Sunflecks and Their Importance to Forest Understorey Plants ROBIN L. CHAZDON I. Summary II. Introduction: Sunflecks as,a Resource III. Measurement of Sunfleck Activity Beneath Forest Canopies . . A. Area-survey Techniques B. Instaneous Sensor Measurements C. Photographic Techniques IV. Sunfleck Activity in Temperate and Tropical Forests . . . . A. B. C. D. Defining Sunflecks Temperate Deciduous Forests Coniferous Forests Tropical Evergreen Forests . .- E. Summary: Generalizations About Sunfleck Activity . . . . V. Photosynthetic Responses to Sunflecks A. Sunflecks and Carbon Gain to Understorey Habitats B. Determinants of Sunfleck Utilization C. Constraints on Sunfleck Utilization in Understorey Habitats D. Sunfleck Regimes and Light Acclimation E. Photosynthesis in Understorey Plants Revisited 2 3 8 8 9 11 12 12 13 14 15 19 20 20 25 32 37 41 VI Seed Germination, Establishment and Growth in Relation to Sun fleck Activity 42 A. Seed Germination and Establishment in Understorey Habitats 43 44 B. Growth of Understorey Plants VII. The Influence of Sunflecks on Reproductive Behavior and Distribu tions of Understorey Species 48 A. Light Availibility, Size Variations and Reproductive Behavior 49 50 50 51 52 B. Vegative and Sexual Reproductive Effort C. Sunflecks, Canopy Gaps and Species Distributions . . . . D. Vertical Distribution of Understorey Species VIII. Conclusions advances in ecological research Vol. 18 ISBN 0-12-013918-9 Copyright © 1988 Acadcmic Press Inc. (London) Limited. rights ofreproduction in anyform reserved 2 SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON significance of sunflecks per se for growth, survivorship, and reproduction of A. The Importance of Sunflecks: Scaling Up From Leaves to Whole Plants B. Directions for Future Research 52 53 Acknowledgements 54 References 54 forest understorey plants. n. INTRODUCTION: SUNFLECKS AS A RESOURCE I. SUMMARY Sunfleck activity has profound effects on ecological processes ranging from The interest of light as an ecological factor arises partly from the great variety of influences which the light climate exercises upon individual organ isms, and also from the very complexity of the light climate itself, which has for long exercised a fascination over those who have been working in this field (Evans, 1966). photosynthesis to microsite distributions. Sunflecks occur when predomi Plants that live beneath forest shade inherit the remnants of the sun's rays nantly direct-beam radiation passes through openings in the forest canopy. after they have filtered through the forest canopy. Those wavelengths of In the forest understorey, sunflecks foster a high degree of spatial and shortwave radiation that are most strongly absorbed by canopy foliage layers temporal variation in light availability. Sunflecks may contribute more than are also the most highly prized by leaves in the understorey, for these are the photosynthetically active wavelengths. In manyforest types, less than 2% of the photosynthetically active radiation (PAR) incident above the canopy may actually reach the forest floor. How is it, then, that so many species of 50% of the daily photon flux density in the understorey of temperate and tropical forests. Although understorey species are usually able to maintain positive carbon balance in the absence of sunflecks, photosynthesis during sunflecks may account for 30-60% of daily carbon gain. Photosynthetic responses to sunflecks involve both short-term (dynamic) plants flourish in the dark recesses of forest undergrowth? and longer-term (induction) responses. Following induction, leaves respond more rapidly to sunflecks. Carbon gainand photosynthetic efficiency during Although we areonly beginning to understand in sufficient detail thelight relations of forest understorey species, it is clear that their ecological and evolutionary successes are largely due to their abilities to capitalize on sunflecks depend strongly on sunfleck duration. When sunflecks are shorter than 40 s, measured carbon gain is often greater than predicted carbon gain based on steady-state photosynthetic responses. This enhancement of photo These patterns come in many different forms, with differing degrees of predictability and exploitability. Spatial patterns of light availability on the patterns of variation in light, the major limitingresourcein most forest types. synthesis has been attributed to post-illumination COj fixation, which forest floor are fundamentally determined by the three-dimensional structure contributes a large proportion of total carbon gain during brief sunflecks, but of the forest, but spatial patterns continually change, according to time of day, season, and latitude. Superimposed on the highly complex spatial pattern are temporal patterns oflight fluctuation, which are affected by cloud cover, atmospheric conditions, and wind. Ifwe examine further the extent of only a small proportion during longer sunflecks. Under natural conditions, photosynthetic utilization of sunflecks may be hindered by a variety of factors including loss of induction during low-light periods, restricted light variation among leaves within a single crown, the added influences of stomatal opening to conserve water loss, photoinhibition, wilting, and high leaf temperatures. There is no evidence that the photosynthetic characteris tics responsible for efficient utilization of sunflecks impose any constraint on efficient utilization of low light. Some evidence does indicate, however, that crown structure and leafdisplay must be considered. Ecological studies of understorey plants need to be concerned with patterns of light variation in photosynthetic adaptation to high light limits photosynthetic efficiency these patterns. during sunflecks. natural habitats and reponses by leaves, individuals, and populations to Perhaps the most striking patterns of light variation within the under At light levels below 20% of full sun, light usually limits growth of understorey species. Accordingly, variation in sunfleck activity among understorey microsites has been correlated with differences in plant growth rates, size, sexual reproduction, and vegetative reproduction. The patchy storey are those created by sunflecks, shafts ofsunlight that penetrate small openings in the canopy (Fig. 1). For many years, plant ecologists and physiologists have recognized the importance of sunflecks to plants in the distribution of some understorey species has, in some cases, been linked to have researchers been able to measure light variation with sufficient resolu tion to evaluate critically its impact on physiological and ecological pro microsite differences in light availability. Integrated organismal responses to changing light conditions make it exceedingly difficult to quantify the forest understorey (Lundegardh, 1922; Evans, 1939). Only recently, however, cesses. In this review, I summarize, and attempt to synthesize, what is now SUNFLECKS AND UNDERSTOREY PLANTS 5 known about sunfleck activity and its importance to understorey plants in temperate and tropical forests. The temporal and spatial scale of light variation are key factors in evaluating the range of physiological and ecological processes affected (Tables 1and 2). Sunfiecks lasting from a few seconds to several minutes may affect photosynthesis rates, stomatal responses, leaf temperature and mor phogenesis, whereas variation in light availability on the scale of weeks to months may lead to differences in plant growth, morphology, survivorship, and reproduction (Table 1). Regardless of temporal scale, the spatial scale of light variation also has important consequences for the types of biological processes affected (Table 2). In this review, I examine the physiological and ecological consequences of sunfleck activity at different spatial and temporal scales. I consider two general classes of understorey plants: (1) species that complete their life-cycle in the forest understorey (shade tolerant herbs, shrubs, or small trees); and (2) seedlings and saplings of canopy tree species, which live in the understorey onlyduring the early stages of their life-history. Mature canopy trees, non-forest species, and agricultural crops, have, in Table 1 Physiological and ecological processes affected by light variation at different timescales with spatial location maintained constant. Adapted from Chazdon (1987). Time scale Process affected Seconds to minutes Transient photosynthetic responses, stomatal responses, leaf temperature, seed germination, Minutes to hours Induction of photosynthetic apparatus, stomatal morphogenesis responses, chloroplast movements, leaf temperature, leaf movements, seed germination, morphogenesis Hours to days Changes in photosynthetic capacity, stomatal responses, leaf phenology, seed germination, Days to weeks Changes in photosynthetic capacity and leaf biochemistry, leaf growth and morphology, plant growth, seedling establishment, survivorship, reproduction Photosynthetic acclimation, whole-plant growth, phenology, canopy structure, leaf morphology, biomass and nutrient allocation, seedling establishment, survivorship, reproduction morphogenesis, photopcriodic responses Weeks to months Months to years Phenology, leaf turnover, whole-plant growth, plant architecture, survivorship, reproduction, nutrient cycling 6 R. L. CHAZDON Table 2 Physiological and ecological processes affected by light variation at different spatial scales with time maintained constant. Adapted from Chazdon (1987). 0 1 c 3 O iC- f/i o -a Spatial unit Process affected c ea X Cells within leaf 3 CC Light scatteringand absorption, chloroplast Part of leaf Leaf o 7 §I B 3 2 5= 5 w .2 V movement, photosynthesis C *0 c Photosynthesis, translocation, stomatal density, photomorphogenesis, herbivory, energy balance Photosynthetic capacity, energy balance, leaf E? 0) .2 '-B c herbivory Crown Nutrient, water, and carbohydrate transport; shading among leaves, age structure of leaves, foliage •o O a Rj — 6 o* « .2 .2 Biomass allocation, establishment, growth, architecture, survival, reproduction, competition Plant population Ecotypic differentiation, age structure, population Communities Succession, regeneration, vegetation structure, species 3 (U u "5 "5 6 2 B ' o 0 •o 1 2 3 distribution, branching pattern Whole plant 4i ^ o s o = i nS 3. Id a. rt 3 e5 tn 2 6 5 3 tC C Ui u 0) 3 &o cS movement, leaf morphology, leaf orientation, competition for light, photomorphogenesis, "2 8 3 C rt <sj C oi u G^ •Ss. *-• 2 G ea •o CO Oi 3 0) o. cx z C/3 CO growth, recruitment B i ® diversity, nutrient cycling, hydrology C rri (N VO I t/T Co 00 T •S JS " X/) 00 t- H o B 3. general, been excluded from this review, although light variation certainly influences these species as well. Just as water and nutrients are environmental resources required in specific ways by different plants, so are sunflecks. If we consider sunflecks as a resource for understorey plants, we can then begin to assess the spatial and temporal distribution of the resource and physiological mechanisms of captureand utilization by different species. Different wavelengths within the radiation spectrum have unique realms of biological influence (Table 3). Total radiation affects leaf energy balance, whereas the ratio of red to far-red quanta controls photomorphogenetic processes mediated by phytochrome (Table 3). By studying sunflecks as a resource, their importance to specific physiological and ecological processes can be considered along with the effects of other critical plant resources. This is not to say that light is always perceived by understorey plants as either background (diffuse) light or as sunflecks. Furthermore, as discussed below, the utilization of sunflecks by leaves is not independent of leaf temperature, plant water status, nor plant nutrition. Rather, the value of regarding sunflecks as a resource lies in the usefulness of applying widely-accepted concepts of resource availability, utilization, and allocation, without implying any exclusive biological signifi cance of sunflecks. B 3. I n •5b G 3. O 3 3. o 00 O B B B g §• T 3 "~- 9 a o o "o •53 V o. c C Ji c o 8 I cd & u JS C c "S "O «1 o g- .S "C o &II D. 5 C >> U o « a> C C ea i: D. c g e 00 o •n o o c c/5 i« k. u I i CG O. B u 2 2 £ w ^ O .22 O Mia I ^ G G B 50 O -G o. a. o 6 g pC c o •a 2 CS ^ '"B G •c I 'S o e? cd u - 1e c4 U •B.2 G 0 .2 •3 S w 1 2 •o G 0 0 X! '•J3 Vl "rt 4-> 0 0 H 13.2 > nj 2 "eS u g (U c 0 H 2 '"B m <a c ^ ^ o1 -S) •T3 o Qi, a> O s o OQ 8 SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON 9 III. MEASUREMENT OF SUNFLECK ACTIVITY BENEATH FOREST CANOPIES a Singapore rainforest understorey, and by Grubb and Whitmore (1967) in lowland and montane rainforests in Ecuador. As noted by Evans (1956), this technique is limited