franz Steiner.
Hobbs, R.J. & Legg, C.J. (1983). Markov models and initial floristic composition in heathland
vegetation dynamics. Vegetatio, 56: 31-43.
Hodge, W.H. (1946). Cushion plants of the Peruvian puna. Journal of the New York Botanical Garden, 47:133-141.
Hodge, W.H. (1960). Yareta -fuel umbellifer of the Andean puna. Economic Botany, 14:113-118.
Hoffstetter, R. (1986). High Andean mammalian faunas during the Plio-Pleistocene. In, F. Vuilleumier & M. Monasterio (eds ), High Altitude Tropical Biogeography. Oxford: Oxford University
Press.
Holdridge, L.R. (1967). Life Zone Ecology. San Jose (Costa Rica): Tropical Science Center.
Holland, P.G. & Steyn, D.G. (1975). Vegetational responses to latitudinal variations in slope angle
and aspect. Journal of Biogeography, 2:179-184.
Hope, A.C.A. (1968). A simplified Monte Carlo significance test procedure. Journal of the Royal
Statistical Society, Series B, 30: 582-598.
Horn, H.S. (1974). The ecology of secondary succession. Annual Review of Ecology and Systematics, 5: 25-37.
Horn, H.S. (1975). Markovian properties of forest succession. In, M.L. Cody & S.M. Otiamond
(eds.), Ecology and Evolution of Communities. Cambridge (Massachussets): Belknap, Harvard
University Press.
Huber, 0. (1987). Neotropical savannas: their flora and vegetation. Trends in Ecology and Evolution, 2: 67-71.
Huey, R. (1978). Latitudinal pattern of between-altitude faunal similarity: mountain passes may be
'higher' in the tropics. American Naturalist, 112: 225-229.
Inouye, R.S. & Schaffer, W.M. (1981). On the ecological meaning of ratio (de Wit) diagrams in
plant ecology. Ecology, 62:1679-1681.
James, D.E. (1973). The evolution of the Andes. Scientific American, 229: 60-67.
References
235
Janzen, D.H. (1967). Why mountain passes are higher in the tropics. American Naturalist, 101:
233-249.
Jenny, H. (1948). Great soil groups in the equatorial regions of Colombia, South America. Soil
Science, 66: 5-28.
Johnson, A.M. (1976). The climate of Peru, Bolivia and Ecuador. In, W. Schwerdtfeger (ed.),
World Survey of Climatology, Vol. 12: Climates of Central and South America. Amsterdam: Elsevier Scientific Publishing.
Jolliffe, RA., Minjas, A.N. & Ruenckles, V.C. (1984). A reinterpretation of yield relationships in replacement series experiments. Journal of Applied Ecology, 21: 277-243.
Jones, H.G. (1983). Plants and Microclimate. Cambridge: Cambridge University Press.
Kaufmann (1977). Soil temperature and drying cycle effects on water relations of Pinus radiata.
Canadian Journal of Botany, 55: 2413-2418.
KOrner, Ch. (1989). The nutritional status of plants from high altitudes: a worldwide comparison.
Oecologia, 81: 379-391.
KOrner, Ch. & Renhardt, U. (1987). Dry matter partitioning and root length/leaf area ratios in herbaceous perennial plants with diverse altitudinal distribution. Oecologia, 74: 411-418.
Uppers, M. (1989). Ecological significance of above-ground architectural patterns in woody
plants: a question of cost-benefit relationships. Trends in Ecology and Evolution, 4: 375-379.
Laegaard, S. (1992). Influence of fire in the grass paramo vegetation of Ecuador. In, H. Balslev &
J.L. Luteyn (eds.), Paramo: An Andean Ecosystem under Human Influence. London: Academic
Press.
Lamotte, M., Garay, I. & Monasterio, M. (1989). Les grands traits du functionnement d'un ecosysteme tropical d'altitude. In: G. Montalenti et a/. (eds), Ecologia (Atti 3 con gresso della Societa
Italiana di Ecologia, Siena, 1987), 1: 61-66.
Larcher, E. (1975). Pflanzenokologische Beobachtungen in die Paramostufe der Venezolanische
Anden. Anzeiger der Osterreichischen Akademie der Wissenschaften Mathematisch-Naturwissenschaftliche Klasse, 11: 194-213.
Larcher, W. (1981). Physiological basis of evolutionary trends in low temperature resistance in vascular plants. Plant Systematics and Evolution, 137: 145-180.
Larcher, W. & Bauer, H. (1981). Ecological significance of resistance to low temperature. In, O.L.
Lange, P.S. Nobel, C.B. Osmond & H. Ziegler (eds.) Physiological Plant Ecology. I. Responses
to the Physical Environment. Berlin: Springer.
Lauer, W. (1976). Zur Hygrischen Hohenstufung Tropischer Gebirge. In, F. Schmithiisen (ed.), Neotropische Oekosysteme. Biogeographica 7. The Hague: Junk.
Lauer, W. (1979). La posici6n de los peramos en la estructura del paisaje de los Andes tropicales.
In, M.L. Salgado-Labouriau (ed.), El Medio Arnbiente Pdramo. Merida (Venezuela): Ediciones
Centro de Estudios Avanzados.
Leggett, J. (1990). Global Warming—The Greenpeace Report. Oxford: Oxford University Press.
Li, P.H. & Palta, J.P. (1978). Frost hardening and freezing stress In tuber bearing Solanum species.
In, PH. Li & A. Sakai (eds.), Plant Cold Hardiness and Freezing Stress. London: Academic
Press.
References
236
Lieth, H. (1975). Primary production of the major vegetation units of the world. In, H. Lieth & R.H.
Whittaker, Primary Productivity of the Biosphere. Berlin: Springer-Verlag.
Liley, M. (1986). Study of Polylepis incana. In, Report of the Cambridge Ecuadorian Andes Expedition 1985. Cambridge: Cambridge University.
Lind, E.M. & Morrison, M.E.S. (1974). East African Vegetation. London: Longman.
Lojtnant, B. & Molau, U. (1982). Analysis of a virgin paramo plant community on Volcan Sumaco,
Ecuador. Nordic Journal of Botany, 2: 567-574.
Lough, T.J., Wilson, J.B., Mark, A.F. & Evans, A.C. (1987). Succession in a New Zealand alpine
cushion community: a markovian model. Vegetatio, 71: 129-138.
Lozano, G. & Schnetter, R. (1976). Estudios ecolOgicos en el Paramo de Cruz Verde, Colombia.
II. Las communidades vegetales. Caldasia, 11: 54-68.
Luteyn, J.L. (1992). Paramos: why study them? In, H. Balslev & J.L. Luteyn (eds.), Paramo: An Andean Ecosystem under Human Influence. London: Academic Press.
Luteyn, J.L., Cleef, A.M. & Rangel, 0. (1992). Plant diversity in paramo: towards a checklist of
paramo plants and a generic flora. In, H. Balslev & J.L. Luteyn (eds.), Paramo: An Andean Ecosystem under Human Influence. London: Academic Press.
Mabberley, D.J. (1986). Adaptive syndromes of Afroalpine species of Dendrosenecio. In, F. Vuilleumier & M. Monasterio (eds.), High Altitude Tropical Biogeography. Oxford: Oxford University
Press.
MacArthur, R.H. (1965). Patterns of species diversity. Biological Reviews, 40: 510-533.
MacArthur, R.H. & Wilson, E.O. (1967). The Theory of Island Biogeography. Princeton (N.J.):
Princeton University Press.
Malloch, A.J.C. (1988). Vespan II: A Computer Package to Handle and Analyse Multivariate
Species Data and Handle and Display Species Distribution Data. Lancaster: University of Lancaster.
Mann, G. (1966). Bases ecolOgicas de la explotaciOn agropecuaria en la America Latina. Serie BiolOgica Monogreica No. 2. Washington: Departamento de Asuntos Cientificos de la Secretaria
General de la OrganizaciOn de Estados Americanos.
Mark, A.F. & Adams, N.M. (1973). New Zealand Alpine Plants. Wellington: A.H. & A.W. Reed.
Meinzer, F. & Goldstein, G. (1985). Some consequences of leaf pubescence in the Andean giant
rosette plant Espeletia timotensis. Ecology, 66: 512-520.
Meinzer, F. & Goldstein, G. (1986). Adaptations for water and thermal balance in Andean giant rosette plants. In, T.J. Givnish (ed.), On the Economy of Plant Form and Function. Cambridge:
Cambridge University Press.
Mena, R & Balslev, H. (1986). ComparaciOn Entre la VegetaciOn de los Paramos y el CinturOn
Afroalpino. Reports from the Botanical Institute, University of Aarhus, No. 12.
Miller, A. (1976). Proyecto de ForestaciOn de los Paramos en el Ecuador. Informe del Proyecto
Conjunto ODA-Servico Forestal, 1972-76. Quito: Ministerio de Agricultura y Ganaderia.
Miller, E.V. & Coleman, N.T. (1952). Colloidal properties of soils from western equatorial South
America. Soil Science Society of America, Proceedings, 16: 239-244.
References
237
Miller, G.A. (1987a). The functional significance of inflorescence pubescence in tropical alpine
species of Puya. In, Rundel, P., Meinzer, F. & Smith, A.R (eds.), Tropical Alpine Environments:
Plant Form & Function. Berlin: Springer-Verlag.
Miller, G.A. (1987b). The Population Biology and Physiological Ecology of Species of Puya
(Bromeliaceae) in the Ecuadorian Andes. Ph.D. Thesis, University of Connecticut.
Miller, G.A. & Silander Jr., J.A. (1991). Control of the distribution of giant rosette species of Puya
(Bromeliaceae) in the Ecuadorian paramos. Biotropica, 23: 124-133.
Mills, K. (1975). Flora de la Sierra, un estudio en el Parque Nacional Cotopaxi 1974/75. Ciencia y
Naturaleza (Quito), 16: 23-43.
Monasterio, M. (1980a). El Paramo de Mucubajf dentro del cuadro general de los paramos Venezolanos. In, M. Monasterio (ed.), Estudios EcolOgicos en los Paramos Andinos. Merida
(Venezuela): Ediciones de la Universidad de los Andes.
Monasterio, M. (1980b). Las formaciones vegetales de los paramos de Venezuela. In, M. Monasterio (ed.), Estudios EcolOgicos en los Péramos Andinos. Merida (Venezuela): Ediciones de la
Universidad de los Andes.
Monasterio, M. (1980c). Los Paramos Andinos como region natural: Caracteristicas biogeograficas generales y afinidades con otras regiones andinas. In, M. Monasterio (ed.), Estudios EcolOgicos en los Paramos Andinos. Merida (Venezuela): Ediciones de la Universidad de Los
Andes.
Monasterio, M. (1986). Adaptive strategies of Espeletia in the Andean desert paramo. In, F Vuilleumier & M. Monasterio (eds.), High Altitude Tropical Biogeography. Oxford: Oxford University
Press.
Monasterio, M. & Reyes, S. (1980). Diversidad ambiental y variaciOn de la vegetaciOn en los 'Aromas de los Andes Venezolanos. In, M. Monasterio (ed.), Estudios EcolOgicos en los Paramos
Andinos. Merida (Venezuela): Ediciones de la Universidad de los Andes.
