Ecologia mediterranea 1999-25(2)

ecologia 

mediterranea 

Revue Internationale d'Ecologie Méditerranéenne 

International Journal ofMediterranean Ecology 

TOME 25 - fascicule 2 - 1999 

188N : 0153-8756

REDACTEURSIEDITORS 

Laurence AFFRE 

Frédéric MEDAIL 

Philip ROCHE 

Thierry TATONI 

Eric VIDAL 

ARONSON 1., CEFE CNRS, Montpellier 

BARBERO M., IMEP, Univ. Aix-Marseille III 

FONDATEUR 1FOUNDER 

Prof. Pierre QUEZEL 

COORDINATRICE 1COORDINATOR 

COMITE DE LECTURE 1ADVISORY BOARD 

BROCK M., Univ. of New England, Armidale, Australie 

CHEYLAN M., EPHE, Montpellier 

DEBUSSCHE M., CEFE CNRS, Montpellier 

FADY B., INRA, Avignon 

GOODFRIEND G. A., Carnegie Inst. Washington, USA 

GRILLAS P., Station Biologique Tour du Valat, Arles 

GUIOT J., IMEP, CNRS, Marseille 

HOBBS R. J., CSIRO, Midland, Australie 

KREITER S., ENSA-M INRA, Montpellier 

Sophie GACHET-BOUDEMAGHE 

TRESORIER 1TREASURER 

Jacques-Louis de BEAULIEU 

LE FLOC'H E., CEFE CNRS, Montpellier 

MARGARIS N. S., Univ. of the Aegan, Mytilène, Grèce 

OVALLE c., CSI Quilamapu, INIA, Chili 

PEDROTTI F., Univ. degli Studi, Camerino, Italie 

PLEGUEZUELOS J. M., Univ. de Grenade, Espagne 

PONEL P., IMEP, CNRS, Marseille 

PRODON R., Labo. Arago, Univ. P. M. Curie, Paris VI 

RICHARDSON D. M., Univ.Cape Town, Afrique du Sud 

SANS F. X., Univ. de Barcelone, Espagne 

SHMIDA A., Hebrew Univ. of Jerusalem, Israël 

URBINATI c., Agripolis, Legnaro, Italie 

ecologia mediterranea 

Faculté des Sciences et Techniques de Saint JérÔme 

Institut Méditerranéen d'Ecologie et de Paléoécologie, case 461 

13397 MARSEILLE Cedex 20 FRANCE 

Tél. : + 33 4 91 28 89 76 - Fax: + 33 4 91 28 8051 

E-mail: ecologia.mediterranea@botmed.u-3mrs.fr 

URL: http://www.ecologia.fst.u-3mrs.fr 

Abonnements / Subscription 

Un an =deux numéros / one year = Iwo issues: 

France: 400 F + 60 F de frais de port 

Europe: 400 F + 80 F de frais de port 

Amérique, Afrique, Asie: 400 F + 120 F de frais de port 

Tous les fascicules de la revue sont disponibles. Pour les commander et pour tout renseignement, nous contacter à 

l'adresse ci-dessus. Ali issues are avai/ab/e. To order or 10 obtain intiJrmation, p/ease contact the address above.

ecologia mediterranea 25 (2),135-146 (1999) 

Caracterizacion de los pinares de Pinus ha/epensis Mill. en el sur 

de la Peninsula Iberica 

Caracterisation des pinedes de Pinus halepensis Mill. du sud de la Peninsule Iberique 

Characterisation of Pinus halepensis Mill. pine forests in southern Iberian Peninsula 

Juan Antonio TORRES*, Antonio GARCIA-FUENTES*, Carlos SALAZAR*, Eusebio CANO* 

& Francisco VALLE** 

* Dpto, BioI. Animal, Vegetal y Ecologia, Fac. C. Experimentales. Universidad de Jaen. E-2307l Jaen-Espafia. Fax: + (34) (953) 

212141; jatorres@ujaen.es 

** Dpto Biologia Vegetal (Botanica). Fac. de Ciencias. Universidad de Granada, E-1800l Granada, Espafia 

RESUMEN 

Estudiamos las comunidades naturales de Pinus halepensis en el sur de Espafia (Andalucia). Se recopila toda la informaci6n 

bibliografica disponible que aporta datos de interes sobre su espontaneidad en toda la cuenca mediterranea, con especial 

referencia a la Peninsula Iberica. Ponemos de manifiesto el caracter edafoxer6filo de este tipo de formaciones en el area de 

estudio con la descripci6n de una nueva asociaci6n y una comunidad vegetal : Junipero phoeniceae-Pinetum halepensis y 

comunidad de Ephedrafragilis y Pinus halepensis. 

Palabras c1ave : Pinus halepensis, pinares aut6ctonos, caracter edafoxer6filo, sur de Espafia 

RESUME 

Cette etude porte sur les communautes naturelles de Pinus halepensis du sud de l'Espagne (Andalusia). Toute I'information 

bibliographique concernant la spontaneite dans le bassin Mediterraneen, particulierement celle de la Peninsule Iberique, est 

compilee. Le caractere edapho-xerophile de ce type de formations dans I'aire d'etude est mis en evidence, et une nouvelle 

association est decrite : Junipero phoeniceaePinetum halepensis, ainsi qu'une communaute 11 Ephedra fragilis et Pinus 

halepensis. 

Mots-c1es : Pinus halepensis, pinMes autochtones, caractere edapho-xerophile, sud de l'Espagne 

ABSTRACT 

A study on the natural communities of Pinus halepensis in southern Spain (Andalusia) is carried out. A compilation of the 

bibliographic information that contributes interesting data on their spontaneity in the Mediterranean basin is previously made, 

pointing out those references where the Iberian Peninsula is treated. 

We emphasize the edaphoxerophilous character of this vegetation type in the study area, and we provide the description of a new 

phytosociological association (Junipero phoeniceae-Pinetum halepensis) and a new plant community (community of Ephedra 

fragilis and Pinus halepensis) 

Key-words: Pinus halepensis, native pine forests, edaphoxerophilous character, southern Spain 

135

Tones et a!. 

ABRIDGED ENGLISH VERSION 

A study on the natural communities of Pinus halepensis 

in the south of Spain (Andalusia) has been carried out. The 

bibliographic information with data of interest about its 

spontaneity in the whole Mediterranean basin (specially in 

the Iberian peninsula) has been compiled. The edaphoxerophilous 

character of this pine-woods in the study area has 

been pointed out, having described a new phytosociological 

association and a plant community: Junipero phaenieeae 

Pinetum halepensis and community of Ephedra jragilis and 

Pinus halepensis. 

The spontaneous formations of Pinus halepensis in the 

Iberian peninsula develop on basic substrata, specially in the 

mediterranean coastal provinces (Catalufia, Comunidad Valenciana, 

Murcia and Baleares), reaching interior zones of 

the Baetic ranges, Iberian System, Ebro valley and eastern 

Pirenees. They mainly appear under Mediterraneanpluviestational-oceanic, 

Mediterranean-xeric-oceanic and 

Mediterranean-pluviestational-continental bioclimates, 

ranging between thermomediterranean and mesomediterranean 

thennotypes and from semi-arid to sub-humid ombrotypes, 

sharing the areas occupied by shrub communities of 

holly-oaks (Quereus caceirera L.), lentiscs (Pistaeia lentiscus 

L.) or savines (Juniperus phoenicea L.) depending on 

the substrata nature. 

The iberian pine-woods of P. halepensis, take the case 

of the iberic southeastern have been scarcely studied, though 

they are one of the most characteristic elements in the plant 

landscape. In spite of the physiognomic importance and their 

widespread area, a large number of authors question the climacic 

role of these woods in the Mediterranean region, even 

stating that P. halepensis does not grow spontaneously in the 

Western Mediterranean. Though there are several bibliographic 

references in which a secondary and anthropic character 

is assigned to this pine in woods and scrubs, it has 

never been regarded as the main species in such communities. 

INTRODUCCION 

Pinus halepensis Mill. es un arbol generalmente 

retorcido, de unos 12-14 m de altura media, 

normalmente aparasolado, aunque puede llegar hasta 

los 22-24 m en las mejores situaciones ecol6gicas. 

Presenta una distribuci6n basicameitte circun 

mediteminea, formando bandas cercanas a la costa 

donde no suele superar los 700-800 m de altitud, a 

excepci6n de ciertas zonas de Africa del Norte donde 

puede llegar hasta los 2000 m (Quezel, 1977, 1980). 

Desde el punto de vista taxon6mico se trata de una 

especie ecol6gica y geneticamente muy pr6xima a 

Pinus brutia Ten. (Biger & Liphschitz, 1991) con el 

que forma un grupo bien definido (grupo Halepenses). 

Ampliamente extendido por todo el Mediterraneo 

occidental, es sustituido hacia el oeste por su 

vicariante Pinus brutia (Figura I). Aunque raramente 

suelen coexistir ambas especies (Akman et aI, 1978), 

en el caso de hacerlo, como ocurre en algunos distritos 

136 

Caracterizach5n de las pinares en el sur de la peninsula iberica 

There are only a few botanists that have pointed out the 

natural and authoctonous character of P. Iwlepensis in some 

territories, either co-dominating with trees and shrubs or 

dominating the community by itself. 

In our study, bearing in mind the results of fossil and 

subfossil registers of the last 15.000 years and the analysis 

and taxonomic identification of charcoals. we show interesting 

data about the spontaneous character of P. halepensis 

in the Baetic area (southern peninsula) where it takes part of 

edaphoxerophilous communities confined to the most thermic 

and dry zones (sunny exposures) due to the lithological, 

geomorphological and climatological territorial factors. The 

widespread calcareous-dolomitic emergences allow the existence 

of a sheer landscape with big rocky blocks where the 

permeability of the substrata reduces the effects of the real 

precipitations in the territory. Furthermore, the karstification 

processes, break and crush the rocks causing hiperxeric environments 

very suitable for the establishing of P. halepensis 

pine-woods. 

These pine-woods can also take place on marls (even 

with a high content of gypsum) in extremely degradated 

soils that accentuate the ombroclimatic xericity.This fact is 

quite common in places with rains shades where the precipitations 

coming from the Atlantic ocean are very reduced 

because of the existence of high mountain boundaries. 

Thus, P. halepensis (as other conifers) constitutes natural 

formations in southern Iberian peninsula, specially in the 

thermomediterranean and mesomediterranean belts, and at a 

lesser extent in the inferior supramediterranean. In some 

cases, a semi-arid territory rich in Neogenous-Quaternarian 

deposits (marls, calcareous marls and conglomerates) with a 

scarce water retention is optimum for the development of P. 

halepensis. In the other hand, a rough geomorphology of 

calcareous-dolomitic ranges may be the determining factor 

for the establishing of these pine-woods, in this case under a 

wider range of ombrotypes (semi-arid, dry and sub-humid). 

de Grecia, en el sureste de Anatolia y en el Lfbano, 

suelen formar hfbridos naturales (Panetsos, 1975). Las 

poblaciones apartadas de Pinus halepensis en Asia 

Menor y Cercano Oriente, al limite este de la Cuenca 

Mediterranea, plantea curiosas cuestiones a cerca del 

modelo de distribuci6n en el pasado y las rutas 

migratorias (Barbero et ai, 1998). 

En la Peninsula Iberica, las formaciones 

espontaneas de Pinus halepensis aparecen sobre 

sustratos basicos, principalmente en las provincias del 

litoral mediterraneo (Cataluna, Comunidad 

Valenciana, Murcia y Baleares), penetrando hacia el 

interior en las sierras Beticas, Sistema Iberico, Valle 

del Ebro y Pirineos orientales. Se distribuye 

mayoritariamente en los bioclimas Mediterraneo 

pluviestacional-oceanico, Mediterraneo xerico 

oceanico y Mediterraneo pluviestacional-continental 

con ciertas penetraciones en el Templado oceanico 

submediterraneo (Rivas-Martinez, I 996a, 1996b). 

Puesto que el factor determinante para su distribuci6n 

ecologia mediterranea 25 (2) - /999

Torres et al. 

parece ser la temperatura, especialmente las minimas 

invernales (Falusi et al., 1984), ocupan tan solo los 

termotipos termomeditemineo y mesomeditemineo 

con ombrotipos que oscilan entre el semiarido y el 

subhumedo, donde convive sobre todo con coscojares 

(Quercus coccifera L.), lentiscares (Pistacia lentiscus 

L.) y sabinares de sabina mora (luniperus phoenicea 

L.), dependiendo de la naturaleza del sustrato. 

Consideraciones sobre el canicter climacico de 

Pinus halepensis 

Los pinares ibericos de pino carrasco, y 

concretamente los del sudeste iberico, han sido objeto 

de escasos estudios botanicos a pesar de tratarse de 

uno de los elementos mas caracteristicos del paisaje 

vegetal de esta parte de la Peninsula Iberica. 

En este sentido cabe mencionar las referencias que 

algunos botanicos clasicos hacen de dichos pinares; 

asi, Huguet del Villar (1916, 1925) pone de manifiesto 

el caracter natural y aut6ctono de Pinus halepensis en 

algunos territorios en los que aparece, donde formaria 

conclimax con la encina (Quercus ilex). Cuatrecasas 

(1929) en su trabajo sobre Sierra Magina hace 

referencia al caracter climacico de estos pinares en las 

vertientes inferiores mas calidas y secas de todo el 

Macizo. Font Quer (1933) en sus observaciones 

botanicas sobre la vegetaci6n de Los Monegros 

constata el papel subalterno de Quercus coccifera en 

Caracterizacion de los pinares en el sur de la peninsula iberica 

su Pinetum halepensis. Laza Palacios (1946) en sus 

estudios sobre la flora y vegetaci6n de las Sierras 

Tejeda y Almijara describe un Pinetum halepensis para 

los pisos semiarido y templado de la clasificaci6n 

fitogeografica de Emberger donde, sin entrar a valorar 

la potencialidad climacica del pino de Alepo, considera 

de gran vitalidad el estado actual del pino carrasco. 

Rivas Goday (1951) al abordar el estudio de la 

vegetaci6n y flora de la provincia de Granada pone de 

manifiesto la espontaneidad de Pinus halepensis sobre 

dolomias. Braun-Blanquet et Bolos (1958) en sus 

estudios del Valle del Ebro al describir la asociaci6n 

Rhamno-Quercetum cocciferae esgrimen dentro de la 

faciaci6n tipica una variante con Pinus halepensis, en 

ombroclima semiarido, que catalogan como una 

comunidad particular donde Quercus coccifera subsiste 

bajo la copa de pinares de Pinus halepensis. Fernandez 

Casas (1972) en su estudio fitografico sobre la Cuenca 

del Guadiana Menor vuelve a resaltar su naturalidad 

sobre sustratos margosos y ombrotipo semiarido. Por 

ultimo, Costa (1987) al describir la vegetaci6n del Pais 

Valenciano insiste en considerar las formaciones de P. 

halepensis como naturales, aunque con un caracter 

secundario (nunca ocupan una posici6n climacica) y 

procedentes de la degradaci6n de carrascales. 

Mas recientemente, Arrojo (1994) pone de 

manifiesto el caracter natural de estos pinares en la 

Sierra de Castril (Granada). 

Figura I. Distribuci6n de Pinus halepensis en la Cuenca Mediteminea (modificado de Barbero et al., 1998) 

Figure 1. Pinus ha1epensis distribution in the Mediterranean basin (modified from Barbero et aI., 1998) 

ecologia mediterranea 25 (2) - 1999 137

Torres et al. 

4< ... .._ 

Caracterizaci6n de los pinares en el sur de la peninsula iberica 

Figura 2. Area de estudio. Localizaci6n biogeogriifica; Provincia Betica, I : Sector Subbetico, la distrito Subbetico-Maginense, lb 

distrito Cazorlense, lc distrito Alcaracense, Id distrito Subbetico-Murciano; 2: Sector Guadiciano-Bacense, 2a distrito Guadiciano-Bastetano, 

2b distrito Serrano-Bacense, 2c distrito Serrano-Mariense, 2d distrito Serrano-Estaciense. 

Figure 2. Study area. Biogeographical location: Baetic province, I. Sub-Baetic sector: Ia Sub-Baetic-Maginense district. I b Cazorlellse 

district. le Alcaracense district, Id Sub-Baetic-Murciano district. 2. Guadiciano-Bacense sector: 2a Guadiciano 

Bastetallo district, 2b Serrano-Bacense district, 2c Serrano-Mariense district, 2d Serrano-Estaciense district. 

De igual forma, Torres et al. (1994) en sus 

estudios sobre los pinares de Pinus halepensis en el 

sector Subbetico consideran el canicter autoctono de 

estas formaciones, criterio igualmente compartido por 

Navarro Reyes (1995, 1997) en las sierras de Baza y 

Las Estancias (Granada). 

De igual forma, diversos autores en otros paises 

del cntorno meditemineo destacan el caracter 

espontaneo de esta especie y su papel en el dinamismo 

de la vegetacion (Emberger, 1939 ; Tregubov, 1963 ; 

LoiseL 1971; Ozenda, 1975; Tomaselli, 1977; 

AchhaL 1986; Quezel , 1986; Quezel et al., 1987; 

Qubel et al., 1988; Fennane, 1988; Quezel & 

Barbero, 1992; Quezel et al., 1992; Barbero et al., 

1998). 

Pese a su importancia fisionomica y a su amplia 

distribucion, muchos autores discuten el papel 

climacico de estos bosques en la region Mediterranea, 

llegando incluso a afirmar que no crece 

espontaneamente en el Mediterraneo occidental. 

Existen muchas referencias concretas de esta especie 

138 

que le asignan siempre un papel secundario y antropico 

en diversas comunidades climacicas boscosas, 0 incluso 

de matorrales, pero en ningun caso se reconoce el papel 

prioritario que pueda tomar el pino carrasco en las 

mismas. Asf, Bolos (1962) considera la presencia de 

Pinus halepensis como un accidente en las 

comunidades vegetales en que aparece. Folch (1981) 

destaca el caracter secundario y antropico del pino 

carrasco y al igual que el anterior autor alude a la falta 

de especies vegetales caracterfsticas. Bolos (1987) en 

sus estudios sobre la vegetacion de Catalufia y la 

Depresion del Ebro plantea el problema de la 

participacion de Pinus halepensis en la estructura y 

dinamismo de la vegetacion propia del dominio del 

Rhamno-Quercetum cocciferae Braun-Blanquet & 

Bolos 1954, donde 10 considera frecuente en la 

maquia climacica, aunque normalmente con una baja 

densidad. Ferreras et a!' (1987) reinciden en su 

caracter secundario debido a la accion del hombre y 

los consideran como etapa serial de otras formaciones 

mas densas y sombrfas. Nuet et a!' (1991) vuelven a 

ecologia mediterranea 25 (2) - 1999

Torres et al. 

referir la falta de un cortejo floristico propio y por 

tanto no llegan a considerlos coma una comunidad 

vegetal. Sanchez-Gomez et Alcaraz (1993) consideran 

al pino carrasco coma especie integrante de la 

vegetacion permanente de roquedos calizos y que 

puede actuar coma primocolonizador de suelos 

degradados. 

Par otro lado las investigaciones paleobotanicas 

tambien contribuyen, a traves de los registros fosiles 0 

subfosiles, sobre todo los de los ultimos 15.000 afios, 

a evidenciar el caracter natural y autoctono de Pinus 

halepensis. En este sentido, los espectros polfnicos 

(Parra, 1983; Stika, 1988; Yll, 1988; Rivera & 

Obon, 1991; Lopez, 1991, 1992) Y sobre todo los 

analisis de maderas carbonizadas con sus precisiones 

taxonomicas e identificacion al nivel especffico 

(Vernet et al., 1983, 1987; Ros, 1988, 1992; 

Bernabeu & Badal, 1990, 1992; Rodriguez Ariza et 

al., 1991 ; Rodriguez Ariza, 1992 ; Rodriguez Ariza & 

Vernet, 1992; Grau, 1993; Dupre, 1988; Hopf, 

1991 ; Badal, 1991, 1995 ; Badal et al, 1994) apartan 

datos interesantes sobre la presencia ancestral de pino 

carrasco en la Peninsula Iberica. Las zonas de Levante 

son las que presentan mayor numero de datos, 

observandose la continuidad de pinG carrasco desde el 

Peniglaciar (Gil Sanchez et aI., 1996). 

Parece, pues, justificado considerar a Pinus 

halepensis, al igual que otras especies de coniferas, 

coma una especie autoctona ampliamente distribuida 

por el mediterraneo Occidental, y con caracter 

permanente 0 edafoxerofilo en aquellos espacios 

ecologicos donde determinados factores limitantes 

(edaficos, litologicos, geomarfologicos y climaticos) 

impiden su colonizacion par la vegetacion esclerofila 

potencial. En nuestra 0pllllOn muchas de las 

afirmaciones que discuten el caracter espontaneo de 

estos pinares se deben a la gran adaptabilidad 

ecologica de esta especie para difundirse a expensas 

de los bosques esclerofilos. 

Los pinares de Pinus halepensis en el sur de la 

Peninsula Iberica : area de estudio 

En el presente trabajo, se estudian las formaciones 

de Pinus halepensis en las zonas Beticas del sur 

peninsular. BiogeogrMicamente, la zona de estudio se 

extiende par los sectores Subbetico y Guadiciano 

Bacense de la provincia carologica Betica, localizados 

ecologia mediterranea 25 (2) - 1999 

Caracterizacion de los pinares en el sur de la peninsula ibhica 

en la zona centro-nororiental de Andalucia (Espafia) 

(Figura 2). 

Desde el punto de vista geologico es posible 

diferenciar dos grandes unidades tectonicas muy 

relacionadas con las tipologias de pinares que 

presentamos en este trabajo: par un lado, las Zonas 

Externas de las Cordilleras Beticas (Prebetico y 

Subbetico), de gran extension en el territorio y 

eminente caracter montafioso, con alturas en algunos 

casos superiores a los 2000 m, donde la litologfa 

responde basicamente a farmaciones calizo 

dolomfticas, en muchos casos incluso kakiritizadas, 

que proporcionan un relieve muy particular de aspecto 

ruiniforme, donde la potencia caliza-dolomia es muy 

variable, llegando incluso en algunos zonas a 

desaparecer completamente las caliza en favor de las 

dolomia, coma es el caso de la Sierras de Castril 

(Granada) y Cazorla-Segura y Las Villas (Jaen). 

Aunque los afloramientos rocosos ocupan una gran 

extension, los suelos que aparecen son de tipo litosol, 

ya que se trata de zonas con fuertes pendientes que 

han sufrido intensos procesos de erosion. 

Por otro lado, la unidad Neogeno-Cuaternaria esta 

representada por la Depresion del Guadiana Menor en 

la que abundan materiales sedimentarios; aparecen 

margas, margo-calizas y margas yesfferas, con 

afloramiento en las cotas mas altas de los materiales 

mas deleznables que permiten una rapida 

evapotranspiracion del agua y por tanto una alta 

xericidad edafica. Los suelos que se desarrollan son de 

tipo cambisoles calcicos y regosoles calcareos, con 

pendientes que oscilan entre el 10 Y el 20% con una 

topografia generalmente colinada. 

MATERIAL Y METODOS 

Los datos geologicos los hemos obtenido a partir 

de la cartograffa del Instituto Geologico y Minero de 

Espafia, hojas de Villacarrillo (Virgili & Fonbote, 

1987), Jaen (Garcfa Duefias & Fonbote, 1986), Baza 

(Vera & Fonbote, 1982) y Granada-Malaga (Aldaya et 

aI., 1980). Para el estudio de suelos hemos seguido a 

Aguilar et al. (1987) y Perez Pujalte & Prieto 

Fernandez (1980). La identificacion de bioclimas y 

pisos bioclimaticos se basan en Rivas-Martinez 

(I996a, 1996b). Para la distribucion biogeogrMica 

seguimos a Rivas-Martfnez et al. (1997). Las series de 

vegetacion han sido determinadas basandonos en la 

obra de Rivas-Martfnez (1987) y para el muestreo de 

139

Torres et al. 

las comunidades, hemos utilizado el metodo 

fitosociol6gico de la Escuela de Zurich-Montpellier 

(Braun-Blanquet, 1979). En la nomenclatura de los 

sintaxones se contemplan las normas del C6digo de 

Nomenclatura Fitosociol6gica (Barkman et aI., 1986), 

mientras que para la denominaci6n y autoria de los 

taxones utilizamos las siguientes obras: Flora Iberica 

(Castroviejo et al., 1986-1998) y Flora Europea (Tutin 

et al., 1964-1980), a excepci6n de Genista cinerea 

subsp. speciosa Rivas Goday et T. Losa ex Rivas 

Martinez, T.E. Diaz, F. Prieto, Loidi et Penas in Los 

Picos de Europa: 268. 1984; Leucanthemopsis 

spathulifolia (Gay) Rivas-Martinez, Asensi, Molero 

Mesa et Valle in Rivasgodaya 6: 42. 1991. 

RESULTADOS 

Las comunidades espontaneas de Pinus halepensis 

en la zona de estudio constituyen formaciones 

edafoxer6filas, limitadas por la litologia, 

geomorfologia y climatologia del territorio alas 

vertientes mas termicas y xer6filas del territorio, 

normalmente coincidentes con las exposiciones mas 

soleadas. Los abundantes afloramientos calizo 

dolomiticos definen un paisaje de escarpes, dolinas, 

lapiaces y grandes bloques rocosos, donde la gran 

permeabilidad de este tipo de sustratos, amortigua las 

precipitaciones reales del territorio. Si a su vez 

sumamos los procesos de karstificaci6n que 

fragmentan y trituran la roca madre, se generan 

medios hiperxericos optimos para el asentamiento de 

pinares de Pinus halepensis. De igual manera, los 

sustratos margosos, a veces con alto contenido en 

yesos y suelos extremadamente degradados, suelen 

acentuar la xericidad ombroclimatica y soportan 

masas de Pinus halepensis, normalmente en zonas de 

sombra de lluvias, donde las precipitaciones 

procedentes del Atlantico se yen muy mermadas por 

los macizos montafiosos del territorio. 

En estos territorios hemos reconocido dos 

formaciones vegetales de gran importancia ecologica 

que describimos como nuevas para la ciencia, donde el 

pino de halepo 0 carrasco (Pinus halepensis) adquiere un 

caracter relevante en la comunidad. 

Junipero phoeniceae-Pinetum halepensis ass. nova 

(Holotypus inv 13, Tabla 1) 

Constituyen pinares abiertos de Pinus halepensis, 

de baja cobertura, normalmente aparasolados, donde 

140 

Caracterizaci6n de los pinares en el sur de la peninsula iberica 

son frecuentes otras gimnospermas como Juniperus 

oxycedrus L. y Juniperus phoenicea L. Destaca la 

presencia de Rhamnus myrtifolius Willk. y Rhamnus 

lycioides L., ambos taxones bien adaptados a los 

suelos esqueleticos carbonatados y que soportan una 

gran desecacion del suelo. 

La composici6n floristica del matorral 

acompafiante, poco diversificada, es rica en 

endemismos beticos que dan matiz corologico a esta 

asociacion; entre ellos destacan Echinospartum 

boissieri (Spach) Rothm., Thymus orospedanus 

Huguet del Villar, Genista cinerea subsp. speciosa y 

Ptilostemon hispanicus (Lam.) Greuter, junto a una 

cohorte de elementos propios de dolomias como son 

Centaurea granatensis Boiss., Convolvulus boissieri 

Steudel, Centaurea boissieri DC. subsp. willkommii 

(Schultz Bip. ex Willk.) Dostal, PterocephaIus 

spathulatus (Lag.) Coulter, Fumana paradoxa 

(Heywood) J. Giiemes, Scorzonera albicans Cosson y 

Leucanthemopsis spathulifolia (Tabla 1). 

La comunidad se extiende por el termotipo 

mesomediterraneo del sector Subbetico, con 

irradiaciones en el sector Guadiciano-Bacense, 

alcanzando los 1400 m de altitud en algunas laderas 

soleadas del Macizo de Magina y Sierra de Castril. Su 

6ptimo aparece bajo ombroclima seco-subhumedo e 

inviernos templados. 

Se desarrolla en crestones y laderas abruptas sobre 

sustratos calcareos, calizas y calizo-dolomias, con 

suelos poco evolucionados de tipo litosol, donde 

adquiere caracter de comunidad permanente. La alta 

xericidad del sustrato y las elevadas temperaturas del 

periodo estival no permiten la entrada de la serie 

potencial de todo el territorio, correspondiente a los 

encinares del Paeonio coriaceae-Querceto 

rotundifoliae sigmetum, con la que contacta 

catenalmente hacia los suelos mas desarrollados. 