by the spectral response of the photocell and shading by The solar radiation incident on the forest canopy is composed of two different forms: direct-beam radiation, and radiation diffused by the earth s the apparatus. Spatial distributions of sunflecks were also investigated by Miller and Norman (1971) and Norman et al. (1971), who proposed a technique for measuring sunfleck lengths using transect lines randomly atmosphere. These two forms penetrate the forest canopy in different ways (Anderson, 1964a; Reifsnyder et al., 1971; Hutchison and Matt, 1977). When thesunisshining, predominantly direct-beam radiation passes through holes in the canopy, dappling the forest floor and leaves with sunflecks (Fig. 1). In contrast, the diffuse sky radiation incident on the forest canopy penetrates all canopy openings, although not always equally. As both forms of radiation pass through vegetation, spectral quality is altered by selective absorption, transmission, and reflection of wavelengths by foliage, branches and boles. Sunflecks come in a wide variety of shapes, sizes, colors and durations; there is no such thing as a "typical" sunfleck, even within an intensely studied understorey microsite. The task of defining precisely what is and is not a sunfleck challenges even the most experienced ecologist. Contributing to the difficulty of describing sunfleck activity is the dual nature of sunflecks; they have both spatial and temporal dimensions. The spatial and temporal dimensions of sunflecks arriving at any particular location are related to the configuration of the forest canopy as seen from that location. Ideally, methods used to measure sunfleck activity should reflect the particular biological phenomenon being investigated. More often, however, definitions of the spatial and temporal dimensions of sunflecks have been limited by the techniques used to measure them. Three general methods have been used in the measurement of sunfleck activity in temperate and tropical forests: areasurvey techniques, instantaneous sensor measurements, and analyses of hemispherical photographs. drawn through a vegetation canopy. An area-survey technique was also used by Ustin et al. (1984) to measure sunfleck size and number in the understorey of a red fir forest. B. Instantaneous Sensor Measurements The use of photoelectric cells to measure instantaneous light conditions within forests has a long history (Atkins and Poole, 1926). It is perhaps within this realm of light measurement that the greatest progress has been made. Using portable data-loggers, it is now possible to record instantaneous light changes as frequently as every 0*01 s, obtain frequency distributions, averages, variances, and other statistical computations over any set of time intervals, and transfer these data directly to magnetic media, or to a microcomputer. A variety of light sensors have been used to measure the distribution of radiation within forests (Pearcy, 1988b; Table 3). These include photoelectric cells (Evans, 1939), net radiometers (Baldocchi et al., 1984), pyranometers (Reifsnyder et al, 1971), and PAR sensors (Biggs et al, 1971; Gutschick et al., 1985). Accordingly, radiation measurements may be made in units of energy (J), radiant flux (Js"' or W), flux density or irradiance (J m"^ s~' or Wm"^), or photonflux density (PFD; nmol m"^ s Table 3). Several important considerations apply to the use of instantaneous sensor measurements for sunfleck descriptions. Each kind of sensor has a character istic spectral and temporal response, which may or may not make it an A. Area-Survey Techniques The first detailed studies of sunfleck activity in forest environments focused on describing spatial distributions based on continuous sampling of irradiance at points on a pre-determined grid within study plots. Evans (1956) described an area-survey technique for measuring the distribution of sun flecks in a Nigerian rainforest. This technique permitted the computation of the areas of sunflecks of different flux density, as measured by a galvan appropriate choice for measuring sunfleck activity (Pearcy, 1988b). In studies of carbon gain during sunflecks, for example, sunfleck activity should be measured in units of photon flux density (Bjorkman and Ludlow, 1972), whereas studies of leaf temperature and energy balance require measure ments of net radiation (Woodward, 1981; Table 3). Moreover, the sampling interval determines the level of sunfleck activity that can be measured. Sunflecks less than 10 s long are not accurately sampled using a sampling interval of 10 s (Chazdon and Fetcher, 1984a; Fig. 2). If significant sunfleck ometer connected to a photoelectric cell. Evans then converted the area scale activity is overlooked because sampling intervals are long relative to sunfleck to an appropriate time scale to estimate the incidence of sunflecks of different flux density on an average day (Evans, 1956). The area-survey technique was subsequently used by Whitmore and Wong (1959) and Evans et al. (1960) in duration, substantial errors in calculating meanor integrated light levels can arise. These errors can be in the positive or negative direction. In the case of the 2-min sequence shown in Fig. 2, the average PFD computed from SUNFLECKS AND UNDERSTOREY PLANTS 11 observations sampled every 10 s was 10% lower than the average computed from 2-s samples. The exclusive use of instantaneous measurements to obtain integrated measurements over longer time periods leads to a substantial loss of information, such as the length and spacing of sunfleck intervals, and peak light intensities during sunflecks. Similarly, although light integrators (Woodward and Yaqub, 1979) provide a cumulative measure of PAR that is correlated with total sunfleck activity, they do not provide information on 500 400- 300- individual sunfleck periods. Furthermore, a single light sensor cannot provide information on the contribution of direct-beam radiation from 200- sunflecks. To measure the relative contributions of direct and diffuse solar radiation during sunflecks, a pair of sensors is required, one with a shadow- I CO CM band that obscures direct radiation (Horowitz, 1969). 100- C. Photographic Techniques In 1924, Robin Hillinvented and constructed a special camera to be used for observations of clouds. This "fish eye" camera was thefirst to becapable of recording an image of a complete hemisphere on a flat plate(Hill, 1924). The o E =L (/) C 500 (D o camera was borrowed by Evans and Coombe (1959), who found it most X U. useful for photographing forest canopies. By orienting the photograph properly and then superimposing a transparency marked with solar tracks, 400- c they were able to see whether direct sunlight could reach the forest floor at any particular time. In 1964 hemispherical photographs were first used to estimate quantitatively the light conditions under forest canopies. Anderson (1964a) pioneered this technique, proposing a method for computing the o -<-• o x: Q. 300- percentage of diffuse light in the open received under the canopy (diffuse site factor) based on hemispherical photographs. She also estimated direct site factors using solar track diagrams to score, hour by hour, the percentage of 200- direct light that could potentially reach the forest floor. 100- 120 Time (s) Fig. 2. Instantaneous measurements of photon flux density (PFD; ^mol m"^ s"') during naturally-occurring sunflecks at (a) 2-s time intervals and (b) subsampled Hemispherical photographs have limited utility for predicting precise sunfleck activity because of variation in weather conditions, transmittance of light through foliage, and penumbral effects (Anderson, 1966; Norman et al., 1971; Anderson and Miller, 1974; Salminen et al., 1983; Chazdon and Field, 1987b). Nevertheless, as an indicator of potential sunfleck duration, they have proven very useful in a wide range of ecological studies of understorey plants (Pearcy, 1983; Ustin et al., 1984; Orozco-Segovia, 1986: Walters and Field, 1987; Chazdon and Field, 1987a; Rich et al., 1987). In contrast to every 10 s. Two-min average PFD computed from 10-s readings was 10% lower than sensor measurements, hemispherical photographs provide a means of assess that computed from 2-s readings. ing Ught conditions over a relatively long period of time (weeks to months) and for a large number ofplants. They also offer the advantage ofestimating diffuse and direct components of the light environment separately. Several 12 R. L. CHAZDON computerized techniques are now being used to analyze hemispherical photographs (Jupp et aL, 1980; Chan et al, 1986; Chazdon and Field, 1987b). Many types of"fish-eye" lenses are currently available (Evans etal., 1975). Photographic analyses should account for lens distortion (if any), areal projection ofa hemispherical image (Herbert, 1988), and leaf angles (Chazdon and Field, 1987a). IV. SUNFLECK ACTIVITY IN TEMPERATE AND TROPICAL FORESTS A. Defining Sunflecks Sunflecks defy generalization. Nevertheless, in order to describe measure ments of sunfleck activity, somegeneral criteria must be established. Because sunflecks are caused by the penetration of predominantly direct-beam solar radiation through openings in the forest canopy, the spectral quality of sunflecks differs from that of diffuse shade light (Coombe, 1957; Federer and Tanner, 1966; Chazdon and Fetcher, 1984b; Lee, 1987). The degree to which sunflecks are composed of direct-beam radiation may vary greatly, however. The mean red:far-red ratio for sunflecks in a wheat canopy was only 15% lower than daylight values (Holmes and Smith, 1977a). In lowland rainfor ests of Panama and Costa Rica, the red:far-red ratio of sunflecks ranged from 0-37 to 1-30, compared to a mean of 1-22-1-33 in a clearing, and 0-350-40 in forest shade (Lee, 1987). Variation in the relative proportions of diff"use and direct light in sunflecks can be partly explained by penumbral efl'ects within forest canopies (Miller and Norman, 1971; Anderson and Miller, 1974). A penumbra is a partial shadow, an "edge effect" of a sunfleck. The probability of penumbras in a forest understorey depends primarily on the size of canopy openings and the canopy height (Norman et aL, 1971; Oker-Blom, 1984). The solar disk subtends an angle of 1/2°at the earth's surface, such that a canopy opening of at least this apparent size is required to transmit full-sun irradiance. In a tall forest with many small openings, sunfleck sizes will be small, and penumbras will be frequent (Anderson and Miller, 1974; Oker-Blom, 1984). In this case, SUNFLECKS AND UNDERSTOREY PLANTS 13 At the other extreme, canopy gaps or widely-spaced canopy trees create relatively large sunflecks, or sun patches, where full-sunlight irradiances are often observed (Young and Smith, 1979). Along the edges of gaps, and in small gaps formed by a branch fall, for example, the frequency of full-sun irradiance is low compared to conditions within larger gaps (Chazdon, 1986). In these intermediate habitats, sunflecks can often be easily distinguished from the slightly elevated levels of background diffuse light. In describing measurements of sunfleck activity, one must keep in mind that each investigator needs to adopt his or her own criteria for distinguishing sunflecks from shade light. For example, the PFD of diffuse light in a lodgepole pine forest in Wyoming (Young and Smith, 1979) is similar to the PFD during sunflecks in a Costa Rican rainforest (Chazdon and Fetcher, 1984a). If sunflecks were universally defined as periods when PFD exceeded 50|imol m"^ s~', this definition would be useless for describing sunfleck activity in the Wyoming forest. Furthermore, investigators often have different concepts of what constitutes a forest understorey habitat relative to a gap. In the following sections, unless otherwise indicated, I have adopted the definitions used by each investigator. Although this approach limits direct comparisons of sunfleck activity among forest types, I believe it is the rnost reasonable way to describe the wide array of sunfleck measurements that have been described. B. Temperate Deciduous Forests The light regime in the understorey of temperate deciduous forests varies dramatically over the year, according to the timing of leaf production and leaf fall in canopy trees and changes in solar evaluation. During winter, when trees are leafless, branches and trunks alone can absorb 50-70% of the incoming solar radiation (Hutchison and Matt, 1977). In a New England hardwood forest, 30% of the daily photon flux incident on the canopy reached the forest floor in April, before the emergence of tree foliage (Curtis and Kincaid, 1984). Based on analyses of hemispherical photographs, Anderson (1964a) found that both diffuse and direct site factors (see p. 