Monasterio, M. & Vuilleumier, F. (1986). High tropical mountain biota of the world. In, F Vuilleumier & M. Monasterio, (eds.), High Altitude Tropical Biogeography. Oxford: Oxford University
Press.
Mooney, H.A. (1974). Plant forms in relation to environment. In, B.R. Strain & W.D. Billings (eds.),
Handbook of Vegetation Science. Part VI. Vegetation and Environment. The Hague: Junk.
Munoz, L., Balslev, H. & De Vries, Tj. (1985). Diversidad de la vegetaciOn en cuatro cuadrantes
en el paramo pajonal del Antisana, Ecuador. Publicaciones del Museo Ecuatoriano de Ciencias Naturales, Arlo 6, 4: 21-33.
Murra, J.V. (1972). El "Control Vertical" de un maximo de pisos ecolOgicos en la economia de las
sociedades Andinas. In, Ortiz, I., Vista de la Provincia de Leon Huenuco (1562), Vol. II. Huanuco (Peru): Universidad Hermillo Valdizan.
Nathaniel, R (1985). Cambridge Ecuador Geological Expedition 1984. Cambridge: published by
author.
Nishikawa, Y. (1990). Role of rhizomes in tussock formation by Carex thunbergii var. appendiculata. Ecological Research, 5: 261-269.
Noble, I.R. & Slatyer, R.O. (1981). Concepts and models of succession in vascular plant communities subject to recurrent fires. In, A.M. Gill, R.H. Groves & I.R. Noble (eds.), Fire and the
Australian Biota. Canberra: Australian Academy of Science.
References
238
Norrington-Davies, J. (1967). Application of diallel analysis to experiments in plant competition.
Euphytica, 16: 391-406.
01Igaard, B. & Balslev, H. (1979). Report on the 3rd Danish Botanical Expedition to Ecuador. Reports from the Botanical Institute, University of Aarhus. No. 4.
Paredes, A.C. (1962). Esquema fisiografico de la vegetaciOn en el Ecuador. Ciencias y Naturaleza
(Quito), 5: 21-30.
Peet, R.K. (1978). Forest vegetation of the Colorado Front Range: patterns of species diversity.
Vegetatio, 37: 65-78.
Pels, B. & Verweij, P.A. (1992). Burning and grazing in a bunchgrass !Aram° ecosyste m: vegetation dynamics described by a transition model. In, H. Balslev & J.L. Luteyn (eds.), P6ramo: An
Andean Ecosystem under Human Influence. London: Academic Press.
Penland, C.W.T. (1941). The alpine vegetation of the southern Rockies and the Ecuadorian Andes.
Colorado College Publications, Studies Series, 32: 4-30.
Perez, F.L. (1987). Soil moisture and the upper altitudinal limit of growth of giant rosettes. Journal
of Biogeography, 14: 173-186.
Pfitsch, W.A. (1988). Microenvironment and the distribution of two species of Draba (Brassicaceae) in a Venezuelan paramo. Arctic and Alpine Research, 20: 333-341.
Polunin, 0. & Stainton, A. (1987). Concise Flowers of the Himalayas. Oxford: Oxford University
Press
Ponce, A. (1984). Ecuadorian Strategy for the Conservation of Wildlands and Wildlife. In, J.A.
McNeely & K.R. Miller (eds.), National Parks, Conservation and Development: the Role of Protected Areas in Sustaining Society. Washington (DC): Smithsonian Institution Press.
Portsch, S. & Hicks, J.L. (1980). A soil conservation program for Ecuador. Journal of Soil & Water
Conservation, 35: 243-244.
Quintanilla, V. (1983). Comparaci6n entre dos ecosistemas tropoandinas: la puna chilena y el
paramo ecuatoriano. Informaciones Geogrâficas , Universidad de Chile, 30: 25-45.
Rada, F., Goldstein, G., AzOcar, A. & Meinzer, F. (1985). Freezing avoidance in Andean giant rosette plants. Plant Cell and Environment, 8: 501-507.
Ramsay, P.M. (unpublished). A Field and Experimental Study of an Environmental Moisture Gradient. B.Sc. Biology Dissertation, University of Wales, Bangor
Ramsay, P.M. (1988). University College of North Wales El 'CAJAS' Expedition, Ecuador, 1985Scientific Report. Bangor: PM. Ramsay.
Rangel, 0. & Aguirre, J. (1987). Estudios ecolOgicos en la Cordillera Oriental de Colombia. III. La
vegetaciOn de la cuenca lacustre de Tota (Boyaca). Caldasia, 15: 263-312.
Rangel, 0. & Franco, R (1985). Observaciones fitoecologicas en varias regiones de vida de la
Cordillera Central de Colombia. Caldasia, 14: 211-249.
Rangel, 0. & Lozano, G. (1986). Un perfil de vegetaciOn entre La Plata (Huila) y el Volcan del Purace. Caldasia, 14: 503-547.
Rauh, W. (1939). Ober polsterfOrmingen Wuchs. Nova Acta Leopoldina N.F. 7, 49: 267-508.
References
239
Raunkiaer, C. (1907). Planterigeta livsformer og deres betydning for geografien. Copenhagen:
Kjobenhavn & Kristiania.
Raunkiaer, C. (1908). Livsformernes statistik som grundlag for biologisk plantegeografi. Botanisk
Tidsskrift, 29: 42-83.
Raunkiaer, C. (1934). The Life Forms of Plants and Statistical Plant Geography. Oxford: Oxford
University Press.
Ricardi, M., Bricerio, B. & Adamo, G. (1987). Sinopsis de la flora vascular del paramo de Piedras
Blancas, Venezuela. Emstia, 44: 4-14.
Robberecht, R., Caldwell, M.M. & Billings, W.D. (1980). Leaf ultraviolet optical properties along
a latitudinal gradient in the arctic-alpine life zone. Ecology, 61: 612-619.
Rossenaar, A.J.G.A. & Hofstede, R.G.M. (1992). Effects of burning and grazing on root biomass
in the paramo ecosystem. In, H. Balslev & J.L. Luteyn (eds.), Paramo: An Andean Ecosystem
under Human Influence. London: Academic Press.
Ruthsatz, B. (1978). Las plantas en cojin de los semi-desiertos andinos del Noroeste Argentino.
Darwiniana, 21: 491-539.
Sakai, A. & Larcher, W. (1987). Frost survival of plants: Responses and Adaptations to to Freezing Stress. Berlin: Springer.
Sandwith, N.Y. (1926). Humboldt and Bonpland's 'finery in Ecuador and Peru. Kew Bulletin, 4:
181-190.
Sarmiento, F. (1987). Desde la Selva... Hasta el Mar: Antologia EcolOgica del Ecuador. Serie
Monografia An() 7 (NOmero 2). Quito: Ediciones Casa de la Cultura Ecuatoriana.
Sarmiento, G. (1986). Ecological features of climate in high tropical mountains. In, F. Vuilleumier
& M. Monasterio (eds.), High Altitude Tropical Biogeography. Oxford University Press.
Schimper, A.F.W. (1898). Pflanzengeographie auf physiologischer Grundlage. Jena.
Schmidt, A.M. & Verweij, RA. (1992). Forage intake and secondary production in extensive livestock systems in paramo. In, H. Balslev & J.L. Luteyn (eds.), Paramo: An Andean Ecosystem
under Human Influence. London: Academic Press.
Schulze, E.-D. (1982). Plant life forms and their carbon, water and nutrient relations. In, O.L.
Lange, PS. Nobel, C.B. Osmond & H. Ziegler (eds.). Physiological Plant Ecology II: Water Relations and Carbon Assimilation. Berlin: Springer-Ver1ag.
Schulze, E.-D., Beck, E., Scheibe, R. & Ziegler, R (1985). Carbon dioxide assimilation and stomatal response of afroalpine giant rosette plants. Oecologia, 65: 207-213.
Schwabe, G.H. (1968). Towards an ecological characterization of the South American continent.
In, E.J. Fittkau, J. lilies, H. Klinge, G.H. Schwabe & H. Sioli (eds.), Biogeography and Ecology
in South America, Vol. 1. The Hague: Junk.
Seibert, R (1983). Human impact on landscape and vegetation in the Central high Andes. In, W.
Holzner, M.J.A. Werger & I. lkusima (eds.), Man's Impact on Vegetation. The Hague: Junk.
Sewell, T.G. (1954). The high-altitude snow-tussock grassland in South Island, New Zealand. New
Zealand Journal of Science and Technology (Section A), 36: 335-364.
Shugart, H.H. & Heft, J.M. (1973). Succession: similarities of species turnover rates. Science,
180:1379-1381.
References
240
Sick, W. (1969). Geographical substance. In, E.J. Fittkau, J. lilies, H. Kling, G.H. Schwabe & H.
Sioli (eds.), Biogeography and Ecology in South America, Vol. 2. The Hague: Junk.
Si!ander Jr., J.A. & Pacala, S.W. (1985). Neighbourhood predictors of plant performance. Oecologia, 66: 256-263.
Siltanen, M., Thurland, M. & Casanova, J. (1987). EvaluaciOn de biomasa en fund& at suelo en
cinco rodales de Polylepis incana. Proyecto Arbolandino (Puno, Peru).
Silvertown, J.W. (1981). Microspatial heterogeneity and seedling demography in species rich
grassland. New Phytolo gist, 88: 117-128.
Silvertown, J.W. (1987). Introduction to Plant Population Ecology. Harlow: Longman Scientific &
Technical.
Simpson, B.B. (1974). Glacial migrations of plants: island biogeographical evidence. Science,
185: 698-700.
Simpson, B.B. (1975). Pleistocene changes in the flora of the high tropical Andes. Paleobiology,
1: 273-294.
Simpson, B.B. (1979). A revision of the genus Polylepis (Rosaceae: Sanguisorbeae). Smithsonian
Contributions to Botany, No. 43.
Sims, P.L. & Singh, J.S. (1978). The structure and function of ten western North American grasslands. III. Net primary production, turnover and efficiencies of energy capture and water use.
Journal of Ecology, 66: 573-597.
Smith, A.P. (1972). Notes on wind-related growth patterns of paramo plants in Venezuela. Biotropica, 4:10-16.
Smith, A.R (1974). Bud temperature in relation to nyctinastic leaf movement in an Andean giant rosette plant. Biotropica, 6: 263-266.
Smith, A.P. (1978). Establishment of seedlings of Polylepis sericea in the paramo (alpine) zone of
the Venezuelan Andes. Bartonia, 45:11-14.
Smith, A.P. (1979). Function of dead leaves in Espeletia schultzii (Compositae), an Andean caulescent rosette species. Biotropica, 13: 39-48.
Smith, A.P. (1980). The paradox of plant height in an Andean giant rosette species. Journal of Ecology, 68: 63-73.
Smith, A.R (1981). Growth and population dynamics of Espeletia (Compositae) in the Venezuelan
Andes. Smithsonian Contributions to Botany, No. 48.
Smith, A.R (1987). Does the correlation of elevation with plant taxonomic richness vary with latitude? Biotropica, 20: 259-261.
Smith, A.R & Young, T.R (1987a). Population biology of Senecio keniodendron, an afroalpine
giant rosette plant. [Cited by Smith & Young (1987b)] In, R Rundel, F. Meinzer & A.R Smith
(eds.), Tropical Alpine Environments: Plant Form and Function. Berlin: Springer Verlag.