Desde el punto de vista sintaxon6mico guarda 

relacion con diversas comunidades paraclimacicas 

donde la sabina mora (Juniperus phoenicea), especie 

extremadamente austera y de marcado caracter 

rupfcola, esta representada de forma constante (Tabla 

3). La ausencia de Pinus halepensis en los sabinares 

del Rhamno lycioidis-Juniperetum phoeniceae Rivas 

Martinez et G. L6pez in G. L6pez 1976, pone de 

manifiesto el claro caracter continental de esta 

asociacion, cuyo optimo se alcanza en los termotipos 

meso- y supramediterraneo de la provincia Castellano- 


Torres et al. 

Orientaei6n 

Pendiente (%) 

Altitud ( 1= 10m) 

Area (I = 10m") 

Cobertura (%) 

Altura media (m) 

N° de orden 

NE 

5 

90 

40 

40 

4 

I 

Caracterfsticas de asociacion y unidades superiores 

Juniperus phoenicea I I 3 I 

Pinus halepensis 2 2 3 3 

Juniperus oxycedrus I + 2 

Rhamnus myrtifolius + + + 

Rhamnus lycioides + + + 

Pistacia terebinthus + + 

Phillyrea angustif()lia + 

Daphne gnidium 

Teuaium fruticans + + 

Buxus sempervirens 

Compafieras 

Rosmarinus ofjicinalis + + + 

Echinospartum boissieri I I 

Bupleurumfruticosum + 

Thymus zygis subsp. gracilis + + + 

Thymus orospedanus 

Teuaium capitatum + 

Phlomis purpurea + + 

Phagnalon saxatile + + 

Cistus albidus + + + 

Brachypodium retusum + + 1 

Vrginea maritima + + 

Genista cinerea subsp. speciosa + + 

Sedum sediforme + + 

Stipa tenacissima + 

Lavandula latifolia 

Ptilostemon hispanicum + 

Thymus mastichina 

Helianthemum crocewn + 

Bupleurum spinosum 

Festuca scariosa 

Helianthemum cinereum subsp. ruhellum 

Vlex parvijlorus 

Centaurea granatensis 

Convolvulus hoissieri 

Centaurea boissieri subsp. willkommii 

Pterocephalus spathulatus 

Fumana paradoxa 

Scorzonera alhicans 

Leucanthemopsis spathulifolia 

Helianthemum frigidulum 

Caracteriwcion de los pinares en el sur de la peninsula ibhica 

E E E E E SE S S SE SE SW S SE E 

90 35 25 10 25 15 70 40 30 30 30 25 40 30 

110 115 115 120 100 110 135 125 11570 110 127 125 130 

30 40 40 40 40 40 10 10 30 10 10 10 40 40 

40 70 60 60 40 60 40 40 60 65 65 60 40 50 

3 3 5 345 3 464 5 5 4 5 

2 3 4 5 6 7 8 9 10 1I 12 13 14 15 

I 

3 

3 

+ 

+ 

+ 

+ 

+ 

2 

+ 

+ 

+ 

+ 

+ 

+ 

132 I 2 

3 2 123 

I + + 

I + 

+ + 

+ 

+ 

+ 

+ 

2 

+ + + + 

I + 2 

+ 

+ 

+ 

+ 

+ 

+ 

+ + 

+ + 

+ + + 

+ 

+ 

+ I 

1 

2 

1 

+ 

+ 

2 2 I 

333 

I 1 + 

I 

I + 

I 

+ 

2 2 2 

I 

+ 

+ 

+ + 

+ 

2 2 

2 3 

+ 

+ 

+ 

+ 

+ 

2 

+ 

+ + 

+ + 

+ + 

+ + 

I 

+ 

+. 

+ 

+ 

+ + 

+ 

+ 

Ademas : Santo!ina canescens + en 3 ; Asphodelus alhus +, Arbutus unedo + en 4; Phlomis lychnitis + en 5; Prunus .lpinoS{l + en 

6 ; Aphyllantes monspeliensis + en 8' Cistus elusii I, Helianthemum cinereum + en 9; Sedum dasyphyllum I y Lactuca 

tenerrima + en 10; Polygala rupestris +, Asperula hirsuta + y Glohularia spinosa + en I1 ; Genista scorpius 2, Paronychia 

suffruticosa + en 12 ; Silene legionensis + en 13; Fumana ericoides + en 14 ; Teucrium pseudochamaepitys + en 15. 

Loealidades: 1,2,3,4 y 5 Cerea Castillo Otinar (Jaen); 6 Barraneo El Lobo (Sierra de Jaen, Jaen); 7 Barraneo Los Cortijuelos 

(Jaen); 8 Barraneo La Canada (Carehelejo, Jaen); 9 Cabra de Sto. Cristo (Jaen); 10 Cerro Cuello de Ventarique (Carehelejo, 

Jaen); II Arroyo Los Miradores (Los Villares, Jaen); 12 Castril-Pozo Alc6n (Granada); 13 Sierra de Castril (Granada); 14 y 15 

Sierra de la Cruz (Sierra de Magina, Jaen). 

Tabla I. Junipero phoenieeae-Pinetum halepensis ass. nova 

Tahle 1. Junipero phoeniceae-Pinetwn halepensis ass. nova 

ecologia mediterranea 25 (2) - 1999 141 

+

Torres et al. Caracterizaci6n de los pinares en el sur de la peninsula iberica 

Orientaei6n N W E N N N NE SE SWNW 

Pendiente (%) 15 10 20 10 40 10 10 30 35 35 

Altitud (l = IOm) 60 70 60 40 48 50 52 67 70 85 

Area (1= 10m 2 ) 40 40 40 40 40 40 40 10 10 10 

Cobertura (%) 40 35 40 60 40 50 50 60 65 65 

Altura media (m) 5 6 4 6 5 5 5 5 5 6 

N° de Orden I 2 3 4 5 6 7 8 9 10 

Caracteristicas de comunidad y unidades superiores 

Pinus halepensis 2 2 2 3 2 2 2 3 3 3 

Ephedrafragilis I 2 2 I 2 I 1 

Juniperus oxvcedrus 2 + 3 1 1 1 2 1 

Rhamnus lycioides + 2 I I 1 I 1 

Olea sylvestris + + + + I 

Juniperus plwenicea 1 

Rhamnus alatemus + 

Phillyrea angustifolia + 

Asparagus acutifolius + + + 

Compafieras 

Rosmarinus officinalis + + 2 

Phlomis purpurea 

Chronanthus biflorus + + 1 

Vlex parvijZorus + 1 + + + + 1 

Cistus albidus 1 I + + 1 1 

Thymus gracilis + + + 2 

Brachypodiulll retusum + + + 1 I 

Stipa tenacissima + 2 + + + 2 + 

Cistus clusii + 1 

Dactylis hispanica + 

Ademas: Echinospartum boissieri y Unum suffruticosum + en 1 ; Daphne gnidium +, Phagnalon saxatile +, Sthaehelina dubia + 

en 2; Phillvrea latifolia, Lonicera hispanica + en 3 ; Bupleurumfruticescens + en 10. 

Loealidades : I Cortijo de Herrera (Pegalajar, Jaen) ; 2 Cerea Castillo de Otifiar (Jaen) ;3 Finea de Otifiar (Jaen) ; 4 Puente Nuevo 

(Pegalajar, Jaen) ; 5,6 y 7 Cerro del Prior (Pegalajar, Jaen) ; 8 Cerea de Guadahortuna (Granada) ; 9 Cabra de Sto. Cristo (Jaen) ; 

10 Pefi6n de Alhamedilla (Granada). 

Maestrazgo-Manchega (sectores Manchego y 

Maestracense), aunque en algunos enclaves 

supramediternineos de este ultimo sector y del sector 

Celtiberico-Alcarrefio llega a enriquecerse en especies 

montanas como Juniperus communis L. 

hemisphaerica (K. Presl) Nyman y Pinus nigra 

Arnold subsp. saizmannii (Dunal) Franco. En las 

sierras Beticas sus relaciones se establecen tanto con 

el Rhamno myrtifoiii-Juniperetum phoeniceae Molera 

Mesa et Perez Raya 1987 de distribuci6n Malacitano 

Almijarense y Rondense occidental (Perez Raya et aI., 

1990), como con el Junipero phoeniceae-Pinetum 

saizmannii Valle, Mota et G6mez-Mercado 1988 de 

areal Subbetico, Guadiciano-Bacense y Malacitano 

Almijarense (Valle et ai, 1989). En ambos casos las 

diferencias florfsticas de nuestra comunidad con ellas 

son muy significativas a nivel de las comunidades 

seriales y en el estrato arb6reo las diferentes especies 

142 

Tabla 2. Comunidad de Ephedrafragilis y Pinus halepensis 

Table 2. Ephedra fragilis-Pinus halepensis community 

del genera Pinus matizan los distintos sintaxones 

(Tabla 3). En la tabla 1, presentamos distintos 

inventarios de esta comunidad, donde recogemos una 

variante mas humeda por la presencia de Buxus 

sempervirens L., localizado en topograffas menos 

accidentadas, y mayor humedad ambiental. 

Recientemente, Perez Latorre et a!. (1998) 

describen la asociaci6n Pino haiepensis-Juniperetum 

phoeniceae Perez Latorre et Cabezudo 1998 para 

caracterizar los pinares-sabinares edafoxer6filos 

dolomfticos termomediterraneos que sustituirfan al 

Chamaeropo-Juniperetum phoeniceae Rivas-Martinez 

in Alcaraz, T.E. Dfaz, Rivas-Martfnez & Sanchez 

G6mez 1989 en el sector Rondefio, aunque sin aportar 

diferencias eco16gicas y florfsticas con respecto a este 

ultimo. Por nuestra parte consideramos que deben de 

sinonimizarse prevaleciendo la asociaci6n de mayor 

antigiledad, en este caso Chamaeropo-Juniperetum 


Torres et al. 

phoeniceae (art. 25, C.P.N.), donde ya en su tipo 

nomenclatural (Rivas-Martinez ex Alcaraz in Cant6, 

Laorga & Belmonte, inv. 3, tabla 34, Opusc. Bot. 

Pharm. Complutensis, 3: 65, 1986) Pinus halepensis 

participa en la comunidad. 

En todo caso presenta marcadas diferencias con 

nuestra comunidad, sobre todo por la ausencia de 

especies de clam caracter term6filo y de 6ptimo 

termomediterraneo como Ceratonia siliqua L., 

Rhamnus lycioides L. subsp. velutinus (Boiss.) 

Nyman, Chamaerops humilis L., Aristolochia baetica 

L. 

Comunidad de Ephedra fragilis y Pinus halepensis 

(Tabla 2) 

Constituyen tambien formaciones abiertas de 

Pinus halepensis de porte mas moderado que en el 

caso anterior que ocupan sustratos blandos, 

principalmente margas, margas yesiferas y 

conglomerados con escasa capacidad de retenci6n de 

agua. En su composici6n floristica es frecuente la 

presencia de Ephedra fragilis Desf., Juniperus 

oxycedrus y Rhamnus lycioides, a veces acompafiados 

de Olea europaea L. var. sylvestris Brot. y en 

ocasiones Juniperus phoenicea. 

Aparece en el piso bioclimatico mesomediterraneo 

inferior semiarido del sector Guadiciano-Bacense 

(distrito Guadiciano-Bastetano), donde ocupa aquellos 

relieves y posiciones mas complejas, como carcavas, 

taludes de cierta pendiente y ramblas sometidos a 

elevada xericidad, con claro caracter topografico. 

Hacia las zonas mas favorables, con mayor 

compensaci6n hidrica y edafica es reemplazado por 

las comunidades climacicas del territorio, coscojares y 

lentiscares del Rhamno lycioidis-Quercion cocciferae 

Rivas Goday ex Rivas-Martinez 1975. Suele irradiar 

hacia las zonas basales de ombrotipo seco del sector 

Subbetico donde la continuidad de estos sustratos 

margosos, sumado al efecto de sombras de lluvias que 

producen la accidentada orografia de toda la unidad 

biogeografica permiten la presencia local del 

ombroclima semiarido. 

La alteraci6n que presenta el territorio por el uso 

del hombre a 10 largo de la historia aconseja, por 

ahora, su tratamiento como comunidad a la espera de 

estudios posteriores mas exhaustivos. 


Caracterizacion de los pinares en el sur de la peninsula iberica 

CONCLUSIONES 

Pinus halepensis, al igual que otras coniferas, 

constituye formaciones naturales en todo el sur de la 

Peninsula Iberica, ampliamente distribuidas por los 

termotipos termo- y mesomediterraneo, y en menor 

medida en el supramediterraneo inferior. 

En unos casos, la presencia de ambientes 

semiaridos ricos en materiales de dep6sitos ne6geno 

cuaternarios (margas, margocalizas y conglomerados) 

con escasa capacidad de retenci6n de agua, 

constituyen medios 6ptimos para el desarrollo de 

Pinus halepensis. En otros, la abrupta geomorfologia 

del territorio, rica en sustratos calizo-domiticos de 

dificil edafizaci6n, actua como factor determinante en 

la instalaci6n de comunidades permanentes de Pinus 

halepensis, en este caso, con una gran amplitud 

ombroclimatica (semiarido-seco-subhtimedo). 

REFERENCIAS 

Achhal, A., 1986. Etude phytosociologique et 

dendrometrique des ecosystemes forestiers du bassin 

versant du N'fis (Haut Atlas central). These Doct. es. 

Sci. Marseille, 204 p. + Annexes. 

Aguilar, J. et aI., 1987. Memoria del mapa de suelos de la 

provincia de Jaen. Escala 1: 200.000. Departamento de 

Edafologia y Quimica Agrfcola. Universidad de 

Granada, 552 pp. 

Akman, Y., Barbero, M. & Quezel, P., 1978. Contribution a 

l'etude de la vegetation forestiere d'Anatolie 

mediterraneenne. Phytocoenologia 5: 1-79. 

Aldaya, F., Vera, J. A. & Fontbote, J. M., 1980. Mapa y 

Memoria explicativa de la hoja 83 (Granada-Malaga) del 

Mapa Geol6gico Nacional E. 1:200.000. I.G.M.E. Serv. 

Public. Min. 1. y Energia. Madrid. 

Arrojo, E., 1994. Cartograffa de la vegetaci6n presente en la 

Sierra de Castrfl: bases para la conservaci6n de las 

comunidades vegeta!es. Tesis Doctoral. Ined. 

Universidad de Granada, 552 pp. 

Badal, E., 1991. La vegetaci6n durante el paleolftico 

superior en el Pais Valenciano y Andalucfa. Resultados 

antracol6gicos. In: Arqueologia medioambiental a traves 

de los macrorestos vegetales. Asociaci6n cultural viva 

(C.S.I.e.). Aula de Ecologia del Ayuntamiento de 

Madrid. Madrid. 

Badal, E., 1995. La vegetaci6n carbonizada. Resultados 

antracol6gicos del Pais Valenciano. In: El Cuaternario 

del Pais Valenciano. Asociaci6n espanola para el 

estudio del Cuatemario. Departamento de Geograffa de 

la Universitat de Valencia. Valencia: 217-226. 

Badal, E., Bernabeu, J. & Vernet, J. L., 1994. Vegetation 

changes and human action from the Neolithic to the 

Bronze Age (7000-4000 BP) in Alicante, Spain, based 

on charcoal analysis. Veget. Hist. Archaebot. 3: 155-166. 

143

Torres et al. 

groupements preforestiers et des matorrals rifains. 

Ecologia Mediterranea 14 (1/2): 77-122. 

Quezel, P. & Barbero, M., 1992. Le pin d'Alep et les 

especes voisines : repartition et caracteres ecologiques 

generaux, sa dynamique recente en France mediterraneenne. 

Forer mediterraneenne 13(3): 158-170. 

Quezel, P.; Barbero, M., Benabid, A. & Rivas-Martfnez, S., 

1992. Contribution a I' etude des groupements forestiers 

et preforestiers du Maroc oriental. Studia Botanica 10: 

57-90. Salamanca. 

Rivas Goday, S., 1951. Contribucion al estudio de la 

Vegetacion y Flora de la provincia de Granada. 

Excursion botanica a la Sierra de Baza y Ziijar. Anal. Re. 

Acad. Farmacia 7: 58-133. 

Rivas-Martfnez, S., 1987. Mapa de Series de Vegetacian de 

Espana escala 1:400.000 y Memoria. I.C.O.N.A. 

Madrid. 

Rivas-Martfnez, S., I996a. Clasificacion bioc1imatica de la 

Tierra. Folia Botdnica Matritensis 16 (33 pags.). Dpto. 

Biologfa Vegetal (Botanica). Universidad de Leon. 

Rivas-Martfnez, S., 1996b. Bioc1imatic map of Europa. 

Cartographic Service, University ofLean. Leon. 

Rivas-Martfnez, S., Dfaz, T.E., F. Prieto, J.A., Loidi, J. & 

Penas, A., 1984. Los Picos de Europa, Ediciones 

Leonesas, Leon. 295 pp. 

Rivas-Martfnez, S., Asensi, A., Molero, J. & Valle, F., 1991. 

Endemismos vasculares de Andalucfa. Rivasgodaya 6: 

5-76. 

Rivas-Martfnez, S., Asensi, A., Valle, F., Molero, J. & Dfez, 

B., 1997. Biogeographical synthesis of Andalusia 

(southern Spain). J. Biogeogr. 24: 915-928. 

Rivera, D. & Obon, e., 1991. Macrorrestos vegetales de los 

yacimientos de la comarca noroeste en los inicios de la 

Edad de los Metales. In: P. Lopez (ed.), El cambio 

cultural del IV al II milenios a. C en la comarea 

noroeste de Murcia. Vol I. e.S.I.C. Madrid: 239-246. 

Rodrfguez Ariza, M.O., 1992. Human-plant relationships 

during the Copper and Bronze Ages in the Baza and 

Guadix basins (Granada, Spain). Bull. Soc. bot. Fr. 139, 

Actual. bot. (2/3/4): 451-464. 

Rodrfguez Ariza, M.O & Vernet, J.L., 1991. Premiers 

resultats paleoecologiques de I' etablissement 

chalcolthique de Los Millares (Santa Fe de Mondujar, 

Almerfa, Espagne) d'apres l'analyse anthracologique. 

In: Waldren, W.H.; J.A. Ensenyal & R.e. Kennard (ed.), 

Iind Deya Conference of Prehistory 1-13, BAR 

International Series, Vol. I. 

Rodrfguez Ariza, M.O, Aguayo, P. & Moreno, E, 1992. The 

environment in the Ronda Basin (Malaga, Spain) during 

recent prehistory based on an anthracological study of 

Old Ronda. Bull. Soc. bot. Fr. 139, Actual. bot. (2/34): 

715-725. 

Ros, M.T., 1988. L'aplicacio de I'analisi antracologica a 

I'arqueologia catalana. Cota Zero 4: 51-60. 

146 

Caracterizacian de los pinares en el sur de la peninsula iberica 

Ros, M.T., 1992. Les apports de I' anthracologie a I' etude du 

paleoenvironnement vegetal en Catalogne (Espagne). 

Bull. Soc. bot. Fr. 139, Actual. bot. (2/3/4): 483-493. 

Sanchez-Gomez, P. & Alcaraz, E, 1993. Flora, vegetacian y 

paisaje vegetal de las Sierras de Segura Orientales. 

Instituto de Estudios Albacetenses. Excma. Diputacion 

de Albacete. Albacete, 459 pp. 

Stika, H.P., 1988. Botanische untersuchungen in der 

Bronzezeitlichen Hohensiedlung Fuente Alamo. 

Madrider Mitteilungen, 29:21-76. 

Tomaselli, R., 1977. Degradacion de la maquia 

mediterranea. In: Quezel, P., Tomaselli, R. & 

Morandini, R. (eds), Bosques y maquia mediterrdneos 

(edicion en castellano de 1982), Ediciones del Serbal, 

Barcelona: 59-134. 

Torres, J. A., Arrojo, E., Cano, E. & Valle, E, 1994. 

Formaciones de Pinus halepensis en el Sector Subbetico. 

In : Universitat de les Illes Balears (ed.), Libro de 

resumenes XV Jornadas de Fitosociologfa, Palma de 

Mallorca: 85. 

Treggubov, V., 1963. Etude des groupement vegetaux du 

Maroc oriental mediterraneen. Bull. Mus. Hist. Nat. 

Marseille, 23: 121-196. 

Tutin, T.G. & aI., 1964-1980. Flora Europaea. Vols. 1-5. 

Cambridge Univ. Press. 

Valle, E, Gomez Mercado, E, Mota, J. & Dfaz De la 

Guardia, e., 1989. Parque Natural de Cazorla, Segura y 

las Villas. Gufa botdnico-ecolagica. Edit. Rueda, 

Madrid, 354 pp. 

Vera, J.A. & Fontbote, J. M., 1982. Mapa y Memoria 

explicativa de la hoja 78 (Baza) del Mapa Geologico 

Nacional E 1:200.000. I.G.M.E. Serv. Public. Min. I. y 

Energfa. Madrid. 

Vernet, J. L.; Badal, E. & Grau, E., 1983. La vegetation 

neolithique du sud-est de I'Espagne (Valencia, Alicante) 

d'apres I'analyse antracologique. CR. Acad. Se. Paris, 

296, Serie 111-48: 669-672. 

Vernet, J. L.; Badal, E. & Grau, E., 1987. L'environnement 

vegetal de l'homme au Neolithique dans le sud-est de 

I'Espagne (Valencia, Alicante): Premiere synthese 

d'apres I' analyse anthracologique. Premieres 

Communautes Paysannes en Mediterranee occidentale, 

Colloque International du CNRS, Montpellier, 1983. 

Paris: 131-136. 

Virgili, e. & Fontbote, J. M., 1987. Mapa y Memoria 

explicativa de la hoja 71 (Villacarillo) del Mapa 

Geologico Nacional E. 1:200.000. I.G.M.E. Serv. Public. 

Min. I. y Energfa. Madrid. 

Yll, El., 1988. Analisis de polen y palinograma. In: Olaria 

(ed.). Cova fosca. Una sentamiento mesoneolftico de 

cazadores y pastores en la serranfa del alto maestrazgo. 

Monograffes de Prehistoria Arqueologfa 

Castellonenques. Diputacion de Castellon. Castellon de 

la Plana, 424 pp. 


Linhart et al. Selective herbivory ofthyme chemotypes by a mollusk and a grasshopper 

INTRODUCTION 

Many Mediterranean ecosystems are characterised 

by a limited availability of nutrients and water. For 

plants, this means that costs of leaf construction are 

high, and leaves must be well protected against the 

herbivores that are found in these challenging habitats 

(Ross & Sombrero, 1991). This need for protection 

has led to the synthesis by many Mediterranean plants 

of a large diversity of secondary compounds, many of 

which are aromatic. These plant species have charac 

teristic smells and tastes, which are often so specific 

that plants can be identified simply by these features, 

without recourse to other botanical characteristics 

such as floral features or leaf morphology. Rosemary 

(Rosmarinus), sage (Salvia), myrtle (Myrtus), bay 

(Laurus), junipers (Juniperus) and scores of other spe 

cies evoke specific odours and tastes. Many plants 

also show a significant amount of genetically-based 

intra-specific variability (Gleizes, 1976; Langenheim, 

1994). Foremost among them are species of the genus 

Thymus (Labiatae) which often have a diversity of 

smells and tastes, associated with the presence of sev 

eral monoterpenes (Adzet et al., 1977; Thompson, in 

press and refs. therein). Such intra-specific variability 

always elicits attention and questions about its origin 

and maintenance. This has been true for various Thy 

mus species and especially for T. vulgaris, a common 

component of dry garrigue-like or matorral ecosys 

tems of southern France and Spain. In T. vulgaris, in 

dividual plants are characterised by the predominance 

of a specific monoterpene (chemotype) in the essential 

oil sequestered in the epidermal glands that cover its 

leaves and stems. As a result, a given plant will have a 

characteristic odour and flavour. The six monoterpe 

nes present in southern France are geraniol (G), linalol 

(L), alpha terpineol (A), thuyanol (U), carvacrol (C) 

and thymol (T). Structurally, the first two are straight 

chains, the next two are cyclic and the last two are cy 

clic and phenolic. The synthesis of all six is under the 

control of a series of independently assorting epistatic 

loci. The details of its biochemistry and genetics are 

described in many publications (e.g. Passet, 1971; 

Vernet et al., 1977; Thompson et al., 1998; Linhart & 

Thompson, 1999). Many populations of thyme contain 

two or more chemotypes and the association between 

chemotypes and various botanical and geographic 

features is also well described, especially in southern 

148 

France (e.g. Gouyon et al., 1986b; Thompson, in 

press). 

The question we wish to address in this study has 

to do with the influences of this variability upon herbivory 

of T. vulgaris by two of its consumer species, 

the grasshopper Leptophyes punctitatissimum (Tetti 

goniidae, Orthoptera) and the slug Deroceras reticu 

latum (Limacidae, Mollusca). 

MATERIALS AND METHODS 

Framework 

The primary issues we addressed were i) whether 

the two herbivores exhibited repeatable patterns of 

selectivity among chemotypes by feeding on some 

chemotypes more than on others, and ii) whether the 

two species fed preferentially on the same chemo 

types. 

Plant material 

Thyme plants offered to the herbivores were clonally 

produced from mother plants, with at least 10 

mother plants per chemotype (A, C, G, L, T, U). 

Artificial diets for slugs 

Gelatine cakes were produced following the meth 

ods of Whelan (1982) and Linhart and Thompson 

(1995), and contained measured weights of chopped 

thyme leaves. 

Feeding trials 

Eight grasshoppers and nine slugs were used in in 

dividual cages. Each animal was presented with two 

replicates of all six chemotypes simultaneously. This 

methodology, while less tidy than the often used binary 

choice trials involving only two items at a time, 

is more realistic, as it reflects more accurately the 

complexities of garrigue vegetation which can contain 

several thyme chemotypes in addition to other 

monoterpene-containing species such as Juniperus 

spp., Rosmarinus, and others. Individual plants were 

4-6 cm tall and had 6 leaves. Artificial diets for slugs 

consisted in food disks measuring 8 mm in diameter, 

and weighing 0.2 g. Each disk contained the leaf 

equivalent to about eight fresh leaves. For grasshop- 



pers, no artificial diet was available, so that this type 

of experiment could not be carried out. 

Containers for individual animals consisted of: 

i) circular, soil-filled glass dishes 15 cm in diameter 

for the mollusks, covered with damp muslin, and 

ii) rectangular planting trays 20x30 cm in size covered 

with metal mesh cages about 5 cm in height for the 

grasshoppers. 

All experiments that followed the protocols described 

here were carried out once. In the case of 

slugs, different animals were used in the two experi 

ments, to prevent the possibility that learning in the 

first trial with the living plants, might influence slug 

behaviour in the second experiment. Some further tri 

als were run, usually with fewer animals, and lasting 

longer periods of time. Because of these differences, 

they were not directly comparable to the experiments 

reported here, and were not analysed statistically. 

However, all observations germane to the results re 

ported here are also noted. 

RESULTS 

Deroceras 

The slugs ate primarily plants of the G and A che 

motypes, and ate little of the U, C, and T chemotypes 

(Table I) of the nine individuals tested, five ate mostly 

G, three ate more A than G and one ate mostly A and 

L. In subsequent tests with fewer animals, G was also 

the chemotype most eaten. 

Tests with gel cakes showed that animals ate more 

of G than the other chemotypes with L also selected, 

and C the least eaten. Once again, there were differ 

ences among individuals, with 4/10 individuals prefer 

ring G, two a combination of G and L, one feeding 

mostly on L, one on U, and two showing no clear 

preference. 

Leptophyes 

The grasshoppers ate all chemotypes, but ate most 

of chemotype T (Table 1). There was marked hetero 

geneity among individuals. In additional runs (data not 

shown) of 8 individuals exposed to 6 chemotypes, 

three showed a strong preference for T while the oth 

ers were less selective in their feeding. 

DISCUSSION 

Both herbivores ate a diversity of thyme chemo 

types but concentrated their feeding on some chemo 

types and were deterred by others. There was a 

general consensus about the most and least eaten che 

motypes for each species, although there was also 

some heterogeneity among individuals. This heteroge 

neity reflects phenotypic variation among individuals, 

observed in most studies of this type, and underscores 

the need to carry out such tests with adequate numbers 

of animals. The two species differed dramatically in 

their preferences, with Leptophyes feeding more on 

thymol (T), while Deroceras fed very little on that 

chemotype, eating primarily plants of geraniol (G) and 

terpineol (A) chemotypes. 