11) varied throughout the year in the understorey of a deciduous forest near Cambridge, UK. The diffuse site factor remained fairly constant at 30% from irradiance during sunflecks will be considerably lower than that of direct beam solar radiation incident above the canopy, and the irradiance and January to April, whereas the direct site factor increased from 3% in January spectral composition will more closely resemble those of shade light. When penumbras are combined with windy conditions, it is often difficult to differentiate sunflecks from fluctuations in background diff"use light. These reached maximum values in April for three consecutive years (Anderson, 1964b). Similarly, in a deciduous forest in Tennessee, maximum amounts of direct beam solar radiation reached the forest floor in early spring, account patterns become even more complicated under partly cloudy conditions, ing for over 90% of the total solar energy received in the understorey during because of cloud edge effects. this period (Hutchison and Matt, 1976, 1977). to a maximum of 19% in April. Monthly irradiance in the understorey site 14 R. L. CHAZDON In the summer, following tree leaf emergence, solar radiation in the understorey decreases to 1-5% of that available above the canopy, and remains low until autumn leaf fall (Hicks and Chabot, 1985). During this period, direct site factors range from 1-3%, and sunflecks become relatively infrequent (Anderson, 1964a,b). Nevertheless, penetrating direct beam solar radiation contributed over 50% of the total radiation budget during the summer monthsin a Tennessee deciduousforest understorey (Hutchison and Matt, 1976, 1977). Measurements of solar radiation penetration in a hard wood stand in Connecticut showed that only 21% of direct-beam radiation reached the forest floor during the summer (Reifsnyder et al., 1971). Sunflecks, which were small and widely scattered, contributed to a high degree of spatial and temporal variation in solar radiation within the understorey; the coefficient of variation for individual 5-min observations was 225% (Reifsnyder et al., 1971). In the understorey of a Michigan mixed hardwood forest during the sunmier, an estimated 45-55% of daily photon flux density (PFD)was contributed by sunflecks exceeding 100 jimol m~^ s~' (Weber et al., 1985). Although PFD was below 50iimolm~^s"' more than 75% of the time, these readings contributed only 35-40% of daily PFD. SUNFLECKS AND UNDERSTOREY PLANTS 15 Young and Smith (1979) made detailed observations on the frequency and duration of sunflecks received by two Arnica species in the understorey of two subalpine coniferous forests in Wyoming. In the lower-elevation lodgepole pine forest, 39% of the sunffecks received were less than 15min long, and only 5% exceeded 60min. The longest sunfleck (or sun patch) lasted 165 min. In contrast, sunflecks in the higher elevation spruce-fir forest were shorter, less frequent, and of lower flux density (Young and Smith, 1979). Both forests are characterized by the occurrence of large sunflecks, often lasting 20-30 min, that receive 40-60% of full sunlight irradiance (Smith, 1985). Apart from the eff'ects of canopy cover, cloud conditions during afternoon periods reduce incident sunlight an estimated 40% over the day (Young and Smith, 1983). Sunffeck activity varied considerably among microsites in red fir forests of California (Ustin et al., 1984). The frequency of readings at relatively low PFD (<75jimol m"^ s"') did not vary significantly among sites, but the frequency of readings above 1025 jimol m~^ s~' varied 3-5-fold. Along a line transect, PFD of sunffecks varied from a minimum of 31 to a maximum of 624|imol m"^ s"'. Among microsites, sunffeck size and frequency varied greatly, although sunffeck size tended to be inversely related to frequency. C. Coniferous forests Sunfleck measurements in the understorey of a redwood forest in Califor nia were made in three sites differing in total canopy cover and exposure The light regime in the understorey of evergreen, coniferous forests also (Powles and Bjorkman, 1981). In a deep shade site with little sunffeck varies greatly during the year, but in this case seasonal variation is primarily activity, daily PFD was 0-73 mol m"^ d"'; the maximum 10min average PFD due to changes in solar elevation rather than to changes in forest canopy measured was 153 ^imol m"^ s~'. At the edge of a gap, 2-3 mol m"^ d"' was received, with as much as 71 % contributed by two sunflecks. The third site, on the south-facing edge of a large clearing received 8-14 mol m"^ d~'; a cover. In general, a higher proportion of direct beam radiation reaches the understorey in coniferous forests compared to deciduous forests (Anderson, 1966; Reifsnyder et al., 1971; Smith, 1985). Diff"use site factors in the understorey of a stand of Pinus syhestris averaged 16*4%, whereas monthly direct site factors varied from 0 in January to a maximum of nearly 30% in June and July (Anderson, 1966). Direct site factors were higher in the summer, when there were moreopportunities for direct sunlight to penetrate openings in the canopy. Sunfleck distributions did not vary in a consistent manner among sites during the year, however. Direct site factors within particular sitesdid not changein parallel throughout the seasons (Anderson, 1966). Sunflecks accounted for over 50% of the total solar radiation beneath a Pinus resinosa stand in Connecticut during the summer (Reifsnyder et al., 1971). In this forest, sunflecks were large and bright, and nearly 25% of direct-beam radiation incident above the canopy reached theforest floor. As in the hardwood forest, sunflecks contributed greatly to light variation within the understorey; the coefficient of variation for individual 5-min observations was 121% (Reifsnyder et al., 1971). major sunfleck at noon contributed as much as 83% of this total. D. Tropical Evergreen Forests Because of their equatorial proximity and tall, dense canopies, understories of tropical evergreen forests receive extremely low levels of diffuse solar radiation on a year-round basis (see review by Chazdon and Fetcher, 1984b). Since the pioneering studies of Evans (1939, 1956), sunffecks have been recognized as an important feature of the light environment within tropical forest understorey habitats. Using a Weston photoelectric cell and galvan ometer, Evans (1939) made many observations of sunffecks in mature Nigerian rainforest and a nearby 14-yr-old secondary forest. In the mature forest understorey, sunffecks were generally oflow irradiance compared with full sunlight, andwere confined to a period of4-5 h in themiddle of theday. Over a 10-hperiod, fewer than 0-1 % of the observations exceeded irradiances more than five times the mean shade irradiance, whereas 5-2% were greater SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON 16 than 20% above themean shade irradiance (Evans, 1939). An estimated 10% ofthe total light energy was attributed to sunflecks (Evans, 1966; Table 4). Light conditions were very similar in the 14-yr-old secondary forest, but the incidence ofhigh-irradiance sunflecks was lower than in the primary forest. Diffuse light irradiance was also lower in thesecondary forest (Evans, 1939). In another Nigerian forest, Evans (1956) calculated that sunflecks contri buted about 70% of the total light energy reaching the forest floor between January and March (Table 4). Using thearea-survey technique, he observed that 20-25% of the area of the forest floor was occupied by sunflecks at midday. On average, sunflecks were visible in a particular spot for about one houra day. As in the other forest, sunflecks were rare during early morning andlate afternoon, when thesolar angle was below 30". Photometer readings in a Brazilian rainforest understorey showed a similar pattern of sunfleck incidence, with sunflecks restricted to a 6 h period in the middle of the day (Ashton, 1958). Extrapolations from similar studies in a tropical rainforest in Singapore showed that, during an entire year, approximately 50% of the total light energy received by understorey plants came from sunflecks (Whitmore and Wong, 1959; Table 4). Considerable variation was observed in the distribu tion of sunflecks and diff'use light in two forest plots (Evans et al., 1960). Bright sunflecks, of equivalent irradiance to full sunlight, were rare (Whit more and Wong, 1959; Evans et al, 1960). In a lowland rainforest in Ecuador, Grubb and Whitmore (1967) found that sunflecks contributed 60% of the light energy on a sunny day (Table 4). Less sunfleck light was received 17 in a montane forest in Ecuador, although diffuse light readings were higher during midday compared to the lowland forest (Grubb and Whitmore, 1967). Patterns of light distribution in the understorey under cloudy conditions werestable and reproducible, according to studies by Evans et al. (1960) in a Singapore rainforest. No correspondence was observed, however, between mean irradiance at the same microsite measured under sunny and cloudy conditions on different days. Evans et al. (1960) therefore concluded that measurements of light distribution on cloudy days could not be used to predict light distribution during sunny conditions. In contrast, Sasaki and Mori (1981a) found that the frequency and illuminance of sunflecks was strongly correlated with the illuminance of diffuse light in a Malaysian Dipterocarp forest. In the darkest understorey microsites, sunflecks were infrequent and of low illuminance, whereas microsites characterized by higher levels of diffuse light received more sunflecks of higher illuminance. Based on measurements of photon flux in a deeply-shaded rainforest microsite in Queensland, Australia, Bjorkman and Ludlow (1972) found that sunflecks contributed 62% of daily photon flux on a clear day (Table 5). Photon flux density during sunflecks reached a maximum of 350 ^mol m"^ s"', equivalent to 20% of PFD above the canopy. Most sunflecks were less than 2 min long. In a considerably brighter understorey site in Queens land, sunflecks contributed 38% of daily PFD (Pearcy, 1987). In the understorey of a Hawaiian rainforest, over 60% of the sunflecks received during the summer were less than 30 s long, and few were over 5 min long Table 5 1 aoie 4 4 Table The percentage of total radiant energy contributed bysunflecks in the understorey of temperate and tropical forests. Definitions of sunfleck activity and the number of The percentage of total photon flux contributed by sunflecks in the understorey of temperate and tropical forests. Definitions of sunfleck activity and the number of days measured are specific to each study. days measured are specific to each study. Percentage total Percentage total Forest type/site Temperate deciduous forest Tennessee, USA (summer) energy Reference Nigeria (1 day) Nigeria (Jan-Mar) Singapore (entire year) Ecuador (1 day) Forest type/site flux Reference Temperate deciduous forest 50 Hutchison and Matt (1976, 1977) 50 Reifsnyder et a/. (1971) 10 70 50 60 Evans (1939, 1966) Evans (1956) Whitmore and Wong (1959) Grubb and Whitmore (1967) Coniferous forest Connecticut, USA (summer) Lowland tropical evergreen forest photon Michigan, USA (summer) Lowland tropical evergreen forest 45-55 Weber et al. (1985) Queensland, Australia (1 day) 62 Bjorkman and Ludlow (1972) (1 day) Hawaii, USA (5 weeks) 12-65 40 10-78 16^ Pearcy (1988a) Pearcy (1983) Chazdon (1986) Costa Rica (3 days) Mexico (1 day) R. L. Chazdon, C. B. Field and R. W. Pearcy (unpublished data) 18 SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON 19 (Pearcy, 1983). These brief sunflecks largely reflect the windy conditions the 16sensors, whichwere located within 0-5 m of each other, total sunfleck characteristic of this forest. The median maximum PFD during sunflecks was duration ranged from 10-6 to 29-2min, and the contribution of sunflecks to daily PFD ranged from 16 to 44% (Table 5). Total sunfleck duration was positively correlated with daily PFD (P<0 001), whereas mean sunfleck 250nmol m"^ s~', with only a small proportion of sunflecks reaching PFD equivalent to full sunlight above the canopy. Arbitrarily defining sunflecks as PFD observations above 150 {imol m~^ s~', the average minutes of sunflecks per day during the summer months were 10-6 and 21 min, for two successive years. On relatively clear days, sunflecks contributed asmuch as80% ofdaily- PFD; over a 5-week measurement period, the estimated contribution of sunflecks decreased to 40% of daily PFD (Table 5). Potential sunfleck incidence, based on hemispherical photographs, ranged from an average of 5min in the winter months to 61 min during the summer months (Pearcy, 1983). Measurements of photon flux in the understorey of a Costa Rican rainforest indicate that sunflecks