Smith, A.R & Young, T.R (1987b). Tropical alpine plant ecology. Annual Review of Ecology and
Systematics, 18: 137-158.
Smith, J.M.B. (1975). Mountain grasslands of New Guinea. Journal of Biogeography, 2: 27-44.
References
241
Smith, J.M.B. (1977). An ecological comparison of two tropical high mountains. Journal of Tropical Geography, 44: 71-80.
Smith, J.M.B. (1978). Relationship of slope aspect to deforestation in highland Papua New Guinea. Queensland Geographical Journal, 4: 69-75.
Smith, J.M.B. & Cleef, A.M. (1988). Composition and origins of the world's tropicalpine floras.
Journal of Biogeography, 15: 631-645.
Smith, J.M.B. & Klinger, L.F. (1985). Aboveground: belowground phytomass ratios in Venezuelan
paramo vegetation and their significance. Arctic and Alpine Research, 17: 189-198.
Snaydon, R.W. (1991). Replacement or additive designs for competition studies? Journal of Applied Ecology, 28: 930-946.
Snaydon, R.W. & Satorre, E.H. (1989). Bivariate diagrams for plant competition data: modifications and interpretation. Journal of Applied Ecology, 26:1043-1057.
Solbrig, O.T. (1960). The South American sections of Erigeron and their relation to Celmisia. Contributions to the Gray Herbarium, 188: 85.
Stadel, C. (1989). The perception of stress by campesinos: a profile from the Ecuadorian Sierra.
Mountain Research and Development, 9: 35-49.
Sturm, H. (1978). Zur 6kologie der andinen Paramoregion. Biogeographica, 14: 1-121.
Sturm, H. & Abouchaar, A. (1981). Observaciones sobre la ecologia del paramo andino de Monserrate. Caldasia, 13: 223 -256.
Sturm, H. & Rangel, 0. (1985). Ecologia de los Nramos Andinos: una VisiOn Preliminar Integrada. Bogota: Universidad Nacional de Colombia.
Svenson, H.K. (1945). The vegetation of Ecuador, a brief review. In, F. Verdoorn (ed.), Plants and
Plant Sciences in Latin America. New York.
Tamm, C.O. (1975). Plant nutrients as limiting factors in ecosystem dynamics. In, Productivity of
World Ecosystems: Proceedings of a Symposium. Washington (D.C.): National Academy of
Sciences.
Tansley, A.G. & Chipp, T.F. (1926). Aims and Methods in the Study of Vegetation. London.
Ter Braak, C.J.F. (1988). CA NOCO-a FORTRA N program for canonical community ordination by
[partial] [detrended] [canonical] correspondence analysis, principal components analysis and
redundancy analysis (Version 2.1). Report LWA-88-02. Wageningen: Agricultural Mathematics
Group.
Th6rhallsdattir, T.E. (1983). The Dynamics of a Grassland Community with Special Reference to
Five Grasses and White Clover Ph.D. Thesis, University of Wales.
Th6rhallsd6ttir, T.E. (1990). The dynamics of a grassland community: a simultaneous investigation of spatial and temporal heterogeneity at various scales. Journal of Ecology, 78: 884-908.
Tieszen, L.L. & Detling, J.K. (1983). Productivity of grassland and tundra. In, 01. Lange, PS.
Nobel, C.B. Osmond & H. Ziegler (eds.), Physiological Plant Ecology IV Ecosystem Processes:
Mineral Cycling, Productivity and Man's Influence. Berlin: Springer-Verlag.
Tol, G.J. & Cleef, A.M. (1992). Nutrient status of a Chusquea tessellata bamboo paramo. In, H.
Balslev & J.L. Luteyn (eds.), Paramo: An Andean Ecosystem under Human Influence. London:
Academic Press.
References
242
Tomlinson, P.B. (1987). Architecture of tropical plants. Annual Review of Ecology and Systematics, 18: 1-21.
Troll, C. (1959). Die tropische Gebirge. Bonner Geographische Abhandlungen, 25:1-93.
Troll, C. (1968). The Cordilleras of the tropical Americas: aspects of climatic, phytogeographical
and agrarian ecology. In, C. Troll (ed.), Geo-ecology of the Mountainous Regions of the Tropical Americas. Bonn: Ferd. aimmlers Verlag.
Usher, M.B. (1979). Markovian approaches to ecological succession. Journal of Animal Ecology,
48: 413-426.
Usher, M.B. (1981). Modelling ecological succession, with particular reference to Markovian models. Vegetatio, 46:11-18.
Van Royen, R (1967). Some observations on the alpine vegetation of Mount Biota (Papua). Acta
Botanica Neerlandica, 15: 530-534.
Van Steenis, C.G.G.J. (1935). On the origin of the Malaysian mountain flora, 2. Altitudinal zones,
general considerations and renewed statement of the problem. Bulletin Jardin Botanique
Buitenzorg, 13: 129-257.
Van Steenis, C.G.G.J. (1939). Ecological observations on the genus Pleiocraterium in Gajoland,
Sumatra. Recueil des Travaux Botaniques Neerlandais, 36: 446-448.
Van der Hammen, T. & Cleef, A.M. (1986). Development of the high Andean paramo flora and
vegetation. In, F. Vuilleumier & M. Monasterio (eds.), High Altitude Tropical Biogeography. Oxford: Oxford University Press.
Vareschi, V. (1970). Flora de los Pâramos de Venezuela. Merida (Venezuela): Universidad de Los
Andes.
Velasquez, A. (1992). Grazing and burning in grassland communities of high volcanoes in Mexico. In, H. Balslev & J.L. Luteyn (eds.), Pâramo: An Andean Ecosystem under Human Influence.
London: Academic Press.
Vera, R. & Lopez, R. (1986). El origen de la cangahua. Paisajes Geograficos (Quito), 16: 21-28.
Verweij, RA. & Beukema, H. (1992). Aspects of human influence on upper-Andean forest line
vegetation. In, H. Balslev & J.L. Luteyn (eds.), PAramo: An Andean Ecosystem under Human Influence. London: Academic Press.
Verweij, RA. & Budde, RE. (1992). Burning and grazing gradients in pàramo vegetation: initial ordination analyses. In, H. Balslev & J.L. Luteyn (eds.), Péramo: An Andean Ecosystem under
Human Influence. London: Academic Press.
Verweij, RA. & Kok, K. (1992). Effects of fire and grazing on Espeletia hartwegiana populations.
In, H. Balslev & J.L. Luteyn (eds.), PAramo: An Andean Ecosystem under Human Influence. London: Academic Press.
Von Humboldt. F.H.A. (1806). Ideen zu einer Physiognomik der GewAsche. Stuttgart: Cotta.
Vuilleumier, F. (1970). Insular biogeography in continental regions. I. The northern Andes of South
America. American Naturalist, 104: 373-388.
Walter, H. (1973). Vegetation of the Earth in Relation to Climate and the Eco-physiological Conditions. London: English Universities Press.
References
243
Walter, H. & Medina, F. (1969). La temperatura del suelo como factor determinante para la caracterizaciOn de los pisos subalpino y alpino en los Andes de Venezuela. Boletin Venezolana de
Ciencias Naturales, 115/116: 201-210.
Wardlaw, I.F. (1969). The effect of water stress on translocation in relation to photosynthesis and
growth. II. Effect during leaf development in Lolium temulentum L. Australian Journal of Biological Sciences, 22:1-16.
Warming, E. (1884). Om Skudbygning, Overvintring og Foryngelse. Copenhagen: Naturhistorisk
Forenings Festskrift.
Warming, E. (1909). Oecology of Plants: An Introduction to the Study of Plant Communities. Oxford.
Watson, M.A. & Cook, C. (1982). The development of spatial pattern in clones of an aquatic
plant, Eichhomia crassipes Solms. American Journal of Botany, 69: 248-253.
Watt, A.S. (1947). Pattern and process in the plant community. Journal of Ecology, 35: 1-22.
Weberbauer, A. (1945). El Mundo Vegetal de los Andes Peruanos. Lima: Ministerio de Agricultura.
Weiner, J. (1982). A neighbourhood model of annual plant interference. Ecology, 63:1237-1241.
Weischet, W. (1969). Klimatologische Regeln zur Vertical-verteilung der Niederschlage in den Tropengebirgen. Die Erde, 100: 287-306.
West, T.L. (1987). The burning bush: exploitation of native shrubs for fuel in Bolivia. In, D.L. Browman, Arid Land Use Strategies and Risk Management in the Andes: a Regional Anthropological Perspective. Boulder: Westview Press.
Whittaker, R.H. (1975). Communities and Ecosystems, 2nd Edn. New York: Macmillan.
Whymper, E. (1892). Travels Amongst the Great Andes of the Equator. London: Murray.
Wilcox, B.R, Bryant, F.C., Wester, D. & Allen, B.L. (1986). Grassland communities and soils of a
high elevation grassland of Central Peru. Phytologia, 61: 231-250.
Williams, E.J. (1962). The analysis of competition experiments. Australian Journal of Biological
Science, 15: 509-525.
Williams, L. (1941). The phytogeography of Peru. Chronica Botanica, 6: 406-410.
Wilson, M.V. & Smida, A. (1984). Measuring beta diversity with presence-absence data. Journal
of Ecology, 72:1055-1064.
de Wit, C.T. (1960). On competition. Verslagen van Landouwkundige Onderzoekingen, 66:1-82.
Young, T.R & Van Orden Robe, S. (1986). Microenvironmental role of a secreted aqueous solution in the Afro-alpine plant Lobelia keniensis. Biotropica, 18: 189-194.
Zar, H. (1984). Biostatistical Analysis. London: Prentice-Hall.
Photographic Plates
Plates
245
Key to Plates on pages 249-254.
Plate 1. The extensive paramo plateau of Cajas, with Laguna Luspa. Note the large,
grass-covered island in the lake. 24 August, 1985.
Plate 2. The Paramo de Guamani, another extensive paramo region. 8 October, 1987.
Plate 3. The paramo of Volcdn Cotacachi, an example of an isolated paramo region
surrounding a volcanic peak. 10 October 1987.
Plate 4. The Collanes valley beneath the crater of El Altar, part of the larger paramo
system on the Altar massif. Note the burning at the far end of the valley.
The El Altar (west) sampling transect was situated on the slopes to the left.
3 September, 1987.
Plate 5. Cushion bog of Distichia muscoides, 4,200 m, Pdramo de Guamani. 9 August,
1987.
Plate 6. Blechnum loxense Shrub Paramo (BS), 3,900 m, Volcdn Tungurahua. 29 August, 1987.
Plate 7. Humid Desert Pdramo (HD), 4,200 m, Volcdn Tungurahua. 28 August, 1987.
Plate 8. Calamagrostis sp. [251] and Chuquiraga jussieui Desert Paramo with Cerastium sp. [200] (CCCD), 4,600 m, Volcdn Chimborazo. 25 October, 1987.
Plate 9. W erneria humilis & Plantago rigida Cushion Paramo (WPC), 4,200 m, Volcdn
Cotacachi. 11 October, 1987.
Plate 10. Transition between Calamagrostis sp. [251] Tussock Grassland with Hypochaeris sonchoides, Halenia sp. [189] and Satureja nubigena (HHCT) and
W emeria humilis and Plantago rigida Cushion Paramo (WPC), 4160 m, Volcan Cotacachi. 11 October, 1987.