Species Chemotype P (3) 

Plant material G L A U C T 

SLUG 

Plants (I) 4.1±0.6 2.7±0.5 3.9±0.6 1.0±0.2 0.9±0.2 1.0±0.3


At first contact, the primary detectable differences 

among thyme plants are those associated with the 

monoterpene contents of their glands. But there are 

also other, more subtle differences among them: these 

include for example pubescence, leaf size, thickness, 

toughness and disposition, and all may affect the 

plants' palatability to various herbivores (Strong et aI., 

1984). For this reason, an artificial diet which reduces 

these differences among food items is an important 

confirmation of the specific role of terpenes in herbi 

vore choice. We were only able to use artificial diets 

using ground leaves for the slugs as no artificial diets 

were available for the grasshoppers. 

The slugs ate somewhat larger amounts of geraniol 

containing cakes, just as they ate more plants of the G 

chemotype than the others. This observation provides 

support for the role of the specific monoterpenes synthesised 

by the plants as determinants of level of her 

bivory. This role was further supported in other 

experiments which involved the snail Helix aspersa 

(Linhart & Thompson, 1995) goats (Capra hircus) and 

sheep (Ovis aries) (Linhart & Thompson, 1999). In all 

three cases, patterns of relative preference of live 

plants characterised by specific chemotypes were the 

same as choices made when feeding on synthetic diets 

supplemented with pure distilled monoterpenes (Lin 

hart & Thompson 1995, 1999). 

The difference in patterns of preference between 

slugs and grasshoppers is notable as it shows that the 

two animals react very differently to the smell and/or 

the taste of the two molecules. These interspecific dif 

ferences are also interesting in combination with another 

comparison made between patterns of herbivory 

by the grasshopper Omocestus viridulus, and the 

mollusk Deroceras reticulatum feeding upon mixtures 

of the graminoid Dactylis glomerata and the legume 

Trifolium repens. The two animals behaved very dif 

ferently, with the grasshopper feeding disproportion 

ately on the common species, with a periodic 

preference for the grass, while the slug fed dispropor 

tionately on the rarer species, with a consistent prefer 

ence for the legume (Cottam, 1985). Such results 

further document the existence of important inter 

specific differences in herbivore behaviour, and sug 

gest that the presence of multiple species of herbivores 

can contribute to small-scale heterogeneity of botani 

cal and biochemical composition within plant communities. 

150 

The patterns of herbivory shown by the slugs in 

our experiment differ markedly from those of Gouyon 

et al. (l986a) who found the animals in their experi 

ments showed preference patterns of A=C>T>U. We 

presume the differences are associated with the fact 

that their slugs had no experience feeding on thyme, 

and also were not exposed to G or L. In any case, the 

differences illustrate how animal origin and experi 

mental design can influence the outcome of a feeding 

trial. 

The inter-specific differences we saw are consis 

tent with other patterns of herbivory of T. vulgaris. 

For example Helix aspersa and the Chrysomelid beetle 

Arima marginata preferred the L chemotype, while 

sheep (Ovis aries) preferred U, while goats although 

less specific in their choices, clearly preferred A, L or 

C to the other three chemotypes either in the form of 

whole plants or in synthetic diets (Linhart & Thomp 

son, 1999). Other plants that show genetically-based 

intra-species variability in chemical defenses have 

also been studied experimentally by exposing diverse 

chemical phenotypes to various herbivores and para 

sites. In the great majority of cases, the herbivores 

show species-specific host selection; that is, there are 

consistent, inter-specific differences in the chemical 

phenotypes they select as food or host plants. This is 

not surprising, as one may well expect that these dif 

ferences reflect the interspecific differences in physi 

ology, metabolism and behaviour that are to be 

expected among organisms as different as molluscs, 

insects, fungi, birds, fish and mammals. These pat 

terns have been observed repeatedly in many plant 

species as diverse as Labiatae, Leguminosae, Graminae, 

Fucoidae, Compositae, and various conifers 

(Simms, 1990; Linhart, 1991; Linhart & Thompson, 

1999; Iwao & Rausher, 1997; Van Der Meijden, 1996; 

Snyder & Linhart, 1998). The consistent nature of 

these observations implies that such species-specific 

herbivory and parasitism can contribute to the mainte 

nance of genetic variability in host defenses via com 

plex patterns of selection. The results also indicate 

that for a given plant species, no one molecule can be 

counted on as a dependable defense against diverse 

herbivores. 

Acknowledgements 

We thank the staff of the Centre d'Ecologie Fonc 

tionnelle et Evolutive (CNRS) in Montpellier, France, 


Linhart et al. Selective herbivary afthyme chematypes by a mallusk and a grasshopper 

and especially C. Collin, D. Couvet, M. Debussche, 

and B. Dommee. Very useful help and advice were 

also provided by J. Lamy and the Chambre d'Agri 

culture de la Drome. Technical assistance was provided 

by L. Bowden, N.B. Linhart, and B. Stojanovic 

Meunier. Financial aid was provided by the U.S. Na 

tional Science Foundation, and the University of Colo 

rado. 

REFERENCES 

Adzet T., Granger R., Passet 1. & San Martin R., 1977. Le 

polymorphisme chimique dans le genre Thymus: sa signification 

taxonomique. Biochem. Syst. Evol., 5: 269 

272. 

Cottam O.A., 1985. Frequency-dependent grazing by slugs 

and grasshoppers. J. Ecol., 73: 925-933. 

Gleizes M., 1976. Biologie des terpenes vegetaux: etude des 

composes monoterpeniques et sesquiterpeniques. Annee 

BioI., 15: 7-127. 

Gouyon P.H., Fort P. & Caraux G., 1986a. Selection of 

seedlings of Thymus vulgaris by grazing slugs. 1. Ecol., 

71: 299-306. 

Gouyon P.H., Vernet P., Guillerm J.L. & Valdeyron G., 

1986b. Polymorphisms and environment. The adaptive 

value of the oil polymorphisms in Thymus vulgaris. L. 

Heredity, 57: 59-66. 

Iwao K. & Rausher M., 1997. Evolution of plant resistance 

to multiple herbivores: quantifying diffuse coevolution. 

Am. Nat., 149: 316-335. 

Langenheim J.H., 1994. Higher plant terpenoids: a phytocentric 

view of their ecological roles. J. Chem. Ecol., 20: 

1223-1280. 

Letchamo W., Saez F. & Stahl-Biskup E., (in press). The 

genus Thymus. 

Linhart Y.B., 1991. Disease, parasitism and herbivory: multi 

dimensional challenger in plant evolution. Trends Ecol. 

Evol., 6: 392-396. 

Linhart Y.B. & Thompson J.O., 1995. Terpene based selective 

herbivory by Helix aspersa (Mollusca) on Thymus 

vulgaris (Labiatae). Oecologia, 102: 126-132. 


Linhart Y.B. & Thompson J.O., 1999. Thyme is of the essence: 

biochemical polymorphism and multi-species 

deterrence. Evol. Ecol. Research, 1: 151-171. 

Passet J., 1971. Thymus vulgaris L. : chemotaxonomie et 

biogenese monoterpenique. These Ooct. d'Etat. Universite 

de Montpellier, Montpellier. 

Ross J.O. & Sombrero c., 1991. Environmental control of 

essential oil production in Mediterranean plants. In: 1.B. 

Harborne & F.A. Tomas-Barberan (eds.). Ecological 

chemistry and biochemistry ofplant terpenoids. Clarendon, 

Oxford, U.K., 83-94. 

Simms E.L., 1990. Examining selection on the multi-variate 

phenotype: plant resistance to herbivores. Evolution, 44: 

1173-1188. 

Sokal R.R. & Rohlf F.J., 1981. Biometry. 2nd ed., Freeman, 

San Francisco, CA. 

Strong O.P., Lawton J.H. & Southwood T.R.E., 1984. Insects 

on Plants. Community Patterns and Mechanisms. 

Harvard Univ. Press Cambridge, MA. 

Thompson J.O., (in press). Population structure, and the 

spatial dynamics of genetic polymorphism in thyme. In: 

Letchamo, W, Saez, F. & Stahl-Biskup, E. (Eds.), in 

press: The genus Thymus. 

Thompson 1.0., Manicacci O. & Tarayre M., 1998. Thirty 

five years of thyme: a tale of two polymorphisms. Bioscience, 

48: 805-815. 

Snyder M.A. & Linhart Y.B., 1998. Subspecific selectivity 

by a mammalian herbivore: geographic differentiation of 

interactions between two taxa of Sciurus aberti and 

Pinus ponderosa. Evol. Ecol., 12: 755-765. 

Van Oer Meijden E., 1996. Plant defense, an evolutionary 

dilemma: contrasting effects of (specialist and generalist) 

herbivores and natural enemies. Ent. Exp. Appl., 80: 

307-310. 

Vernet P., Gouyon P.K. & Valdeyron G., 1977. Genetic 

control of the oil content in Thymus vulgaris L.: a case 

of polymorphism in a biosynthetic chain. Genetica, 69: 

227-231. 

Whelan R.J., 1982. An artificial medium for feeding choice 

experiments with slugs. J. Appl. Ecol., 19: 89-94. 

151

ecologia mediterranea 25 (2),153-161 (1999) 

Patterns and correlates of exotic and endemic plant taxa in the 

Balearic islands 

, 1 _ 

Montserrat VILA & Irma MUNOZ 

Centre de Recerca Ecologica i Aplicacions Forestals, Universitat Autonoma de Barcelona, 08193 Bellaterra, Barcelona, Spain 

1 

Author for correspondence: Tel: 93-5811987, Fax: 93-5811312, e-mail: vila@cc.uab.es 

ABSTRACT 

We analysed the taxonomy and biogeography of endemic and exotic plants in the Balearic Islands (Spain), one of the "hot-spots" 

of the Mediterranean Basin. Richness, diversity and density (number of taxallog 1o area) of exotic taxa (species and subspecies) is 

higher than that of endemic taxa. Mallorca is the island with the highest number of endemic and exotic taxa. On average, exotic 

and endemic taxa are little abundant or rare and represent 8.4% and 6% of the total flora, respectively. The taxonomic distribution 

of both exotic and endemic species is not random: Solanaceae, Amaranthaceae, Iridaceae and Euphorbiaceae are overrepresented 

families within the exotic taxa. Plumbaginaceae, Labiatae and Rubiaceae are over-represented among endemics. 

While most exotic taxa are therophytes, chamaephytes are the dominant life-form among endemics. Exotics are mainly found in 

cultivated areas, in disturbed and ruderal communities, while most endemics are located in rocky habitats. Coastal communities 

display a great proportion of endemic taxa (35.94 %), and are little represented by exotic taxa (5.94 %). It is in this habitat where 

most effort should be addressed in order to preserve both endemic and non-endemic native vegetation. 

Key words: commonness, conservation of islands, endemism, Mallorca, Pithyusic Islands, rarity 

RESUME 

Nous avons analyse la taxonomie et la biogeographie des plantes endemiques et exotiques des lies Baleares (Espagne), l'un des 

"hot-spots" du bassin mediterraneen. La richesse, la diversite et la densite (nombre de taxallog lO zones) des taxa exotiques (especes 

et sous-especes) sont plus elevees que celles des taxa endemiques. Majorque est 1'lIe ayant le plus grand nombre de taxa endemiques 

et exotiques. En moyenne, les taxa exotiques et endemiques sont peu abondants ou rares, et representent 

respectivement 8,4% et 6% de la flore totale. La repartition taxonomique des especes exotiques et endemiques n'est pas aleatoire 

: les Solanaceae, Amaranthaceae, Iridaceae et Euphorbiaceae sont les families sur-representees parmi les exotiques. Les 

Plumbaginaceae, Labiatae et Rubiaceae sont sur-representees parmi les endemiques. Tandis que la plupart des taxa exotiques 

sont des therophytes, les chamaephytes constituent la forme de vie dominante parmi les endemiques. Les especes exotiques se 

rencontrent principalement dans les zones cultivees et les communautes perturbees et ruderales, alors que la plupart des endemiques 

s'observent dans les habitats rocheux. Les communautes c6tieres affichent une grande proportion de taxa endemiques 

(35,94 %), mais peu de taxa exotiques (5,94 %). Ce sont dans ces biotopes littoraux que la plupart des efforts devraient etre entrepris 

afin de preserver la vegetation indigene, endemique et non-endemique. 

Mots-ch\s : frequence, conservation des lies, endemisme, Majorque, lies Pithyuses, rarete 

153

Vilcl & Mulio:: 

INTRODUCTION 

There is a great interest in understanding the proc 

esses that shape the ecology of endemic and intro 

duced species (e.g. Gaston, 1994; Drake et a!., 1989, 

respectively). This concern requests a strong previous 

knowledge of the diversity and abundance patterns of 

these taxa (McIntyre, 1992; Schwartz, 1993; Cowling 

& Samways, 1995). On the other hand, preservation 

of endemic species and control of introduced species 

are two main goals of conservation programmes 

world-wide that often are simultaneously carried out 

(Usher, 1986; Houston & Scheiner, 1995; Schieren 

beck, 1995). Thus research on the distribution patterns 

of endemics and exotics at the regional level and 

taxonomic, biological and ecological affinities of both 

groups of taxa is imperative to highlight hypothesis to 

be experimentally tested, and also to advance on basic 

knowledge for conservation practices. 

Islands have high levels of plant endemism. The 

best examples are found in big isolated islands such as 

Madagascar with 12.000 species 80 % of wich are en 

demic or New Zealand with 82 % endemic species. 

Endemics are also common in small islands: Canary 

Islands (612 endemics), Mauricio Island (280 endem 

ics), Madeira Islands (129 endemics) (Lean & Hin 

richsen, 1990). In the Mediterranean Basin, the 

Tyrrhenian Islands are one of the 10 "hot-spots" of 

species diversity and have almost 20 % of endemic 

plant taxa (Medail & Quezel, 1997). 

Endemic plant species are especially vulnerable in 

islands (Eliasson, 1995). For example, more than 90 

% of endemic plants in Sta. Elena, Ascension Island 

and Lord Howe island are endangered (Lean & Hin 

richsen, 1990). Intrinsic causes of vulnerability are 

related to the characteristics of insular species such as 

biological simplicity and reduced dispersal (Carlquist, 

1965; Eliasson, 1995; Cody & McCoverton 1996; 

Schiffman 1997). However, the main causes of such 

vulnerability are overexploitation, deforestation, 

habitat destruction, alteration of regional hydrological 

cycles, water pollution and species introductions. 

Islands are very vulnerable to biological invasions 

(Loope & Mueller-Dombois, 1989; Atkinson & Cam 

eron, 1993; McDonald & Cooper, 1995). The percent 

age of exotic plant species is very high in islands, i.e. 

Hawaii (44 %), New Zealand (40 %), British Islands 

(43 %), Ascension Island (83 %) (Vitousek et a!., 

1997). This high fraction of exotic species may be re- 

154 

Diversity ofexotic and endemic plants in the Balearic Islands 

lated to higher invasibility of islands due to higher 

number of releases and propagules per unit area, a 

lack of biotic mechanisms controlling invasion, the 

existence of unsaturated communities, high distur 

bance regimes, higher susceptibility to the effects of 

invaders than similar mainland areas (D'Antonio & 

Dudley, 1995) and also to a large perimeter-area ratio 

than that for continents (Lonsdale, 1999). 

The Balearic Islands have a high species diversity 

(Simon, 1994) and are rich in endemic taxa (G6mez 

Campo et a!', 1984). Endemic plants have been stud 

ied with regard to evolutionary origin based on cyto 

taxonomy analysis (Cardona & Contandriopoulos, 

1979; Contandriopoulos & Cardona, 1984). However, 

the diversity and distribution of endemic taxa has not 

been quantified and compared to that of the exotic 

component. In this study we analyse several aspects of 

the diversity and distribution of both the endemic and 

the exotic component of the Balearic flora. The ques 

tions are: I) Are endemic and exotic taxa similarly 

abundant? 2) Are there taxonomic and life-history 

similarities between endemic and exotic taxa? 3) In 

which communities are endemic and exotic taxa lo 

cated? The biogeographic origin of exotics is also 

commented. We base our study on a bibliographic 

survey. 

METHODS 

Area of study 

The Balearic Islands are the most eastern islands 

of the Mediterranean Basin and belong together with 

Corsica, Sardenia and Sicily to the Tyrrhenian Islands. 

They originated after the geologic drift and posterior 

rotation of 30° of the Cyrno-Sardinian plate at the end 

of the Oligocene and early Miocene from the adjacent 

coast of Provence, Languedoc and NE Catalonia. 

Materials are calcareous from the Triassic to the terti 

ary except in Menorca, where the geology is more 

heterogeneous and contains silicic esquists. The flora 

is typically Mediterranean dominated by evergreen 

sclerophyllous shrubs and forests. The Balearic Is 

lands form two different groups in terms of their geol 

ogy and endemism. The eastern Balearic Islands or 

Gymnesias Islands (Mallorca and Menorca) that have 

Tyrrhenian affinities, and the western Balearic Islands 

or Pithyusic Islands (Eivissa and Formentera) that 

have an Iberian and North African affinity. 


VilCI & Munoz Diversity ofexotic and endemic plants in the Balearic Islands 

Region Area (km2) No. endemics (%) No. endemics/loglO area Family diversitya Abundance b 

Balearic Islands 5014 89 (6) 24 2.9 2.2 

Mallorca 3655.9 70 (5.8) 19.6 2.6 2.3 

Menorca 701.8 39 (4.4) 13.7 2.5 2.6 

Pithyusic Islands 623.3 26 (2.3) 9.3 2.9 2.4 

Pithyusic Islands = Eivissa and Formentera. 

aCalculated as the Shannon index = -EPix1nPi where Pi is the number of endemic taxa in family i divided by the total number of endemic taxa. 

b Median value of 1= RRR (very rare), 2 = RR (rare), 3 = R (not abundant), 4 = C (relatively common), 5 = CC (common) and 6 = CCC (very 

common) according to Bolos et al (1993) nomenclature. 

Table I. Numbers of endemic taxa of the Balearic Islands 

Family No. endemic taxa (%)1 No. native non-endemic taxa (%)1 X 2 

PI umbaginaceae 14 (15.7) 18(1.3) 627.7 * 

Fabaceae 12 (13.5) 132 (9.7) 1.5 ns 

Asteraceae 9 (10.1) 151(11.1) 0.0 ns 

Labiatae 7 (7.8) 46(3.4) 6.5 * 

Umbelliferae 6 (6.7) 52 (3.8) 2.5 ns 

Scrophulariaceae 4 (4.5) 46 (3.4) 0.3 ns 

Rubiaceae 4 (4.5) 24 (1.8) 4.2 * 

Caryophyllaceae 3 (3.4) 59 (4.3) 0.3 ns 

Euphorbiaceae 3 (3.4) 26 (1.9) 1.0 ns 

Ranunculaceae 3 (3.4) 29 (2.1) 0.65 ns 

I number oftaxa in the family/total number oftaxa ofthe respective group 

* p< 0.05, ns = not significant. X 2 compares the no. endemic taxa (%) with the no. native non-endemic taxa (%). 

If significant, the family is over-represented. 

Table 2. The ten largest families of endemic taxa in the Balearic Islands 

Lifeform 

Therophytes 

Hemicryptophytes 

Nanophanerophytes 

Macrophanerophytes 

Phanerophytes 

Chamaephytes 

Geophytes 

% endemics 

5.6 

27.0 

13.5 

1.1 

44.9 

7.9 

% exotics 

37.9 

17.7 

5.6 

11.3 

5.6 

12.1 

9.7 

Table 3. Percentage of lifeforms of endemic and exotic taxa in the Balearic Islands 

Distribution and abundance of the endemic taxa 

Of the 89 endemic taxa, 5 are also endemic in other 

north-eastern regions of Spain: Medicago arborea 

subsp. citrina (Papilionaceae) also endemic to Co 

lumbretes islands (Valencia), Asplenium petrarchae 

subsp. majoricum (Polypodiaceae), Asperula cynan 

chica subsp. paui (Rubiaceae) and Saxifraga corsica 

subsp. cossoniana (Saxifragaceae) are also present in 

Valencia and Limonium gobertii (Plumbaginaceae) is 

also endemic to Catalonia. 

Mallorca is the island with most endemic taxa, and 

with the highest density and family diversity of endemic 

taxa. Endemic taxa have low abundance (R) or 

are rare (RR) in all islands (Table I). 

156 

Most endemic plants occur on rocky habitats (64 

%) in the mountaintops mainly in non coastal sites 

(39.36 %). Coastal communities also display a great 

proportion of endemic plants (35.11 %). Only 7 taxa 

are located in ruderal communities. There are 10 en 

demic taxa in shrublands and only two in forests (Ta 

ble 4). 

Diversity and origin of the exotic taxa 

The Balearic Islands display 124 exotic taxa dis 

tributed in 46 families that represent 8.4% of the total 

flora. Only one exotic species is a gymnosperm, 17 

are monocotyledons and 106 are dicotyledons. Den 

sity and family diversity for exotics are 33.5 and 3.31 

respectively (Table 5). 


Vi/a & Muiioz Diversity ofexotic and endemic plants in the Balearic Islands 

Family No. exotic taxa (%)1 No. native taxa (%)1 

Asteraceae 17 (13.7) 160 (10.8) 

Fabaceae 14 (11.3) 144 (9.9) 

Solanaceae 9 (7.2) 16 (Ll) 

Poaceae 7 (5.6) 160 (10.8) 

Amaranthaceae 6 (4.8) 9 (0.6) 

Brassicaceae 6 (4.8) 59 (4) 

Euphorbiaceae 6 (4.8) 29 (2.2) 

Iridaceae 6 (4.8) 11 (0.7) 

Labiatae 5 (4) 53 (3.4) 

Chenopodiaceae 4 (3.2) 30 (2.0) 

J number oftaxa in the family/total number oftaxa ofthe respective group 

*** p< 0.001, * p< 0.05, ns = not significant. X 2 compares the no. exotic taxa (%) 

with the no. native taxa (%). If significant, the family is over-represented. 

Table 6. The ten largest families of exotic taxa in the Balearic Islands 

Region 

Non tropical America 

Tropical America 

Mediterranean region 

Africa 

Asia 

Middle East 

Submediterranean region 

Tropical 

Oceania 

Macaronesia 

Unknown 

In the Balearic Islands, the most represented exotics 

have an American origin like for other Mediterranean 

Basin regions (Di Castri, 1989; Groves & Di 

Castri, 1991), and the families with the largest number 

of exotic taxa belong also to the largest families 

world-wide i.e. Asteraceae, Fabaceae, Poaceae 

(Weber, 1997; Daehler, 1998; Pysek, 1998). However, 

some families were over-represented. This may partly 

be explained by deliberate and reiterated introductions 

of certain taxa and by specific features of these taxa, 

making them more invasive. It may also reflect the 

identity of naturalised plants, e.g. the Amaranthaceae 

contain many weeds in agroecosystems. 

There is not a strong overlap of habitats occupied 

by endemic and exotic taxa. While endemic taxa are 

located in more isolated pristine habitats, exotic taxa 

are located in most disturbed habitats, except for 

coastal communities where an important proportion of 

both taxa co-occur. These habitats, are the ones where 

endemics will be most threatened by invasion by exotics. 

For example, invasion of Carpobrotus edulis is 

very high in the Mallorcan coast and is threatening 

several endemic Limonium spp. (PANDION, 1997). 

Moreover, in coastal habitats the human influence is 


No. taxa 

26 

18 

26 

13 

10 

10 

5 

4 

2 

1 

19 

% taxa 

19.4 

13.4 

19.4 

9.7 

7.5 

7.5 

3.7 

3.0 

1.5 

0.7 

14.2 

Table 7. Biogeographic origin of exotic taxa of the Balearic Islands 

1.2 ns 

0.3 ns 

99.0 *** 

2.8 ns 

78.6 *** 

0.3 ns 

5.0 * 

60.7 *** 

0.1 ns 

0.9 ns 

the strongest, and this increases the rate of species introduction 

and threat to endemics. 

Isolated edaphic systems appear to be major endemic 

centres (G6mez-Campo et aI., 1984). Besides 

crop fields, most exotic taxa were found in anthropogenic 

habitats (dumps, roadsides). Water courses are 

especially prone to invasion by exotic plants because 

they act as effective corridors providing a route for the 

dispersal of water-borne propagu1es (de Waal et aI., 

1994). Very few exotics succeed in closed forest and 

shrublands. Low disturbance levels may prevent invasion 

of closed forest and shrub1ands (Hobbs & Huenneke, 

1992). 

Often the term rarity is confused with that of endangered 

taxa but they are not synonyms (Kruckeberg 

& Rabinowitz, 1985). It is important to notice that in 

average both endemic and exotic taxa are rare. However, 

putative mechanisms of rarefaction are different 

in both group of taxa and their fate may also discourse 

in opposite directions. At the human scale, we have 

witnessed changes from common to rare in native taxa 

and from rare to common in exotic taxa (Kruckeberg 

& Rabinowitz, 1985). The rarity status of exotic taxa 

does not prevent their invasion status either. Because 

159

Vila & MuilOZ 

rarity depends on geographical range, habitat speci 

ficity and population size, one exotic taxa can be re 

stricted to a small geographical area but be very 

abundant or vice-versa, have widespread small popu 

lations. In both cases its presence may be considered 

invasive and may have an impact on the native biota 

and ecosystem functioning (Rabinowitz et al., 1986). 

Endemics and exotics are two faces of the same 

coin because management of both taxa have strong 

conservation implications. Five percent of the overall 

flora of the Balearic Islands is seriously endangered 

(Mus & Mayol, 1993). In addition, new introduced 

species are becoming naturalised (Fraga & Pallicer, 

1998). Because low-altitude areas are both fairly rich 

in threatened endemic taxa and exotic taxa there is a 

need to tackle conservation priorities in these habitats 

and to reduce main threats which are from more to 

less important: tourism, fire, overgrazing, urbanisation 

and cropping (Medail & Quezel, 1997). 

Conservation priorities should be strongly en 

forced to those taxa that are evolutionary unique 

(Williams et al., 1994). In the Balearic Islands, relict 

taxa (paleoendemics) should be conserved because 

they are those that have gone through major distur 

bances and environmental changes. Paleoendemics 

include species represented by monospecific genera 

(Naufraga balearica), morphologically isolated spe 

cies (Pimpinella bicknellii, Daphne rodriguezii), and 

those without clear affinities (Helichrysum ambiguum, 

Hypericum balearicum, Paeonia cambessedesii) some 

of which are located in coastal habitats also invaded 

by exotic species. 

More studies like this one should be conducted at 

other regions in order to have a global assessment of 

the diversity and distribution of endemic and exotic 

taxa. Next research step should focus on the experi 

mental study of the ecological mechanisms that con 

trol the establishment of plants of conservation 

interest (Mack, 1996), such as the studies undertaken 

with Cyclamen balearicum (Affre et al., 1995) or the 

ones currently analysed by Traveset et al. (unp. data) 

on endemics, in order to have ecological bases for 

conservation efforts as for example the Ligusticum 

huteri Porta reintroduction program (Vicens, 1998). 


We thank F. Lloret, A. Traveset, E. Weber, P. Py 

sek and D. Jeanmonod for their valuable comments on 

160 


earlier drafts of the manuscript. Partial financial sup 

port was provided by the Comissi6 Interdepartamental 

de Ciencia i Tecnologia (CIRIT) de la Generalitat de 

Catalunya. 

REFERENCES 

Affre L., Thompson J. D. & Debussche M. 1995. The reproductive 

biology of the Mediterranean endemic Cyclamen 

balearicum Willk. (Primulaceae). Bot. J. Linnean Soc.. 

118: 309-330. 

Atkinson I. A. E. & Cameron E. W. 1993. Human influence 

on the terrestrial biota and biotic communities of New 

Zealand. Trends Ecol. Evol., 8: 447-451. 

Bolos 0., Vigo J., Masalles R.M. & Ninot J.M. 1993. Flora 

manual dels Pai"sos Catalans. Ed. Portic, Barcelona, 

1247 p. 

Bonafe F. 1980. Flora de Mallorca. 4 vol Ed. Moll, Palma. 

Cardona M. A. & Contandriopoulos J. 1979. Endemism and 

evolution in the islands of the western mediterranean. 