may contribute from 10 to 78% of daily PFD on clear days (Chazdon and Fetcher, 1984a; Chazdon, 1986; Table 5). The relative proportion of daily PFD contributed by sunflecks increased as thePFD of background difluse radiation decreased (Chazdon, 1986). During sunflecks, PFD rarely exceeded 500nmol m'^ s"'. In this forest, sunflecks tend to be more frequent in the morning, because of afternoon cloud-cover and rain (R. L. Chazdon, unpublished data). Sunfleck activity can vary greatly overa small spatial scale. In a Mexican rainforest, meansunfleck incidence (observations above 50ixmol m"^ s"') for a single day ranged from 0 to 42 min among 16 leaves within the same understorey plant (Chazdon et al., 1988). Mean minutes of sunflecks per day in four understorey plants from the same site ranged from 4 to 22 min per day. Average potential minutes of sunflecks per day for these same plants, based on hemispherical photographs, ranged from 11 to 64 min. Measure ments of sunfleck activity along a 21 m line transect showed that minutes of sunflecks per day varied from 7 to 33 min among 16 sensors spaced 15 cm apart (R. L. Chazdon and C. B. Field, unpublished data). Daily PFD varied among sensors by a factor of 5. Sunfleck activity was so localized that daily patterns of PFD among sensors at distances greater than 0-6 m were not significantly correlated (Chazdon et al., 1988). Detailed sunfleckmeasurements in a closed-canopy site in this same forest length was negatively correlated with the total number of sunflecks received {P < 0-01, R. L. Chazdon, C. B. Field, and R. W. Pearcy, unpublished data). In the understorey of a Queensland rainforest, sites only 0-5 m apart can show two-fold variation in daily PFD and the number of sunflecks received (Pearcy, 1988a). Within a 5-m radius, daily PFD ranged from 0-47 to 1-5 mol m~^ d~and thepercentage contributed bysunflecks ranged from 12 to 65% (Table 5). Maximum PFD during sunflecks was almost always below full-sun levels; only 1% of the sunflecks measured exceeded 1200nmol m"^ s"'. Within sapling crowns of two canopy tree species in a Costa Rican lowland forest, total minutes of sunfleck activity per day ranged from 2 to 106 for Dipteryx panamensis and from 1 to 94 for Lecythis ampla (Oberbauer et al., 1988). For both species, total sunfleck exposure per day averaged 18-20 minutes. £. Summary: Generalizations About Sunfleck Activity The studies described above clearly illustrate the tremendous variation in sunfleck activity within and among forest types. Despite this variation, estimates of the percent of total light energy contributed by sunflecks are remarkably similar in temperate and tropical forests (Table 4). Sunflecks contribute 50% of the total light energy received during the summerin both coniferous and deciduous temperate forests, and 50-70% in tropical ever green forests (Table 4). The percentage of total PFD contributed by sunflecks is also quite similar in temperate deciduous forests during summer and tropical forests (Table 5). The two long-term estimates, 45-55% for a temperate deciduous forest during the summer (Weber et al., 1985), and 40% for a five-week period in a Hawaiian subtropical forest (Pearcy, 1983), are consistent with the range of lOO^mol m ^ s ', whereas less than 2% exceeded 500|i.mol m ^ s variation measured among sensors within a single day in three different tropical forests (Table 5). No yearly estimates of the contribution of sunflecks to total PFD are available for comparison. These studies, carried out independently by different researchers using a variety of techniques and assumptions, clearly illustrate the importance of sunflecks as sources of radiant energy and photosynthetically active radiation in all forest types. To the extent that sunflecks are a consequence of forest canopy structure, some generalizations can be made with regard to the frequency, duration, and peak intensities of sunflecks in diff"erent forest types during sunny Furthermore, sunflecks were clumped in their temporal distribution. Among periods. Tall, multi-layered forest canopies have many openings that are using 16 sensors placed in a two-dimensional array showed that sunflecks were brief, of low intensity relative to full sun, and extremely variable on a spatial scale of 0-25 m^ or less. On average, 56% of the sunflecks received were less than 4 s long, and over 90% were less than 32 s long (R. L. Chazdon, C. B. Field, and R. W. Pearcy, unpublished data). Mean sunfleck length was 13 s. Nearly 64% of the sunflecks had a maximum PFD below 20 R. L- CHAZDON smaller than the diameter of the solar disk from the perspective of the forest floor. Thus, penumbral effects often dominate in the understorey. Sunflecks, when they occur, are brief(usually less than 1min long), extremely localized (less than 0-5 min length or width), and tend to have peak PFD well below full-sunlight irradiance. Sunflecks tend to be clustered in distribution, and are rare during early morning and late afternoon. In more open forest types and leafless forests (temperate or tropical deciduous forests), sunflecks are longer in duration (often up to 10min) and occupy a larger area on the forest floor (up to several m in length or width). During sunfleck periods, peak PFD will frequently reach full-sun irradiance. Moreover, "sunpatches" will also tend to have higher diffuse PFD between sunfleck periods than more shaded microsites. Within a particular understorey habitat, the number of sunflecks received per day tends to be negatively correlated with the mean length of sunflecks. Daily PFD is positively correlated with the total minutes of sunflecks received. Further generalizations can be made by incorporating known weather and atmospheric conditions within each forest. To the extent that overcast skies, rain, and wind are predictable, microsite variation in the frequency and duration of sunflecks of different peak intensity can be estimated for any particular spatial scale. It is important to note, however, that descriptions of sunfleck activity are specific to the spatial and temporal scale at which they are measured. V PHOTOSYNTHETIC RESPONSES TO SUNFLECKS Sunflecks are a common and important feature of the light environment in all forest understorey habitats. In light-limiting habitats, the extent to which sunflecks influence plant growth, reproduction, and microsite distribution ultimately depends on their importance for leaf carbon gain. In this section, I review what is known about carbon gain during sunflecks in natural habitats and constraints on photosynthetic responses to sunflecks. These studies provide a basis for deeper understanding of the ecological significance of sunflecks for forest understorey plants. generalize results to other species or microsites. Simulations using modelled photosynthetic responses, when combined with field data, provide a means to extrapolate from an initially limited set of results. The most powerful approach involves the collection of detailed field measurements, modelling of leaf responses, and subsequent computer simulations using field-collected light and photosynthesis data. 1. Field Studies To determine the importance of sunflecks for daily carbon gain in forest understorey plants, continuous measurements of PFD and COj exchange are needed. Accurate instantaneous measurements of gas exchange during naturally fluctuating light conditions further require that instrument res ponses are faster than the physiological responses of leaves. In addition, large quantities of data must be recorded quickly and stored for subsequent analysis. Because of thesetechnical difficulties, few detailed measurements of daily patterns of COj exchange in forest understorey plants have been made (Table 6). The first measurements of gas exchange under natural sunfleck conditions in a Queensland rainforest understorey showed that sunflecks contributed substantially to daily carbon gain (Bjorkman et ai, 1972b). Nevertheless, leaves of the understorey herbs Alocasia macrorrhiza and Cordyline rubra were capable of achieving positive daily carbon balance in the absence of sunflecks. Both COj assimilation and stomatal conductance increased rapidly in response to sunflecks. Although some sunflecks did -2 exceed the light saturation point for both species (about 100 jimol m"^ s" ), the efficiency of light use during both clear and overcast days was high (Bjorkman et al., 1972b). Based on data presented by Bjorkman et al. Table 6 Percentage of daily carbon gain contributed by sunflecks in natural habitats basedon field measurements. Weather conditions and number of days measured varied among sites and species. Percentage daily carbon Forest type/species A. Sunflecks and Carbon Gain in Understorey Habitats Two general approaches have been used to evaluate the relative proportion of total carbon gain attributed to sunflecks in forest understorey plants: field studies and computer simulations. Ideally, photosynthetic measurements should be made on plants growing under natural patterns of light variation. Field studies, however detailed, are frequently limited by the inability to 21 SUNFLECKS AND UNDERSTOREY PLANTS gain Reference •Temperate deciduous forest Acer saccharum (summer) Tropical rainforest Euphorbiaforbesii (1 day) Claoxylon sandwicense (1 day) Argyrodendron peralatum (1 day) 35 Weber et al. (1985) 60 Pearcy and Calkin (1983) Pearcy and Calkin (1983) 40 32 Pearcy (1987) R. L. CHAZDON 22 (1972b), Weber etal. (1985) calculated that, in the absence of sunflecks, total carbon gain for Alocasia would be reduced by about 10%. In the subalpine understorey species Arnica cordifolia, photosynthetic characteristics differed between plants growing in relatively shaded and sunlit areas (Young and Smith, 1980): in shaded sites, photosynthesis during sunflecks was twice that during shaded periods. Carbon gain in understorey species is not always positively correlated with daily PFD. Photosynthetic rates ofArnica latifolia ina mixed spruce-fir forest were often light-saturated during sunflecks (Young and Smith, 1983). Oncloudy days, net carbon gain ofA. latifolia was 37% greater than on clear days, despite a 30% decrease in daily PFD. Diffuse light levels on cloudy days were often higher than on sunny days. Therefore, photosynthetic rates approached light saturation on cloudy days, partially accounting for the observed increase in daily carbon gain. Pearcy and Calkin (1983) studied field gas exchange of two tree species in the understorey of a Hawaiian forest. In the absence of sunflecks, both Euphorbiaforbesii and Claoxylon sandwicense wereable to maintain positive rates of CO2 assimilation over virtually the entire day because of low light compensation points. Photosynthesis during sunflecks, however, accounted for a substantial fraction of daily carbon gain. Photosynthetic responses to brief sunflecks were very rapid, and were effectively integrated by the gasexchange apparatus. Sunflecks contributed an estimated 60% of the carbon gain in Claoxylon on a relatively clear day, and 40% of the carbon gain in Euphorbia on a less clear day (Table 6). At light saturation, net photosyn thesis of Euphorbia was 50-60% higher than that of Claoxylon, and, consequently. Euphorbia appeared to utilize longer sunflecks more efficiently than Claoxylon (Robichaux and Pearcy, 1980; Pearcy and Calkin, 1983). On the other hand Claoxylon had higher rates of CO^ assimilation under diffuse light in the understorey, such that growth rates of the two species were similar (Pearcy and Calkin, 1983; Pearcy, 1983). In a Puerto Rican forest understorey, sunflecks increased photosynthetic rates of the shrub Piper treleaseanum by a factor of 5-8 over rates measured under diffuse light conditions (Lawrence, 1984). Photosynthesis was barely above the light compensation point under diffuse illumination, emphasizing the extreme importance of sunflecks for daily carbon gain in this species (Lawrence, 1984). Field studies of gas exchange in seedlings of Acer saccharum in the understorey of a mixed hardwood forest in Michigan showed that, during summer, approximately 35% of daily carbon gain occurred during sunflecks exceeding 50nmol m"^s~' (Weberet al., 1985; Table 6). Fluxes of diffuse and direct radiation in this forest were an order of magnitude greater than those SUNFLECKS AND UNDERSTOREY PLANTS 23 in the Hawaiian forest studied by Pearcy (1983) and Pearcy and Calkin (1983). During the day, COj assimilation closely tracked variation in PFD. Weber et al. (1985) calculated that daily carbon gain of Acer seedlings would be reduced by about 5% in the absence of sunflecks. Pearcy (1987) measured the diurnal pattern of CO2 assimilation of seedlings of Argyrodendron peralatum in the understorey of a Queensland rainforest (Fig. 3). The daily photon flux was about 3% of that received by leaves in the canopy, whereas daily carbon gain was nearly 10% of that of canopy leaves. Photosynthesis during sunflecks contributed 32% of the daily carbon gain (Table 6). 