Plate 11. Calamagrostis sp. [251] Tussock Grassland with V iola humboldtii (VCT),
3,700 m, Paramo de Daldal. 25 September, 1987.
Plate 12. Neurolepis elata Bamboo Paramo (NB), 3,800 m, eastern slopes of El Altar.
10 August, 1989.
Plate 13. Calamagrostis. sp. [251] and Espeletia pycnophylla Tussock Grassland with
V iola sp. [192] (PCE), 3,700 m, Volcdn Chiles. 22 October, 1987.
Plate 14. Calamagrostis sp. [251] Tussock Grassland with Paspalum tuberosum and
Cluysactinium acaule (PCCT), 3,100 m, Ofia. 16 September, 1987.
Plate 15. Calamagrostis sp. [251] and Chuquiraga jussieui Desert Paramo (CCD),
4,150 m, Volcan Chimborazo. 25 October, 1987.
Plates
246
Plate 16. Stem Rosette growth form: Espeletia pycnophylla ssp. angelensis, 3,700 m,
Volcan Chiles. 22 October, 1987.
Plate 17. Basal Rosette growth form: Puya sp., 3,500 m, Cajas. 11 September, 1987.
Plate 18. Tussock growth form: Cortaderia sericantha, 4,100 m, Paramo de Daldal. 19
August, 1987.
Plate 19. Acaulescent Rosette growth form: Senecio repens, 4,050 m, Paramo de Guamani. 8 August, 1987.
Plate 20. Cushion growth form: A zorella cogmbosa, 4,300 m, Paramo de Daldal. 18
August, 1987.
Plate 21. Upright Shrub growth form: Pentacalia andicola, 4,050 m, Volcan Chiles. 21
October, 1987.
Plate 22. Prostrate Shrub growth form: Disterigma empetrifolium, 4,100 m, Volcan
Chiles. 21 October, 1987.
Plate 23. Erect Herb growth form: Bartsia laticrenata, 4200 m, Volcan Cotacachi. 11
October, 1987.
Plate 24. Prostrate Herb growth form: Satureja nubigena, 3,900 m, Paramo de GuamaM. 8 August, 1987.
Plate 25. Trailing Herb growth form: Lobelia tenera, 3,600 m, Pdramo de Daldal. 16
July, 1987.
Plate 26. Paramo fire near Laguna Luspa, Cajas. 27 August, 1985.
Plate 27. Paramo fire near Laguna Luspa, Cajas. 27 August, 1985.
Plate 28. Result of overburning: the loss of soil. Pdramo near Laguna Yantahuayco,
Cajas. 18 August, 1985.
Plate 29. Fire in Espeletia Pdramo, 3,700 m, Volcan Chiles. 22 October, 1987.
Plate 30. Thermochrom crayons and pyrometers used in the experimental fire at
Cajas. 13 September, 1985.
Plate 31. Recently burned paramo at 3,750 m in the Daldal valley. This area was
mapped and the transition experiment transects located in the areas between the tussocks (Figure 4.4). 4 July 1987.
Plate 32. Burned Calamagrostis tussock with recently germinated seedlings of Rumex
acetosella, approximately 15 weeks after the fire. 3,750 m, Paramo de Daldal. 25 September, 1987.
Plates
247
Plate 33. Intertussock regeneration at 3,750 m, approximately 15 weeks after the fire.
Paramo de Daldal. 25 September, 1987.
Plate 34. Burned paramo at 3,950 m, 123 days after a fire. The transition experiment
transects at this altitude were located in this area. Paramo de Daldal. 30
October, 1987.
Plate 35. Transition zone between Plantago rigida cushions and Calamagrostis tussocks after a recent fire, 4,000 m, Paramo de Daldal. 30 October, 1987.
Plate 36. Plantago rigida cushion smothering a Calamagrostis tussock. Note the moss
growing on the top of the cushion, where the Plantago has already begun to
break down. Paramo de Daldal. 3 July, 1987.
Plate 37. W emeria humilis cushion growing over Calamagrostis sp. [251] tussock. 4200
m, Volcan Cotacachi. 11 October, 1987.
Plate 38. The upper Daldal valley, showing the location of the four productivity study
exclosures (B-E). 22 August, 1987.
Plate 39. Part of the Alao valley. Exclosure A of the productivity study was located on
the valley floor, near the Rio Mao at 3,100 m. 19 August, 1987.
Plate 40. Exclosure B of the productivity study, 3,250 m, Daldal valley. 5 July, 1987.
Plate 41. Exclosure C of the productivity study, 3,450 m, Daldal valley. 22 August,
1987.
Plate 42. Exclosure D of the productivity study, 3,750 m, Paramo de Daldal. 5 July,
1987.
Plate 43. Exclosure E of the productivity study, 3,950 m, Paramo de Daldal. 20 August, 1987.
Plate 44. Tussock and Cushion Paramo, 4,150 m, Paramo de Guamani. The collection
area for Calamagrostis sp. tussock material for the greenhouse competition
experiments. Small tussocks of Calamagrostis A can be seen amongst the
mats of Plantago rigida. Shrubby vegetation of Loricaria ilinissae and Diplostephium rupestre can be seen in the background. 12 November, 1987.
Plate 45. Tussock Paramo, 4,000 m, Paramo de Guamani. The collection area for Calamagrostis sp. tussock material for the greenhouse competition experiments. Large tussocks of Calamagrostis B dominate the landscape. Other
prominent plants are Puya clava-herculis and Senecio chionageton. 12 November, 1987.
Plate 46. Cattle grazing on the upper part of the Cajas sampling transect on Soldados,
Cajas. Note the typical location of the Polylepis woodland at 3,800 m: beneath a cliff on a rocky substrate. 12 September, 1987.
Plates
248
Plate 47. Local farmer collecting Chuquiraga jussieui for fuelwood (the load shown
lasts approximately five days), 4,300 m, Volcan Chimborazo.
Plate 48. Severe soil erosion, 4,150 m, Volcan Cotacachi. 11 October, 1987.
249
Plates
2.
1.
3
5.
6.
Plates
9.
10.
11
12
13
14
15.
16.
250
251
Plates
17.
18.
20
19.
21.
23.
22.
24.
252
Plates
25.
26.
28.
27.
30.
32.
Plates
33.
34.
37.
38.
39.
40.
253
254
Plates
41.
42.
43
44.
IMI•nrn
att ,
-
-r
•
A I
-
..P.4.4047,
ANN/
oat'
45.
46.
47.
48.
-
:,AY4*
Appendix 1.
Vascular Plant Species
Appendix 1. Vascular Plant Species
256
The species referred to in the text have been cross-referenced with voucher material deposited in herbaria in the UK and in Ecuador (as described in the text). The
following list allows the species names and codes used in the text to be linked to the
pressed specimens in the herbaria. The voucher collection numbers relate to those of
the Paramos '87 collection by P.M. Ramsay & P.J. Merrow-Smith, 1987, unless otherwise stated. A small number of species were only found in sterile form and were not
collected. Some of the code numbers were not used.
Code Scientific Name, Family, Voucher Collection Numbers and Growth Form
1
Bomarea caldasii (H.B.K.) Willd. (Alstroemeriaceae). Voucher Collection N 2 68, 827. Growth form:
Trailing Herb.
2
Bomarea glaucescens Baker (Alstroemeriaceae). Voucher Collection N2 231, 348, 648, 778. Growth
form: Erect Herb.
3
Blechman loxense (H.B.K.) Hieron. (Blechnaceae). Voucher Collection N 2 74, 272. Growth form: Stem
Rosette.
4
Lobelia tenera H.B.K. (Campanulaceae). Voucher Collection N2 22, 838. Growth form: Trailing Herb.
5
Cerastium danguyi Macbr. (Caryophyllaceae). Voucher Collection N 2 801. Growth form: Trailing Herb.
6
Cerastium floccosum Benth. (Caryophyllaceae). Voucher Collection N2 253, 811. Growth form: Prostrate Herb.
7
Cerastium mollissimum Pair. (Caryophyllaceae). Voucher Collection N 2 395, 732, 1060. Growth form:
Prostrate Herb.
8
Stellaria leptopetala Benth. (Caryophyllaceae). Voucher Collection N2 294, 366, 532, 733. Growth
form: Trailing Herb.
9
Baccharis alatemoides H.B.K. (Compositae). Voucher Collection N 2 669. Growth form: Upright Shrub.
10
Baccharis alpina H.B.K. (Compositae). Voucher Collection N 2 138, 256, 363, 499, 977. Growth form:
Prostrate Shrub.
11
Baccharis genistelloides H.B.K. (Compositae). Voucher Collection N 2 633, 993. Growth form: Erect
Herb.
12
Baccharis genistelloides H.B.K. (Compositae). Voucher Collection N2 556. Growth form: Erect Herb.
13
Baccharis genistelloides H.B.K. (Compositae). Voucher Collection N2 84,288. Growth form: Erect
Herb.
14
Baccharis humifusa H.B.K. (Compositae). Voucher Collection N2 215. Growth form: Prostrate Shrub.
15
Bidens andicola H.B.K. (Compositae). Voucher Collection N 2 46, 560, 612. Growth form: Prostrate
Herb.
16
Chrysactinium acaule (H.B.K.) Wedd. (Compositae). Voucher Collection N 2 452, 565, 609. Growth
form: Acaulescent Rosette.
17
Chuquiraga jussieui Grnel. (Compositae). Voucher Collection N2 442. Growth form: Upright Shrub.
18
Conyza uliginosa (Benth.) Cuatr. (Compositae). Voucher Collection N 9 467, 923. Growth form: Upright Shrub.
Appendix 1. Vascular Plant Species
257
19
Cukithun adscendens Benth. (Compositae). Voucher Collection N2 382, 1000. Growth form: Erect
Herb.
20
Cu!cilium nivale H.B.K. (Compositae). Voucher Collection N 2 252. Growth form: Acaulescent Rosette.
21
Culcitium ovatum (Sch.) Blake (Compositae). Voucher Collection N2 148, 276, 969. Growth form:
Erect Herb.
22
Cukitium rufescens H. & B. (Compositae). Voucher Collection N2 754, 934. Growth form: Basal Rosette.
23
Diplostephium hartwegii Hieron. (Compositae). Voucher Collection N2 490. Growth form: Upright
Shrub.
24
Diplostephium oblanceolatum Blake (Compositae). Voucher Collection N 2 624. Growth form: Upright Shrub.
25
Diplostephium rupestre (H.B.K.) Wedd. (Compositae). Voucher Collection N 2 152, 757, 966. Growth
form: Upright Shrub.
26
Erigeron pinnatus Turcz. (Compositae). Voucher Collection N2 988. Growth form: Acaulescent Rosette.
27
Espeletia pycnophylla Cuatr. ssp. angelensis Cuatr. (Compositae). Voucher Collection N 2 849. Growth
form: Stem Rosette.
28
Gnaphallurn ? luteo-album L. (Compositae). Voucher Collection N 2 640, 836, 1064. Growth form:
Erect Herb.
29
Gnaphaliurn pensylvanicum Willd. (Compositae). Voucher Collection N 2 112, 268, 365, 918, 1043.
Growth form: Erect Herb.
30
Gnaphalium antennarioides DC. (Compositae). Voucher Collection 1\1 2 947. Growth form: Erect
Herb.