In: Bramwell D. (ed.), Plants and Islands. Academic 

Press, London: 133-169. 

Carlquist S. 1965. Island Biology: a natural history of the 

islands ofthe world. Nat. Hist. Press, Garden City. 

Cody M. L. 1986. Diversity, rarity, and conservation in 

mediterranean-climate regions. In: Soule M. E. (ed.), 

Conservation biology. The science ofscarcity and diversity. 

Sinauer Ass., Massachusetts: 122-152. 

Cody M. L. & McCoverton J. 1996. Short-term evolution of 

reduced dispersal in island plant populations. 1. Ecol., 

84: 53-61. 

Contandriopoulos J. & Cardona M. A. 1984. Caractere 

original de la flore endemique des Baleares. Botanica 

Helvetica, 94: 101-131. 

Cowling RM., Rundel P.W., Lamont B.B., Arroyo M.K. & 

Arianoutsou M. 1996. Plant diversity in mediterraneanclimate 

regions. Trends Ecol. Evol., I I: 362-366. 

Cowling R. M. & Hilton-Taylor C. 1997. Phytogeography, 

flora and endemism. In: Cowling R. M., Richardson D. 

M. & Pierse S. M. (eds.). Vegetation ofSouthern Africa, 

Cambridge University Press, Cambridge: 43-61. 

Cowling R. M. & Samways M. J. 1995. Predicting global 

patterns of endemic plant species richness. Biodiv. Letters, 

2: 127-131. 

Daehler C. C. 1998. The taxonomic distribution of invasive 

angiosperm plants: ecological insights and comparison 

to agricultural weeds. Bioi. Conserv., 84: 167-180. 

D'Antonio C. M. & Dudley T. L. 1995. Biological invasions 

as agents of change on islands versus mainlands. In: 

Vitousek P. M., Loope L. L. & Adsersen H. (eds.). Islands: 

biological diversity and ecosystem jitllction. 

Springer-Verlag, Berlin: 103-121. 

De Waal L. c., Child L. E., Wade P. M., Brock J. H. 1994. 

Ecology and management of invasive riverside plants. 

John Wiley & Sons, Chichester. 

Di Castri, F. 1989. History of biological invasions with special 

emphasis on the Old World. In: Drake J., Mooney 

H.A., Di Castri F. et al. (eds.). Biological invasions: a 

global perspective. John Wiley & Sons, NewYork: 1-30. 

Drake J.A., Mooney RA., Di Castri F., Groves K.H., Kruger 

F. S., Rejmanek M. & Williamson M. 1989. Bio- 


Vila & Mufioz 

logical invasions. a global perspective. Scope 37. John 

Wiley & Sons, New York. 

Eliasson, U. 1995. Patterns of diversity in island plants. In: 

Vitousek P. M., Loope L. L. & Adsersen H. (eds). Islands: 

biological diversity and ecosystem function 

Springer-Verlag, Berlin: 35-50. 

Fraga P. & Pallicer X. 1998. Notes florfstiques de Menorca. 

Butlletf de la Instituci6 Catalana d'Historia Natural, 66: 

35-40. 

Gaston K. J. 1994. Rarity. Chapman & Hall, London. 205p. 

G6mez-Campo L., Bermudez-De- Castro L., Casiga M. J., 

Sanchez-Yelamo M. D. 1984. Endemism in the Iberian 

Peninsula and Balearic Islands. Webbia, 38: 709-714. 

Groves R. H. & Di Castri F. 1991. Biogeography of Mediterranean 

invasions. Cambridge University Press, Cambridge. 

485 p. 

Hobbs R. J. & Huenneke L. F. 1992. Disturbance, diversity 

and invasion: implicatiosn for conservation. Conserv. 

Bioi., 6: 324-337. 

Houston D. B. & Schreiner E. G. 1994. Alien species in National 

Parks: drawing lines in space and time. Conserv. 

Bioi., 9: 204-209 

Jeanmonod D. 1998. Les plantes introduites en Corse: impact, 

menaces et propositions de protection de la flore 

indigene. Biocosme Mesogeen, 15: 45-68. 

Kruckeberg A. R. & Rabinowitz D. 1985. Biological aspects 

of endemism in higher plants. Ann. Rev. Ecol. Syst., 16: 

447-479. 

Lean G. & Hinrichsen D. 1990. Atlas of the environment. 

Arrow Books, London. 192 p. 

Lonsdale W. M. 1999. Global patterns of plant invasions 

and the concept of invasibility. Ecology, 80: 1522-1536. 

Loope L. L., Mueller-Dombois D. 1989. Characteristics of 

invaded islands, with special reference to Hawaii. In: 

Drake J. A., Mooney H. A., Di Castri F., Groves K. H., 

Kruger F. S., Rejmanek M. & Williamson M. (eds.), 

Biological Invasions. A Global Perspective. Scope 37, 

John Wiley & Sons, New York: 257-281. 

Mack R. N. 1996. Predicting the identity and fate of plant 

invaders: emergent and emerging approaches. Bioi. 

Conserv.,78: 107-121. 

McDonald 1. A. W. & Cooper J. 1995. Insular lessons for 

global biodiversity conservation with particular reference 

to alien invasions. In: Vitousek P. M., Loope L. L. 

& Adsersen H. (eds.). Islands: biological diversity and 

ecosystem function. Springer-Verlag, Berlin: pp. 189 

204. 

McIntyre S. 1992. Risks associated with the setting of conservation 

priorities from rare plant species lists. BioI. 

Conserv., 60: 31-37. 

Medail F. & Quezel P. 1997. Hot-spots analysis for conservation 

of plant biodiversity in the Mediterranean basin. 

Ann. Missouri Bot. Gard., 84: 112-127. 

Medail F. & Verlaque R. 1997. Ecological characteristics 

and rarity of endemic plants from Southeast France and 

Corsica: implications for biodiversity conservation. Bioi. 

Conserv., 80: 269-281. 



Mus M. & Mayol J. 1993. Plans de Conservaci6 dels vegetals 

amenaf;ats ales Balears. Documents tecnics de 

Conservaci6. Govern Balear. Direcci6 General 

d'Estructures Agr. i Medi Natural. Palma de Mallorca. 

PANDION 1997. Memoria del projecte de protecci6 dels 

Limonium de Punta Carregadors i control de poblacions 

de Carpobrotus edulis. Govern Balear. Conselleria 

de Medi Ambient, Palma de Mallorca. 

Pia V., Sastre B. & Llorens Ll. 1992. Aproximaci6 al 

cataleg de laflora vascular de les Illes Balears. Universitat 

de les Illes Balears. Jardf Botanic de S6ller 

(MBCN), Palma de Mallorca. 

Prance G. T. & Elias T. S. 1977. Extinction isforever. New 

York Botanical Garden, New York. 

Pysek P. 1998. Is there a taxonomic pattern to plant invasions? 

Oikos, 82: 282-294. 

Rabinowitz D., Cairns S. & Dillon T. 1986. Seven forms of 

rarity and their frequency in the flora of the British isles. 

In: Soule M. E. Conservation biology. The science of 

scarcity and diversity. Sinauer Ass., Massachussets: 

182-205. 

Raunkiaer C. 1934. The life-forms of plants and statistical 

plant geography. Clarendon Press, Oxford. 

Rejmanek M. & Randall J. M. 1994. Invasive alien plants in 

California: 1993 summary and comparison with other 

areas in North America. Madrofio, 41: 161-177. 

Romo A. 1994. Flores si/vestres de las Baleares. Ed. Rueda, 

Madrid. 

Schierenbeck K. A. 1995. The threat to the California flora 

from invasive species; problems and possible solutions. 

Madrofio,42: 168-174. 

Schiffman P. M. 1997. Wing reduction in island Coreopsis 

gigantea achenes. Madrm!o, 44: 394-396. 

Schwartz M.W. 1993. The search for pattern among rare 

plants: are primitive species more likely to be rare? BioI. 

Conserv.,64: 121-127. 

Sim6n J. C. 1994. La flora vascular espafiola: diversidad y 

conservaci6n. Ecologfa (ICONA), 8: 203-225. 

Solbrig O. T. 1994. The complex structure of the taxonomic 

system. In: Peng C. 1. & Chou C. H. (eds.). Biodiversity 

and Terrestrial Ecosystems. Institute of Botany, Academia 

Sinica Monograph Series, 14, Taipei: 7-14. 

Usher M. B. 1986. Invasibi1ity and wildlife conservation: 

invasive species on nature reserves. Phi/os. Trans. R. 

Soc. Lond., B 314: 695-710. 

Vicens M. 1998. PIa de recuperaci6 de Ligusticum huteri 

Porta. Govern Ba1ear. Conselleria de Medi Ambient, 

Ordenaci6 del Territori i Litoral, Palma de Mallorca. 

Vitousek P. M., D'Antonio C. M., Loope L. L., Rejmanek 

M. & Westbrooks R. 1997. Introduced species: a significant 

component of human-caused global change. 

New Zealand J. Ecol., 21: 1-16. 

Weber E. 1997. The alien flora of Europe: a taxonomic and 

biogeographic review. J. Veg. Sci., 8: 565-572. 

Williams P. H., Vane-Wright R. 1. & Humphries C. J. 1994. 

Measuring biodiversity for choosing conservation areas. 

In: Lasalle J. & Gau1d 1. D. (eds.). Hymenoptera and 

biodiversity. CAB Internat. Wallingford : 309-328. 

161

Aidoud et al. Changements edaphiques le long d'un gradient d'intensif1.5 de paturage dans une steppe d'Algerie 


The aim of the present paper is to emphasise the effects 

of overgrazing on soil through a set of characters. The study 

was conducted in an arid steppe (mean annual rainfall: 260 

mm). The study area is located in the western part of the 

steppic high-plains of Algeria. This region has been subject 

to land degradation due to overgrazing which is one of the 

main causes of desertification in arid rangelands. During the 

two last decades, vegetation has been marked by important 

changes that have been pointed out through a long-term 

monitoring of phytomass, primary net productivity and specific 

composition. 

Steppic landscape is dominated by pediment surfaces 

gently sloped «2%) subject to erosion by wind. Soils are 

shallow over a limestone crust at 15 to 30 cm depth. The 

study was undertaken using a grazing gradient method. The 

chosen transect includes three levels of grazing intensity: 

(I) a non-grazed exclosure PO of the pristine system that had 

been, 20 years ago, a typical steppe with an alfa-grass pure 

stand under homogeneous environmental conditions; (2) a 

field PI of about 500 ha under controlled grazing with 0.25 

unit/ha stocking rate; (3) the rest of the steppic open rangelands 

P3 in which the stocking rate has rapidly increased 

reaching 0.50-0.75 unit/ha in 1990. Perennial plant cover, 

sand cover, soil organic matter and texture are the main indicators 

used to assess soil degradation. The main results are 

the following: 

(I) at the transect scale: 

- soil features within the non-grazed exclosure (PO) remain 

similar to those of the pre-existing system. The dominant 

INTRODUCTION 

La desertification peut etre definie comme une de 

gradation des terres sous climat aride (s.l.). Cette defi 

nition simple, adoptee par la Conference des Nations 

Unies pour l'Environnement et le Developpement en 

1992, cache en fait une grande complexite qui se reflete 

a travers le debat actuel autour de ce concept 

(Hellden, 1988; Thomas & Middleton, 1993; Dodd, 

1994; Hutchinson, 1996; Thomas, 1997). Dans ce 

debat, le caractere reversible ou irreversible des chan 

gements demeure une question centrale (Mainguet, 

1994, 1998). Un indicateur incontestable 

d'irreversibilite, considere comme ultime changement, 

est la degradation du sol (Friedel, 1992 ; Floret et al., 

1992 ; Mainguet, 1994 ; Milton et al., 1994). Les sols 

sous climat aride sont pauvres en matiere organique 

(Evenari, 1985) ce qui confere au systeme une faible 

resilience vis-a-vis des changements naturels ou in 

duits par l'homme (Albaladejo et al., 1998). Aussi, les 

dommages causes dans ces milieux sont-ils difficiles a 

reparer (Milton et aI., 1994). 

Sous climat aride, la desertification agit par stades 

successifs (Milton et al., 1994). L'un des premiers 

164 

perennial Stipa tenacissima covers 36 ± 5%, the sand layer 

is almost completely lacking, soil organic matter rate is 1.7 ± 

0.3% and clay and fine-silt rate is 29 ± 5%; 

- on heavily grazed area (P2), the difference with PO is 

highly significant (p

Aidoud et al. Changements edaphiques le long d'un gradient d'intensite de paturage dans une steppe d'Algerie 

homogene au depart, ont montre une dynamique di 

rectionnelle regressive de la vegetation, dont la cause 

principale est le surpaturage par les ovins (Aidoud, 

1989, 1994). Des changements ont ete mis en evi 

dence, le long d'un transect d'intensite de paturage, 

par la baisse de la biomasse de l'espece perenne dominante 

(Aidoud & Touffet, 1996) et par des change 

ments de composition specifique de la vegetation ainsi 

que des caracteres de surface du sol (Aidoud, 1994; 

Slimani, 1998). Le transect utilise represente un gra 

dient comportant trois etats de paturage : nul, contr61e 

et libre. II constitue une situation experimentale in 

natura permettant d'approcher des changements dans 

le temps par la transposition de variations analysees 

dans l'espace (Pickup, 1992). 

Le present travail, qui s'integre a cette approche, a 

pour objectif de verifier, le long de ce meme transect, 

en se basant sur les principaux caracteres du sol sous 

climat aride, l'hypothese d'un changement edaphique 

significatif. 

CADRE D'ETUDE ET METHODOLOGIE 

Le plateau de Rogassa, couvrant pres de 30 000 ha, 

est situe a une trentaine de km au nord de la ville d'EI 

Bayadh. Les principales caracteristiques ecologiques 

sont donnees dans le tableau I. Les donnees meteoro 

iogiques placent le site dans l'etage mediterraneen 

aride moyen a hiver froid au sens d'Emberger. Jus 

qu' au debut des annees 1980, le paysage vegetal etait 

une nappe alfatiere de plaine consideree parmi les plus 

denses et les plus homogenes d' AIgerie et ou Stipa 

tenacissima L. assurait pres de 95% du couvert vege 

tal global. La vegetation du plateau de Rogassa a ete 

integree aux groupements d'alfa purs selon l'analyse 

realisee dans le bassin du Chott Chergui du Sud 

Oranais (Aidoud-Lounis, 1989). Exploitee en paturage 

colIectif et libre, cette steppe supportait, dans son etat 

preexistant, une charge pastorale moyenne de 0,25 

unite ovine par hectare. La charge a augmente tres ra 

pidement atteignant 0,5 a 0,75 uo/ha en 1990 (Aidoud 

& Touffet, 1996). 

Le terme « plateau» est attribue au site etudie en 

raison de la topographie homogene qu'imprime le gla 

cis d'erosion sub-horizontal ou legerement ondule 

(versant-glacis) du quaternaire ancien (Pouget, 1980). 

Situee a une profondeur de 20 a 30 cm, la croOte cal 

caire zonaire, feuilIetee et tres dure en surface, est ca 

racteristique de ce type de glacis (RueIlan, 1971). A la 


surface du sol, dominent les elements grossiers (gra 

viers) et la pellicule de gla


et 4,22 pour I' axe 2), Cette ordination, conforme au 

gradient de paturage, souligne I'importance de ce fac 

teur sur les caracteres traites. 

Le relcve situe a l'extremite negative de l'axe 2 se 

detache nettement du groupe P2. Les donnees de ce 

releve, non representatif car localise dans une depres 

sion (daya), n'ont pas ete integrees aux analyses ci 

apres. 

Les caracteres du couvert vegetal et du sol, le long 

du transect, sont resumes dans le tableau 2. Les ca 

racteres de surface qui varient le plus significative 

ment sont le couvert vegetal et le voile sableux. Le 

couvert dcs plantes perennes diminue de 36 a 16 %, 

soit plus de la moitie, entre la parcelle mise en defens 

(PO) et le terrain surpature (P2). La diminution du 

couvert vegetal global est de 40 % mais a cependant 

une valeur indicatrice moindre en raison de la plus 

grande fluctuation des especes annuelles. Dans 

l'espace intcrstitiel ou sol nu, le voile sableux, prati 

quement absent dans PO, atteint, dans P2, un taux de 

couverture de pres de 50 % et une epaisseur de pres de 

8 cm en moyenne. 

Ces changements edaphiques le long du transect se 

traduisent, correlativement et tres nettement, par une 

baisse des argiles et Iimons fins et du taux de matiere 

organique. Les variations des caracteres edaphiques, 

mesurees sous la touffe et dans le sol interstitiel, sont 

significatives. La variation est cependant nettement 

plus importante en considerant la couche de surface. 

La baisse globale du taux d'argiles et limons fins 

est de 43 % entre PO et PI. Elle est de 87 % dans la 

couche superficielle du sol sous la touffe et de 84 % 

entre lcs touffes. La diminution, moindre dans la 

deuxieme couche (47 et 28 %), reste cependant signi 

ficative, Exprimee selon les classes definies par Pou 

get (1980), la texture passe de «grossiere» a « tres 

grossiere » sous la touffe et de « moyenne» a « tres 

grossiere » dans l'espace entre les touffes. 

Les valeurs obtenues dans P I (sous paturage 

controle) sont a la fois intermediaires et significative 

ment differentes de celles correspondant a PO et P2. 

Elles demcurent cependant, comme illustre par la fi 

gure I, plus proches de la situation de mise en defens. 

La baisse de matiere organique dans le sol est glo 

balement dc 38%. Elle suit a peu pres la meme evolu 

tion que celle des argiles et limons fins, mais dans des 

proportions legerement inferieures. Cette diminution 

est, respectivement pour le sol sous la touffe et le sol 

168 

interstitiel, de 63 et 59 % pour la couche superieure et 

de 25 et 35 % en profondeur. 

A l'echelle stationnelle, c'est dans la zone surpatu 

ree que les differences sont hautement significatives 

(p


(1982), l'isohumisme qui est souvent reconnu dans ce 

type de sol ne serait qu'apparent. En effet, la decom 

position relativement active de la litiere dans cette 

steppe (Bessah, 1998), suggere un apport organique 

essentiellement lie aux parties aeriennes de l'alfa. La 

faible contribution a la biomasse tota1e des organes 

souterrains (Aidoud, 1989), en accord avec la synthese 

dressee par Barbour (1981) pour les zones arides en 

general, peut conforter cette hypothese. Ceci doit ce 

pendant etre relativise, car les processus fonctionnels 

en cause doivent etre consideres plus en terme de turn 

over de la matiere organique qu'en terme de biomasse. 

Les changements edaphiques observes le long du 

gradient peuvent s'expliquer par la reduction du cou 

vert de l'alfa. La deterioration d'attributs vitaux eda 

phiques (Aronson et al., 1995) du systeme, tels que la 

texture ou la matiere organique, a pu etre evaluee ainsi 

apres une periode relativement courte de moins de dix 

ans. Cette rapidite est confirmee en Espagne, par Al 

baladejo et al. (1998) qui montrent qu'apres environ 5 

ans la reduction de la couverture vegetale peut se re 

percuter significativement sur les caracteristiques du 

sol par la baisse de la matiere organique et la deterioration 

des proprietes physiques. 

AI'echelle stationnelle, la vegetation steppique est 

constituee de touffes separees par un espace interstitiel 

a couverture vegetale tres faible. Dans la parcelle 

protegee PO, les resultats montrent une difference si 

gnificative entre touffe et espace interstitiel pour les 

caracteres edaphiques, en particulier ceux de la couche 

de surface. lis confirment ceux obtenus notamment 

par Puigdefabregas & Sanchez (1996), Cerda (1997) 

et Domingo et al. (1998), montrant le role que la 

touffe d'alfa joue, dans le systeme, par ses aptitudes a 

intercepter et a retenir l'eau et les sediments. La rela 

tion entre la baisse de vigueur des touffes d'alfa et la 

sensibilite a l'erosion a ete montree par simulation 

(Sanchez & Puigdefabregas, 1994). Ajoutees ases ca 

pacites adaptatives a l'aridite (Nedjraoui, 1990; Ai 

doud, 1992 ; Pugnaire & Haase, 1996), ces proprietes 

font de l'alfa une « espece clef de voute » du systeme, 

au sens de Aronson et al. (1995). A l'oppose, dans la 

zone surpaturee P2, le deperissement des touffes en 

trai'ne la destruction de leurs proprietes edaphiques et 

en particulier du role protecteur contre l'erosion. La 

difference plus faible entre la touffe et l'espace inters 

titiel dans la zone P2 reflete la destruction de la structure 

du systeme, tendance qui semble se refleter a tra 

vers la composition de la vegetation (Aidoud, 1994). 


Dans le Sud-Oranais, cet etat correspond acelui de 

l'ensemble des anciennes steppes d'alfa de plaine. Le 

niveau de degradation atteint constitue un stade irre 

versible qui peut indiquer d'un point de vue edaphi 

que, un seuil fondamental entre paturage et surpatu 

rage. Notons que 1'« irreversibilite » est employee ici 

dans la sens Oll la restauration s.s. (Aronson et al., 

1995) du systeme preexistant est devenue tres peu 

probable, voire impossible. Dans la logique pastorale 

actuelle, la zone PI sous paturage controle, malgre les 

changements edaphiques observes, constitue le sys 

teme de reference presentant les normes a atteindre 

par rehabilitation (Aronson et al., 1995). 

CONCLUSION 

Les changements edaphiques, analyses atravers le 

taux de matiere organique et la texture dans les steppes 

d' alfa de plaine, sont significatifs. lis montrent 

1'importance des phenomenes de desertification dans 

ce type de steppes qui semblent les plus affectees en 

Algerie. Les resultats obtenus soulignent tout l'interet 

de l'approche utilisant le gradient d'intensite de patu 

rage, partant d'une situation de steppe soustraite au 

paturage durant une vingtaine d'annees, acelle Oll la 

steppe d'origine a pratiquement disparu par surpatu 

rage. La disparition des principales fonctions edaphi 

ques des touffes d'alfa suite a leur deperissement 

confirment les changements observes sur la biomasse 

(Aidoud & Touffet, 1996) ainsi que sur la diversite et 

la composition floristique (Aidoud, 1994; Slimani, 

1998). Les changements edaphiques restent cependant 

moins importants en profondeur qu'en surface, tradui 

sant ainsi une certaine inertie du systeme. Malgre 

l'intensite de degradation, les attributs edaphiques, 

dans le pire des cas, n'ont pas atteint les niveaux ob 

serves dans un desert et une rehabilitation demeure 

donc possible. Celle-ci reste cependant tributaire de 

l'efficacite et de la rapidite des mesures de gestion du 

rable des paturages qui doivent etre prises adifferents 

niveaux de decision. 

Remerciements 

Nous remercions D-Y. Alexandre, I-B. Bouzille, 

B. Clement et F. Roze du Service d'Ecologie (Univ. 

Rennes 1) pour leurs corrections et conseils, ainsi que 

tous les collegues chercheurs et techniciens de 1'Unite 

de Recherche sur les Ressources Biologiques Terres- 

169


Nedjraoui D., 1990. Adaptation de ['alfa aux conditions stationnelles: 

contribution a I 'etude du jCJ/lctionnement de 

l'ecosyste,ne steppique. These de Doctorat d'Etat. Universite 

des Sciences et de la Technologie Houari Boumediene, 

Alger. 256 p. 

Pickup G., 1992. Dynamics of rangeland ecosystems. In: 

Gaston A., Kernick M. & Le Houerou H.N. (eds.), Actes 

du 4eme Congres International des Terres de Parcours, 

Montpellier, 22-26 avril 1991 : 1066- 1069. 

Pouget M., 1980. Les relations sol-vegetation dans les steppes 

Sud-algeroises. Trav. Doc. ORSTOM. 116: 1-555. 

Pugnaire F.T. & Haase P., 1996. Comparative physiology 

and growth of two perenial tussock grass species in a 

semi-arid environment. An/l. Bot., 77 : 81-86. 

Puigdefabregas J. & Sanchez G., 1996. Geomorphological 

implications of vegetation patchiness on semi-aride slopes. 

In : Anderson M.G. & Books S.M. (eds.), Advances 

in Hillslope processes, 2 : 1028-1060. 

Rognon P., 1993. La desertification : un risque majeur ? Secheresse,4 

: 74-74. 

Ruellan A., 1971. Contribution a la connaissance des sols 

des regions mediterraneennes. Les sols a profil calcaire 


differencie des plaines de la basse Moulouva (Maroc 

oriental). ORSTOM Editions, Paris. 302 p. 

Sanchez G. & Puigdefabregas J., 1994. Interactions of plant 

growth and sediment movement on slopes in a semi-arid 

environment. Geomorphology, 9 : 243-260. 

Slimani, H. 1998. La desertijication de la steppe d'alfa (Stipa 

tenacissima L.) des Hautes Plaines occidentales 

d'Algerie : ejfets du paturage sur la vegetation et le sol. 

These de Magister, Univ. Sci. Technol. Houari Boumediene, 

AIger. 123 p. 

Thomas D.F.G. & Middleton N.J., 1993. Desertilreation: 

exploding the myth. John Wiley & Sons, Chichester. 

194 p. 

Thomas D.S.G., 1997. Science and the desertification debate. 

J. Arid Environ., 37 : 599-608. 

UNEP, 1992. World Atlas ofdesertijrcation. UNEP, Nairobi 

& Edward Arnold, London. 

Verstraet M.M. & Schwartz S.A., 1991. Desertification and 

global change. In : Henderson-Sellers A. & Pitman AJ. 

(eds), Vegetation and climate interactions in semi-arid 

regions. Kluwer Academic Publishers, Belgium: 3-13. 

171

Abd El-Ghani Soil variables affecting the vegetation ofinland western desert ofEgypt 

INTRODUCTION 

Desert vegetation in Egypt is by far the most important 

and characteristic type of the natural plant life. 

It covers vast areas and is formed mainly of xero 

phytic shrubs and subshrubs. Monod (1954) recog 

nised two types of desert vegetation, namely 

contracted and diffuse. Both types refer to permanent 

vegetation which can be accompanied by ephemeral 

(or annual) plant growth depending on the amount of 

precipitation in a given year. Kassas (1966, 1971) 

added a third type "accidental vegetation" where pre 

cipitation is so low and falls so irregularly that no 

permanent vegetation exists. It occurs mainly as con 

tracted patches in runnels, shallow depressions, hol 

lows, wadis and on old dunes with coarse sand. 

Accidental vegetation consists of species which are 

able to perform an annual life cycle: potential annuals 

(sensu Haines, 1951), or potential perennials (sensu 

Bornkamm, 1987), but can likewise continue growing 

as long as water persists in the soil. Thomas (1988) 

identified these plants as those with episodic growth 

strategies linked to immediate water availability. Re 

cently, Springuel (1997) classified the accidental 

vegetation in south eastern Egypt into three groups: 

run-off dependent vegetation in the main wadi chan 

nels, run-on dependent vegetation of playa formation, 

and rain dependent vegetation on levelled plains of 

sand sheets. 

In a survey of the vegetation units in the Western 

Desert of Egypt, outside the Oases, Bornkamm and 

Kehl (1990) distinguished five desert zones along a 

precipitation gradient. Besides the well known 

semidesert and full desert zones in the very north, 

three zones of extreme desert show a significant differentiation 

(Figure I). Both extreme desert zones III 

and IV support the growth of accidental vegetation, 

where the precipitation in the former amounts to 5-10 

(20) mm year I whereas in the latter is 1-5 mm year-I. 

On the other hand, extreme desert V in the very south 

is practically void of vegetation where precipitation is 

proposed to be less than 1 mm year-I. Typical acci 

dental vegetation types in the Western Desert of Egypt 

are: Zygophyllum coccineum with Salsola imbricata 

subsp. imbricata, Stipagrostis acutiflora with Zilla 

spinosa as well as stands of Salsola imbricata subsp. 

imbricata with Fagonia arabica. However, ground 

water-dependent vegetation in all the three extreme 

desert zones exists too: Zahran (1972) and Abd EI- 

174 

Ghani (1981, 1985) in large oases (Siwa, Bahariya and 

Farafra), and in small oases and depressions (Bir Saf 

saf, Bir EI-Shab, Bir Tarfawi and Qara): EI-Hadidi 

(l980b), Bornkamm (1986), Abd EI-Ghani (1992). 