2. Computer Simulations When it has not been feasible to measure diurnal variation in CO2 assimila tion in natural understorey habitats, the importance of sunflecks for daily carbon gain has been investigated using computer simulations based on actual or modelled photosynthetic responses, together with data on light variation. Gross (1982) used this approach to estimate the importance of sunflecks for carbon gain in Fragaria virginiana. Based on studies of dynamic photosynthetic responses to step-changesin PFD (Gross and Chabot, 1979; see below), he showed that sunflecks often made a significant contribution to carbon gain. When the canopy was in full leaf, appoximately 50% of daily carbon gain was attributed to sunflecks. The closer the mean low light level was to the light compensation point, the more important sunflecks became to daily carbon gain. Single sunflecks of similar duration to those observed in temperate deciduous forest were found to contribute a rather small percent age of daily carbon uptake if they occurred infrequently, even if the photon flux due to the sunfleck was quite high (Gross, 1982). Chazdon (1986) used computer simulations to estimate the significance of lightvariation to total carbon gain in three species of rainforest understorey palms. These simulations were based on steady-state CO2 assimilation rates measured in the field and daily patterns of PFD measured in a wide range of understorey and gap microsites in a Costa Rican lowland rainforest. Daily carbon gain was positive under most understorey conditions, even in the absence of sunflecks. When midday diffuse PFD was below 5|imol m"^ s"', however, sunflecks provided the light energy needed to maintain positive carbon balance. Although daily carbon gain was linearly related to daily PFD when sunflecks were absent or infrequent, daily carbon gain was not a simple function ofdaily PFD when sunflecks contributed over 50% ofdaily photon flux. In the latter case, daily carbon gain was 33-35% lower than when the same daily PFD was achieved through higher flux densities of diffuse light. On days with sunny periods, simulations indicated that sun- SUNFLECKS AND UNDERSTOREY PLANTS Argyrodendron peralafum flecks accounted for 15-60% of daily carbon gain. In accordance with the findings of Gross (1982), the percentage of daily carbon gain contributed by sunflecks increased as midday diffuse PFD decreased (Chazdon, 1986). understorey lonn 800 E 25 600 o h 4UU B. Determinants of Sunfleck Utilization =3 •w vni) lXJiU n u. D- 0 Sunfiecks are a critical resource for forest understorey plants. Although understorey plants are usually able to maintain a positive carbon balance in the absence of sunflecks, light is the major environmental factor limiting 6 — growth and reproduction in deeply-shaded understorey environments. We might therefore expect understorey plants to utilize sunflecks efficiently. tn C* 4 E Analyses of steady-state light responses can off'erlimited insight into patterns "o 2 < 0 of sunfleck utilization. Based on determinations of a leaf's light compensa tion point and light saturation point, it is possible to predict photosynthetic responses to known intensities of diffuse and directradiation(Harbinson and Woodward, 1984). Estimates of photosynthetic responses during sunflecks based on steady-state rates can be grossly misleading, however, particularly when averaged light measurements are used (Gross, 1982; 1984; but see McCree and Loomis, 1984). Steady-state photosynthetic responsessimply do 100 w N 75 E 7) F E 50 not apply when light is fluctuating rapidly. Ultimately, studies of transient photosynthetic responses are needed to determine how sunflecks of different 25 frequency, duration, intensity, and temporal distribution are utilized by C7> leaves (see review by Pearcy, 1988a). 0 For the present discussion, it is useful to distinguish between two types of transient photosynthetic responses; induction reponses and photosynthetic 400 — 1 k- o dynamics. Induction responses involve relatively slow (from several minutes o ."snn to over an hour) increase in COj assimilation in leaves equilibrated in darkness or low light following a sudden step-wise increase in light (Rabinowitch, 1956; Marks and Taylor, 1978; Pearcy et ai, 1985; Chazdon and 200 Pearcy, 1986a; Kirschbaum and Pearcy, 1988). These induction responses affect the "readiness" of a leaf to respond to short-term light fluctuations a. cT 30 30 T k_ p o o I—" 20 20 AW 10 0 J 10 o that dynamic responses are independent of the state of induction of a leaf; on £ the contrary, dynamic responses are highly dependent on whether or not a leaf has undergone induction (Chazdon and Pearcy, 1986a,b and see below). Rather, induction and photosynthetic dynamics are both transient responses I 10 12 14 16 18 Time { h) Fig. 3. Daily course of photon flux density (PFD; nmol s"'), CO, assimilation, stomatal conductance, internal COj pressure, and leaf temperature for a seedling of Argyrodendron peralatum in a Queensland rainforest. From Pearcy (1987), with permission of the publisher. (Chazdon and Pearcy, 1986a,b). Photosynthetic dynamics, on the other hand, involve short-term photosynthetic responses (seconds long) to light fluctuatioJ55, such assunfiecks (Gross, i986). This distinclian docs not!mp)'y that occur on different time-scales. Transient photosynthetic responses are best understood as dynamic responses superimposed on the background of leaf induction state. 26 SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON 1. Photosynthetic Induction A typical induction response is shown for Alocasia macrorrhiza in Fig. 4(a). Following a long period at 10^mol m~^ s"', PFD was suddenly increased to 400nmol m"^ s"'. In this example, it took 35min for the leaf to reach the steady-state rate at saturating PFD (Chazdon and Pearcy, 1986a). Induction in rainforest understorey species may take from 20 to over 60 min (Pearcy et al, 1985; Chazdon and Pearcy, 1986a). Although the time period for induction may vary somewhat between species, the induction requirement for maximum photosynthetic rates is an intrinsic feature of photosynthesis in all plants (Rabinowitch, 1956; Walker, 1981). I 27 Recent studies of photosynthetic induction in the rainforest understorey species Alocasia macrorrhiza (Chazdon and Pearcy, 1986a) indicate that, during the first 5-10 min following the light increase, increases in CO2 assimilation are primarily limited by biochemical factors, such as the activity of the carboxylating enzyme ribulose-l,5-bisphosphate carboxylase (RuBPCase). Generally, photosynthetic limitations by CO2 diffusion into the mesophyll become more important during the later phases of induction. Studies of induction in other rainforest understorey species confirm the importance of biochemical limitations during the early phases (Pearcy et al.y 1985; Chazdon and Pearcy, 1986a). The nature and extent of stomatal limitation during inductionmayvaryamong species and indifferent environ mental conditions. In the C4 species Euphorbia forbesii, internal COj pres sures decreased from 300 to a minimum of 60-80 jibar following the light increase, suggesting that stomatal limitation may play a greater role in the V) I induction response of this species (Pearcy et al., 1985). E The degree to which stomatal conductance limits COj assimilation im "o E mediately following the light increase may also vary according to initial 3 stomatal conductance. Kirschbaum and Pearcy (1988) found that, in low- c o light grown plants Alocasia, wheninitial conductances were low(resulting in internal CO2 pressures below 100 ^bar), increases in internal COjpressure 400 800 1200 1600 2000 Time (s) were largely responsible for increases in CO2 assimilation during the first 10 min following the light increase. When stomatal conductances are low, a failure to partition transpiration between stomatal and cuticular paths can result in a significant overestimation of internal CO2 pressure (Kirschbaum and Pearcy, 1988). Field studies initially suggested, and laboratory studies later confirmed, that constant high light is not required to effect induction. Studies of the Hawaiian understorey species Euphorbiaforbesii and Claoxylon sandwicense showed that, during a sequence of artificial sunflecks (lightflecks) 1min in CN I o length, maximum CO2 assimilation rates increased with successive lightflecks E c o 0 1 'ot 01 < Time (s) Fig. 4. The time course of photosynthetic induction in Alocasia macrorrhiza grownin low light, (a) Induction during a step-change in photon flux density from 10 to 400nmol s"' (b) Induction during 60-s lightflecks (400|imol m"^ s"') separated by 2min of low light (lOnmol m s"'). From Chazdon and Pearcy (1986a), with permission of the publisher. (Pearcy et al., 1985). In the Australian rainforest species Alocasia macro rrhiza and Toona australis, leafinduction stateincreased 2-to 3-fold during a sequence of five 30or 60 s lightfleck separated by 2min oflow light (Chazdon and Pearcy, 1986a; Fig.4(b)). For Alocasia, low PFD and sunfleck PFD were 10 and 500nmol m"^ s"', respectively, whereas for Toona, low PFD and sunfleck PFD were 15 and 1200 ^mol m"^ s"'. The rate of induction during the 60s lightfleck sequence was not substantially different from the rate observed during constant illumination at the same high-light level (Chazdon and Pearcy, 1986a). Thus, sunflecks that occurearly in a series can increase leaf "readiness" to respond to subsequent sunflecks. Once leaves have undergone induction, and are returned to low light, they do not remain indefinitely in a state of photosynthetic "readiness". In low- 28 R. L. CHAZDON light grown leaves of Alocasia, the loss of induction in low light followed a negative exponential function with a half-time of approximately 25min (Chazdon and Pearcy, 1986a). Complete induction loss required over 60 min of exposure to constant low light. The rate of induction loss in high-light SUNFLECKS AND UNDERSTOREY PLANTS 29 were less than 40 s long, carbon gain was 20-80% higher than that estimated from steady-state photosynthetic rates during the high- and low-light per iods. Photosynthetic responses to 5 s intervals of high- and low-light also grown leaves of Toona australis was considerably faster (Chazdon and showed substantial enhancement when light-saturating PFD was used. Enhancement of carbon gain during brief lightflecks and flashing light was Pearcy, 1986a). attributed to post-illumination COj fixation, which contributed a large 2. Photosynthetic Dynamics and Carbon Gain During Sunflecks proportion of total carbon gain during brief sunflecks, but only a small proportion during long sunflecks. Pearcy et al. (1985) further observed that Few studies have examined photosynthetic dynamics of forest understorey species. Because of the difficulties in making accurate, rapid measurements and in carefullycontrolling and replicating experimental conditions, most of these studies have been carried out under laboratory conditions. Several studies on photosynthetic dynamics during high frequencies of flashing light have shown that photosynthetic rates are often higher than rates predicted from steady-state responses (Rabinowitch, 1956; McCree and Loomis, 1969; photosynthetic responses to lightflecks were strongly influenced by whether or not leaves had been previously exposed to high light. Following a 2 h exposure to low light, the carbon gain achieved during a 1 min lightfleck was only 44.5% and 47.3% of expectedcarbon gain for Euphorbiaand Claoxylon, respectively. They attributed the lower carbon gain to inactivation of the photosynthetic apparatus (induction loss) during the long exposure to low light. Pollard, 1970; Kriedemann et al., 1973). These studies, however, are insuf Studies by Chazdon and Pearcy (1986b) and Pearcy et al. (1987a) also ficient to assess photosynthetic responses to sunflecks in understorey plants, because they were not conducted using shade-grown understorey species, and they did not consider photosynthetic responses to measured frequencies showed that in low-light grown Alocasia, carbon gain and photosynthetic efficiency during lightflecks were greatly aflected by leaf induction state, lightfleck length, and lightfleck PFD. Net carbon gain during lightflecks at of light variation in natural habitats. saturating PFD (530nmol m"^ s"') increased with the steady-state PFD applied before the lightfleck sequence (Fig. 5(a)). Increases in lightfleck PFD from 25 to 120(imol m"^ s"' also led to highercarbon gain during lightflecks presented following