31
Gnaphalium coarctaturn Willd. (Compositae). Voucher Collection N 2 538, 823. Growth form: Erect
Herb.
32
Gnaphalium dysodes Spreng. (Compositae). Voucher Collection N 2 116. Growth form: Erect Herb.
33
Gnaphalium gnaphaloides (Kunth.) Beauv. (Compositae). Voucher Collection N 2 737. Growth form:
Erect Herb.
34
Gynoxys baccharoides (H.B.K.) Cass. (Compositae). Voucher Collection N2 312. Growth form: Upright Shrub.
35
Gynarys buxifolia (H.B.K.) Cass. (Compositae). Voucher Collection N 2 674. Growth form: Upright
Shrub.
36
Gynoxys cuicochensis Cuatr. (Compositae). Voucher Collection N2 647. Growth form: Upright Shrub.
37
Gynoxys filliginosa (H.B.K.) Cass. (Compositae). Voucher Collection N 2 1124. Growth form: Upright
Shrub.
38
Gynazys ndniphylla Cuatr. (Compositae). Voucher Collection N2 504. Growth form: Upright Shrub.
39
Hieracium frigidum Wedd. (Compositae). Voucher Collection N 2 80, 305, 352, 643, 673, 685, 799.
Growth form: Erect Herb.
Appendix 1. Vascular Plant Species
258
40
Hypochaeris sessiliflora H.B.K. (Compositae). Voucher Collection N2 306, 955, 960, 985. Growth
form: Acaulescent Rosette.
41
Hypochaeris sonchoides H.B.K. (Compositae). Voucher Collection N2 126, 714, 764, 995. Growth
form: Acaulescent Rosette.
42
Loricaria complanata (Sch.Bip.) Wedd. (Compositae). Voucher Collection N 2 489. Growth form: Upright Shrub.
43
Loricaria thuyoides (Lam.) Sch.Bip. (Compositae). Voucher Collection N2 304, 340. Growth form:
Upright Shrub.
44
Lucilia lehmanii Hieron. (Compositae). Voucher Collection N2 200. Growth form: Acaulescent Rosette.
45
Lucilia radians (Benth.) Cuatr. (Compositae). Voucher Collection N2 989. Growth form: Cushion.
4.6
Oritrophium hieracioides (VVedd.) Cuatr. (Compositae). Voucher Collection 1•12 334. Growth form:
Acaulescent Rosette.
47
Oritrophium limnophilum (Sch.Bip.) Cuatr. (Compositae). Voucher Collection N 2 233, 333, 994, 1030.
Growth form: Acaulescent Rosette.
48
Oritrophium peruvianum (Lam.) Cuatr. (Compositae). Voucher Collection N 2 163, 260, 347, 765, 846.
Growth form: Acaulescent Rosette.
49
Oritrophium peruvianum (Lam.) f. intermediurn Cuatr. (Compositae). Voucher Collection N 2 458,
564, 620. Growth form: Acaulescent Rosette.
50
Perezia pungens (H.B.K.) Less. (Compositae). Voucher Collection N2 850. Growth form: Erect Herb.
51
Pentacalia aff. andicola Turcz. (Compositae). Voucher Collection N2 779, 891, 1067. Growth form:
Upright Shrub.
52
Pentacalia arbutifolius H.B.K. (Compositae). Voucher Collection NQ 220, 285. Growth form: Upright
Shrub.
53
Senecio chionageton Wedd. (Compositae). Voucher Collection N2 150, 436, 948. Growth form: Erect
Herb.
54
Senecio lingulatus (Schlechtd.) Cuatr. (Compositae). Voucher Collection N 2 493. Growth form: Erect
Herb.
55
Senecio pimpinellifolia H.B.K. (Compositae). Voucher Collection N 2 65, 286, 394. Growth form: Acaulescent Rosette.
56
Senecio repens DC. (Compositae). Voucher Collection N 2 183, 329. Growth form: Acaulescent Rosette.
57
Pentacalia stuebellii Hieron. (Compositae). Voucher Collection N 2 883, 921. Growth form: Upright
Shrub.
58
Senecio teretifolius (H.B.K.) DC. (Compositae). Voucher Collection N 2 1005. Growth form: Upright
Shrub.
59
Sonchus ? oleraceus L. (Compositae). Voucher Collection N2 44. Growth form: Erect Herb.
60
Stevia sp. (Compositae). Voucher Collection N 2 634. Growth form: Erect Herb.
Appendix 1. Vascular Plant Species
259
61
Taraxacum officinale Weber (Compositae). Voucher Collection N2 70. Growth form: Acaulescent Rosette.
62
V emonia sp. (Compositae). Voucher Collection N 2 902. Growth form: Upright Shrub.
63
W emeria ? crassa Blake (Compositae). Voucher Collection N 2 999. Growth form: Acaulescent Rosette.
64
W emeria humilis H.B.K. (Compositae). Voucher Collection N 2 186, 362, 726, 759, 866. Growth form:
Cushion.
65
W emeria nubigena H.B.K. (Compositae). Voucher Collection N 2 177, 780. Growth form: Acaulescent
Rosette.
66
W emeria pumila H.B.K. (Compositae). Voucher Collection N 2 869. Growth form: Cushion.
67
Diplostephium glutinosum Blake (Compositae). Voucher Collection N 2 219, 322, 583. Growth form:
Upright Shrub.
68
Lepidium sp. (Cruciferae). Voucher Collection N2 736, 838. Growth form: Erect Herb.
69
Cares crinalis Boon (Cyperaceae). Voucher Collection N2 359, 476, 525, 607, 644, 1026, 1058. Growth
form: Acaulescent Rosette.
70
Carer lemanniana Boott (Cyperaceae). Voucher Collection N 2 747. Growth form: Tussock.
71
Cares pichinchensis H.B.K. (Cyperaceae). Voucher Collection N 2 125, 419, 491, 876. Growth form:
Thssock.
72
Cares 07:y 11cl:a Boott (Cyperaceae). Voucher Collection N 2 161, 360, 385, 533, 815, 959, 1025. Growth
form: Acaulescent Rosette.
73
Oreobolus goeppingeri K. Svessenguth (Cyperaceae). Voucher Collection N 2 327, 632, 688, 702, 926.
Growth form: Cushion.
74
Oreobolus obtusangulus Gaud. (Cyperaceae). Voucher Collection N 2 225. Growth form: Cushion.
75
Rhynchospora macrochaeta Steud. (Cyperaceae). Voucher Collection N 2 508, 699, 859. Growth form:
Tussock.
76
Rhychospora cf. ruiziana Boeck. (Cyperaceae). Voucher Collection N2 550, 617, 906, 1056. Growth
form: Basal Rosette.
77
Uncinia cf. hamata (SW .) Urb. (Cyperaceae). Voucher Collection NQ 378. Growth form: Acaulescent
Rosette.
78
Uncinia phleoides Pers. (Cyperaceae). Voucher Collection N2 739. Growth form: Thssock.
79
Disterigma empetnfollum (H.B.K.) Drude (Ericaceae). Voucher Collection N 2 85, 287, 342, 417, 694,
774, 958. Growth form: Prostrate Shrub.
80
Gentiana sedifolia H.B.K. (Gentianaceae). Voucher Collection 1\1 2 120, 280, 316, 569, 629, 867.
Growth form: Prostrate Herb.
81
Gendanella ? corallina Gilg. (Gentianaceae). Voucher Collection N 2 622. Growth form: Erect Herb.
82
Gentianella ? foliosa (H.B.K.) Fabris (Gentianaceae). Voucher Collection 1\1 2 862. Growth form:
Erect Herb.
Appendix 1. Vascular Plant Species
260
83
Gentianella cernua (H.B.K.) Fabris (Gentianaceae). Voucher Collection N2 991. Growth form: Erect
Herb.
84
Gentianella gracilis (H.B.K.) Fabris (Gentianaceae). Voucher Collection N2 554. Growth form: Erect
Herb.
85
Gentianella hirculus (Griseb.) Fabris (Gentianaceae). Voucher Collection N 2 449. Growth form:
Erect Herb.
86
Gentianella hyssoptfolia (H.B.K.) Fabris (Gentianaceae). Voucher Collection N 2 464, 637, 671.
Growth form: Erect Herb.
87
Gentianella nummalanfolia Griseb. (Gentianaceae). Voucher Collection N 2 713, 766. Growth form:
Erect Herb.
88
Halenia weddelliana Gilg. (Gentianaceae). Voucher Collection N 2 335, 892. Growth form: Erect
Herb.
89
Geranium multipanitum Benth. (Geraniaceae). Voucher Collection N 2 123, 319. Growth form: Prostrate Herb.
90
Geranium reptans Kunth. (Geraniaceae). Voucher Collection N 2 18, 299, 364, 730. Growth form: Prostrate Herb.
91
Geranium sibbaldioides Benth. (Geraniaceae). Voucher Collection N2 122, 318, 625, 687, 775, 925.
Growth form: Prostrate Herb.
92
A ciachne flagelhfera Laegaard (Gramineae). Voucher Collection N2 719, 762, 863. Growth form:
Cushion.
93
,4nthoxanthum odoratum L. (Gramineae). Voucher Collection N2 249, 540. Growth form: Erect Herb.
94
Bromus lanatus Kunth (Gramineae). Voucher Collection N 2 291, 320, 523, 524, 591, 720. Growth
form: Erect Herb.
95
Cortaderia nitida (Kunth) Pilger (Gramineae). Voucher Collection N2 105. Growth form: Tussock.
96
Holcus lanatus L. (Gramineae). Voucher Collection N2 248. Growth form: Erect Herb.
97
Paspalurn tuberosurn Mez. (Gramineae). Voucher Collection N 2 103, 325. Growth form: Prostrate
Herb.
98
Onhrosanthus chimboracensis (H.B.K.) Baker (Iridaceae). Voucher Collection N2 83, 593, 1069.
Growth form: Erect Herb.
99
Sisyrinchium jamesoni Baker (Iridaceae). Voucher Collection N 2 167, 267, 323, 802, 932, 1023.
Growth form: Thssock.
100
Sisyrinchium tinctorium H.B.K. (Iridaceae). Voucher Collection N 2 547, 559, 619, 695. Growth form:
Tussock.
101
Luzula gigantea Desv. (Juncaceae). Voucher Collection N 2 301, 964. Growth form: Tussock.
102
Luzula racemosa Desv. (Juncaceae). Voucher Collection N 2 251, 341, 721, 793. Growth form: Acaulescent Rosette.
103
Satureja nubigena (Kunth.) Brig. (Labiatae). Voucher Collection N 2 121, 300, 795, 885. Growth form:
Prostrate Herb.
Appendix 1. Vascular Plant Species
261
104
Stachys elliptica Kunth. (Labiatae). Voucher Collection N 2 66, 380. Growth form: Erect Herb.
105
Lupinus ? purdieanus C.P (Leguminosae). Voucher Collection N2 273. Growth form: Prostrate
Shrub.
106
Lupinus ? sarmentosus Desr. (Leguminosae). Voucher Collection N 2 328, 492, 641, 788, 924, 1041.
Growth form: Prostrate Herb.
107
Lupinus microphyllus Desr. (Leguminosae). Voucher Collection N2 798. Growth form: Prostrate
Shrub.