Although our knowledge of the growth of acci 

dental vegetation in Egypt has considerably increased 

during the last two decades (Alaily et aI., 1987; 

Bornkamm, 1987; Springuel et aI., 1990), much less is 

known about this vegetation in quantitative terms. The 

present study aims at describing the floristic composi 

tion of the accidental type of vegetation growing in 

parts of the Western Desert of Egypt and analysing the 

distribution of species in relation to certain environ 

mental factors by applying the muItivariate analysis 

techniques. 

STUDY AREA 

The present study has been conducted in two con 

secutive extreme desert zones (sensu Bornkamm & 

Kehl, 1990), where the accidental type of vegetation 

exists. Data is from two transects: the northern one 

extends for a distance of about 150 km; half-way 

along Siwa Oasis-Mersa Matruh road, and represents 

the extreme desert zone III (Figure 1). This transect is 

principally located in the inland part of the Middle 

Miocene plateau that rises to about 100 m above the 

depression floor (reaches 20 m below sea level). The 

southern transect extends for a distance of about 140 

km, along the Dakhla-Farafra road and represents the 

extreme desert zone IV. It is located in the middle 

limestone plateau (500 m above sea level). The north 

ern transect lies in the Libyan Desert while the southern 

one in the Nubian Desert (EI-Hadidi, 1980a). In 

general, the landscape of the northern transect is a part 

of the Central Sahara, whereas the southern transect is 

a part of the Southern Sahara (Schiffers, 1971). 

According to WaIter and Breckle (1984) the study 

area lies in the zone of subtropical arid deserts. The 

temperature regime is characterised by mild winters 

and very hot summers. Whereas average January tem 

perature remains rather constant between 12°C and 

14°C, the July mean rises to approximately 31°C. The 

absolute maxima of the southern region of the study 

area may reach 49°C. Precipitation is erratic, variable 

and unpredictable with frequent long dry periods. Zahran 

and Willis (1992) reported that the mean annual 

rainfall ranges from 9.6 mm year -1 in Siwa Oasis (the 

nearest station to the northern transect) to nearly I mm 



I 

5 EA 

Figure I. Map showing the five vegetation zones of the 

Western Desert of Egypt (after Bornkamm & Kehl, 1990), 

indicating the position of the two studied transects: 

T n = Northern transect, T s = Southern transect 

year-I in Dakhla Oasis (the nearest to the southern 

transect). The climatic gradient along the N-S direc 

tion in the study area is obvious (Table 1). The Pluviothermic 

Quotient (Emberger, 1955) for Siwa Oasis is 

about 1.43, while that of Dakhla Oasis is nearly zero 

indicating extreme aridity. 

METHODS 

The phytosociological survey of the study area was 

carried out during several visits in 1986-88 and 1995 

96. Stands were randomly chosen at locations where 

either dense vegetation or change in species composition 

was encountered. A total of 144 stands (each of a 

size of 25 x 25 m) were sampled: 83 in the northern 

transect and 61 in the southern one. In each stand, 

density (individuals/lOO m 2 ) and frequency (occur 

rences/lOO quadrats) of the present species were esti 

mated using fifty I x 2 m 2 randomly located quadrats. 

Plant cover (m/100 m) was determined by the line 

intercept method (Canfield, 1941), using five parallel 

lines distributed randomly across the stand. Relative 

importance value (IV) for each species in each stand 


was calculated by the sum of its relative density, fre 

quency and cover (it has a maximum value out of 

300). Voucher specimens of each species were col 

lected, identified and deposited in the Herbarium of 

Cairo University (CAI). Taxonomic nomenclature is 

according to Tackholm (1974), and updated following 

Boulos (1995, 1999). 

For each sampled stand, four soil samples were 

collected from profiles of 0-50 cm and pooled together 

to form one composite sample. Soil texture was de 

termined with the hydrometer method and CaC03 by 

Collin's calcimeter. Organic matter content and soil 

moisture were estimated by drying and then ignition at 

600°C for 3 hours. Soil-water extracts (1: 2.5) were 

prepared for the determination of electric conductivity 

and pH using conductivity-meter and pH-meter, re 

spectively. 

Stand-species data matrix was classified using the 

importance values of perennial species by Two-way 

indicator species analysis (TWINSPAN; Hill, 1979) 

(Table 2). Due to the relative paucity of most stands 

(generally between 5 and 14 species), classification by 

TWINSPAN was stopped at the third level so that the 

size of stands would demonstrate ecological meaning 

though their floristic structure. All the default settings 

were used for TWINSPAN, except the pseudo-species 

cut levels were altered to 0, 2, 5, 10,20,40,50,60 and 

80, and the number of indicator species was four per 

class. 

Canonical correspondence analysis (CANOCO: ter 

Braak, 1988 & 1990) was used for the same set of 

data. Rare species were downweighted to reduce dis 

tortion of the analysis. Detrended Correspondence 

Analysis (DCA: Hill & Gauch, 1980) was applied to 

check the magnitude of change in species composition 

along the first ordination axis (i.e. gradient length in 

standard deviation units). Direct gradient analysis, 

Canonical Correspondence Analysis (CCA) were used 

in order to examine the relationships between the floristic 

composition of the sampled stands and the esti 

mated soil variables. An exploratory CCA was 

performed using all studied edaphic variables, fol 

lowed by a CCA with forward selection. The CCA 

with forward selection was evaluated by examining 

the canonical coefficients (significance assessed by 

approximate t-tests) and the intraset correlations (ter 

Braak, 1986). 

175


Station 

Mersa Matruh 

Siwa 

Dakhla 

Temperature (QC) 

Mean Min Mean Max 

12.4 24.7 

4.1 38.0 

5.8 39.3 

Relative Humidity (%) 

Mean Min Mean Max 

51 67 

42 61 

19 42 

Rainfall 

(mm annuar l ) 

144.0 

9.6 

0.7 

Evaporation 

(mm daft) 

8.3 

11.5 

18.4 

Table I. Climatic characteristics (average 1931-1978) of three stations distributed in the study area (after Zahran, 1972; Zahran & 

Willis, 1992) 

TWINSPAN group I 11 III IV V VI VII VIII 

Group size 7 16 45 19 5 23 24 5 

Prosopis farcta (P fr) 121 2 1 

Phoenix dactylifera (P df) 15 9 5 

Calotropis procera (C pr) 1 12 1 

Tamarix nilotica (T ni) 30 109 19 4 23 6 1 

Stipagrostis plumosa (S pi) 2 6 3 6 I 9 

Pulicaria incisa (P cr) 1 1 I 12 4 8 5 

Heliotropium digynum (H dg) 5 I 7 3 2 4 

Alhagi graecorum (A gr) 77 22 2 3 2 

Citrullus colocynthis (C cl) 1 2 2 I 

Fagonia bruguieri (F br) 1 9 3 

Zygophyllum album (Z al) 12 4 

Traganum nudatum (T nd) 1 2 

/mperata cylindrica (I cy) 5 1 

Nitraria retusa (N rt) 11 1 

Launaea nudicaulis (L nu) 5 2 

Hyoscyamus muticus (H mu) 1 I 

Zygophyllum coccineum (Z co) 8 69 14 5 4 5 

Salsola imbricata subsp. imbricata (S im) 18 62 10 

Cornulaca monacantha (C mo) 1 15 142 5 

Fagonia arabica (F ar) 1 I 71 5 6 1 

Pulicaria crispa (P cr) 36 4 2 5 

Monsonia nivea (M nv) 2 1 3 2 I 

Fagonia indica (F in) 4 

Sarcocorniafruticosa (S fr) 1 

Calligonum polygonoides subsp. comosum (Ccm) 1 3 1 

Zilla spinosa subsp. spinosa (Z sp) 10 2 9 4 2 

Acacia tortilis subsp. raddiana (A rd) 1 19 6 2 

Astragalus trigonus (A tr) 3 2 3 2 2 

Anabasis articulata (A ar) 3 13 14 1 15 

Atriplex leucoclada (A lu) 109 4 2 

Herniaria hemistemon (H he) I 

Deverra tortuosa (D tr) 14 39 86 

Randonia africana (R af) 107 26 19 

Carduncellus mareoticus (C mr) 2 14 

Farsetia aegyptia (F ag) 3 1 

Capparis spinosa var. aegyptia (C sp) 12 96 75 

Zilla spinosa subsp. biparmata (Z bi) 3 7 33 

Helianthemum lippii (H 1p) 1 1 6 

Salsola villosa (S vr) 4 2 

Ephedra alata (E a1) 3 

Table 2. Species composition of the 144 stands in the two transects, arranged in order of occurrence in the eight TWINSPAN 

groups. Entries in bold are characteristic species in each group. Species abbreviations displayed in Figures 2, 3 and 5 are given 

between parentheses. 

176 ecologia mediterranea 25 (2) - /999


Soil variable TWINSPAN groups F- ratio P 

I 11 III IV V VI VII VIII 

EC (mS cm-I) 0.5±0.3 2.8±0.7 0.8±0.9 0.9±0.7 2.2±0.2 0.7±0.7 0.5±0.3 1.1±0.9 19.0 0.0001 

pH 8.0±0.6 8.8±0.4 7.9±0.5 7.9±0.4 7.7±0.5 7.7±0.5 7.9±0.4 7.6±0.4 1.3 0.3 

CaC0 3 (%) 9.6±2.6 11.3±4.9 15.0±5.8 13.5±5.3 21.0±8.1 14.8±6.0 15.1±5.5 19.1±3.8 4.6 0.0001 

MC (%) 1.5±0.3 2.0±0.6 2.12±0.7 2.3±0.5 2.5±0.2 2.6±0.8 2.9±0.7 3.3±0.7 5.4 0.0001 

OM(%) 1.5±0.2 0.3±0.4 0.21±0.3 0.2±0.3 0.1±0.1 0.1±0.1 O.I±O.I 0.1±0.06 29.5 0.0001 

Sand (%) 93.5±1.0 93.3±1.8 94.5±4.1 92.4±2.0 91.2±0.9 91.5±1.0 88.4±16.9 90.4±0.9 12.5 0.0001 

Silt + clay (%) 6.5±1.4 6.7±1.8 5.5±2.1 7.6±2.0 8.8±0.9 8.4±1.0 11.6±1.1 9.6±1.0 8.6 0.0001 

Table 3. Mean values, standard deviation errors and ANOVA F values of the soil variables in the stands supporting the 8 vegetation 

groups obtained by TWINSPAN. EC =electric conductivity, CaC0 3 =calcium carbonate, MC =moisture content and 

OM = organic matter 

Eigenvalues 

DCA 

CCA 

Species-environment 

correlation coefficients 

DCA 

CCA 

1 

0.830 

0.474 

0.724 

0_781 

Axis 

2 

0.496 

0.376 

0.468 

0.871 

3 

0.377 

0.292 

0.409 

0.658 

Table 4. Comparison of the results of ordination by DCA and CCA: eigenvalues and species-environment correlation coefficients 

for the first three axes are demonstrated 

Canonical coefficient Intra-set correlations 

Soil variables Axis I Axis 2 Axis 3 Axis I Axis 2 Axis 3 

EC -0.48* 0.88* 0.01 -0.50 0.85 0.35 

pH -0.13 -0.16* -0.10 -0.22 0.07 0.06 

CaC03 -0.08 0.02 -0.21 0.06 -0.08 -0.37 

MC 0.28* 0.09 0.10 0.49 -0.10 0.24 

OM -0.33* -0.19* 0.88* -0.41 -0.28 0.85 

Sand 0.12 -0.06 0.01 -0.09 -0.07 0.005 

Silt + clay 0.53* 0.29* 0.25 0.67 0.53 0.27 

Table 5. Canonical coefficients and the intra-set correlations of soil variables with the first three axes of CCA 

* = t-values > 2.3 (only indicative of coefficient strength) 

EC =electric conductivity, CaC03 =calcium carbonate, MC =moisture content and OM =organic matter 

The exploratory CCA was evaluated using the intraset 

correlations, since the canonical coefficients 

were unstable due to inclusion of highly correlated 

variables. 

Eight soil variables were included: electric conductivity, 

pH, CaC03 , soil moisture, organic matter, 

sand, silt and clay. The significance of differences 

between the different vegetation groups, as to their 

edaphic variables, was tested by ANOVA. CCA axes 

were evaluated statistically by means of a Monte 


Carlo permutation test (ter Braak, 1988). All the statistical 

techniques were according to student SYSTAT 

software (STUSTATW 5: Berk, 1994). 

RESULTS 

Sixty plant species related to 19 families of the angiosperms 

and one of the gymnosperms were recorded 

in this study. They constituted 40 perennial and 20 annual 

species: Cruciferae and Chenopodiaceae (13.3% 

177

Abd EI-Ghani Soil variables affecting the vegetation ofinland western desert ofEgvpt 

each), Zygophyllaceae (11.7%), Caryophyllaceae, 

Compositae and Leguminosae (10.0% each), while the 

other 14 families share 31.7%. Chamaephytes are the 

most abundant life form and constituted 40.0% of the 

total flora of the study area, followed by therophytes 

(33.3%), hemicryptophytes (15.0%) and phanero 

phytes (11.7%). The most common perennials re 

corded were: Prosopis farcta, Tamarix nilotica, 

Fagonia arabica, Zygophyllum coccineum, Salsola 

imbricata subsp. imbicata, Cornulaca monacantha, 

Alhagi graecorum, Atriplex leucoclada, Randonia af 

ricana, Deverra tortuosa and Capparis spinosa var. 

aegyptia may be considered as leading dominants and 

characteristic species. Each of these species attains a 

maximum importance value (IV) of more than 140 

(out of 300 for all species in a stand), and a mean of 

more than 60. Common but less important perennials 

are Phoenix dactylifera, Pulicaria crispa, Anabasis 

articulata, Zilla spinosa subsp. spinosa, Stipagrostis 

plumosa and Pulicaria incisa. Common annuals in 

clude: Trigonella stellata, Zygophyllum simplex, Co 

tula cinerea, Eremobium aegyptiacum, Schouwia 

thebaica and Paronychia arabica subsp. arabica. 

The 144 stands were classified into eight vegeta 

tion groups according to TWINSPAN technique (Fig 

ure 2). The first level of the dendrogram separates all 

the stands into two main groups. The first group com 

prises 57 stands found mainly in the northern transect, 

and the second comprises 87 stands sampled from 

both transects. Table 3 summarises the mean values 

and the standard deviations of the measured soil vari 

ables in the eight groups derived from TWINSPAN. 

Generally, pH shows the least variation among 

groups. It can also be noted that whereas levels of 

lime and fine materials attain their highest values in 

the groups of the northern transect, the organic matter 

content reaches its highest levels in those of the south 

ern transect. 

The identified vegetation groups are named after 

the characteristic species as follows: Prosopis farcta 

Tamarix nilotica (lower part of the southern transect 

in the vicinity of the lowlands of Dakhla Oasis); 

Tamarix nilotica-Alhagi graecorum (southern tran 

sect, high salinity levels favour the growth of some 

halophytic species, e.g. Nitraria retusa and Zygo 

phyllum album); Zygophyllum coccineum-Salsola imbricata 

subsp. imbricata (in runnels and depressions 

of the southern transect); Cornulaca monacantha 

Fagonia arabica (larger catchment areas of the north- 

178 

ern transect, some species of this group do not pene 

trate into other groups of the southern transect); 

Atriplex leucoclada (lower part of the northern tran 

sect); Randonia africana-Deverra tortuosa (the silty 

runnels, and occupying a distance of about 20 km of 

the middle part of the northern transect); Deverra 

tortuosa-Capparis spinosa var. aegyptia (occupying a 

distance of about 30 km in the upper stretches of the 

middle part of the northern transect); Capparis spi 

nosa var. aegyptia-Zilla spinosa subsp. biparmata (in 

the low depressions between km 185 and km 198 

along the lower part of the northern transect). 

The four DCA axes explain 11.7%, 7.0%, 5.3% 

and 3.4% of the total variation in the species data, re 

spectively. This low percentage of variance explained 

by the axes is attributed to the many zero values in the 

vegetation data set. Table 4 shows that the eigenvalue 

for the first DCA axis was high indicating that it cap 

tured the greater proportion of the variation in species 

composition among stands, but the species 

environment correlation coefficients were low for 

DCA axes. 

DCA ordination of the perennial species (Figure 3) 

shows that species with high positive scores on axis I 

are found mainly in stands of the southern transect, 

and ordinated close to the right end-point: Prosopis 

farcta, Phoenix dactylifera, Nitraria retusa, Sarcocor 

nia fruticosa and Imperata cylindrica. The species po 

sitioned on the other end of this axis include: 

Capparis spinosa var. aegyptia (negative end), De 

verra tortuosa, Randonia africana, Ephedra alata and 

Helianthemum lippii. These and many other species 

are commonly found in the upper and middle parts of 

the northern transect. In the centre of axis I there are 

many species found throughout the investigated area 

with no preference to any geographical aspect (e.g. 

Zygophyllum coccineum, Tamarix nilotica, Alhagi 

graecorum, Fagonia arabica and Cornulaca monacantha). 

Along axis 2, species with high positive 

scores include: Atriplex leucoclada and Zilla spinosa 

subsp. biparmata. These species are recorded from the 

lower part of the northern transect, whereas species on 

the other end are of common occurrence in the middle 

and upper parts. To compare the classification and 

DCA ordination results, the recognised eight 

TWINSPAN groups are superimposed. Significance 

correlations of soil variables with the first four DCA 

axes (results not shown) revealed greater correlations 

along axis I than the higher order axes. 

ecalogia mediterranea 25 (2) - 1999


and salinity, while the second axis was defined by salinity 

and organic matter content. This fact is also evi 

dent in the ordination diagrams (Figures 4 & 5). 

Contribution of salinity, fine sediments, organic mat 

ter and moisture content, as indicated by the forward 

selection in the CCA program, to the variation in spe 

cies data were 29.4%, 29.0%, 22.1 % and 11.3%, re 

spectively. 

CCA ordination diagram for the species scores and 

canonical coefficients scores of the soil variables is 

presented in Figure 4. Three species grouping are evi 

dent. The first was highly associated with organic 

matter and sand, and includes species such as Proso 

pis farcta, Salsola imbricata subsp. imbricata, Fagonia 

bruguieri, Zygophyllum coccineum and Cornulaca 

monacantha. The corresponding TWINSPAN groups 

were I, III and IV (Figure 5). A second group associ 

ated with salinity and pH and includes: Tamarix ni 

lotica, Alhagi graecorum, Nitraria retusa, Atriplex 

leucoclada and Calotropis procera. Vegetation groups 

located here are II and V. The third group was closely 

associated with moisture content, fine materials and 

lime. The associated vegetation groups are VI, VII and 

VIII, and include: Deverra tortuosa, Randonia afri 

cana, Zilla spinosa subsp. biparmata and Atriplex leu 

coclada. 

DISCUSSION 

The vegetation groups which resulted from the ap 

plication of TWINSPAN in the study area, may be 

related in its northern transect to the Salsolion tetrandrae 

of habitats with soils derived from chalks and 

marls and rich in gypsum and soluble salts and the 

Anabasion articulatae arenarium, Hammada 

Anabasion articulatae (Zohary, 1973), and Thy 

melaeion hirsutae (Tadros & Atta, 1958) of the pro 

gressively less saline habitats. The associations 

belonging to these alliances and their characteristic 

species have repeatedly been recorded as abundant in 

ecological studies of specific habitats in the western 

Mediterranean coastal region of Egypt (Migahid et aI., 

1971), and in the north-western Negev, Israel (Tielb6rger, 

1997). TWINSPAN groups of the southern 

transect can be inferred to the alliance Zygophyllion 

coccinei (El-Sharkawy et al., 1984) with their charac 

teristic species are commonly recorded in the communities 

of the southern part of the Western Desert of 

Egypt (El-Hadidi, 1980b; Boulos, 1982). According to 

180 

the detailed phytosociological survey by Bornkamm 

and Kehl (1990), they suggested one new order: Pi 

turanthetalia tortuosi to comprise all the plant com 

munities of the Western Desert of Egypt. Within this 

order the northern communities can be attributed to 

the alliance Thymelaeion hirsutae (Eig, 1946) and the 

southern one to the alliance Zygophyllion coccinei. 

The results obtained in this study largely corroborate 

the latter suggestion. Whereas some species, e.g., 

Atriplex leucoclada, Deverra tortuosa, Randonia afri 

cana, Capparis spinosa var. aegyptia and Zilla spi 

nosa subsp. biparmata are confined to the northern 

transect, and Prosopis farcta, Fagonia bruguieri, Tra 

ganum nudatum and Salsola imbricata subsp. imbricata 

to the southern one, some other species exhibit 

wide ecological amplitude of tolerance through their 

distribution in both transects. Consequently, it can be 

suggested that the vegetation of the study area repre 

sents a gradual transition from the southern Nubian 

communities (of the Nubian Desert sensu El-Hadidi, 

1980a) in the south-western part to those characteristic 

of the northern Mediterranean coast (Libyan Desert 

sensu El-Hadidi, 1980a). The transitional character is 

clearly indicated by the fact that a group of species 

reaches its southern limit, and another group reaches 

its northern limit of distribution. 

The habitat investigated in this study is a relatively 

simple one, in which the species have to withstand 

harsh environmental conditions. This is not only re 

flected by the preponderance of annuals, but also by 

the presence of several highly-adapted, drought 

resistant species (Abdel-Razik et aI., 1984). In this re 

spect, the vegetation along the two studied transects is 

similar to that of gorges (karkurs) of Gebel Uweinat 

and some neighbouring areas of south-western Egypt 

(Boulos, 1982; Bornkamm, 1986). A major difference 

between the two corresponding habitat types is the 

high degree of scarceness of precipitation which may 

fall once every seven to ten years in Uweinat, and up 

to twenty years or more in Gilf Kebir and the neigh 

bouring areas. The vegetation cover of the latter land 

scape is less than 1% in the extreme desert (Stahr et 

aI., 1985). Thus, often just one species reaches domi 

nance forming monotypic stands. However, mono 

dominant stands in this study are not as common as 

those codominated by more than one species. 

Along gradients of decreasing precipitation, vegetation 

varies from grasslands to shrublands (Westoby, 

1980). 



The relative advantage of shrubs over grasses when 

water is limiting, as in the study area, can be explained 

by their extensive root systems which are capable to 

utilise water stored in different soil depths, whereas 

grasses utilise the transient water stored in the upper 

soil synchronic with precipitation pulses. The upper 

dry layer of the surface deposits acts as a protective 

layer, moisture is stored in subsurface layers, and the 

underlying sandstone provides added water storage 

capacity. The presence of a subsurface layer that is 

permanently wet is well-known phenomenon in the 

Egyptian Deserts (Kassas & Batanouny, 1984). As 

presented in the results, the dominance of shrubby 

plant species over the grasses is evident. 

Chamaephytes constitute 40% of the floristic compo 

sition, followed by therophytes. The dominance of 

both chamaephytes and therophytes over other life 

forms seem to be a response to the hot dry climate and 

human and animal interferences. A comparison of the 

life-form spectrum of the same 5° of the northern 

latitude in the corresponding Eastern Desert of Egypt 

(25°N - 30 0 

N), Abd EI-Ghani (1998) showed more 

therophytes (38.3%) and hemicryptophytes (22.0%) 

and less chamaephytes (29.0%). 

The vegetation that characterise the study area can 

be divided, according to TWINSPAN technique, into 

eight vegetation groups: Prosopis farcta-Tamarix nilotica, 

Tamarix nilotica-Alhagi graecorum, Zygo 

phyllum coccineum-Salsola imbricata subsp. 

imbricata, Cornulaca monacantha-Fagonia arabica, 

Atriplex leucoclada, Randonia africana-Deverra tor 

tuosa, Deverra tortuosa-Capparis spinosa var. aegyp 

tia and Capparis spinosa var. aegyptia-Zilla spinosa 

subsp. biparmata. Some species: Atriplex leucoclada, 

Anabasis articulata, Capparis spinosa var. aegyptia 

and Randonia africana common to the western Medi 

terranean coastal belt (Ayyad & Ammar, 1974; Abdel 

Razik et aI., 1984) are found in the less arid sites of 

the northern transect where the silty soil decreases 

water infiltration. A group of salt-tolerant plants in 

cluding Nitraria retusa, Tamarix nilotica, Sarcocornia 

fruticosa and Alhagi graecorum are found in the dry 

saline sites of the southern transect, and form phyto 

genic mounds of variable size. Alhagi graecorum is a 

widely distributed species that seems to grow in dif 

ferent habitats (Kassas, 1952). It is also considered as 

a groundwater-indicating plant (Girgis, 1972). According 

to Kassas and Girgis (1965), the growth of the 

desert scrub Nitraria retusa represents the highest tol- 

182 

erance to soil salinity conditions, and a penultimate 

stage in the successional development. The plant 

reaches its northernmost limit of distribution around 

Qara Oasis on the south-western edge of Qattara Depression 

(Abd EI-Ghani, 1992) as well as in Bahariya 

Oasis (Abd EI-Ghani, 1981). Nitraria retusa, how 

ever, has not been recorded beyond Latitude 28°N in 

Egypt (M. Kassas, pers. comm.). Further studies con 

cerning the distribution of this plant in the country is 

recommended. Prosopis farcta, Imperata cylindrica 

and Salsola imbricata subsp. imbricata are commonly 

found in the dry sandy plains along the southern tran 

sect. Restriction of Imperata to the high sandy plains 

is apparently due to the inability of this species to 

reach the capillary fringe of the groundwater which is 

fairly close to the surface (Rikli, 1943). The species is 

considered as facultative halophyte mainly occurring 

on sandy soil with slight salt content. Thus, this habi 

tat may represent a transitional zone between the less 

arid and dry saline habitats. The xero-psammophytes 

Fagonia arabica, Cornulaca monacantha, Zilla spi 

nosa subsp. spinosa, Calligonum polygonoides subsp. 

comosum, Pulicaria incisa, Citrullus colocynthis and 

Heliotropium digynum were found in dry non-saline 

sandy sites along both transects where infiltration is 

higher and water accumulated in deeper layers, and 

the soil is highly fertile. This group of species are 

more widely distributed in Egypt and neighbouring 

countries (Batanouny, 1979; Zahran & Willis, 1992; 

Frankenberg & Klaus, 1980; Wojterski, 1985). 

Ayyad (1976) pointed out that physiographic and 

edaphic factors have greatly affected the distribution 

of plant communities in the Western Desert of Egypt. 

DCA and CCA analyses of the vegetation and soil 

data in the present study indicated the relative posi 

tions of species and sites along the most important 

ecological gradients. DCA axis 1 may represent a 

geographical trend in the floristic data set. The main 

reason for this may be a gradient in the local harsh 

climate within the study area. Both ordination tech 

niques emphasised that salinity, fine sediments, or 

ganic matter and moisture content are the significant 

factors controlling the distribution of the vegetation in 

this region. This has been reported in other studies 

such as those of EI-Ghareeb and Hassan (1989), El 

Demerdash et al. (1995) and Shaltout et aI., (1997). 

The soil texture gradient in arid desert environments 

results in a corresponding gradient of available soil 

moisture. Therefore, moisture content is probably one 



of the most effective physical factors leading to vege 

tation variations from north to the south of the study 

area. The organic matter content plays an important 

role as a key element in soil fertility. Sharaf El Din 

and Shaltout (1985), and Abd EI-Ghani (1998) indi 

cated the role of soil organic matter in the Egyptian 

arid desert ecosystems. 


This study would not have been possible without 

the support of the "Alexander von Humboldt-Stiftung, 

Germany" for which I am extremely grateful. The 

following people and institutions are acknowledged 

with appreciation for their contribution: R. Bornk 

amm, F. Alaily, H. Kehl, F. Darius (all TU-Berlin) for 

their continuous support, fruitful discussions and 

valuable information on the soil structure; Institut flir 

Mathematik und Informatik (Freie Universitiit-Berlin) 

for statistics and many other facilities; S. EI-Banna 

(Department of Lands & Underground Water Re 

sources, Cairo University) for confirming and com 

menting on the soil analysis and S. T. Michael for his 

kind assistance in soil analysis. I thank two anony 

mous referees for their keen revision and suggestions 

to improve an earlier version of this paper. 

REFERENCES 

Abd EI-Ghani M.M., 1981. Preliminary studies on the 

vegetation ofBahariya Oasis, Egypt. Unpublished M.Sc. 

Thesis, Cairo University. 309 p. 

Abd EI-Ghani M.M., 1985. Comparative study of the vegetation 

of Bahariya and Farafra Oases and the Faiyum 

region, Egypt. Unpublished Ph.D. Thesis, Cairo University. 

464 p. 