induction (Fig. 5(b)). Net carbon gain attributed to the Photosynthetic responses to sudden increases and decreases in PFD similar to those during naturally occurring sunflecks were studied in Fragaria virginiana, the common wild strawberry of the eastern USA (Gross and Chabot, 1979). Leaves grown at low- and high-light levels were subjected to step-increases and decreases in PFD. The sudden change in PFD was always followed by a time-lag before a change in COj assimilation was first observed. Time-lags of about 10s were observed over all the PFD increases and decreases, but were somewhat longer when PFD was initially very low. Following the time-lag, leaves responded rapidly to decreases in PFD and more slowlyto increases in PFD. For light increases, the time constants (time required to reach 65% of the increment to the new steady-state rate) were less then 60 s, whereas time constants for light decreases were from 1 to 5 s. Pearcy and Calkin (1983) investigated photosynthetic dynamics during a step-change from shade light to 700Kimol m"^ s"' in Euphorbia forbesii and Claoxylon sandwicense. Increases in CO2 uptake briefly lagged behind the light change and then increased rapidly to the new steady-state rate. Responses were virtually complete within 60s. When light decreased to initial levels the rate of response was similar, except for a post-illumination CO2 release in Claoxylon, which is characteristic of C3 species. Laboratory studies on photosynthetic dynamics of Euphorbia and Claoxy lon by Pearcy et al. (1985) showed that, following induction, carbon gain during lightfiecks depended strongly on lightfleck length. When lightflecks lightfleck increased with lightfleck length, but the efficiency of light use decreased with lightfleck length (Fig. 5). An index of light utilization efficiency during lightflecks was calculated by comparing integrated carbon gain during a lightfleck with predicted carbon gain based on steady-state rates at the low- and high-light levels (Chazdon and Pearcy, 1986b). Efficiency following a 2h period at low light ranged from 110% for 5-s lightflecks to 60% for 40-s lightflecks. Following induction, these efficiencies increased to 160% and 100%, respectively (Pearcy et al, 1987a). Regardless of leaf induction state, however, photosynthetic efficiency decreased with lightfleck length. Similar responses to lightflecks were observed in leaves of Toona australis and Alocasia grown in high light, but efficiency was almost always below 100% (Chazdon and Pearcy, 1986b; Pearcy et al., 1987a). Enhancement of carbon gain during sunflecks does not require that leaves receive light-saturating PFD, especially when leaves have not undergone photosynthetic induction. Photosynthetic efficiency during 5-s lightflecks before induction exceeded 100% for shade-grown Alocasia, even when sunfleck PFD was below lOO^imol m"^ s~' (Chazdon and Pearcy, 1986b). Following induction, however, the degree of enhancement during 5-s light flecks increased as the PFD of lightflecks increased. Recent studies by SUNFLECKS AND UNDERSTOREY PLANTS —2 530 Mmol m 200.. 120pmolm —2 —2 50 jjmol m Sharkey et al. (1986) indicate that the observed build-up of pools of triosephosphates during 5-s lightflecks following induction in Alocasia was suffi cient to account for the enhancement of carbon gain due to post-illumination CO2 fixation. They hypothesize that extensive grana stacking, large intrathylakoid space, and high levels of chlorophyll in low-light grown plants enable significant post-illumination ATP synthesis, which is required to produce ribulose-1,5-bisphosphate (RuBP) from accumulated triose-phosphates. This hypothesis is consistent with observed differences between predicted and observed photosynthesis during brief sunflecks at different PFD (Chazdon and Pearcy, 1986b; Pearcy et al., 1987a). —1 s 8 —1 —1 s 160-- I (0 25 >jmo! m 120-- CM —2 a —1 I E CM o o 3. Stomatal Responses to Sunflecks Patterns of stomatal opening and closure during fluctuating light conditions can be affected by endogenous rhythms (Gregory and Pearse, 1937), leaf o E 3. 50 'w' 60 C *o 31 200 a> B c o jQ 120 pmol m 160-- —2 o —1 O o <D water status (Davies and Kozlowski, 1975), the length of the previous dark period (Brun, 1972), PFD during high-light periods (Woods and Turner, 1971), and leaf induction state (Chazdon and Pearcy, 1986a). In general, stomatal responses tend to lag behind changes in irradiance and CO, uptake (Pearcy et al., 1985). Woods and Turner (1971) studied the time required to reach equilibrium stomatal conductance following a light changein four tree species. Stomatal opening was always faster than closure, regardless of the magnitude of the light change. Stomatal opening took from 3 to 20min, whereas closure required from 12 to 36min. In three of the four species, 120-- 50 >jmol m 80-. 25 >imol m 40-- —2 —2 8 8 —1 —1 stomatal opening and closing was faster when the magnitude of the light change was greater. The most shade-tolerant species, Fagus grandifolia, had the fastest rates of stomatal opening and closure. In a comparative study of seedlings of six hardwood species, Davies and Kozlowski (1975) also found that the three most shade-tolerant species had the fastest stomatal responses to increases in irradiance. These studiessuggestthat relatively rapid stomatal responses to light fluctuations served to maximize photosynthesis during 0 50 60 Lightfleck length (s) Fig. 5. Netcarbon gain (^mol COj m"^ s"') as a function of lightfleck lengthfor lowlight grown leaves of Alocasia macrorrhiza. (a) Responses to lightflecks at 530 ^moI s"' presented following equilibration of leaves at each of four different light levels, (b) Responses to lightflecks of diff"erent PFD in leaves following induction. In s"'. Data are from Chazdon and Pearcy all cases, low-light PFD was lO^mol (1986b). sunfleck periods. Field measurements of diurnal variation in stomatal conductance of forest understorey speciesshow that, over long timeperiods, stomatal conductance follows changes in PFD (Bjorkman et al., 1972b; Young and Smith, 1979, 1983; Elias, 1983; Pearcy and Calkin, 1983; Masarovicova and Elias, 1986; Pearcy, 1987). Stomatal responses to sunflecks, however, are often consider ably slower than responses in CO, assimilation (Knapp and Smith, 1987; Weber et al., 1985). In seedlings of Argyrodendron in a Queensland forest understorey, Pearcy (1987) noted that peak stomatal conductances were not reached until after a sunfleck had passed. During sunflecks, internal CO^ pressures decreased from 240 to 200iibar. In the subalpine understorey species Arnica cordifolia, decreases instomatal conductance during simulated 32 R. L. CHAZDON SUNFLECKS AND UNDERSTOREY PLANTS cloud cover lagged behind photosynthetic decreases by about 4 min (Knapp and Smith, 1987). Not all understorey species exhibit changes in stomatal conductance during sunflecks, however. In a northern hardwood forest, 1987b), far less is known about how these constraints operate under fluctuating environmental conditions, such as during sunflecks. In this section, I examine what is known about these constraints and their relative stomatal conductances of leaves of Viola blanda exposed to 10 min of saturating PFD did not differ significantly from those of shaded leaves (Curtis and Kincaid, 1984). importance for carbon gain during sunflecks in plants from different understorey habitats. 1. Leaf Induction State Although stomatal conductance often fluctuates with PFD during the day, in many understorey species stomatal conductance tends to remain high under low-light conditions (Mooney et al., 1983). During periods of diffuse light, stomatal conductance in Argyrodendron remained above 25mmol m"^ s"' (Pearcy, 1987; Fig. 3). Measurements of stomatal conductance in other forest understorey plants confirm that, even under extremely low diffuse PFD, stomata typically remain open (Bjorkman et al., 1972b; Mooney et al., 1983; Pearcy and Calkin, 1983; Chazdon, 1984). Because of the apparent inability ofstomata to respond as rapidly as do COjassimilation pathways to 33 Laboratory studies suggest that, when sunflecks are few and far between, the efficiency of light utilization during some sunflecks may be partially con strainedby leaf induction state (Chazdon and Pearcy, 1986a,b). Field studies 1 by Pearcy (1987), however, show little evidence that induction limited photosynthesis of Argyrodendron seedlings during sunflecks in a Queensland understorey. In this case, sunflecks were distributed throughout the day, such that leaves maintained a high induction state (Fig. 3). Daily PFD measure ments for eight sensors in this site showed that 70% of the sunflecksoccurred light fluctuations, photosynthetic utilization of sunflecks may be impeded if stomata are closed during low-light conditions. Therefore, a high stomatal conductance relative to COj assimilation rate under shaded conditions may within one minute of the preceding sunfleck, and only 5% of the sunflecks were preceded bylow-light periods ofanhour orlonger (Pearcy, 1988a). The temporal patterning of sunflecks in anyparticular microsite will influence the serveto ensure that photosynthesis during sunflecks is not limited by internal degree to which leaves remain in a state of "readiness" to respond to CO2 pressures (Mooney et al., 1983; Pearcy, 1983; Fig. 3). A high ratio of internal CO2 pressure to ambient COj pressure could also be advantageous sunflecks. Sunflecks are often clumped in their distribution (Pearcy, 1983, 1988a), creating the potential for rapid leaf induction during a series of because of the resulting increased quantum yield for CO2 uptake (Pearcy, 1987). Pearcy (1987) calculated that, in a C3 plant, maintaining internal COj pressures at 320 rather than 220 ^ibar at a leaftemperature of 25"C should closely spacedsunflecks. On dayswhen sunshine isinterrupted byovercast or cloudy skies for several hours, the first sunflecks hitting a leafmay not yield result in a 14% increase in quantum yield. An interesting exception to this case is the Hawaiian C4 species Euphorbia forbesii. In this species, stomatal Although shorter sunflecks are utilized with greater efficiency, longer sun as much carbon gain as similar sunflecks occurring later in the sequence. flecks are more effective for photosynthetic induction. After a series of five 30-s lightflecks separated by 2min of low light, leaf induction state of conductance was very low under diffuse PFD, but increased rapidly in reponse to sunflecks. Nevertheless, stomatal conductance was found to limit Alocasia was only 48% of that measured for fully-induced leaves, whereas photosynthetic responses during at least some sunflecks (Pearcy and Calkin, during a similar sequence using 60-s lightflecks, relative induction state 1983). C. Constraints on Sunfleck Utilization in Understorey Habitats Photosynthetic utilization of sunflecks may be constrained by a variety of factors including loss of induction during low-light periods, restricted stomatal opening to conserve water, photoinhibition, wilting, and high leaf temperatures during prolonged high-light periods. Under field conditions, photosynthesis is influenced by a combination of environmental factors, of which PFD is only one. We are far from understanding how these different environmental factors affect carbon gain during sunflecks and during the intervening low-light periods. Although we have learned a great deal about constraints on carbon gain in natural understorey habitats (Pearcy et al.. reached 75% (Chazdon and Pearcy, 1986a). 2. Water-use Efficiency |f In forest understorey conditions, light is usually a more important limiting factor than water stress. As discussed above, stomatal limitations to COj uptake are rarely observed in field studies of forest understorey species. Under drought conditions, or during prolonged sunfleck exposures, how ever, regulation of stomatal opening to conserve water may impose limi tations on CO2 uptakeduring sunflecks. Stomatal responses to light increases in the sun-loving species Pelargonium slowed when plants were subjected to water stress, and responses to light decreases became faster (Willis and Balasubramaniam, 1968). Similar changes in stomatal responses during water stress were observed in seedlings of six hardwood tree species (Davies 34 SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON 35 conductance decreased by over 50%. This response serves to reduce water loss when evaporative demand is high, and to maximizephotosynthesis when evaporative demand is low. Thus, it is unlikely that stomatal conductance restricts CO2 assimilation during sunflecks under humid conditions. During the dry season, however, relative humidity often drops below 90% in the tropical rainforest understorey (Fetcher et ai, 1985). Under these conditions, decreased stomatal conductances in this species may impose a strong limitation on carbon gain during sunflecks. and Kozlowski, 1975). During a summer drought in a temperate deciduous forest, rates of COj assimilation decreased in Impatiens parviflora and Aegopodium podagraria (Masarovicova and Elias, 1986). Studies of stomatal responses to light fluctuations simulating natural cloud cover show that the stomata of the subalpine understorey species Arnica cordifolia responded rapidly to light increases and decreases in a manner similar to COj assimilation (Knapp and Smith, 1987). This species routinely undergoesseverewiltingduring extended periods of full sunlight (Young and Smith, 1979). Coupling of stomatal opening and COj assimilation resulted in a 30% increase in water-use efficiency compared to values calculated with a constant high-light rate of stomatal conductance (Knapp and Smith, 1987). 