108
Lupinus ramosissimus Benth. (Leguminosae). Voucher Collection N 2 106. Growth form: Upright
Shrub.
109
Lupinus smithianus Kunth. (Leguminosae). Voucher Collection N 2 987. Growth form: Prostrate
Herb.
110
Trifolium repens L. (Leguminosae). Voucher Collection N 2 89. Growth form: Prostrate Herb.
111
V icia ? andicola H.B.K. (Leguminosae). Voucher Collection N2 144, 738. Growth form: Trailing
Herb.
112
V ida ? setifolia H.B.K. (Leguminosae). Voucher Collection N 2 841. Growth form: Trailing Herb.
113
Pinguicula calyptrata H.B.K. (Lentibulariaceae). Voucher Collection N 2 628, 684, 913. Growth form:
Acaulescent Rosette.
114
Tofieldia sessilijlora Hook (Melanthiaceae). Voucher Collection N 2 621, 670. Growth form: Erect
Herb.
115
Brachyotum alpinum Cogn. (Melastomataceae). Voucher Collection N2 434. Growth form: Upright
Shrub.
116
Brachyotum cf. confertum (Bonpl.) Triana (Melastomataceae). Voucher Collection N 2 611. Growth
form: Upright Shrub.
117
Brachyotum leclifolium (Desr.) Triana (Melastomataceae). Voucher Collection N2 825. Growth form:
Upright Shrub.
118
Plan tago major L. (Plantaginaceae). Voucher Collection N 2 63. Growth form: Acaulescent Rosette.
119
Plantago rigida H.B.K. (Plantaginaceae). Voucher Collection N 9 213, 379, 526. Growth form:
Cushion.
120
Montana crassifolia H.B.K. (Polygalaceae). Voucher Collection 1•19 905. Growth form: Upright Shrub.
121
Rumex acetosella L. (Polygonaceae). Voucher Collection N 2 216, 935. Growth form: Prostrate Herb.
122
A nemone jamesonii Hook. (Ranunculaceae). Voucher Collection N2 487. Growth form: Prostrate
Herb.
123
Ranunculus peruvianus Pers. (Ranunculaceae). Voucher Collection 1\19 338, 450, 848, 1018. Growth
form: Prostrate Herb.
124
Lachemilla ? rupestris (H.B.K.) Rothm. (Rosaceae). Voucher Collection N 2 337, 686, 929, 1045.
Growth form: Prostrate Herb.
125
Lachemilla andina (Perry) Rothm. (Rosaceae). Voucher Collection N 2 227. Growth form: Prostrate
Herb.
Appendix 1. Vascular Plant Species
262
126
Lachemilla galioides Benth. (Rosaceae). Voucher Collection N 2 1046. Growth form: Acaulescent
Rosette.
127
Lachemilla hispidula (Perry) Rothm. (Rosaceae). Voucher Collection N 2 257, 361, 717. Growth
form: Acaulescent Rosette.
128
Lachemilla holosericea (Perry) Rothm. (Rosaceae). Voucher Collection N 2 718. Growth form: Acaulescent Rosette.
129
Lachemilla nivalis H.B.K. (Rosaceae). Voucher Collection N 2 506, 768, 877, 878. Growth form:
Acaulescent Rosette.
130
Lachemilla orbiculata R. & P (Rosaceae). Voucher Collection /42 303, 368. Growth form: Prostrate
Herb.
131
Lachemilla pinnata R. & R (Rosaceae). Voucher Collection /4 2 884. Growth form: Prostrate Herb.
132
Nenera granadensis (Lf.) Druce (Rubiaceae). Voucher Collection N2 262, 522, 963. Growth form:
Prostrate Herb.
133
Relbunium hypocarpium (L.) Hemsl. (Rubiaceae). Voucher Collection N 2 289, 293, 638, 741, 1063.
Growth form: Trailing Herb.
134
Bartsia laticrenata Benth. (Scrophulariaceae). Voucher Collection N2 330, 771, 855. Growth form:
Erect Herb.
135
Calceolaria fenuginea Cav. (Scrophulariaceae). Voucher Collection N 2 307. Growth form: Upright
Shrub.
136
Ourisia chamaedryfolia Benth. (Scrophulariaceae). Voucher Collection N2 881. Growth form: Acaulescent Rosette.
137
Pedicularis incurva Benth. (Scrophulariaceae). Voucher Collection N 2 513, 631. Growth form: Acaulescent Rosette.
138
V eronica setpyllifolia L. (Scrophulariaceae). Voucher Collection N2 534. Growth form: Erect Herb.
139
A zorella aretoides H.B.K. (Umbelliferae). Voucher Collection N2 210, 353, 783, 944. Growth form:
Acaulescent Rosette.
140
A zorella corynzbosa (R. & P) Pers. (Umbelliferae). Voucher Collection N2 376. Growth form:
Cushion.
141
A zorella crenata (R. & P) Pers. (Umbelliferae). Voucher Collection N 2 887, 1053. Growth form:
Acaulescent Rosette.
142
A zorella pedunculata (Spreng.)
form: Cushion.
143
Eryngium humile Cav. (Umbelliferae). Voucher Collection N 2 228, 266, 566, 613, 837, 937, 980.
Growth form: Acaulescent Rosette.
144
Hydocotyle bonplandii A. Rich (Umbelliferae). Voucher Collection N2 250. Growth form: Prostrate
Herb.
145
Niphogeton dissecta (Benth.) F. Macbr. (Umbelliferae). Voucher Collection N.° 119, 709, 787, 879,
890. Growth form: Acaulescent Rosette.
M. & C.
(Umbelliferae). Voucher Collection N 2 134, 271. Growth
Appendix 1. Vascular Plant Species
263
146
Oreomyrrhis andicola (Kunth.) Hook f. (Umbelliferae). Voucher Collection N° 141, 297, 357, 729,
819. Growth form: Acaulescent Rosette.
147
V aleriana aretioides H.B.K. (Valerianaceae). Voucher Collection No 756, 763. Growth form: Cushion.
148
V aleriana rigida R. & R (Valerianaceae). Voucher Collection N° 808, 1040. Growth form: Cushion.
149
V aleriana adscendens Turcz. (Valerianaceae). Voucher Collection N° 711. Growth form: Acaulescent
Rosette.
150
V aleriana bonplandiana Wedd. (Valerianaceae). Voucher Collection N° 218, 275, 389, 561, 616, 691,
745, 880. Growth form: Upright Shrub.
151
V aleriana bracteata Benth. (Valerianaceae). Voucher Collection No 332, 494. Growth form: Acaulescent Rosette.
152
V aleriana alyptfolia ssp. alypifolia (Valerianaceae). Voucher Collection N° 971. Growth form: Prostrate Shrub.
153
V aleriana microphylla H.B.K. (Valerianaceae). Voucher Collection N o 355, 446, 809, 998, 1029.
Growth form: Upright Shrub.
154
V aleriana plantaginea H.B.K. (Valerianaceae). Voucher Collection N o 952. Growth form: Basal Rosette.
155
V iola lzumboldtii Tr. & Fl. (Violaceae). Voucher Collection N° 29. Growth form: Prostrate Herb.
156
V iola 'avails Benth. (Violaceae). Voucher Collection No 414, 1037. Growth form: Acaulescent Rosette.
157
A stragalus geminiflorus H.B.K. (Leguminosae). Voucher Collection N° 982. Growth form: Cushion.
158
Geranium sp. (Geraniaceae). Voucher Collection N o 472. Growth form: Prostrate Herb.
159
Geranium sp. (Geraniaceae). Voucher Collection N o 1004. Growth form: Prostrate Herb.
160
Geranium sp. (Geraniaceae). Voucher Collection N° 716, 888. Growth form: Prostrate Herb.
161
Hvericum sp. (Guttiferae). Voucher Collection N o 314, 495, 562, 661, 676, 941, 1021. Growth form:
Erect Herb.
162
Unidentified species (Melastomataceae). Voucher Collection N° 462, 664. Growth form: Prostrate
Herb.
163
Unidentified species (Melastomataceae). Voucher Collection N° 639. Growth form: Prostrate Herb.
164
Oxalis sp. (Oxalidaceae). Voucher Collection N° 521. Growth form: Prostrate Herb.
165
Bartsia sp. (Scrophulariaceae). Voucher Collection N o 498, 615, 690, 936, 1022, 1051. Growth form:
Erect Herb.
166
Bartsia sp. (Scrophulariaceae). Voucher Collection N o 816. Growth form: Erect Herb.
167
Bartsia sp. (Scrophulariaceae). Voucher Collection N o 263. Growth form: Erect Herb.
168
Castilleja sp. (Scrophulariaceae). Voucher Collection N o 222, 331, 772, 981. Growth form: Erect
Herb.
169
Castilleja sp. (Scrophulariaceae). Voucher Collection N o 946. Growth form: Erect Herb.
Appendix 1. Vascular Plant Species
264
170
Castilleja sp. (Scrophulariaceae). Voucher Collection I sT2 265, 672. Growth form: Erect Herb.
171
Castilleja sp. (Scrophulariaceae). Voucher Collection N2 354, 460, 553, 817, 1019. Growth form:
Erect Herb.
172
V eronica sp. (Scrophulariaceae). Voucher Collection N2 535. Growth form: Erect Herb.
173
A phanactis jamesonia Wedd. (Compositae). Voucher Collection N 2 189, 317, 938. Growth form:
Acaulescent Rosette.
174
Cotula ? mexicana (DC.) Cabr. (Compositae). Voucher Collection N2 17, 370, 939. Growth form:
Acaulescent Rosette.
175
Unidentified species (Alliaceae). Voucher Collection N2 810. Growth form: Erect Herb.
176
Nototriche jamesonii A.W. Hill (Malvaceae). Voucher Collection N2 997. Growth form: Cushion.
177
As for Code N2 157.
178
Unidentified species (Family not known). Voucher Collection N2 984. Growth form: Cushion.
179
Puya clava-herculis Mez & Sodiro (Bromeliaceae). Voucher Collection N 2 516. Growth form: Basal
Rosette.
180
Puya cf. pygmaea L.B. Smith (Bromeliaceae). Voucher Collection N2 655. Growth form: Basal Rosette.
181
Puya sp. (Bromeliaceae). Voucher Collection N2 662. Growth form: Basal Rosette.
182
Unidentified species (Cruciferae). Voucher Collection N 2 735. Growth form: Erect Herb.
183
Not used.
184
Cardamine sp. (Cruciferae). Voucher Collection N2 367. Growth form: Erect Herb.
185
Pemettya sp. (Ericaceae). Voucher Collection N 2 814, 953. Growth form: Prostrate Shrub.
186
Gentianella sp. (Gentianaceae). Voucher Collection N 2 1017. Growth form: Erect Herb.
187
Halenia sp. (Gentianaceae). Voucher Collection N 2 570. Growth form: Erect Herb.
188
Halenia sp. (Gentianaceae). Voucher Collection /%12 642, 689. Growth form: Erect Herb.
189
Halenia sp. (Gentianaceae). Voucher Collection N 2 712, 760. Growth form: Erect Herb.
190
Not used.
191
Unidentified species (Umbelliferae). Voucher Collection N2 731. Growth form: Erect Herb.
192
V iola sp. (Violaceae). Voucher Collection N 2 927. Growth form: Acaulescent Rosette.