Abd EI-Ghani M.M., 1992. Flora and vegetation of Qara 

Oasis, Egypt. Phytocoenologia, 21: 1-14. 

Abd EI-Ghani M.M., 1998. Environmental correlates of species 

distribution in arid desert ecosystems of eastern 

Egypt. J. Arid. Environ., 38: 297-313. 

Abdel-Razik M., Abdel-Aziz M. & Ayyad M., 1984. Environmental 

gradients and species distribution in a transect 

at Omayed (Egypt). J. Arid. Environ., 7: 337-352. 

Alaily F., Bornkamm R., Blume H.-P., Kehl R. & Zielinski 

H., 1987. Ecological investigations in the Gilf Kebir 

(SW- Egypt). Phytocoenologia, 15: 1-20. 

Ayyad M., 1976. The vegetation and environment of the 

western Mediterranean coastal land of Egypt. IV. The 

habitat of non-saline depressions. J. Ecol., 64: 713-722. 

Ayyad M. & Ammar M.Y., 1974. Vegetation and environment 

of the western Mediterranean coastal land of 

Egypt. n. The habitat of inland ridges. J. Ecol., 62: 509 

523. 


Batanouny K.H., 1979. The desert vegetation in Egypt. 

Cairo Univ. Afr. Rev., Special Publ., I: 9-37. 

Berk K.N., 1994. Data analysis with Student SYSTAT Windows 

Edition. Course Technology, Inc. 

Bornkamm R, 1986. Flora and vegetation of some small 

oases in S-Egypt. Phytocoenologia, 14: 275-284. 

Bornkamm R., 1987. Growth of accidental vegetation on 

desert soils in SW Egypt. Catena, 14: 267-274. 

Bornkamm R. & Kehl H., 1990. The plant communities of 

the western desert of Egypt. Phytocoenologia, 19: 149 

231. 

Boulos L., 1982. Flora of Gebel Uweinat and some neighbouring 

regions of south-western Egypt. Candollea, 37: 

257-276. 

Boulos L., 1995. Flora ofEgypt. Check list. Al Hadara Publishing, 

Cairo. 283 p. 

Boulos L., 1999. Flora of Egypt. vol.l. Al Hadara Publishing, 

Cairo. 417 p. 

Canfield R., 1941. Application of the line interception 

method in sampling range vegetation. 1. Forest., 39: 

288-294. 

Eig A., 1946. Synopsis of the phytosociological units of Palestine. 

J. Bot. Jerusalem Ser., 3 (I): 225-312. 

EI-Demerdash M.A., Hegazy A.K. & Zailay A.M., 1995. 

Vegetation-soil relationships in Tihamah coastal plains 

of Jazan region, Saudi Arabia. J. Arid. Environ., 30: 

161-174. 

EI-Ghareeb R & Hassan LA., 1989. A phytosociological 

study on the inland desert plateau of the Western Desert 

of Egypt at EI-Hammam. J. Arid Environ., 17: 13-21. 

EI-Hadidi M.N., 1980a. An outline of the planned flora of 

Egypt. Taeckholmia. Add. Ser., I: 1-12. 

EI-Hadidi M.N., 1980b. Vegetation of the Nubian Desert. 

In: Wendorf F. & Schild R. (eds), Prehistory ofthe eastern 

Sahara. Academic Press, New York: 345-351. 

EI-Sharkawy H.M., Fayed A.A. & Salama F.M., 1984. 

Vegetation of inland desert wadis in Egypt. V. Wadi 

Qassab. Feddes Repert., 95: 561-570. 

Emberger L., 1955. Une classification biogeographique des 

climats. Recueil de Travaux des Laboratoires de Botanique, 

Geologie et Zoologie de la Faculte des Sciences 

de l'Universite de Montpellier, 7: 3-43. 

Frankenberg P. & Klaus D., 1980. Atlas der pflanzenwelt 

des Nordafricanischen Trockenraums. Arheit. Geogr. 

Inst. (Univ. Bonn), A133: 1-237. 

Girgis W.A., 1972. Plant indicators in the Egyptian deserts. 

Desert. Inst. Bu ll., A.R.E., 21: 511-525. 

Haines RW., 1951. Potential annuals of the Egyptian Desert. 

Bull. Inst. Des. Egypte, 1(2): 103-118. 

Hill M.O.,1979. TWINSPAN - A FORTRAN program Fir 

arranging multivariate data in an ordered two-way table 

of classification of individuals and attributes. Ithaca. 

NY, Cornell University. 90 p. 

Hill M.O. & Gauch H.G., 1980. Detrended Correspondence 

Analysis: An improved ordination technique. Vegetatio, 

42: 47-58. 

Kassas M., 1952. On the distribution of Alhagi maurorum in 

Egypt. Proc. Egypt. Acad. Sci., 8: 140-151. 

Kassas, M., 1966. Plant life in deserts. In: Hills E.S. (ed.), 

Arid Lands. UNESCO, London, Mathuen/Paris. p. 145 

180 

183

Abd EI-Glwni Soil variables affecting the vegetation ofinland western desert ofEgypt 

Kassas, M., 1971. Die Pflanzenwelt. 2.Teil, Pflanzenleben in 

der ostlichen Sahara. In: Schiffers H. (ed.), Die Sahara 

wzd ihre Randgebiete. I. Physiogeographie. Weltforum 

Verlag, Munchen : 477-497. 

Kassas M. & Batanouny K.H., 1984. Plant Ecology. In: 

Cloudsley-Thompson J.L. (ed.), Sahara Desert. Pergamon 

Press, Oxford: 77-90. 

Kassas M. & Girgis W.A., 1965. Habitat and plant communities 

in the Egyptian desert. VI. The units of a desert 

ccosystem. J. Ecol., 53: 715-728. 

Migahid A.M., Batanouny KH. & Zaki M.A.F., 1971. Phytosociological 

and ecological study of a sector in the 

Mediterranean coastal region in Egypt. Vegetatio, 23: 

113-134. 

Monod Th., 1954. Mode contracte et diffus de la vegetation 

saharienne. In: Cloudsley-Thompson J.L., (ed.), Biology 

ofDesert. Institute of Biology, London: 35-44. 

Rikli M., 1943. Das Pjlanzenkleid der Mittelmeerliinder. 

Band I. Verlag-Huber, Berlin. 436 p. 

Schiffers H., 1971. Allgemines. In: Schiffers H., (ed.), Die 

Sahara wzd ihre Randgebiet. l. Phytogeographie. Weltforum-Verlag, 

Munchen: 19-36. 

Shaltout K.H., EI-Halawany E.F. & EI-Garawany M.M.. 

1997. Coastal lowland vegetation of eastern Saudi Arabia. 

Biodiv. Conserv., 6: 1027-1040. 

Sharaf El Din A. & Shaltout KH., 1985. On the phytosociology 

of Wadi Araba in the Eastern Desert of Egypt. 

Proc. Eg.vpt Bot. Soc., 4: 1311-1325. 

Springuel I. 1997. Vegetation, land use and conservation in 

the south eastern Egypt. In: Barakat H.N. & Hegazy 

A.K. (eds). Reviews in Ecology: Desert Conservation 

and De\'elopment. Metropole, Cairo: 177-206. 

Springuel I., Sheded M. & Abed W., 1990. Plant growth in 

relation to a rain incident in Wadi Agag, south Egypt. 

Vegetatio, 90: 159-165. 

Stahr K., Bornkamm R., Gauer J. & Kehl H., 1985. Die 

Veranderung von Boden und Vegetation am Obergang 

von Halbwuste zur Vollwuste zwischen Mittelmeer und 

184 

Quattara-Depression in Agypten. Geookodynamik, 6: 

99-120. 

Tackholm V., 1974. Students' Flora of Eg:"pt (2 nd Edn.). 

Cairo University Press, Cairo. 888p. 

Tadros T.M. & Atta B.M., 1958. Further contribution to the 

sociology and ecology of halophilous communities of 

Mareotis (Egypt). Vegetatio, 8: 137-160. 

ter Braak C.J.F., 1986. Canonical correspondence analysis: a 

new eigenvector technique for multivariate direct gradient 

analysis. Ecology, 67: 1167-1179. 

ter Braak c.J.F., 1988. CANOCO - A FORTRAN program 

or canonical community ordination bv [partial] [detrended] 

[canonicall correspondence analysis, principal 

components analysis and redundancy anal.vsis, Version 

2.1., Agricultural Mathematics Group, Wageningen. 95 

p. 

ter Braak c.J.F., 1990. Update notes: CANOCO version 3.1. 

Agricultural Mathematics Group, Wageningen. 35 p. 

Thomas D.S.G., 1988. The biogeomorphology of arid and 

semi-arid environments. In: Viles H.A. (ed.), Biogeomorpholog}'. 

Basil B1ackwelL Oxford: 193-221 

Tielborger K, 1997. The vegetation of linear desert dunes in 

the north-western Negev, Israel. Flora, 192: 261-278. 

WaIter H. & Breckle S.-W., 1984. Okologie der Erde, Band 

2. Spezielle Okologie der Tropischen und Subtropischen 

Zonen. Gustav Fischer, Stuttgart. 461 p. 

Westoby M., 1980. Elements of a theory of vegetation dynamics 

in arid rangelands. Israel 1. Bot., 28: 169-194. 

Wojterski T.W., 1985. Guide de I'excursion internationale 

de phytosociologie Algirie du Nord. Goltze, Gottingen. 

274 p. 

Zahran M.A., 1972. On the ecology of Siwa Oasis, Egypt. 

Egypt. J. Bot., 15(2): 223-242. 

Zahran M.A. & Willis, A.J., 1992. The Vegetation ofEgypt. 

Chapman & Hall, London. 424 p. 

Zohary M., 1973. Geobotanical Foundations of the Middle 

East. 2 vols, Gustav Fischer-Verlag, Stuttgart. 739 p. 


Besso!l et al. 


Ccllulose represents a considerable part of the organic 

matter and StifJa tellocissimo content ranges from 42 to 50%. 

Thus, thc degradation of cellulose is a good indicator of the 

biological activity of an ecosystem. 

The aim of this study was to measure the microbial 

fauna activity and specially cellulolytic activity. in two alfa 

steppes corresponding to a gradient of aridity, and study the 

effects of the alfa rhizosphere and depth level on this activity. 

All cxpcriments were conducted ill vitro at different 

temperature and moisture levels. 

Thc study area is located in the western high plains of 

Algcria, where Stipa tenacissimo is a perennial tussock 

grass. The most productive rangelands of the region have 

been degraded by successive dry years and increasing grazing 

pressurc. Local climate is semi-arid at Rogassa station 

and arid at Mekalis. The mean annual rainfall in the region 

studied only reaches 200 - 300 mm with high variability. 

Other characteristics of this zone have been summarised in 

tables I and 2. 

Soil samples were collected in each station underneath 

and bctween several tufts of alfa, at different depths. To es- 

INTRODUCTION 

La degradation de la cellulose est reconnue comme 

un bon indicateur de l'activite biologique des sols 

(Dommergues & Mangenot, 1970; Berg & Roswall, 

1973; Labroue & Lascombes, 1975; Roze, 1986). 

Les steppes a alfa du Sud-Oranais (Algerie) sont en 

voie de degradation tres rapide. Entre 1978 et 1990, la 

phytomasse verte de I'alfa est passee de 1750 a 75 kg 

Ms / ha selon Aidoud et Touffet (1996). Ces auteurs 

attribuent cette degradation essentiellement au surpa 

turage qui s'ajoute a plusieurs annees de secheresse 

successives (Djellouli, 1990). Dans ces milieux en 

voie de dcsertification, le taux de matiere organique 

dans le sol est considere comme I'un des attributs vi 

taux essentiels (Aronson et al., 1995) et sa reduction, 

un indicateur pertinent de desertification. La minerali 

sation de la matiere organique a ete prise comme in 

dice de l'activite microbiologique des sols (Djellali, 

1981) ; cependant, I' activite cellulolytique n' a pas fait 

I'objet d'etude particuliere. Afin de mieux compren 

dre le fonctionnement des sols de ce type de steppe, le 

present travail est une contribution al'evaluation de la 

cellulolyse dans les sols alfatiers. Notre objectif a ete 

de mettre au point un protocole selon la structure lo 

cale de I' espece dominante et le degre d' aridite. Dans 

ce but, deux stations situees dans des conditions eco 

logiques differentes ont ete choisies afin de mesurer 

186 

Activite ceLlulolytique in vitro de .1'01.1' de deux steppes cl alfcl d'Algerie 

timate the degradation rate of cellulose, a technique based 

on the in vitro degradation on filter papers was used. The 

disappearing of cellulose was determined by following the 

ponderal changes of the filter paper residues. 

The results (Figures 2 and 3) demonstrate that: 

- the intensity of degradation is higher in the semi-arid than 

in arid conditions. The higher contents in organic matter 

(plant biomass) and other nutrient as nitrogen, phosphorus 

and calcium (Table 2) in the semi-arid could explain this 

behaviour. 

- the cellulolytic activity decreases with depth, except under 

tufts of alfa which affect the distribution of micro-organisms 

along the profile. 

- the rate of cellulose degradation is more important under 

the tufts than between them in both studied sites. This is 

probably because most indicators of soil fertility (organic 

matter. total and available nitrogen. phosphorus and cations, 

number and activity of microbial and arthropod 

decomposers) are much higher in the soil under tufts than 

elsewhere, as previously shown by Noy-Meir (1985). 

les variations de I'activite cellulolytique et de verifier 

I'importance des effets des conditions ecologiques a 

differentes echelles spatiales. 

MATERIEL ET METHODES 

Stations 

L'etude a ete realisee sur les steppes d'alfa des 

hautes plaines du Sud-Oranais dans deux stations: 

la station de Rogassa se trouve a 1090 md'altitude 

au nord ouest d'EI Bayadh, sur un glacis de pente tres 

faible ; cette station se place dans I' etage bioclimati 

que semi-aride inferieur a hiver froid (Nedjraoui, 

1990). Le sol, sur croute calcaire (a environ 30 cm de 

profondeur), est de type brun calcaire et de texture sa 

blo-limoneuse. Il presente des elements grossiers et de 

nombreuses racines verticales et horizontales (Ned 

jraoui, 1990). Le recouvrement global de la vegetation 

(steppe a alfa pur) est de 55%, sa richesse specifique 

est de 29 taxons (Bessah, 1998) ; 

la station de Mekalis se situe a 36 km au nord de 

Aln Sefra, a 1300 m d' altitude. Elle repose sur un gla 

cis de pente tres faible, inferieure a2%. Elle est situee 

dans I' etage bioclimatique aride superieur, variante a 

hiver froid. Le sol a croute calcaire est assez profond 

(40 cm ou plus) et sa texture est essentiellement sa 

bleuse. 

ecologio mediterranea 25 (2) - /999

Bessah et al. Activite cellulolytique in vitro de sols de deux steppes Cl alfa d'Algerie 

L'activite cellulolytique a ete suivie in vitro selon 

le protocole suivant : dans des boites de Petri steriles 

on introduit environ 100 g de terre, sans tamisage pre 

alable. Le papier filtre servant de substrat cellulosique 

a ete humecte puis mis en sandwich entre deux cou 

ches de terre. Les boites ont ete mises aincuber a dif 

ferentes temperatures (5, 15 et 30° C) et humidites 

(5%, 15% et 30%). La perte d'eau appreciee par pesee 

a ete compensee periodiquement par l'apport d'eau 

distillee. Apres 90 jours d'incubation, les papiers fil 

tres ont ete recueillis puis nettoyes selon la methode 

de Rashid et Schaefer (1984). Le poids de cellulose 

pur restant apres incubation a ete calcule comme la 

difference de deux pesees: une premiere realisee 

apres sechage a 105°C et une deuxieme apres incine 

ration a500°C. 

RESULTATS 

Le papier filtre est passe par differentes etapes de 

degradation sous I'effet des micro-organismes cellu 

lolytiques. Cette evolution s'est manifestee par 

I'apparition de taches jaunes, grises, brunes et meme 

rougeatres. 

Entre les touffes d'alfa 

L'activite cellulolytique mesuree in vitro de la mi 

croflore est plus intense pour le sol de la station semi 

aride que pour celui de la station aride et ceci, aussi 

bien pour les echantillons preleves en surface qu'en 

profondeur. En effet, le pourcentage de degradation 

varie de 20 a 85% pour la premiere station et de 11 a 

42% pour la seconde. Cette cellulolyse diminue avec 

la profondeur du sol dans les deux stations. Dans la 

station semi-aride, les taux de degradation enregistres 

(Figure 2) sont de 22 a85% en surface et de 20 a70% 

en profondeur. Dans la station aride, ces taux varient 

de 15 a42% dans I'horizon superficiel et de 11 a 38% 

en profondeur. 

Sous les touffes 

En surface, l'activite cellulolytique est plus mar 

quee pour le sol de la station semi-aride que pour celui 

de la station aride, avec des taux de degradation va- 

188 

riant de 9 a 94% pour la premiere station et de 13 a 

63% pour la seconde. 

Pour la station semi-aride, cette cellulolyse dimi 

nue avec la profondeur du sol. Le pourcentage de de 

gradation passe de 9 a 94% dans l'horizon de surface 

et de 9 a47% en profondeur. Par contre pour la station 

aride l'activite cellulolytique ne diminue pas avec la 

profondeur du sol: elle a meme tendance aaugmenter 

puisque le taux de degradation est en moyenne de 13 a 

53% en surface et de 17 a63% en profondeur (Figure 

3). 

L'influence du taux d'humidite et de la tempera 

ture sur I'activite cellulolytique en fonction des 

conditions d'incubation connait la meme evolution 

sous les touffes et entre les touffes: le maximum 

d'activite cellulolytique pour la station semi-aride est 

atteint a 30% d'humidite quelle que soit la temperature 

d'incubation. Pour la station aride, un maximum 

d'activite a ete observe a 15% d'humidite et pour des 

temperatures d'incubation de 15 et 30°C. A 5°C et 5% 

d'humidite, le papier filtre semble peu altere, si ce 

n'est dans sa coloration, les evolutions ponderales 

etant alors peu marquees. 

DISCUSSION 

L'activite cellulolytique des sols de steppes aalfa, 

mesuree in vitro, apparait relativement faible par rapport 

acelle des sols des regions temperes. Par exem 

pie, des mesures realisees dans les memes conditions 

pour les sols de la foret d'Orsay (Rashid & Schaefer, 

1984) ont donne des taux de degradation de I' ordre de 

90% (mull) et de 58% (anmoor) au bout d'un mois 

d'incubation, alors que dans notre cas ces valeurs 

n'ont ete atteintes qu'au terme de trois mois 

d'incubation en conditions moyennes. 

L'intensite de l'activite cellulolytique de la micro 

flore depend des conditions stationnelles, microstationnelles 

(sol sous la touffe et entre touffes ), de la 

profondeur du sol et des conditions d'incubation reali 

sees au laboratoire. D'apres nos resultats, elle semble 

etre plus importante dans la station semi-aride que 

dans la station aride, ce qui peut etre lie aux caracte 

ristiques climatiques et edaphiques des stations. 

L'aridite de la station de Mekalis serait non seulement 

climatique, mais egalement edaphique en raison de 

caracteres pedologiques (Pouget, 1980). 


Bessah et al. Activite cellulolytique in vitro de sols de deux steppes a alfa d'AIgerie 

Les resultats expriment la capacite de reaction et 

de resistance des micro-organismes vis-a-vis de ces 

conditions ecologiques qui deviennent vraisembla 

blement limitantes le long du gradient d'aridite crois 

sant caracterisant les deux stations. 

La biomasse vegetale marque une baisse impor 

tante entre les deux stations. La phytomasse aerienne a 

ete evaluee a 8040 kg Ms / ha dans la station semi 

aride et a 3525 kg Ms / ha dans la station aride (Ned 

jraoui, 1990). Cette difference peut, outre les condi 

tions ecologiques, etre attribuee en partie a l'intensite 

d'exploitation. La biomasse vegetale represente la 

source principale de matiere organique (Pouget, 1980 ; 

Kadi-Hanifi, 1998). La diminution de celle-ci entraine 

une baisse du taux de matiere organique dans le sol. 

Ce taux passe en moyenne de 1% dans la station semi 

aride a 0,1 % dans la station aride. Or I'incorporation 

de la matiere organique dans le sol influe sur la stabi 

lite structurale et donc sur I'aeration du sol et sa per 

meabilite (Duchaufour, 1970; Duthil, 1973; Guckert 

et al., 1975 ; Kadi-Hanifi, 1990, 1998). Elle stimule 

egalement le developpement de la microflore (Dommergues 

& Mangenot, 1970; Le Houerou, 1995). 

L'effet favorable de la matiere organique sur les ger 

mes cellulolytiques a ete mis en evidence dans des 

sols au Maroc par Bryssine (1967). A partir de ces 

travaux, nous pouvons supposer que la diminution de 

la matiere organique en fonction de I'aridite du climat 

(Pouget, 1980) est a I'origine de la faible activite de la 

microflore dans la station aride. 

La cellulolyse plus importante dans le cas de la 

station semi-aride peut etre influencee egalement par 

d'autres constituants chimiques du sol. En effet, Du 

chaufour (1970), Roze (1986) et Widden et al., (1988) 

ont mis en relation la sensibilite de I'activite cellulo 

Iytique et la richesse du sol en azote: la cellulolyse est 

plus rapide dans les milieux bien pourvus en cet ele 

ment, elle est par contre lente dans les milieux pauvres 

en azote. De plus, Berg et Rosswall (1973) ont montre 

que le phosphore est I' element qui favorise le plus la 

cellulolyse, I' azote et le calcium venant ensuite. Nos 

resultats concordent avec ces donnees bibliographi 

ques: la cellulolyse est plus active dans la station 

semi-aride dont les sols sont notablement mieux pour 

vus en ces nutriments que ceux de la station aride (Tableau 

2). 

L'activite cellulolytique est plus marquee dans les 

echantillons de sol preleves sous les touffes (sol rhi 

zospherique) que dans ceux preleves entre les touffes 

190 

d'alfa (sol nu). Cette plus forte activite doit etre liee a 

la quantite de matiere organique plus elevee sous la 

touffe et au microclimat ameliorant les conditions 

edaphiques creees par la touffe. En effet, la plupart 

des indicateurs de la fertilite du sol dans les zones ari 

des (matiere organique, azote total, phosphore, nom 

bre et activite des micro-organismes) sont plus 

importants sous les touffes qu'entre celles-ci (Garcia 

Moya & McKell, 1970; Tiedmann & Klemmedson, 

1973 ; Charley & West, 1975; Barth & Klemmedson, 

1978 ; Doescher et al., 1984 ; Noy-Meir, 1985 ; Klo 

patek, 1987; Bolton et al., 1990). Ainsi, les racines 

peuvent enrichir le sol en matiere organique et stimu 

ler leur colonisation par les micro-organismes rhizos 

pheriques (Warembourg, 1975 ; Hailer & Stolp, 1985 ; 

Gorissen, 1994). Les produits neoformes de la litiere 

racinaire, tres importante chez ]' alfa, et les produits 

simples issus de I'exsudation racinaire (polysacchari 

des), favorisent la proliferation de bacteries et de 

champignons (Dommergues & Mangenot, 1970; Zeriahene, 

1987). De plus, Djellali (1981), Ali-Haimoud 

(1982) et Kihal (1986) ont demontre que la rhizos 

phere de I'alfa presente une action positive sur la mi 

croflore des sols de steppes. 

La profondeur du sol influe egalement sur 

I' activite cellulolytique. Entre les touffes, I'activite de 

la microflore diminue avec la profondeur dans la sta 

tion aride comme dans la station semi-aride. Cette dif 

ference d'activite entre surface et profondeur serait en 

relation avec la distribution verticale des micro 

organismes. Ces derniers se trouvent en majorite dans 

la couche superficielle du sol, mieux aeree et plus ri 

che en substances nutritives et leur nombre diminue 

progressivement avec la profondeur (Boullard & Mo 

reau, 1962 ; Gaucher, 1968 ; Dommergues & Mange 

not, 1970). Les memes resultats ont ete obtenus dans 

les landes bretonnes (Roze, 1986) oll l'activite cellu 

lolytique mesuree in situ est de plus en plus reduite en 

fonction de la profondeur du sol. Les travaux de Djel 

lali (1981) et Ali-Haimoud (1982) confirment ce fait 

pour des sols alfatiers d'AIgerie. 

Sous les touffes, I'activite cellulolytique augmente 

avec la profondeur du sol dans la station aride. Par 

contre, sous conditions semi-arides, cette activite cel 

lulolytique diminue avec la profondeur du sol. Cette 

difference de fonctionnement de la microflore en 

fonction de la profondeur est probablement liee a la 

morphologie racinaire de I'alfa. En effet, les racines 

d'alfa presentent des modifications qui leur permettent 


Bessah et al. Activite cellulolytique in vitro de sols de deux steppes Cl alfa d'Algerie 

de mieux s'adapter aux conditions hydriques et edaphiques 

de leur biotope (Zeriahene, 1987; Aidoud, 

1989; Nedjraoui, 1990). En fonction de l'aridite, les 

racines deviennent de plus en plus longues et epais 

ses: 33,76 cm de longueur et 2,5 mm de diametre 

pour l'alfa des regions pre-sahariennes, 24,5 cm de 

longueur et I mm de diametre pour I'alfa provenant 

des regions semi-arides (Nedjraoui, 1990). 

Dans la station semi-aride, les sols sont peu pro 

fonds et la croute calcaire est proche de la surface, 

constituant ainsi un obstacle au developpement des 

racines en profondeur. Celles-ci parviennent difficile 

ment a franchir la croute calcaire et forment un reseau 

dense au niveau de I'horizon superficiel. Le develop 

pement racinaire se fait donc lateralement. Dans la 

station aride, les sols a texture essentiellement sa 

bleuse sont assez profonds. Les racines sont essen 

tiellement pivotantes et penetrent profondement dans 

le sol a la recherche de zones plus humides. Ainsi la 

repartition des germes cellulolytiques suit celle des 

racines ; leur activite diminue avec la profondeur dans 

la station semi-aride et augmente avec celle-ci dans la 

station aride. Cette derniere observation rejoint celle 

de Labroue et Lascombes (1975) qui ont trouve dans 

les sols de I' etage alpin une activite cellulolytique in 

situ plus importante en profondeur qu'en surface. Se 

Ion ces auteurs, cette intensite de l'activite cellulolyti 

que serait liee a l'enracinement profond des graminees 

colonisant la station etudiee. 

L'intensite de l'activite cellulolytique varie egalement 

en fonction de l'humidite. En effet, les resultats 

obtenus apres incubation a differents taux d'humidite 

montrent que les micro-organismes caracteristiques de 

la station semi-aride, probablement adaptes a un taux 

d'humidite relativement important, reagissent mieux a 

30% d'humidite. Par contre, dans la station aride ou 

les micro-organismes sont probablement adaptes a une 

certaine secheresse du milieu, la reaction est maximale 

a 15% d'humidite. 

L'activite cellulolytique est maximale parfois a 

30° et parfois a 15°C dans les deux stations. Ces re 

sultats seraient lies aux exigences de la microflore vis 

a vis de la temperature. Dommergues et Mangenot, 

(1970) signalent une activite des bacteries cellulolytiques 

a partir de 15-20°C avec des valeurs optimales de 

28-30°C. Le minimum d'activite cellulolytique est ob 

serve a 5°C. Nous avons deja montre (Roze, 1986) 

que des temperatures trop basses induisent une faible 

activite cellulolytique. 


CONCLUSION 

Cette etude a permis de quantifier l'activite ·cellulolytique 

dans les sols de steppes et de rechercher les 

principaux facteurs de sa variation. Elle aborde un 

volet particulier du fonctionnement qui vient s'ajouter 

aux travaux anterieurs realises dans les steppes a alfa 

par Aidoud (1989), Nedjraoui (1990) et Kadi-Hanifi 

(1998). Les resultats obtenus ont montre que l'activite 

cellulolytique diminue avec I'aridite et la profondeur 

du sol. La morphologie de la touffe d'alfa semble modifier 

la distribution verticale de la microflore cellu 

lolytique. Le taux d'humidite constitue un facteur 

essentiel qui n'est pas independant de la structure de 

la touffe d'alfa. 

L'activite cellulolytique met en relief des differen 

ces entre les stations qui avaient deja ete evoquees : il 

s'agit donc d'un bon indicateur de la degradation des 

sols. Cet indicateur devra etre confronte a d'autres, 

tels que la proteolyse et la mineralisation de I'azote, 

afin de determiner son niveau de pertinence. 