3. Photoinhibition During sunflecks, PFD may suddenly increase 200-fold over diffuse light levels. If light intensities during sunflecks exceed light saturation for long periods, photoinhibition may occur. Shade-grown plants are highly suscep tible to photoinhibition because they have low light-saturatedphotosynthetic capacities (Boardman, 1977; Bjorkman, 1981). In the shade fern Pteris cretica, exposure of fronds to PFD greater than 300nmol m"^ s"' caused a continuous decrease in COj assimilation with time (Hariri and Prioul, 1978). Despite a number of laboratory studieson photoinhibition in shade plants (Kozlowski, 1957; Bjorkman et ai, 1972a; Powles and Thome, 1981), few Thus, although water-use efficiency of this species does not remain constant under fluctuating light conditions, coupled stomatal and photosynthetic responses ensure that water-use efficiency is high during sunfleck exposures. Water-use efficiency of Arnica latifolia increased seven-fold on cloudy days compared to clear days (Young and Smith, 1983). Furthermore, daily carbon gain was 37% higher on cloudy days. On clear days, long sunflecks (sunpatches) led to increased leaf temperatures and leaf-to-air water vapor differences, higher transpiration rates, and lower water-use efficiency. Although the physiological basis for the decreased carbon gain on clear days is not known, lower xylem pressure potentials on clear days may cause decreases in photosynthesis in A. latifolia. Studies ofcordifolia by Young and Smith (1979) showed that increased water loss may indirectly reduce photosynthesis during prolonged sunflecks through a decrease in xylem studies have documented photoinhibition during sunflecks under field con ditions. The understorey herb Oxalis oregana, common in the deeply shaded redwood forests of northern California, exhibits leaflet movements in response to sunflecks when PFD exceeds 300-400 (imol m"^ s~' (Bjorkman and Powles, 1981). Leaflet movement was sensitive only to wavelengths between 375 and 490 nm, which is characteristic of blue-light-induced phototropic movements. These leaflet movements significantlyreduced PFD incident on the leaflet surface; PFD decreasedfrom an initial value of 1590to 295nmol m~^ s"' for two leaflets and 492jimol m"^ s"' for the third leaflet, pressure potential followed by decreased stomatal conductance. Studies of seasonal variation in stomatal responses to light in Douglas fir saplings show that, during autumn, winter and early spring, stomatal conductance was weakly related to PFD (Meinzer, 1982). During these periods of plentiful soil moisture, maximization of carbon gain may be more based on calculations of leaf angle and azimuth (Powles and Bjorkman, important thanregulation ofwater-use efficiency. In thesummer, when water conservation was most important, stomatal opening was tightly coupled with 1981). When leaflet movement was restrained during an 18-min sunfleck having an average PFD of 1500nmol m"^ s~\ COj assimilation in diffuse PFD, even under low-light conditions. Under field conditions, sudden light following the sunfleck was reduced by 30%. Decreases in photosyn thesis following the sunfleck could not be attributed to stomatal conduc changes in both PFD and vapor-pressure deficit (VPD) may occur simulta neously during a sunfleck. Dynamic stomatal responses of Douglas fir saplings to step-changes in VPD were in the opposite direction to that tance, because both stomatal conductance and internal COj pressure were predicted: step-increases in VPD resulted in increases in stomatal conduc higher than before the sunfleckexposure. Furthermore, variable fluorescence significantly declined when Oxalis leaves were exposed to intense PFD tance. Although these VPD responses resulted in increased rates of water loss, they served to enhance the speed of stomatal opening during brief similar to natural sunflecks. This decrease was associated with photoinhibi- tory inactivation of the photosystem II reaction centers. In the same understorey site, Powles and Bjorkman (1981) observed that leaves of sunflecks (Meinzer, 1982). In the Mexican understorey species Piper hispidum, stomatal responses to humiditywereverystrong, whereas responsesto PFD were weak (Mooney et Trillium ovatum growing next to Oxalis also suffered a 40% reduction in variable fluorescence following exposure to an intense sunfleck. Leaves of Trillium are incapable of the protective movements exhibited by Oxalis. al., 1983). When relative humidity decreased from 95 to 85%, stomatal •A 36 R. L. CHAZDON Measurements of COj assimilation of Oxalis showed that the decrease in PFD resulting from leaflet movements did not result in lower photosynthetic rates during the sunfleck. Leaves of Oxalis reached light saturation at relatively low PFD; at lOOnmol m"^ s"' leaves attained 90% of their light saturated rate. Leaflet movements caused by sunflecks of intermediate PFD appeared to adjust leaf angles so that PFD incident on leaflets was suffi ciently high to permit light saturation, but sufficiently low to avoid photoinhibition (Powles and Bjorkman, 1981). Species that do not possess the protective mechanisms against photoinhibition shown by Oxalis may suffer reductions in CO2 assimilation following intense sunflecks. In the absence of morefield observation, however, the role of photoinhibition during sunflecks in restricting subsequent utilization and direct PFD remains unknown. 4. Leaf Temperature and Water Relations Leaf temperatures during sunflecks may affect photosynthesis directly, through temperature dependence of enzymatic reactions, or indirectly, through effects on stomatal conductance and water relations. During an intense sunfleck, leaf temperature may rise as much as 18°C above air temperature (Ellenberg, 1963; Rackham, 1975; Young and Smith, 1979). In some cases, these temperatures may cause permanent heat damage and leaf necrosis, as has been observed in Mercurialis perennis in a deciduous woodland near Cambridge, UK (Rackham, 1975). Normally, leaf tempera tures on the forest floor areslightly below air temperature and are unlikely to exceed air temperature, even during small sunflecks (Rackham, 1975; Chiariello, 1984). In the subalpine understorey species Heracleum lanatum, temper atures of shaded leaves were 2-5'C below air temperature, whereas sunlit leaves were 3-5''C above air temperature (Young, 1985). Although leaf temperatures may increase during sunflecks, thermal damage resulting from an unusually intense sunfleck is a rare phenomenon in the understorey. In a Hawaiian understorey, leaf temperatures of Euphorbia forbesii and Claoxylon sandwicense may increase S'C during the first 50 s of typical sunflecks, andmay reach 30"'C during long sunflecks (Robichaux and Pearcy, 1980). Leaf temperatures during sunflecks are closer to the photosynthetic SUNFLECKS AND UNDERSTOREY PLANTS 37 through transpiration. Leaf water potential and pressure potential both declined rapidly for the first 2min of the sunfleck and then remained constant for the remainder of the sunfleck. Consequently, shoot extension rates declined significantly during the sunfleck. Observations of 24 sunflecks over a 2-d period showed that in 75% of the cases, the post-sunfleck rate of shoot extension was less than the pre-sunfleck rate. Changes in photosynthe tic rates during the sunfleck were then calculated based on modelled responses of CO2 uptake to flux resistances to CO2 transfer, assuming an instantaneous change in photosynthetic rate following the light increase. Predicted photosynthesis during the sunfleck reached 90% of the maximum light-saturated rate, a substantial increase over rates in diffuse light. In the understorey species Arnica cordifolia and A. latifolia, exposures to long-term sunflecks resulted in elevated leaf temperatures and transpiration rates, which often led to various degrees of wilting (Young and Smith, 1979). Microhabitats occupied by A. cordifolia received more frequent, longer, more intense sunflecks than those occupied by A. latifolia. Leaf temperatures of A. latifolia were generally well below air temperatures, even during sunflecks, and xylem water potentials for A. latifolia remained much higher during most of the day than for A. cordifolia. Consequently, midday wilting occurred more frequently in A. cordifolia. A. cordifolia, however, maintained turgor following sunfleck exposures of up to 165 min, whereas A. latifolia permanently wilted after 90 min exposure (Young and Smith, 1979). Photo synthesis of A. cordifolia remained positive, even after plants had wilted (Young and Smith, 1980). Similar responses were observed for six other understorey species from the same subalpine habitat (Smith, 1981). During sunflecks, decreases in xylem water potential led to midday wilting for four of the seven species studied. No stomatal closure was observed during sunflecks in any of the species, however. The increase in stomatal conductance and transpiration during sunflecks may be advantageous for two reasons. Higher rates of photosyn thesis were permitted, and excessively high leaftemperatures were avoided. Following sunfleck periods, plants rapidly regained turgor and xylem water potential returned to pre-sunlit levels. temperature optimum of Euphorbia, however, conferring a carbon gain advantage over Claoxylon during sunflecks. Changes in leaf temperature, water relations, and shoot extension were followed in Circaea lutetiana during a 7-min sunfleck in a Fagus sylvatica woodland in England (Woodward, 1981). Leaf temperature and transpir ation rose rapidly during the first minute of exposure, but transpiration subsequently declined following stomatal closure. Convective (sensible) heat transfer increased and remained high during the entire 7-min period. Because of stomatal closure, radiation could not bedissipated by latent heat transfer D. Sunfleck Regimes and Light Acclimation Plants living in exposed habitats exhibit diff'erent photosynthetic properties than exhibited by those living in shaded conditions (Bohning and Bumside, 1956; Boardman, 1977; Bjorkman, 1981). Moreover, many plants have the capacity to shift photosynthetic responses following a change in growth conditions. These acclimatory responses are generally in a direction that improves growth under the new environmental conditions. Unlike short- 38 SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON term fluctuations in COj assimilation during sunflecks, acclimatory changes occur on a time-scale from days to weeks (Gross, 1986; Table 1). Leaves at different stages of expansion may show differential abilities to acclimate to a change in light conditions (Pearce and Lee, 1969; Jurik et al., 1979). In the understorey species Fragaria virginiana, acclimation potential was greatest during early stages of leaf expansion, and decreased as expansion was completed (Jurik et ai, 1979). Typically, light acclimation of photosynthesis is measured by comparing steady-state lightresponses of plants grown under different light conditions (Bjorkman and Holmgren, 1963). Light conditions during growth are carefully controlled, and are usually maintained constant over the entire photoperiod. Some studies, however, have investigated acclimation of mor phological and photosynthetic characteristics under conditions where peak lightlevels and photoperiods were varied to yield different integrated as well as instantaneous PFD (Nobel, 1976; Chabot et al., 1979; Nobel and Hartsock, 1981). These studies elegantly demonstrated that photosynthesis and leaf structure were determined by integrated PFD, rather than by peak PFD. Within the forest understorey, daily PFD is positively correlated with the total minutes of sunflecks received and with the relative contribution of sunflecks to daily PFD (R. L. Chazdon, C. B. Field, and R. W. Pearcy, unpublished data). The potential therefore exists for light acclimation in response to consistent microsite variation in sunfleck activity. As discussed previously, variation in sunfleck activity occurs on both temporal and spatial scales. Understorey microsites that receive few minutes of sunflecks during one month may receive significantly longer sunfleck exposures several months later. Similarly, some microsites within the forest understorey will tend to receive more PFD from sunflecks than others (Chazdon, 1986). To what extent does light acclimation occur within the understorey in relation to spatial or temporal variation in sunfleck activity? 