193
Ribes sp. (Grossulariaceae). Voucher Collection N 2 854. Growth form: Upright Shrub.
194
V aleriana sp. (Valerianaceae). Voucher Collection N2 990. Growth form: Cushion.
195
A zorella sp. (Umbelliferae). Voucher Collection N 2 868. Growth form: Cushion.
196
Sibthorpia repens (Mutis) Kuntze (Scrophulariaceae). Voucher Collection N 2 511, 961. Growth form:
Prostrate Herb.
Appendix 1. Vascular Plant Species
265
197
Cerastium sp. (Caryophyllaceae). Voucher Collection N 2 454, 536. Growth form: Trailing Herb.
198
Cerastium sp. (Caryophyllaceae). Voucher Collection N 2 728. Growth form: Prostrate Herb.
199
Cerastium sp. (Caryophyllaceae). Voucher Collection N 2 743, 871. Growth form: Prostrate Herb.
200
Cerastium sp. (Caryophyllaceae). Voucher Collection N 2 979. Growth form: Prostrate Herb.
201
V icia sp. (Leguminosae). Voucher Collection N 2 381. Growth form: Trailing Herb.
202
Lupinus sp. (Leguminosae). Voucher Collection N 2 146. Growth form: Prostrate Herb.
203
Lupinus sp. (Leguminosae). Voucher Collection N2 824. Growth form: Upright Shrub.
204
Unidentified species (Compositae). Voucher Collection N2 746. Growth form: Prostrate Herb.
205
Unidentified species (Umbelliferae). Voucher Collection N2 527. Growth form: Prostrate Herb.
206
Cotula sp. (Compositae). Voucher Collection N 2 1001. Growth form: Acaulescent Rosette.
207
Sisyrinchium sp. aff. alatum Hook. (Iridaceae). Voucher Collection N 2 481. Growth form: Tussock.
208
Monnina sp. (Polygalaceae). Voucher Collection N2 830. Growth form: Upright Shrub.
209
Lachernilla sp. (Rosaceae). Voucher Collection N2 336. Growth form: Prostrate Herb.
210
Lachemilla sp. (Rosaceae). Voucher Collection N2 653. Growth form: Prostrate Herb.
211
V iola sp. (Violaceae). Voucher Collection N 2 557. Growth form: Prostrate Herb.
212
Neurolepis elata (Kunth) Pilger (Gramineae). Voucher N 2 436 from Ramsay, Evans & Buckland
1989 Collection. Growth form: Tussock.
213
Ranunculus guzmanii H.B.K. (Ranunculaceae). Voucher Collection N2 856. Growth form: Prostrate
Herb.
214
Corer sp. (Cyperaceae). Voucher Collection N2 542. Growth form: Acaulescent Rosette.
215
Distichia muscoides Nees & Meyen (Juncaceae). Voucher Collection N2 179. Growth form: Cushion.
216
Equisetum bogotense H.B.K. (Equisetaceae). Voucher Collection N 2 813, 1052. Growth form: Erect
Herb.
217
Ophioglossum crotalophoroides Walt. (Ophioglossaceae). Voucher Collection N2 133, 255, 943, 1156.
Growth form: Acaulescent Rosette.
218
Plantago linearis H.B.K. (Plantaginaceae). Voucher Collection N 2 58, 558, 812, 1028, 1059. Growth
form: Acaulescent Rosette.
219
Baccharis caespitosa (R. & P) Pers. (Compositae). Voucher Collection N 2 1224. Growth form: Prostrate Shrub.
220
Poa cucullata Hack. (Gramineae). Voucher Collection N 2 1225. Growth form: Erect Herb.
221
Not used.
222
Huperzia hypogoea B. 011g. (Lycopodiaceae). Voucher Collection NQ 1231. Growth form: Erect
Herb.
Appendix 1. Vascular Plant Species
266
223
Unidentified species (Family not known). Voucher Collection N 2 520. Growth form: Acaulescent
Rosette.
224
Unidentified species (Family not known). Voucher Collection N 2 623. Growth form: Upright Shrub.
225
Eudema nubigena H.B.K. (Cruciferae). Voucher Collection N 2 727, 1185. Growth form: Cushion.
226
Unidentified species (Family not known). Voucher Collection N 2 734. Growth form: Erect Herb.
227
Unidentified species (Family not known). Voucher Collection N 0 1061. Growth form: Erect Herb.
228
Eriosorus sp. (Filicopsida). Voucher Collection N2 264, 872. Growth form: Basal Rosette.
229
Thelypteris sp. (Thelypteridaceae). Voucher Collection N2 292, 1066. Growth form: Basal Rosette.
230
A splenium sp. (Aspleniaceae). Voucher Collection N 2 270. Growth form: Acaulescent Rosette.
231
Lysipomia montioides H.B.K. (Campanulaceae). Voucher Collection N2 212. Growth form: Prostrate Herb.
232
Unidentified species (Compositae). Voucher Collection N2 873. Growth form: Prostrate Herb.
233
Diplostephium sp. (Compositae). Voucher Collection N2 933. Growth form: Upright Shrub.
234
Draba sp. (Cruciferae). Voucher Collection N2 715. Growth form: Erect Herb.
235
Draba sp. (Cruciferae). Voucher Collection N 2 1062. Growth form: Erect Herb.
236
A grostis sp. (Gramineae). Voucher Collection N 2 548, 940. Growth form: Erect Herb.
237
.A grostis sp. (Gramineac). Voucher Collection N 2 537. Growth form: Erect Herb.
238
A grostis sp. (Gramineae). Voucher Collection N 2 411. Growth form: Erect Herb.
239
A grostis sp. (Gramineae). Voucher Collection N 2 864. Growth form: Erect Herb.
240
A grostis sp. (Gramineae). Voucher Collection NQ 530, 551, 660, 679, 680, 698, 931, 1047. Growth
form: Erect Herb.
241
A grostis sp. (Gramineae). Voucher Collection N 2 910. Growth form: Erect Herb.
242
A grostis sp. (Gramineae). Voucher Collection NQ 377, 416, 501, 531, 659, 822, 992, 1008. Growth
form: Erect Herb.
243
A grostis sp. (Gramineae). Voucher Collection N 2 261, 284, 326, 514, 700. Growth form: Erect Herb.
244
A grostis nigritella Pilg. (Gramineae). Voucher Collection N 2 254, 344, 421, 722, 786, 865, 986. Growth
form: Erect Herb.
245
Not used.
246
Calamagrostis sp. (Gramineae). Voucher Collection N 2 707, 761. Growth form: Erect Herb.
247
Calatnagrostis sp. (Gramineae). Voucher Collection N 2 1031. Growth form: Erect Herb.
248
Not used.
249
Calamagrostis sp. (Gramineae). Voucher Collection N 2 549, 692, 696. Growth form: Erect Herb.
Appendix 1. Vascular Plant Species
267
250
Calamagrostis sp. (Gramineae). Voucher Collection Ng 384. Growth form: Erect Herb.
251
Calamagrostis sp. (Gramineae). Voucher Collection Ng 321, 509, 580, 657, 697, 724, 725, 777, 893,
894, 920, 954, 956, 972, 1007, 1036, 1057. Growth form: Tiissock.
252
Not used.
253
Elymus attenuatum (H.B.K.) R. & S. (Gramineae). Voucher Collection N g 740. Growth form: Erect
Herb.
254
Festuca sp. (Gramineae). Voucher Collection N g 343, 442, 1035. Growth form: Tussock.
255
Festuca sp. (Gramineae). Voucher Collection N Q 502, 896. Growth form: TUssock.
256
Festuca sp. (Gramineae). Voucher Collection N g 510, 742. Growth form: Tussock.
257
Muhlenbergia angustata (Pres!) Kunth (Gramineae). Voucher Collection Ng 834. Growth form: Thssock.
258
A fuhlenbergia ligularis (Hack.) Hitchc. (Gramineae). Voucher Collection N g 369, 541. Growth form:
Erect Herb.
259
Paspalum sp. Mez. (Gramineae). Voucher Collection N g 568. Growth form: Prostrate Herb.
260
Paspalum sp. Mez. (Gramineae). Voucher Collection N g 908. Growth form: Prostrate Herb.
261
Poa sp. (Gramineae). Voucher Collection N g 345, 356, 723, 1033. Growth form: Erect Herb.
262
Poa sp. (Gramineae). Voucher Collection N g 346, 500, 505, 744, 821, 875, 1032. Growth form: Erect
Herb.
263
Poa sp. (Gramineae). Voucher Collection N g 383. Growth form: Erect Herb.
264
Poa sp. (Gramineae). Voucher Collection N g 390, 889. Growth form: Erect Herb.
265
Foci sp. (Gramineae). Voucher Collection N g 425. Growth form: Erect Herb.
266
Foci sp. (Gramineae). Voucher Collection N 2 543. Growth form: Erect Herb.
267
Poa sp. (Gramineae). Voucher Collection N Q 544. Growth form: Erect Herb.
268
Poa sp. (Gramineae). Voucher Collection N g 807. Growth form: Erect Herb.
269
Poa sp. (Gramineae). Voucher Collection N g 820. Growth form: Erect Herb.
270
Stipa sp. (Gramineae). Voucher Collection Ng 586, 658. Growth form: Tussock.
271
Trisetum spicatum (L.) Richt. (Gramineae). Voucher Collection N g 1034. Growth form: Erect Herb.
272
Hypericum sp. (Guttiferae). Voucher Collection N g 840. Growth form: Prostrate Herb.
273
Hypericum sp. (Guttiferae). Voucher Collection N g 313, 486, 748, 915, 1050. Growth form: Upright
Shrub.
274
Not used.
275
Hypericum sp. (Guttiferae). Voucher Collection N 2 474, 475, 587, 957. Growth form: Prostrate Herb.
276
Hypericum sp. (Guttiferae). Voucher Collection Ng 826. Growth form: Upright Shrub.
Appendix 1. Vascular Plant Species
268
277
Jamesonia sp. (Hemionitidaceae). Voucher Collection N 2 1176. Growth form: Erect Herb.
278
Jamesonia alstonii A.F. Tryon (Hemionitidaceae). Voucher Collection N 2 388, 497, 627, 1049.
Growth form: Erect Herb.
279
Jamesonia pukhra Hook. & Grey. (Hemionitidaceae). Voucher Collection N2 907, 930. Growth
form: Erect Herb.
280
Jamesonia robusta Karst. (Hemionitidaceae). Voucher Collection N 2 675. Growth form: Erect Herb.
281
Elaphaglossum sp. (Lomariopsidaceae). Voucher Collection N 2 805, 1048. Growth form: Erect
Herb.
282
Elaphaglossum sp. (Lomariopsidaceae). Voucher Collection N 2 269. Growth form: Erect Herb.
283
Elaphaglossum sp. (Lomariopsidaceae). Voucher Collection N 2 1065. Growth form: Erect Herb.
284
Huperzia compacta (Hook.) B. 011g. (Lycopodiaceae). Voucher Collection N 2 515. Growth form:
Erect Herb.
285
Lycopodium sp. (Lycopodiaceae). Voucher Collection N 2 803. Growth form: Erect Herb.
286
Lycopodium sp. (Lycopodiaceae). Voucher Collection N 2 942. Growth form: Erect Herb.
287
Lycopodium sp. (Lycopodiaceae). Voucher Collection N 2 1038. Growth form: Erect Herb.