BIBLIOGRAPHIE 

Aidoud A., 1989. Contribution Cl l'etude des ecosystemes 

steppiques patures des hautes plaines algero-oranaises. 

These de Doctorat, Alger, 240 p. 

Aidoud A., & Touffet J., 1996. La regression de I'alfa (Stipa 

tenacissima L.), graminee perenne, un indicateur de desertification 

des steppes algeriennes. Secheresse, 7: 187 

193. 

Ali-Haimoud A., 1982. Contribution al'etude des sols alfatiers. 

Fixation de I'azote asymbiotique. Effet du paillage 

sur cette activite. These, Magister, Alger, 112 p. 

Aronson J., Floret c., Le Floc'h E., Ovalle c., Pontanier R., 

1995. Restauration et rehabilitation des ecosystemes degrades 

en zones arides et semi-arides. Le vocabulaire et 

les concepts. In: L'homme peut-il refaire ce qu'il a defait? 

John Libbey Eurotext, Paris, p. 11-29. 

Barth R.c., & Klemmedson J.O, 1978. Shrub induced spatial 

patterns of dry matter, nitrogen and organic carbon. 

Soil Sci. Soc. Am. J., 42: 804- 809. 

Berg B., & Roswall T., 1973. Decomposition ofcellulose in 

laboratory pot experiment. Swedish IBP tundra Biome 

Projet Tech Report, 14: 207- 212. 

Bessah R., 1998. De la decomposition des litieres ala mineralisation 

de l'azote dans les steppes a alfa du Sud oranais. 

Memoire de Magister, Alger, 120 p. 

Bolton H., JR., Smith J.L., & Wildungr E., 1990. Nitrogen 

mineralization potentials of shrub steppe soils with different 

disturbance histories. Soil Sci. Soc. Am. J. 54: 

887- 891. 

Boullard B., & Moreau R. , 1962. Sol, microflore et vegetation. 

Equilibre biochimique et concurrences biologiques. 

Masson, Paris, 168 p. 

191

Bessah et al. ActiviM cel/ulolytique in vitro de sols de deux steppes a alfa d'AIgerie 

Bryssine I., 1967. La richesse des sols marocains en germes 

cellulolytiques en fonction des conditions ecologiques. 

Ann. Inst. Pasteur, Paris, 115: 542-548. 

Charley J.L., West N.E., 1977. Micropatterns of nitrogen 

mineralization activity in soils of some shrub dominated 

semi-deserts ecosystems of Utah. Soil BioI. Bioch., 9: 

357- 365. 

Djellali N., 1981. Contribution a I'etude microbiologique 

des sols a alfa. Mineralisation de I'azote et du carbone 

dans le sol de l'alfa de Ain Oussera. Effet du pail/age 

sur cette activite. These, Magister, Alger, 105 p. 

Djellouli Y., 1990. Flare et climat en AIgerie septentrionale. 

Determinisme climatique des especes. These, Doctorat, 

Alger, 296 p. 

Doescher P.S., Miller R.F, & Winward A.H., 1984. Soil 

chemical patterns under eastern Oregon communities 

dominated by big bage brush. Soil, Sci. Soc. Am. 1., 48: 

659-663. 

Dommergues Y., & Mangenot E, 1970. Ecologie microbienne 

du sol. Masson, Paris, 783 p. 

Duchaufour P., 1970. Pedologie : constitutions et proprietes 

du sol. Masson, Paris, 459 p. 

Duthil J., 1973. Elements d'ecologie et d'agronomie, tome 

3, Bailliere, Paris, 651 p. 

Garcia-Moya E., & McKell C.M., 1970. Contribution of 

shrubs to the nitrogen economy of a desert wash plant 

community. Ecology, 51: 81-88. 

Gaucher G., 1968. Traite de pedologie agricole. Dunod, Paris, 

578 p. 

Gorissen T.. 1994. Microbial activity and soil carbon dynamics 

in a changing climate. Change, 21: 6-8. 

Guckert A., Chore T.H., & Jacquin E, 1975. Microflore et 

stabilite structurale des sols. Rev. Ecol. BioI. Sol, I: 211 

223. 

Hailer T., & Stolp H., 1985. Quantitative estimation of root 

exsudation of maize plants. Plant and Soil, 86: 207-211. 

Kadi-Hanifi H., 1990. Etude phytoecologique des formations 

aalfa dans le sud oranais. Biocenose, 1-2: 37-67 

Kadi-Hanifl H., 1998 L'alfa en AIgerie : syntaxonomie, relations 

milieu-vegetation, dynamique et perspectives 

d'avenir. These, Doctorat, Alger, 267 p. 

192 

Kihal BM., 1986. Contribution cl l'etude de la lIlicroflore 

des sols a/fatiers au cours de l'annee et recherche des 

micro-organismes cel/ulo/ytiques. These, Magister, 

Oran, 105 p. 

Klopatek J.M., 1987. Nitrogen mineralization and nitrification 

in mineral soil of Pinyon-Juniper ecosystem. Soil 

Sci. Soc. Am. J., 52: 453-457. 

Labroue L., & Lascombes G., 1975. L'activite cellulolytique 

dans les sols de quelques stations de l'etage alpin du pic 

du Midi de Bigorre. Rev. Ecol. BioI. Sol, I: 181-192. 

Le Houerou RN., 1995. Degradation, regeneration et mise 

en valeur des terres seches d'Afrique. In : L 'holllme peW 

il re/aire ce qu'il a de/ait ? J. Libbey Eurotext, Paris: 

65-102. 

Nedjraoui D., 1990. Adaptation de l'a/jcl (Stipa tenacissima 

L.) aux conditions stationnel/es. Contribution a I'etude 

du fonctionnement de l'ecosysteme steppique. These, 

Doctorat, Alger, 256 p. 

Noy-Meir I., 1985. Desert ecosystem structure and function: 

In: M. Avenari, I. Noy-Meyer, & D.W. Goodhall (Ed.). 

Ecosystems of the world, vol 27: hot deserts and arid 

shrub lands. 93-103, Elsevier, Amsterdam. 

Pouget M., 1980. Les relations sol- vegetation dans les steppes 

sud algeroises. ORSTOM, Paris, 555 p. 

Rashid GH., & Shaefer R., 1984. Activite cellulolytique 

dans deux sols d'une catena de la fixet d'Orsay. Rev. 

Ecol. Bioi. Sol, 4: 431-438. 

Roze E, 1986. Le cycle de I'azote dans les landes bretonnes. 

These, Doctorat, Rennes, 292 p. 

Tiedman A.R., & Klemmedson J.Q., 1973. Effect of mesquite 

on physical and chemical properties of the soil. J. 

Range Manag., 26: 27- 29. 

Warembourg ER., 1975. Application des techniques radioisotopiques 

it l'etude de l'activite ecologique dans la rhizosphere 

des plantes. Rev. Ecol. BioI. Sol, I: 261- 272. 

Widden P., Cunningham 1., & Breil B., 1988. Decomposition 

of cotton by Trichoderma species. Influence of 

temperature, soil type and nitrogen levels. Can. J. 

Micro., 35: 469- 473. 

Zeriahene N., 1987. Etude du systeme racinaire de I'a/jcl 

(Stipa tenacissima L.) en relation avec I'adaptation au 

xerophytisme. These, Magister, Oran, 113 p. 


Guiot et al. 

INTRODUCTION 

During the last decades, transfer functions have 

been used to reconstruct past environments from ter 

restrial data (pollen, diatoms, tree-rings, insects, mol 

luscs ... ) as well as from ocean data (foraminifera, 

diatoms, dinoflagelates ... ). The methods primarily 

used were based on modern distribution of one or sev 

eral species in a climatic domain (lversen, 1944; 

Grichuk, 1984; Atkinson et al., 1987; Zagwijn, 1994; 

Fauquette et al., 1998) or on a statistical calibration of 

modern assemblages in terms of climatic variables, 

using (i) multivariate statistical analyses (lmbrie & 

Kipp, 1971; Gasse & Tekaya, 1983; Ko

Guiot et al. 

Pollen pft 

Boreal summergreen 

Boreal conifer 

Cool temperate conifer (including cte!) 

Temperate summergreen (including tsl) 

Temperate summergreen (warmer) 

Warm temperate evergreen (including wtel) 

Warm temperate evergreen (warmer) 

Alpine-arctic shrubs 

Cool grass/shrub 

Heath 

Warm grass/shrub 

Desert forbs 

Glacial maximum climate by inverse vegetation modelling 

Code R 

bs 0.85 

bec 0.84 

ctc 0.81 

ts 0.79 

ts2 0.55 

wte 0.85 

wte2 0.76 

aa 0.80 

cogs 0.79 

h 0.41 

wags 0.79 

df 0.55 

Table I. Results of the transfer translating NPP into pollen pft scores. We use 8 neurones in the intermediate layer and a logsigmoid 

at the intermediate as well as at the output layer. The 13 inputs are transformed each into 5 classes by "fuzzification" (see 

Guiot et al., 1996). R is the correlation between observations and predictions on the 1245 spectra used for calibration. 

Our final goal is to reconstruct these bioclimatic 

variables that constrain the vegetation by a kind of in 

version of the model. 

The vegetation model 

BIOME3.5 is a process-based terrestrial biosphere 

model which includes a photosynthesis scheme that 

simulates acclimation of plants to changed atmos 

pheric CO 2 by optimisation of nitrogen allocation to 

foliage and by accounting for the effects of CO 2 on net 

assimilation, stomatal conductance, leaf area index 

(LAI) and ecosystem water balance. It assumes that 

there is no nitrogen limitation. 

The inputs of the model are soil texture, CO 2 rate, 

absolute minimum temperature (Tmin), monthly mean 

temperature (T), monthly total precipitation (P) and 

monthly mean sunshine (S) i.e. the ratio between the 

actual number of hours with sunshine over the poten 

tial number (with no clouds). From these input vari 

ables, BIOME3.5 computes bioclimatic variables 

(growing degree-days, ratio of actual evapotranspira 

tion to evapotranspiration at the equilibrium and pre 

cipitation minus evapotranspiration) that represent 

energy and water constraints on the vegetation. Then, 

the model calculates the maximum sustainable leaf 

area index and the NPP (in kg m- 2 y(l) for the pfts 

able to live in such an input climate. Competition 

among pfts is simulated by using the optimal NPP of 

each pft as an index of competitiveness. We use the 10 

pfts present in our modern pollen dataset: temperate 

broad-leaved evergreen (tbe), temperate summergreen 

(ts), temperate evergreen conifer (tc), boreal evergreen 

(bec), boreal deciduous (bs), temperate grass (tg), 

woody desert plant type (wd), tundra shrub type (tus), 

196 

cold herbaceous type (clg), lichen/forb type (If). The 

pollen pfts are slightly different in such a way that 

some of these pfts are sometimes subdivided. 

We retain as predictors the NPP of these 10 pfts, 

the total annual net absorbed photosynthetically active 

radiation (APAR in MJ.m- 2 .yr- 1 ) and the LAI of the 

dominant pft (with a maximum NPP). For a better dis 

crimination, we use also an additional predictor re 

lated to the rank of the pft in the above list, so that 

forest pfts have a lower rank than the herbaceous pfts. 

Transfer function (TF) from BIOME3.5 to pollen 

BIOME3.5 is run, on the 1245 modern pollen 

samples, to estimate the 13 predictors described in 

section 2.1. Then we use a backpropagation neural 

network technique (as in section 1) to calculate a set 

of non-linear relationships between the 12 pollen 

derived pft scores and the 13 predictors. The correla 

tion coefficients are greater than 0.55 and often 

greater than 0.8 (see Table 1) except for the heaths 

(mainly Calluna) which is a really ubiquitous pft. 

The likelihood function (LH) 

The inverse modelling problem consists in finding 

the value of an input parameter vector for which the 

model output fits as much as possible a set of obser 

vations (Mosegaard & Tarantola, 1995). Here, the pa 

rameters of the model are a set of climatic variables 

which constrain properly the pollen pft scores. 

The crudest approach is the exhaustive sampling, 

where all the points in a dense grid, covering the 

model space, are visited. This method is not be rec 

ommended if the number of parameters is high. An 


Guiot et al. 

alternative approach is the Bayesian approach which 

describes the "0 priori information" we may have on 

the parameter vector by a probability density. Then it 

combines this information with the information pro 

vided by the comparison of the information provided 

by the model to that provided by the observations in 

order to define a probability density representing the a 

posteriori information. The likelihood function, which 

roughly measures the fit between observed data and 

data predicted by the model, links the a posteriori 

probability to the a priori one. 

To give a maximum weight to cases where the cor 

rect biome is predicted, we adopt a likelihood function 

LH which is proportional to the difference between 

predicted and observed pft and also to the difference 

between predicted and observed biomes. More details 

are given in Guiot et al. (in press). 

Definition of the parameter vector 

The parameters of interest in this study are tem 

perature and precipitation. We add to the modern 

monthly temperatures a term L'l T; and we multiply 

the modern monthly precipitation by a given !1P 

J 

(all 

being either positive or negative). It is not necessary to 

sample independently the 12 monthly parameters be 

cause they are linked by the seasonal cycle. Thus, we 

vary first the January and July climatic parameters; 

then, we deduce the value for the other months 

U=I, ...,l2) by linear interpolation (Guiot et aI., in 

press). 

The sunshine parameter cannot be processed inde 

pendently from temperature and precipitation as it re 

lates more or less strongly to the latter ones (see Table 

4 in Guiot et al., in press). 

Consequently, for each parameter vector (L'lTJan, 

L'lTJul. L'lPJan, L'lPJul ), we calculate 12 monthly temperature 

values (Tj+L'lT j , j=1,12), 12 monthly precipitation 

values (Pj*L'lPj, j=I,12) and 12 monthly sunshine val 

ues which are then used as input into the vegetation 

model. 

Monte-Carlo sampling (Metropolis-Hastings algorithm) 

Let us consider a multi-dimensional mathematical 

domain where each dimension represents a parameter 

range. A vector of parameters is an element of the 



multi-dimensional domain. The Metropolis-Hastings 

(MH) algorithm is an iterative method which browses 

the domain of the parameters according to an accep 

tance-rejection rule (Metropolis et aI., 1953 ; Hast 

ings, 1970). Strictly speaking, the MH algorithm is not 

an optimisation method of the posterior joint density 

function, but a method for browsing the prior defini 

tion domain of the parameter vector in order to simu 

late its posterior distribution. The histograms of the 

parameters are incremented according to the candidate 

or the actual value, according to the fact that it is ac 

cepted or not. These histograms are estimates of the 

posterior probability distribution of the parameters. 

The MH algorithm was applied with the LH func 

tion presented in section 2.3 and with a multivariate 

uniform distribution as a prior of the hyper-parameter. 

LH is not sensu stricto a likelihood function in the 

probability sense, but it has the form; numerous em 

pirical tests have shown its suitability in our applica 

tion. 

RESULTS 

Before applying the method to LGM pollen data, a 

test was performed on 591 pollen data selected among 

the 1245 spectra of the modern data set. These 591 

spectra are more or less equally distributed in Europe 

and Eurasia. 

Validation with modern data 

The CO 2 concentration was set to 340 ppmv. The 

input parameters were allowed to vary within the fol 

lowing ranges: 

- L'lTJan : [-10, IO]OC in terms of deviations of the 

observed value 

- L'lTJul : [-10, IO]OC in terms of deviations of the 

observed value 

- L'lPJan : [-60, 60]% of the modern value 

- L'lPJul : [-60, 60] % of the modern value 

The number of iterations is set to 2400 for each 

sample. The value of LH converges most of the time 

towards -3. We select the 10% iterations giving the 

highest LH (note that LH is defined as negative). 

Among them we select the iterations which simulates 

the most frequent biome. Only these selected itera 

tions (generally between 500 and 1000) are used for 

197

Guiot et al. 

building the a posteriori probability distribution. We 

calculate then the probability distribution of the four 

input parameters (L1TJan , L1TJu1 , L1PJan , L1P Jul ). 

For a given site, the results are summarised as 

follows: (I) we look for the mode of the distribution; 

(2) if it is positive we calculate the probability that it 

is significantly larger than 0 by summing the prob 

abilities above 0 (if it is negative, we proceed in the 

same way with the probability to be negative); (3) if 

that probability is less than 75%, the reconstructed 

value is zero, otherwise it is equal to the mode. For the 

next step, we retain that reconstructed values which 

are distributed in function of the reconstructed biome. 

Figure 2 shows the box plots of the reconstructed 

anomalies of temperature and precipitations along a 

gradient from the coldest to the warmest biome (the 

three last biomes being the driest ones). 

Most of the medians are zero, which shows that 

the mean bias is null. There are a few exceptions: 

- L1TJu1 is underestimated for tundra and overestimated 

for warm steppes, which is maybe due to the 

fact that the modern dataset is dominated by samples 

taken in very cold tundra and very warm steppes; 

- L1TJan is generally overestimated for warm mixed 

forest maybe because, in our modern dataset, that bi 

ome mainly occurs with mild winters; 

- L1PJu1 is overestimated for temperate deciduous 

forest, which is hard to interpret as that biome is not 

frequently reconstructed in our dataset (the corre 

sponding pollen assemblages are often similar with 

those sampled in cool mixed forest); 

- L1PJan is reconstructed as zero everywhere be 

cause it corresponds to probabilities less than 75%. 

We have also to consider the interquartile interval 

which is related to dispersion of the reconstructions 

around the median. A weak dispersion around a null 

median means that the parameter is extremely well 

reconstructed: it is the case for L1TJan in tundra, cool 

steppes and cool mixed forest, L1PJan everywhere, L1PJul 

in the driest biomes (warm mixed forest, xerophytic 

vegetation, steppes). The dispersion is generally weak, 

which encourages us to apply the method to the fossil 

data. 

The final objective of the method is to reconstruct 

the bioclimatic variables which drive the model. To 

have an idea of its precision, we use the root mean 

squared error (RMSE) statistic. It is calculated over all 



the modern samples and over the two dominant biomes 

of the LGM (Table 2). 

The last glacial maximum 

After validation, the method can be applied to the 

71 fossil spectra of Europe and Eurasia. The first ex 

periment is based on the LGM CO 2 concentration (200 

ppmv: Barnola et aI., 1987). Figure 3 shows the geo 

graphical distribution of the three major bioclimatic 

variables. The western part of the continent seems to 

have been colder than the eastern part. The moisture 

variable (EIPE) does not seem to have been much 

lower than present except in some sites. These results 

must be compared with those obtained with the same 

data by Peyron et al. (1998) and Tarasov et al. (1999). 

For that, we divide the continent into four regions: 

Mediterranean (9 sites of Peyron), western Europe 

(the other 6 sites of Peyron), northern Eurasia (the 

sites of Tarasov at north of 53°N), Southern Eurasia 

(the sites of Tarasov at south of 53°N). We do not mix 

the reconstructions of the two papers because there is 

a difference in the number of variables reconstructed. 

We compare also these reconstructions to a second 

experiment based on 340 ppmv of CO2• The results 

are summarised as box plots in Figure 4. 

Figure 4 shows well the west-east gradient of 

MTCO and GDD5. It shows also the good agreement 

between the results of the pft method and those of the 

two experiments (200 and 340 ppmv). For E/PE, the 

agreement is also good except for western Europe, 

where the vegetation model is able to produce cool 

steppes with a small decrease of available water (and a 

large decrease of temperature), while, for the pft 

method, it is necessary to decrease EIPE to values less 

than 65%. 

The box plots of Figure 4 are not enough detailed 

to compare the results between 200 and 340 ppmv 

CO 2 • Indeed, it has been shown (Guiot et aI., in press) 

that the response of pollen pft to climatic change 

could appear as multimodal distribution. To analyse 

these responses in more detail, we have selected a few 

sites in each of the four zones and we have represented 

these distributions for the three bioclimatic pa 

rameters (Figure 5). 

MTCO: there is no significant deviation between 

the two distributions except in Mediterranean region 

where anomalies of -25°C are more probable under 

low CO 2 than higher anomalies; 

199

Guiot et al. Glacial maximum climate by inverse vegetation modelling 

70 0 

70 0 

70 0 

MTCO : deviations from modern value (OC) 

• 

Pro.b=75% 10 

N 

GODS : deviations from modern value (OC*day) 

Prob=75% 2000 

• 

N 

N 

• Pro.b=75% 

Figure 3. Reconstruction of three bioclimatic parameters using inverse modelling of pollen spectra for the last glacial maximum 

ecologia mediterranea 25 (2) - 1999 201 

10go Ii 

5 

o 

1000 

o

Madon ct a1. Influence ofstudy scale on the characterisation afp/ants community organisation 

INTRODUCTION 

Patterns of species distributions, at the community, 

the landscape, or the region level first focused mainly 

on the response of the species to an environmental 

gradient (see in particular Clements, 1916, 1936; 

Oleason, 1917, 1926; Whittaker, 1951, 1967; Curtis, 

1959, and see also the recent synthetical model of 

Collins et al., 1993). An environmental gradient, or 

complex gradient (Whittaker, 1967) is the factors as a 

whole -including biotic factors- changing in space and 

governing the species distribution. 

In later studies, the pattern analysis of species dis 

tributions took factors other than environmental axes 

into account, what Whittaker (1975) called « noise» 

in quantitative analysis was carefully studied. In par 

ticular, at the community level, patch dynamics (Levin 

& Paine, 1974; Whittaker & Levin, 1977; Connell, 

1978, 1979) has had a considerable impact on the un 

derstanding of the community functioning. According 

to this concept, a plant community is a mosaic of 

patches of differing successional stages. The existence 

of these successional stages would be caused by dis 

turbances. The hierarchical organisation of the com 

munity, or of the region (Alien, 1987; Kolasa, 1989; 

Pickett et al., 1989) develops this concept by consid 

ering (Kolasa's model) the habitat as hierarchically 

heterogeneous. For example, patches are made of 

smaller patches; the latter are made of even smaller 

patches. The species are distributed in these patches 

according to their specialisation level. The core 

satellite hypothesis (Hanski, 1982 , 1991) distin 

guishes core species - frequent and abundant - from 

satellite species - sparse and rare. A few abundant 

species are infrequent - urban species - and a few rare 

species are frequent - rural species. 

This body of theories (hierarchical organisation 

and core-satellite hypothesis) predicts that species 

which constitute the largest category would be the 

least frequent ones (to be found in less than 10 % of 

sites). Discrepancies appear for other maxima: Ko 

lasa's model (1989) predicts other peaks of species 

number which are weaker and weaker towards the 

highest frequencies (Figure IA), while Hanski's 

model (1982) predicts only a second maximum, for 

the most frequent species (Figure ID). These latter 

models loosely relate to gradients. Hanski (1982) 

stresses that her theory is applicable when sites are in 

similar habitats, but Collins et al. (1993) introduce an 

206 

environmental gradient (Figure 2) in their hierarchical 

continuum concept, a synthesis between the individu 

alistic hypothesis (Oleason, 1917, 1926), the hierar 

chical structure ofcommunity (Kolasa, 1989), and the 

core-species hypothesis (Hanski, 1982, 1991). 

The purpose of this study is to appreciate the effect 

of sampling scale on predictive and explanating value 

of various models at the level of the community. The 

impact of scaling is examined through two different 

problems: 

- the detection of ecological gradients through the 

means of Correspondence Analysis (CA); 

- the validation of one of the two contradictory 

models: core-satellite hypothesis or Kolasa' s model. 

About the first point, the area-richness curves 

show the key-role of the area investigated for meas 

uring richness (e.g. Arrhenius, 1921; Connor & 

McCoy, 1979; Williamson, 1988; Rey Benayas et al., 

1999). Then one must consider how the census of new 

species influences the detection of ecological gradi 

ents, as sample size increases. According to the hierarchical 

organisation theory, patches of various sizes 

could generate changing interpretation for gradients 

by changing scale. Several cases can occur: (i) the 

detected ecological gradients are the same at all the 

scales, in the same order of importance; (ii) they are 

the same but in different orders; (iii) they are not the 

same. The ordination of samples on axes must also be 

considered. It can vary as the size of the sample var 

ies. 

The second point has been studied by varying the 

level of organisation. Hanski (1982, 1991) originally 

put forward the core-satellite hypothesis for regional 

distribution patterns, then Ootelli & Simberloff 

(1987), Collins & Olenn (1990) proved that commu 

nity-level data and small-scale study data supported 

this model. But what about scaling at the same level 

(here the community)? Coliins & Olenn (1990) le 

gitimately suspect a strong influence of scaling on fre 

quency measures, but this must be tested. 

METHODS AND STUDY SITE 

Study site 

The study was carried out in a Mediterranean 

limestone grassland. This grassland lies on the western 

ridge of the Massif du Ventoux (Provence, France) at 

a height of 835 m. 

ec%gia mediterranea 25 (2) - 1999

Madon et al. Influence ofstudy scale on the characterisation ofplants community organisation 

Axis I Axis 2 Axis 3 Species number 

Quadrats A 14,65 11,36 10,01 56 

Quadrats B 18,16 12,40 10,14 61 

Quadrats C 17,21 11,12 9,98 76 

Quadrats D 16,07 11,46 11,18 86 

Table I. Relative inertia partition on the three first axes of CA for the four sampling scales 

Positive contributions Negative contributions 

A B C D A B C D 

Petrorhagia prolifera 99 Argyrolobium zanonii 43 

Centaurea IJanicu/ata 58 Teucrium montanum 38 36 49 

Koeleria vallesiana 62 Fumana procumbens 50 38 29 37 

Crupina vulr;aris 38 52 40 Genista hispanica 51 42 74 81 

Trinia Rlauca 82 69 38 31 Leuzea conifera 33 28 

Cerastium arvense suffruticosum 100 74 53 46 Aster sedifolius 56 44 

Echium vu/gare 40 Ame/anchier ovalis 54 40 

Lactuca perennis 60 Quercus humilis 74 

Anthericum /i/iaRo 54 31 Dactvlis f!lomerata 28 

Bupleurum baldense 51 35 34 Potentilla hirta 62 

Sedum acre 57 32 

Helianthemum nUl7lmu/arium 41 

Asperula cynanchica 38 

Arl7leria arenaria 72 74 

Genista X martinii 39 

Table 2. Highest plant species contribution for axis I to the different scales 

Quadrats A B C D 

Axes 1 2 1 2 1 2 1 2 

A I 1.000 

2 0.033 1.000 

B 1 0.758 0.399 1.000 

2 -0.104 0.543 0.134 1.000 

C I 0.532 0.290 0.681 0.196 1.000 

2 0.176 0.080 0.110 -0.237 -0.123 1.000 

D I 0.348 0.080 0.406 0.055 0.638 -0.185 1.000 

2 -0.021 0.015 0.018 0.015 -0.042 -0.001 0.068 1.000 

Table 3. Correlations between axes 1 and 2 on the different scales (correlations for axes I in bold) 

The positive pole is characterised for each scale by 

species of xeric grasslands (Thero-Brachypodietalia), 

in particular with some annuals (Petrorhagia prolif 

era, Crupina vulgaris, Bupleurum baldense). The 

negative pole is characterised for each scale by species 

which are less xerophytic and often belong to more 

mature communities. 

Axis I appears to represent a gradient linked to the 

water balance. The analysis of the plot-points (Figure 

I) indicates that their projection onto axis I of each 

CA is closely linked to their geographic arrangement 

along a north-south axis. The northern plots corre 

spond to the xeric pole. This gradient can be explained 

by the topography, the northern plots being more ele- 


vated than southern ones by about I meter. Alterna 

tively the mass effect (Shmida & Ellner, 1984) from 

the ridge itself should not be dismissed. The xeric spe 

cies on the ridge can scatter into the windward portion 

of the grassland, thus increasing the xericity gradient. 

Axis I leads to the same interpretations, irrespec 

tive of the scale of data collections. This result is in 

keeping with the correlations between the axes I, 

which are relatively strong (Table 3). These correla 

tions are discussed below. 

Axis 2 always presents a weak inertia (Table I), as 

regards the number of species. It has not been possible 

to interpret it at any scale. Additionally, axes 2 at the 

different scales show weak correlations between one 

209


another (Table 3), except between scales A and B. 

However, concerning CA of quadrats D, plots which 

are close in space are most of the time also close on 

factorial axis 2. Similarly, we were unable to interpret 

axis 3. 

These results indicate that only one physical gradi 

ent can be identified at the different scales we examined. 

The organisation of the plot-points is slightly 

different on axis I according to the size of the 

quadrats (plot 9 in particular shows an important move 

along axis I). 