288
Lycopodium sp. (Lycopodiaceae). Voucher Collection N 2 258. Growth form: Erect Herb.
289
Lycopodium sp. (Lycopodiaceae). Voucher Collection N 2 373, 758, 860. Growth form: Erect Herb.
290
Huperzia columnaris B. 011g. (Lycopodiaceae). Voucher Collection N 2 484. Growth form: Erect
Herb.
291
Lycopodium sp. (Lycopodiaceae). Voucher Collection N 2 635. Growth form: Prostrate Herb.
292
Lycopodium clavatum L. ssp. contiguum(n) B. 011g. (Lycopodiaceae). Voucher Collection N 2 545,
555, 903, 636. Growth form: Erect Herb.
293
Lycopodium sp. (Lycopodiaceae). Voucher Collection N 2 463, 804. Growth form: Prostrate Herb.
294
As for Code N2 292.
295
Lycopodium sp. (Lycopodiaceae). Voucher Collection N 2 278. Growth form: Prostrate Herb.
296
Lycopodiwn magallanicum (P Beauv.) Sw. (Lycopodiaceae). Voucher Collection N 2 315, 517, 1044.
Growth form: Prostrate Herb.
297
Lycopodium sp. (Lycopodiaceae). Voucher Collection N 2 350, 909, 1054. Growth form: Prostrate
Herb.
298
Epilobium denticulatum R. & P. (Onagraceae). Voucher Collection N2 296. Growth form: Erect
Herb.
299
A ltensteinia ? fimbriata H.B.K. (Orchidaceae). Voucher Collection N 2 277, 351, 962, 1068. Growth
form: Acaulescent Rosette.
300
Myrosmodes sp. (Orchidaceac). Voucher Collection N 2 708, 1024. Growth form: Erect Herb.
301
Plantago sp. (Plantaginaceae). Voucher Collection NQ 1010. Growth form: Acaulescent Rosette.
Appendix 1. Vascular Plant Species
269
302
Plantago sp. (Plantaginaceae). Voucher Collection N2 978. Growth form: Cushion.
303
Ranunculus sp. (Ranunculaceae). Voucher Collection N2 339, 832, 1020. Growth form: Prostrate
Herb.
304
Ranunculus sp. (Ranunculaceae). Voucher Collection N 2 283, 710. Growth form: Prostrate Herb.
305
A rcytophyllum sp. (Rubiaceae). Voucher Collection N 2 916. Growth form: Cushion.
306
A rcytophyllum aristatum Standley (Rubiaceae). Voucher Collection N2 519. Growth form: Prostrate
Shrub.
307
A rcytophyllum filiforme (R. & P.) Standley (Rubiaceae). Voucher Collection N2 496. Growth form:
Cushion.
308
,4rgtophyllum vemicosum Standley (Rubiaceae). Voucher Collection N2 630, 665. Growth form:
Upright Shrub.
309
Galium sp. (Rubiaceae). Voucher Collection N 2 1039. Growth form: Trailing Herb.
310
Relbunium croceum (R. & P.) Sch. (Rubiaceae). Voucher Collection N2 806, 983. Growth form: Trailing Herb.
311
Relbunium hirsutum (R. & P.) Schum. (Rubiaceae). Voucher Collection N 2 466, 1055. Growth form:
Trailing Herb.
312
V aleriana sp. (Valerianaceae). Voucher Collection N2 874. Growth form: Cushion.
313
V aleriana ? niplzobia Brici. (Valerianaceae). Voucher Collection N 2 870. Growth form: Acaulescent
Rosette.
314
X yris subulata R. & P. var. subuiata(Xyridaceae). Voucher Collection N 2 546, 618, 693. Growth form:
Tussock.
315
Bromus pitensis H.B.K. (Gramineae). Voucher Collection N2 797, 895. Growth form: Erect Herb.
316
Gentianella sp. (Gentianaceae). Voucher N2 425 from Ramsay, Evans & Buckland 1989 Collection.
Growth form: Erect Herb.
317
As Code N2 128.
318
Oreobolus sp. (Cyperaceae). Voucher Collection N2 518, 626. Growth form: Cushion.
319
Carex sp. (Cyperaceae). Voucher Collection N 2 529. Growth form: Acaulescent Rosette.
320
Carer sp. (Cyperaceae). Voucher Collection N 2 358. Growth form: Acaulescent Rosette.
321
Carex sp. (Cyperaceae). Voucher Collection N2 375. Growth form: Acaulescent Rosette.
322
Carex sp. (Cyperaceae). Voucher Collection N 2 418. Growth form: Acaulescent Rosette.
323
Isoetes sp. (Isoetaceae). Voucher not collected. Growth form: Acaulescent Rosette.
324
Not used.
325
Not used.
326
V iola sp. (Violaceae). Voucher not collected. Growth form: Prostrate Herb.
Appendix 1. Vascular Plant Species
270
327
As for Code N2 225.
328
Uncinia sp. (Cyperaceae). Voucher Collection N2 349. Growth form: Acaulescent Rosette.
329
Uncinia sp. (Cyperaceae). Voucher Collection N2 386, 800, 1027. Growth form: Acaulescent Rosette.
330
Unidentified species (Family not known). Voucher Collection N 2 681. Growth form: Erect Herb.
331
Culcitium sp. (Compositae). Voucher Collection N2 901. Growth form: Basal Rosette.
332
Culcitium sp. (Compositae). Voucher Collection N2 485. Growth form: Basal Rosette.
333
Erigeron sp. (Compositae). Voucher Collection N 2 259. Growth form: Acaulescent Rosette.
334
Loricaria sp. (Compositae). Voucher Collection N 2 844. Growth form: Upright Shrub.
335
Unidentified species (Ericaceae). Voucher Collection N 2 281. Growth form: Prostrate Shrub.
336
Unidentified species (Ericaceae). Voucher Collection N 2 552. Growth form: Prostrate Shrub.
337
Unidentified species (Ericaceae). Voucher Collection N 2 666. Growth form: Prostrate Shrub.
338
Unidentified species (Ericaceae). Voucher Collection N 2 667. Growth form: Prostrate Shrub.
339
Unidentified species (Ericaceae). Voucher Collection N 2 678. Growth form: Prostrate Shrub.
340
Unidentified species (Ericaceae). Voucher Collection N 2 828. Growth form: Upright Shrub.
341
Unidentified species (Gramineae). Voucher Collection N 2 528. Growth form: Erect Herb.
342
Jamesonia sp. (Hemionitidaceae). Voucher Collection N2 767, 861. Growth form: Erect Herb.
343
Unidentified species (Juncaceae). Voucher Collection N 2 590. Growth form: Tussock.
344
Luzula sp. (Juncaceae). Voucher Collection N2 1002. Growth form: Acaulescent Rosette.
345
Unidentified species (Orchidaceae). Voucher Collection N 2 1190. Growth form: Erect Herb.
346
Plantago sp. (Plantaginaceae). Voucher Collection N 2 274. Growth form: Cushion.
347
Paspalum sp. (Gramineae). Voucher Collection N2 324. Growth form: Prostrate Herb.
348
Unidentified species (Gramineae). Voucher Collection N 2 996. Growth form: Erect Herb.
349
Unidentified species (Gramineae). Voucher Collection N 2 503. Growth form: Erect Herb.
350
A grostis sp. (Gramineae). Voucher Collection NQ 539. Growth form: Erect Herb.
351
A grostis sp. (Gramineae). Voucher Collection N2 567. Growth form: Erect Herb.
352
Festuca sp. (Gramineac). Voucher Collection N2 1006. Growth form: Erect Herb.
353
Stipa sp. (Gramineae). Voucher Collection 1\1 2 1009. Growth form: Tussock.
354
ReIbunium sp. (Rubiaceae). Voucher not collected. Growth form: Trailing Herb.
355
V accinium sp. (Ericaceae). Voucher not collected. Growth form: Prostrate Shrub.
356
Puya sp. (Bromcliaceae). Voucher not collected. Growth form: Basal Rosette.
Appendix 1. Vascular Plant Species
271
357
V eronica sp. (Scrophulariaceae). Voucher not collected. Growth form: Erect Herb.
358
V iola sp. (Violaceae). Voucher not collected. Growth form: Prostrate Herb.
359
Oxalis sp. (Oxalidaceae). Voucher not collected. Growth form: Prostrate Herb.
360
Unidentified species (Family not known). Voucher not collected. Growth form: Erect Herb.
361
Puya sp. (Bromeliaceae). Voucher not collected. Growth form: Basal Rosette.
362
Unidentified species (Orchidaceae). Voucher not collected. Growth form: Acaulescent Rosette.
363
Unidentified species (Family not known). Voucher not collected. Growth form: Erect Herb.
Appendix 2.
Example Chi Square
Calculation for
Transition Probabilities
(Chapter 4, p.139)
272
Appendix 2. Example Chi Square Calculation
273
The x2 test will determine the probability that the pattern of replacement observed
is completely random. If for each species pair, the species present at time 1 is called
the ith species and the species present at time 2 thef h species, then the null hypothesis states that "species i will be replaced by species j in that proportion which the
total replacements made by species] contribute to the overall number of changes" or:
Eij = E (nir— nu) x (nrj — njj)
E (nrj— njj)
where 'r' represents all species other than i or j, 'nit' the total number of times
species i is followed by all other species, 'me the total number of quadrats occupied
by species i at time 1 and time 2, `nrf the total number of times species] follows all
other species, and 'nit the total number of quadrats occupied by species] at time 1
and time 2. Put another way, the expected value is:
=
Total number of quadrats x Total number of quadrats
h
vacated by ill' species
invaded by/ species
Grand Total of All Changes
provided the diagonal terms (the species replacing themselves) in the matrix are subtracted before making the calculation.
Most of the species involved in the data were rare and to avoid bias in the x2 ,
values those species with an expected value less than 5 were not subjected to a x`
test. The rarer species were treated as a group to overcome this problem. Yates' correction for continuity was applied.
Using the transition matrix for unburned vegetation at 3,750 m in the Pâramo de
Daldal given in Table 4.4 (p.149), Paspalum sp. replaced A zorella pedunculata three
times. The total number of quadrats vacated by A zorella pedunculata (excluding the
diagonal) was 86.75 and the total number of quadrats invaded by Paspalum sp. (excluding the diagonal) was 34. The grand total of all changes (again excluding the diagonals) was 368.75.
274
Appendix 2. Example Chi Square Calculation
Therefore, the expected number of replacements of A zorella pedunculata by
Paspalum sp. is given by:
E=
86.75 x34
368.75
E=
8.00
The Chi Square test requires the following information:
A zorella pedunculata replaced by
Other species
Paspalum sp.
All Species
Observed
3
83.75
86.75
Expected
8
78.75
86.75
The x2 calculation (with Yates' correction applied) is:
X
2
=
E (Observed - Expected - 0.5)2
Expected
So, for the example given:
X
2
=
2
(3 - 8 - 0.5) ± (83.75 - 78.75 - 0.5)2
8
X
x
2 =
2
=
78.75
3.781 + 0.257
4.038
From x2 tables, the probability that this result was due to random replacement lies
between 0.05 and 0.01. Therefore, the null hypothesis is rejected: A zorella pedunculata was replaced by Paspalum sp. less frequently than chance would predict.