Interpretation of species contributions by their pat· 

terns of distribution 

If we examine the species with a strong contribu 

tion for axis I (Table 2), we can make out 3 groups: 

Species characterising the poles (positive or nega 

tive) only on a fine scale (Petrorhagia prolifera, Ar 

gyrolobium zanonii... ). We hypothesise that these 

species are simply more scattered in one environment 

than another (at one geographic pole than another). In 

a limited sampling, th€y would tend to appear in one 

type of environment only, in such a way that they 

largely contribute to determine axis 1. On the con 

trary, when sampling scale increases, they would also 

appear in the other environment and their contribution 

would decrease, i.e. they would take a lesser part in 

the discrimination between one environment and the 

other. 

Species characterising the poles at all the scales 

(Cerastium arvense subsp. suffruticosum, Genista his 

panica... ). We suggest they are almost absent from 

quadrats of one geographic pole whatever the scale, at 

least at the selected scales. 

Species characterising the poles only on a large 

scale (Sedum acre, Amelanchier ovalis... ). These spe 

cies are exclusively present in one environment but 

they are represented by individuals or groups of individuals 

quite sparse in this environment. Conse 

quently, they tend to appear only in larger sample size 

only. 

The analysis of species presence-absence data 

among the plots at the different scales enabled us to 

clearly see the interpretation proposed for each point 

above. A few species have a high contribution only on 

middle scales (Echium vulgare, Asperula cynanchica... 

). That means that their distribution is linked to 

the gradient, but that the frequency gap from one pole 

210 

to the other is of relatively little importance. Let us 

stress the fact that axis 1 retains its meaning across 

scales. Table 3 shows, however, that the correlations 

of axes 1 between one another are better between two 

successive scales. 

Interpretation of plot-points distribution in factorial 

planes by species patterns 

The organisation of the plot-points in the factorial 

planes 1-2 is very characteristic (Figure 1): the active 

point of a group of 4 plots (the other 3 being non ac 

tive) clearly shows a trend to be the furthest from the 

origin, whatever the CA. For an active point repre 

senting a quadrat of a given size, we can underline: 

The smallest quadrats (non-active) tend to be 

placed closer to the origin for they contain less often 

the species with a strong contribution. Indeed, we 

verified above that some of these species don't occur 

in most of the small quadrats because they are too 

scattered. 

The largest quadrats (non-active) also tend to come 

closer to the origin for, on the contrary, they all tend 

to contain some species with a strong contribution. 

Indeed, some of the « discriminating» species at the 

concerned scale are no more «discriminating» at a 

larger scale (see above). 

Patterns of species frequency 

We have represented in histograms the distribution 

of the taxa among classes of frequency (presences in 

the plots) (Figure 4). It appears clearly that on the 

smallest scale (quadrats A), collections are dominated 

by scarce species (lowest frequency). Secondary 

maxima are visible, but their distribution depends on 

the classes of frequencies used. When the observation 

scale goes up, a second peak emerges in the class of 

most frequent species. Thus, in larger squares, there is 

a large proportion of species present in one quadrat 

only, and a large proportion of species present in all 

the quadrats. 

On all the scales, there is a very good positive cor 

relation between abundance and frequency of species 

(Figure 5; quadrats A: r=0.81; quadrats B: r=0.79; 

quadrats C: r=0.81; quadrats D: r=0.82; for each test 

p


This property is consistent with expectations of 

Hanski's (1982) and Kolasa's (1989) models. 

DISCUSSION 

Detection of the gradients 

In our data, only one ecological factor was de 

tected, and it always corresponds to axis 1 whatever 

the study scale. In these data set, a small grain is thus 

sufficient to detect the ecological gradient. At the ex 

amined scales, the organisation of the vegetation in 

patches (Whittaker & Levin, 1977) does not influence 

strongly the detection of a gradient whatever the scale. 

Nevertheless, correlations between axes 1 are weaker 

and weaker as the difference of scale increases. This 

underlines the risks there are to introduce different 

size plotting in a multi-dimensional analysis. 

In our study, the organisation of the plot-points 

which is slightly different on axis 1 according to the 

size of the quadrats can be linked to the existence of 

patches. These moves correspond to a mosaic organi 

sation of the vegetation. In any case, patches of vari 

ous sizes are visible in the physiognomy of the 

community. For example, some Crassulaceae and an 

nuals are confined in microhabitats of 1 to 3 dm 2 • 

These microhabitats cannot be detected with the grain 

of resolution we have used. They are linked to distur 

bances and stress (Madon & Medail, 1997), and life in 

them is very difficult because of ranges of tempera 

ture, drought and the possible accumulation of litter 

(Fowler, 1988; Bergelson, 1990; Ryser, 1993). They 

are more close at the north part of the grassland than 

at the south part. 

Models of distributions offrequency 

Figure 4 clearly shows that, on a small scale, the 

results are consistent with Kolasa's model (1989): one 

mode for infrequent species and other secondary 

modes weaker and weaker towards classes of strong 

frequency. On the contrary, towards the large scales, 

the results become consistent with Hanski's model 

(1982) for the communities: two important modes for 

the two extreme classes of frequency. This result 

shows that even at the same level of organisation 

212 

(within the community), simply by changing the study 

scale, data can support one or the other model. 

On a small scale (quadrats of 1 m 2 ), Collins & 

Glenn (1990) found a second maximum in the group 

of high frequency. But in this study, the quadrats were 

adjacent: numerous species have a strong probability 

to occur in most of the quadrats. Indeed, the higher 

similarity of geographically close plots has been ac 

knowledged for a long time (Curtis, 1959; Whittaker, 

1972; Barbour et aI., 1980). Shmida & Ellner (1984) 

attribute this feature to mass effect, but recognise that 

the existence of cryptic gradient is not to be dismissed. 

Williams (1950) and Coilins and Glenn (1990) 

suggested the possibility of a « saturation» of large 

quadrats in a great number of species. Figure 5 

(quadrats D) clearly confirms this suggestion: the spe 

cies plotted on a small scale tend to move towards the 

right on the large scale graph; therefore a certain 

number is present in all the quadrats (


frequency 

among 

quadrats .& 

100% :-,1"..- saturation 

"""" _ 

quadrats A 

quadrats B 

quadrats C 

quadrats D 

gradient 

Figure 6. Influence of quadrat size on the character core or satellite of a species. The size of quadrats increases from A to D. At a 

small scale (quadrats A), the species can be unfrequent (satellite or urban) and at a larger scale (quadrats D), the species can saturate 

quadrats (it is core or rural). 


We are grateful to M. Roux for his support during 

the study, P. Roche & T. Tatoni for their detailed re 

view and two anonymous reviewers for their correc 

tions. 

REFERENCES 

Alien T. F. H. 1987. Hierarchical complexity in ecology: a 

non-euclidean conception of the data space. Vegetatio, 

69: 17-26. 

Arrhenius O. 1921. Species and area. J. Ecol., 9 : 95-99. 

Barbour, M. H., Burk, 1. H. & Pitts, W. D. 1980. Terrestrial 

Plant Ecology. Benjamin Cummings, Men10w Park, 

California, U.S.A. 

Bergelson, 1. 1990. Life after death: site pre-emption by the 

remains of Poa annua. Ecology, 71: 2157-2165. 

Clements, F. E. 1916. Plant succession: an analysis of the 

development of vegetation. Carnegie Inst. Wash. Pub!. 

n° 242. 

Clements, F. E. 1936. Nature and structure of the climax. J. 

Ecol., 24: 552-584. 

Collins, S. L. & Glenn, S. M. 1990. A hierarchical analysis 

of species' abundance patterns in grassland vegetation. 

Am. Nat., 135: 633-648. 

Collins, S. L., Glenn, S. M. & Roberts, D. W. 1993. The hierarchical 

continuum concept. J. Veg. Sci., 4: 149-156. 

Connell, 1. H. 1978. Diversity in tropical rain forests and 

coral reefs. Science, 199: 1302-1310. 

Connell, 1. H. 1979. Tropical rain forests and coral reefs as 

open non-equilibrium systems. In: Anderson R. M., 

Turner, B. D. & Taylor, L. R. (eds.) Populations Dynamics: 

20th Symposium ofthe British Ecological Society, 

B1ackwell, Oxford: 141-163. 

Connor E.F. & McCoy ED. 1979. The statistics and biology 

of the species-area relationship. Am. Nat., 113 : 791-833. 

Curtis, 1. T. 1959. The vegetation of Wisconsin. Univ. Wisconsin 

Press, Madison. 

Fowler, N. 1988. What is a safe site?: neighbor, litter, germination 

date and patch effects. Ecology, 69: 1731 

1740. 

Frontier S. & Pichod-Viale D. 1991. Ecosystemes: structure, 

fonctionnement, evolution. Masson, Paris. 

Gleason, H. A. 1917. The structure and development of the 

plant association. Bull. Torrey Bot. Club, 44: 463-481. 

Gleason, H. A. 1926. The individualistic concept of the plant 

association. Bull. Torrey Bot. Club, 53: 7-26. 

Gotelli, N. 1. & Simberloff, F. D. 1987. The distribution and 

abundance of tallgrass prairie plants: a test of the coresatellite 

hypothesis. Am. Nat., 130: 18-35. 

Hanski, I. 1982. Dynamics of regional distribution: the core 

and satellite hypothesis. Oikos, 38: 210-221. 

Hanski, I. 1991. Single-species metapopulation dynamics: 

concepts, models and observations. Bioi. 1. Linn. Soc., 

42: 17-38. 

Kerguelen, M. 1993. Index synonymique de la flore de 

France. Museum National d'Histoire Naturelle, Paris. 

Kolasa, 1. 1989. Ecological systems in hierarchical perspective: 

breaks in community structure and other consequences. 

Ecology, 70: 36-47. 

Levin, S.A. & Paine, R.T. 1974. Disturbance, patch formation 

and community structure. Proc. Nati. Acad. Sci. 

U.SA, 71: 2744-2747. 

Madon, O. & Medail, F. 1997. The ecological significance 

of annuals on a Mediterranean grassland (Mt Ventoux, 

France). Plant Ecol., 129: 189-199. 

Magurran, A. E. 1988. Ecological diversity and its measurement. 

Croom Helm, Princeton Univ. Press, Princeton.Pickett, 

S.T.A., Kolasa, 1., Armesto, 1.1. & Collins, 

S.L. 1989. The ecological concept of disturbance and its 

expression at various hierarchical levels. Oikos, 54: 129 

139. 

Rey Benayas, 1.M., Colomer, M.G.S. & Levassor, C. 1999. 

Effects of area, environmental status and environmental 

variation on species richness per unit area in Mediterranean 

wetlands. J. Veg. Sci., 10 : 275-280. 

Ryser, P. 1993. Influences of neighbouring plants on seedling 

establishment in limestone grassland. 1. Veg. Sci., 4: 

195-202. 

ecologia mediterranea 25 (2) - 1999 213


Shmida, A. & Ellner, S. 1984. Coexistence of plant species 

with similar niches. Vegetatio, 58: 29-55. 

Ter Braak, C.l.F. 1987. Ordination. In: longman, R.H.G., 

Ter Braak, C.l.F., & Van Tongeren, O.F.R. (eds.) Data 

analysis in community and landscape ecology. Pudoc, 

Wageningen: 91-173. 

Whittaker, R. H. 1951. A criticism of the plant association 

and climatic climax concepts. Northw. Sci., 25: 17-31. 

Whittaker, R. H. 1967. Gradient analysis of vegetation. Bioi. 

Rev., 42: 207-264. 

Whittaker, R. H. 1972. Evolution and measurement of species 

diversity. Taxon, 21: 213-251. 

214 

Whittaker, R. H & Levin, S. A. 1977. The role of mosaic 

phenomena in natural communities. Theor. Pop. Bioi., 

12: 117-139. 

Williams, C. B. 1950. The application of the logarithmic 

series to the frequency of occurrence of plant species in 

quadrats.1. Ecol., 38: 107-138. 

Williamson M. 1988. Relationship of species number to 

area, distance and other variables. In : A. A. Myers & 

P.S. Giller (eds.). Analytical biogeography. Chapman & 

Hall, London: 91-115. 

ecologia mediterranea 25 (2) - /999

Plant systematics: a phylogenetic approach 

W.S. JUDD, C.S. CAMPBELL, E.A. KELLOGG & P.P. TEVEN 

Sinauer associates, Inc., 464 p. (1999) 

Ce livre est une synthese de 

tous les travaux recents en matiere 

de phylogenie ainsi que de leurs 

applications directes dans la systematique 

des plantes. Le resultat 

est vraiment exceptionnel, compte 

tenu de la multitude des travaux 

ponctuels actuellement publies, 

tant d'un point de vue moleculaire 

que morphologique (cette derniere 

approche retrouvant un certain essor 

du fait des limitations de la 

precedente). Ce Iivre se presente 

comme une sorte de precis de systematique 

moderne articule en 

deux parties. Le lecteur se familiarisera 

dans un premier temps avec 

les techniques generales de la 

phylogenie, avant de decouvrir les 

clades degages de ces approches. 

Nous commen

tyledones et des Laurales avec 

d'autres groupes longtemps consideres 

comme basaux au sein des 

Angiospermes. Citons egalement 

I'admission de la plupart des ex 

Scrophulariaceae au sein des 

Plantaginaceae, ou des Anthirrinaceae, 

seIon les avis, ainsi que la 

fragmentation des Liliales - Liliaceae 

en un grand nombre de families 

et genres. 

Ces nombreux exemples eloignes 

des croyances classiques peuvent 

evidement donner naissance a 

certaines critiques plus ou moins 

legitimes. Quelle valeur peut-on 

vraiment accorder aux groupes 

Ghilem MANSION 

Institut de Botanique de Neuchatel (Suisse) 

consideres dans ce livre, dans la 

mesure ou n'ont ete envisagees que 

des approches partielles, avec des 

genes et des caracteres classiques 

variants selon les etudes? 

Nous pouvons conclure avec 

les propos des auteurs eux-meme, 

conscients des limites de cet ouvrage: 

"Les etudiants apprecieront 

{'abandon des rangs classiques... 

Avec ces bases, ils pourront avancer 

de maniere sure dans la comprehension 

de la diversite des 

plantes" ou encore "...dans ce livre, 

nous pensons que certains des 

groupes decrits seront peut-etre 

Biology and Wildlife of the Mediterranean Region 

J. BLONDEL & J. ARONSON 

Oxford Univ. Press, Oxford, 328 p. (1999) 

S'engager dans la redaction 

d'un ouvrage sur la biodiversite du 

monde mediterraneen, etait a la 

fois courageux et ambitieux; les 

auteurs le reconnaissent, comme 

ils indiquent en preface avoir volontairement 

axe leurs discussions 

et illustrations sur les questions et 

les groupes qui s'inscrivaient dans 

le domaine de leurs activites, tout 

en tentant de rester clair et d'eviter 

un jargon hermetique aux nonspecialistes. 

11 faut reconnaitre 

qu'ils y sont parvenus, et que cet 

ouvrage se lit facilement et avec 

interet. Je pense que le but souhaite 

par les auteurs est atteint, et 

qu'en quelques 300 pages, ils sont 

arrives a dresser un bilan clair et 

quasi-exhaustif des problemes relatifs 

a la biodiversite du monde 

mediterraneen. 

C'est volontairement que les 

auteurs restent sur des thematiques 

216 

souvent generales, et, ils le soulignent, 

au niveau de la semivulgarisation, 

car ils ne veulent 

nullement faire oeuvre de specialistes. 

Ceci assurera certainement 

une plus large diffusion a cet ouvrage, 

meme si le scientifique reste 

parfois sur sa faim. 

Les bilans et les exemples 

choisis illustrent clairement les caracteres 

majeurs du monde mediterraneen, 

depuis son individualisation 

climatique et biologique, en 

passant par sa mise en place actuelle 

et ses traits de vie, et en 

s'ouvrant sur les challenges que 

posent son avenir. Le role de l'action 

humaine est fort justement 

souligne, comme l'indique joliment 

le titre du chapitre 8 " Humans as 

sculptors of mediterranean landscape". 

11 n'est question ici, ni de resumer 

un ouvrage extremement 

modifies, voire detruits, dans les 

annees avenir". 

Quoiqu'il en soit, malgre ces 

bemols, ce livre peut etre considere 

sans retenue comme une reference 

fondamentale pour le botaniste du 

prochain millenaire. 

Ajoutons une bibliographie 

pour le moins monumentale, a la 

fois tres recente et complete, qui 

justifie a elle seule I'achat de ce 

livre, deux appendices tres interessants 

et un CD Rom richement illustre 

et nous obtenons une bible 

dans le domaine, avec en prime un 

excellent rapport qualite/prix. A se 

procurer absolument ! 

dense, ni de le critiquer, car je ne 

puis que souscrire, dans leurs 

grandes lignes, et souvent dans le 

detail, aux conclusions des auteurs 

qui, contrairement a un certain 

nombre de leurs predecesseurs, 

donnent une idee claire et juste des 

criteres fondamentaux du monde 

mediterraneen et de sa diversite. 

Cependant, je ne puis que regretter 

quelques oublis ou imprecisions. 

C'est ainsi, par exemple, que 

si definir les limites de la region 

mediterraneenne pose bien des 

problemes, mais reste necessaire, 

les auteurs ne contribuent guere a 

regler cette question. lis proposent 

dans les figures lA, 1.5 et 1.6, des 

solutions variables, excluant le S-E 

tunisien pour la premiere, et incluant 

toutes les bordures meridionales 

de la mer Noire qui en fait 

s'integrent dans la zone euxinienne 

(pontique) a Fagus orientalis et 


Rhododendron ponticum dont le 

c1imat et la flore n'ont rien de mediterraneen. 

La carte de la figure 

1.6 dont je recuse toute paternite, 

comme cela est indique dans la legende, 

est la plus critiquable. Par 

exemple, la region saharo-arabe 

grignote une grande partie du 

Maghreb, dont les Atlas, alars que 

la province pontique cette fois englobe 

les Hauts Plateaux anatoliens 

et s'etend jusque sur les rivages de 

la Cilicie. De meme, I'idee des 

quatre quadrants (fig. lA) dans le 

monde mediterraneen est interessante, 

mais les limites proposees 

restent critiquables, surtout pour la 

peninsule italienne et ses lies annexes, 

ecartelees entre trois d'entre 

eux. 

Du point de vue historique, 

mon passe helas reduit et lointain 

de biospeleologue me fait toutefois 

regretter que cette facette n'ait pas 

ete analysee, pour rendre compte 

de certaines mises en place pre 

Miocene, a partir de la repartition 

d'especes cavernicoles sur lesquelles 

existe une bonne documentation. 

Parmi les "habitats" caracteristiques 

du monde mediterraneen, 

un bon bilan a ete dresse, encore 

que la distinction entre foret 

sclerophylle et foret laurifoliee aurait 

gagne aetre etoftee et discutee 

en raison de leur signification 

ecologique fort differente. De 

meme, les mares tranSltOlres, un 

des joyaux du monde mediterraneen, 

selon le mot de Braun 

Blanquet, avec les adaptations remarquables 

de leur faune et de leur 

flore, ne sont pas evoquees. 

Au niveau des traits de vie, les 

donnees relatives a I'origine de la 

sc1erophyllie et de son rOle dans le 

c1imat mediterraneen, de meme 

que le probleme du haut pourcentage 

de therophytes dans la flore, 

sont analysees, mais je demeure 

pour ma part sur mes incertitudes, 

quant aux explications et aux resultats 

exposes. Apropos du role 

et des modes de dissemination des 

especes et en particulier des diaspores, 

I'exemple des lies macaronesiennes 

(Canaries et Madere), 

aurait gagne aetre developpe. 

Les auteurs, a juste titre, ont 

mis I'accent sur les donnees historiques 

et le role de la region mediterraneenne 

en tant que capital 

biologique de base pour l'ec1osion 

des civilisations neolithiques. lis 

montrent une culture litteraire et 

histarique qui merite d'etre soulignee, 

et ont rendu en particulier a 

Theophraste, un juste hommage en 

tant qu' "inventeur" de la specificite 

du monde mediterraneen; 

Pierre QUEZEL 

Institut Mediterraneen d'Ecologie et de Paleoecologie, Marseille 

LOUIS-JEAN 

avenue d'Embrun, 05003 GAP cedex 

Tel. : 04 92 53 17 00 

Depot legal : 94 - Fevrier 2000 

Imprime en France 

mais pourquoi alars ne pas aVOlr 

au moins evoque son remarquable 

chapitre sur la biologie du figuier. 

.. 

Meme si l'on pourrait citer encore 

quelques imprecisions ou 

quelques oublis, il n'en reste pas 

moins que le travail de BLONDEL 

& ARONSON, restera pour longtemps, 

une reference et une mine 

de renseignements, pour tous ceux 

qui aiment la region circummediterraneenne, 

et etudient sa 

faune et sa flore. Ce n'est pas I'un 

de leur moindre merite, que d'avoir 

su regrouper et integrer a la fois 

des groupes animaux et vegetaux, 

parfois fort differents quant aleurs 

significations histariques et a leurs 

reactions face aux contraintes 

ecologiques mediterraneennes. Les 

auteurs ont donc su dresser un bilan 

coherent, tout en soulignant la 

richesse biologique incomparable 

du monde mediterraneen, et le role 

qu'il a joue dans l'eclosion de nos 

civilisations. Ils terminent fort 

justement en soulignant les gravissimes 

menaces de tous ardres qui 

pesent sur lui et I'assaillent de 

toutes parts, et qui font certainement 

de cette region, a l'heure actuelle, 

une des plus menacees de la 

planete, "a microcosm of world 

problems". 

ec%gia mediterranea 25 (2) - /999 217

Ecologia Mediterranea publishes original research report and 

reviews in the fields of fundamental and applied ecology of 

Mediterranean areas. except descriptive articles or articles ahout 

systematics. The editors of Ecologia Mediterranea invite 

original contrihutions in the fields of: bioclimatology. 

biogeography, conservation biology. population hiology. 

genetic ecology. landscape ecology. mierohial ecology. vegetal 

and animal ecology, ecophysiology, palaeoecology, 

palaeoclimatology. but not marine ecology. Symposium 

proceedings, review articles, methodological notes. hook 

reviews and comments on recent papers in the journal are also 

puhlished. Manuscripts arc reviewed hy appropriate referces. or 

hy members of the Editorial Board, or by the Editors 

themselves. The final decision to accept or reject a manuscript 

is made by the Editors. Please send 3 copies of the manuscript 

to the editors. When an article is accepted, the authors should 

take the reviewers' comments into consideration. They must 

send back to the journal Editorial Office their corrected printed 

manuscript (onc copy) and include the corresponding floppy 

disk (as far as possible: 3.5" PC, Word 7 or .RTF) within 3 

months. The authors arc asked to check the conformity between 

printed and computerised versions. Enclose the original 

illustrations. Corrected proofs must he returned to the journal 

Editorial Office without delay. Books and monographs to he 

reviewed must be submitted to the Editors. 

Manuscript preparation 

Manuscripts (typewritten with double spacing and A4 paper 

size) should preferahly be written in French or English. but 

Spanish and Italian are accepted. If the language is not English. 

you should include an English short version and English titles 

for figures and tables. 

The manuscript must bc complete: French and English titles, 

author(s) and address(es). French and English abstracts (at 

least), an English short version (only if it is not the language 

used in the article), key-words. text. rcrerences. 

acknowledgements, figures and tahles. For research papers, the 

text should normally consist of 4 sections: introduction. 

methods, results and discussion. 

When typing the manuscript, please distinguish titles clearly 

from other paragraphs. Titles and subtitles should not he 

numbered. Avoid using letters to number subtitles. Use lowercase 

letters for names. Do not underline any words. In English. 

there is one space after punctuation. none before. 

Copy editing of manuscripts is performed by the journal. 

Each author's address should be specified. The first time. please 

state the complete address of the correspondent author to whom 

the proofs should be sent. 

Abstract, key-words, abridged version 

Ahstracts should be no longer than 300 words. The English 

abridged version should not exceed onc page (1000 words). Do 

not use more than six key-words (translated in the abstraet's 

language). Key-words should not be present in the title. 

Instructions to authors 

References 

All puhlications quoted in the text should be presented in a list 

of references following the text. The list of references should be 

arranged alphahetically hy author and chronologically per 

author. You should abbreviate the titles of periodicals in the list 

of references (except if you arc not sure of it). Make sure that all 

citations and references correspond. Use the following system 

for references: 

- journal article: 

Andow D.A.. Karieva P.. Levin S.A. & Okubo A.. 1990. Spread 

of invading organisms. .J. Ecol.. 4 : 177-188. 

- book: 

Harper J.L., 1977. Population hiology of' plants. Academic 

Press, London. 300 p. 

- book section: 

May R.M., 1989. Levels of organisation in ecology. In : Cherret 

J.M. (cd.). Ecological concepts. Blackwell Scientific Public.. 

Oxford: 339-363. 

- conference proceedings: 

Grootaert P., 1984. Biodiversity in insects, speeiation and 

behaviour in Diptera. In: Hoffmann M. & Van del' Vekcn P. 

(eds.), Proceedings oj'the sVl11posiul11 on " Riodiversitv: studv, 

exploration, conservation ". Ghcnt, 18 November 1992: 121 

141. 

Citations in-text 

The words « figures» and « tables» announced in-tcxt should 

he written in extenso and in lower-case letters. In the text. refer 

to thc author's name and year of publication (followed by pages 

only if it is a quotation). If a publication is written by more than 

two authors. the name of the first author should be used 

followed hy « et al. » (this indication, however, should never be 

used in the list of references: first author and co-authors should 

be mentioned). Examples: «Since Dupont (1962) has shown 

that. .. ». or « This is in agreement with previous results (Durand 

et Ill., 1990 ; Dupond & Dupont, 1(97) ... ». 

Abbreviation, Latin words 

Explanation of a technical abbreviation is required when first 

used. Intcrnational convention codes for nomenclature should 

be used. Latin words should he in italics (et al.. a priori, ete.). 

particularly for taxonomic classifications (the first time, please 

state the author's name: for example. Olea elfropllea Linnee). 

Figures and tables 

All illustrations should he submitted separately and they should 

be preceded by thc figure and table legends on a separate page. 

Figures and tables should he sent «rcady to be printed ». 

without need to be reduced (so their size should be 16 x 22 cm 

or 8 x 22 cm maximum). All the illustrations being in-text 

should he quoted, numbered in sequence and should have a 

legend. Computerised tables' columns should not be represented 

by signs ( : or I ). 

Reprints 

Twenty-five reprints will he supplied frce of charge. 8y request. 

additional reprints can be orclered with a charge.

ecologia mediterranea 

SOMMAIRE 

Tome 25 fascicule 2,1999 

J. A. TORRES, A. GARCIA-FUENTES, C. SALAZAR, E. CANO & F. VALLE - Caracterizaci6n de 

los pinares de Pil1l1s halepellsis Mill. en el sur de la Peninsula lberica 135 

Y. B. LlNHART. L. CHAOUNI-BENABDALLAH, J.-M. PARRY & J. D. THOMPSON - Selective 

herbivory of thyme chemotypes by a mollusk and a grasshopper 147 

M. VfLA & I. MUNOZ - Patterns and correlates of exotic and endemic plant taxa in the Balearic islands 153 

A. AIDOUD, H. SLLMANI, F. AlDOUD-LOUN1S & J. TOUFFET - Changements edaphiques le long 

d'un gradient d'intensite de piiturage dans une steppe d'Algerie 163 

M. ABD EL-GHANI - Soil variables affecting the vegetation of inland western desert of Egypt 173 

R. BESSAH, F. ROZE & D. NEDJRAOUI - Activite cellulolytique ill vilro de sols de deux steppes it 

alfa (Slipa tellllcissima L.) d' Algerie 185 

J. GUIOT, F. TORRE, R. CHEDDADJ. O. PEYRON, P. TARASOV, D. JOLLY, J.O. & KAPLAN 

The climate of the Mediterranean Basin and of Eurasia of the last glacial maximum as reconstruclcd by 

inverse vegetation modelling and pollen data 193 

O. MADON, S. GACHET & M. BARBERO - Innuence of study scale on the characterisation of plants 

community organisation in a Mediterranean grassland (Mont Ventoux, France) 205 

Analyses d'ouvrages 215 

Indexe dans BIOSIS et PASCAL (CNRS)

Ecologia mediterranea 1999-25(2)

Create successful ePaper yourself

Delete template?

Save as template?