1. Spatial Variation To answer this question, comparisons of photosynthesis and leaf structure must be made within forest microsites that have demonstrably different sunfleck regimes. Such a study was done by Young and Smith (1980) on the subalpine understorey species Arnica cordifolia. This species exhibits con siderable phenotypic plasticity in photosynthetic characteristics, leaf struc ture, and water relations. Sun and shade plants occur in relatively open and densely shaded areas, respectively. The sun plants received nearly twice as much energy and photon flux during sunfleck periods, and had photosynthe tic capacities 2-5 times greater than shade plants. Sun plantsalso had greater 39 stomatal conductances, higher light saturation points, higher photosynthetic temperature optima, greater water-use efficiency, smaller leaf area, thicker leaves, higherspecific leafmass, and less chlorophyll perdrymass compared with shade plants (Young and Smith, 1980). Plants of Aster acuminatus grown undercanopy gap(approximately 25% of full sun) and understorey conditions (approximately 3% of full sun) differed significantly in maximum photosynthetic rates measured during May, but not in July, when photosynthetic rates of both groups of plants declined (Pitelka and Curtis, 1986). Photosynthetic differences were even more pronounced when plants were grown under low- and high-light conditions in growth chambers. An increasing number of laboratory studies indicate that, compared to plants from relatively open habitats, forest understorey plants exhibit less potential for light acclimation (Bjorkman, 1981; Bazzaz and Carlson, 1982; Langenheim et al., 1984). In a comparison of six rainforest species in the genus Piper, Chazdon and Field (1987a) found that photosynthetic capacity showed little variation among leaves of understorey plants, despite high variation in light availability among leaf microsites. Plants in a nearby I clearing exhibited considerable variation in photosynthetic capacity among leaves in relation to leaf light environment. Other studies, however, have i shown that acclimation potential is not always clearly correlated with i successional status or ecological conditions (Osmond, 1983; Fetcher et al., I 1987; Walters and Field, 1987). In a study of light acclimation of tropical tree seedlings, Kwesiga et al. (1986) found that seedlings grown under light with a reduced red:far-red ratio had higher maximum rates of photosynthesis and higher quantum efficiency than seedlings grown under high red:far-red ratios. Integrated PFD was maintained constant. The low red:far-red ratio used in the experiment was similar to values measured in diffuse light in a Costa Rican rainforest understorey, although daily PFD was considerably higher (Chaz don and Fetcher, 1984b; Lee, 1987). In contrast, Corre (1983) found no significant difference in photosynthetic characteristics of herbaceous species grown under different red: far-red ratio. Light acclimation has traditionally been measured by changes in steadystate photosynthetic responses. No field studies have addressed the issue of acclimatory responses in photosynthetic dynamics, such as photosynthetic efficiency during sunflecks. In laboratory investigations, high-light grown leaves Alocasia macrorrhiza and Toona australis exhibited lower photosyn thetic efficiency during sunflecks compared to low-light grown leaves of Alocasia (Chazdon and Pearcy, 1986b; Pearcy et al., 1987a). Moreover, highand low-light grown leaves of Phaseolus exhibited different capacities for 40 R. L. CHAZDON post-illumination COj fixation during sunflecks (Sharkey et al., 1986). Until further study, we do not know whether microsite variation in sunfleck activity can effect dynamic photosynthetic responses. SUNFLECKS AND UNDERSTOREY PLANTS 41 that light acclimation potential in all but spring ephemerals was restricted only when light availability increased dramatically (Fonteno and McWilliams, 1978). Based on measurements of seasonal variation in sunfleck activity and diffuse light penetration in deciduous forests, it is unlikely that 2. Seasonal Variation daily PFD will increase throughout the summer growth season within the In deciduous forest understoreyhabitats, the availability of diffuse and direct radiation varies greatly over the year (see Subsection IV.B). Deciduous herbaceous species in these forests show a variety of growth patterns in relation to seasonal variation in light availability (Salisbury, 1916; Sparling, 1967). Photosynthetic lightresponses of thesespecies strongly reflect the light conditions prevailing during leaf development (Sparling, 1967; Taylor and Pearcy, 1976; Kawano et al, 1978; Hicks and Chabot, 1985; Masarovicova and Elias, 1986). Spring ephemerals develop their leaves under conditions of high lightavailability, before leafexpansion in the canopy. In April, leavesof the springephemeral Erythronium americanum exhibited photosynthetic light responses similar to herbs found in open habitats (Sparling, 1967;Taylor and Pearcy, 1976). As summer began, and the forest canopy closed, light- understorey, except in the event of a tree or branch fall. saturated photosynthetic rates and light saturation points of Erythronium declined to less than 50% of early spring values. Rates of dark respiration, however, remained high (Taylor and Pearcy, 1976). Decreases in COj assimilation were correlated with decreased RuBP carboxylase activity (Taylor and Pearcy, 1976). Similar seasonal changes in photosynthetic responses and leaf biochemistry were found in the spring-active Anemone raddeana in a Japanese deciduous forest (Yoshie and Yoshida, 1987). In another group of deciduous herbs, leaf expansion occurs during or after canopy closure, in early May. A representative of this group. Trillium grandiflorum, exhibited photosynthetic rates more typical of shade species. These rates also declined throughout the summer, and reached a minimum in July (Taylor and Pearcy, 1976). Species such as Parthenocissus quinquefolia and Solidagoflexicaulis, in which leaf expansion occurred during midsum mer, had the lowest light-saturated photosynthetic rates and RuBP carboxy Evergreen woodland herbs also exhibit variation in photosynthetic light responses, which parallel seasonal changes in light availability (Kawano et al., 1983; Yoshie and Kawano, 1986). In Pachysandra terminalis, one-year- old leaves rapidly increased photosynthetic capacity following snow melt, and reached a yearly maximum in late April. Photosynthetic capacity then decreased to a minimum in July, when light availability in the understorey was lowest. In mid-August, photosynthetic capacity increased again, reach ing a second peak in early October, when canopy leaves were senescing. Subsequently, photosynthetic capacity declined through the winter months. Current-year leaves appeared in early June, and photosynthetic capacity increased throughout the rest of the summer to a maximum in late Sep tember. Stomatal conductance varied in parallel throughout the year for all leaves (Yoshie and Kawano, 1986). During early spring, over-wintering leaves of Pachysandra exhibited acclimation to high light availability, and did not show any evidence of photoinhibition (Yoshie and Kawano, 1986). These acclimatory responses, however, occur over a relatively long period of gradually increasing light availability from mid-March to early April. In contrast,laboratory studies of light acclimation expose plants to sudden, dramatic changes in irradiance. Results from these laboratory studiesmaytherefore not apply to thegradual seasonal changes in PFD that occur in deciduous forest understories. The semi-evergreen herb Hepatica acutiloba underwent major changes in chlorophyll and carotenoid content, and photosynthetic unit size in an oak forest understorey from April through October (Harvey, 1980). These changes were presumably associated with efficient light utilization as light lase activity of the herb species studied (Taylor and Pearcy, 1976). Seasonal variations in photosynthetic activity observed within the species studied were parallel to variations among species (Sparling, 1967; Taylor and availability decreased. In contrast, spring ephemerals in the same forest did not exhibit the same degree of plasticity in allocation to light-absorbing pigments when light availability decreased (Harvey, 1980). Pearcy, 1976; Hicks and Chabot, 1985). Rates of dark respiration also decreased among species from spring to summer, enabling positive carbon balance to be maintained as light availability decreased (Taylor and Pearcy, E. Photosynthesis in Understorey Plants Revisited 1976). These patterns suggest that deciduous understorey herbs possess a generalized acclimatory response to decreasing light availability from spring to fall. Leaves of Fragaria virginiana developed in low light were not able to acclimate completely to highlight,whereas high-light grown leaves were able to acclimate completely to low light (Jurik et al., 1979). These results suggest Forest understorey plants exhibit a variety of photosynthetic characteristics that enable them to maintain positive carbon balance under extremely low PFD (Boardman, 1977; Bjorkman, 1981). Among these arelow rates ofdark respiration and high quantum efficiency under low light. Leaf anatomy, biochemistry, and chloroplast structure of understorey plants have also been 42 SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON interpreted as adaptations for maximizing the efficiency of light utilization at lowirradiances under steady-stateconditions (Bjorkman, 1968; Goodchild et ai, 1972; Nobel, 1976; Caemmerer and Farquhar, 1981). As we begin to accumulate information on sunfleck activity and its importance for leafcarbon gain in understorey habitats, these photosynthetic characteristics may need to be interpreted more broadly. Recent studies suggest that growth in low light mayeffect changes in the regulation of pool sizes of Calvin cycle intermediates that ultimately control the efficiency of light use during transient sunflecks (Sharkey et al., 1986). Extensive grana stacking in chloroplasts of low-light grown leaves may create a "capaci tance" in the photosynthetic system that allows for transient build-up of a proton gradient for ATP formation following a sunfleck (Sharkey et al., 1986; Pearcy et ai, 1987a). Both factors lead to substantial post-illumination COj fixation, and enhancement of photosynthetic efficiency during brief sunflecks (Pearcy et ai, 1987a). In many understorey habitats, the great majority of sunflecks are very brief, and of fairly low PFD (Figs 2, 3;see Section IV). It is during these brief sunflecks that light is most efficiently utilized. Furthermore, if the time interval betweensunflecks is not too long, leaves of forest understorey plants will remain at fairly high induction states. There is no evidence that the photosynthetic characteristics responsible for efficientutilization of sunflecks impose any constraint on efficient utilization of low intensities of diffuse light. Some evidence does indicate, however, that photosynthetic adaptation to high irradiance imposes constraints on photosynthetic efficiency during sunflecks (Chazdon and Pearcy, 1986b; Sharkey et al., 1986). An accurate picture of the photosynthetic characteristics of forest under storey plants must incorporate transient photosynthetic responses. It is true that extremely low levels of diffuse light predominate over 75% of the time in 43 genetic effects apart from the effects of light quantity on growth processes (Holmes and Smith, 1977b; Morgan and Smith, 1978; Corre, 1983). To the extent that growth is light-limited, relatively small differences in light availability may have significant effects on plant growth (Shirley, 1929; Blackman and Rutter, 1946; Hughes, 1966). Relatively open understorey microsites will, on average, receive more direct PFD from sunflecks as well as higher levels of diffuse PFD (Young and Smith, 1979; Sasaki and Mori, 1981; Chazdon, 1986). When the incidence of diffuse and direct PFD are correlated, it is difficult, if not impossible, to determine whether microsite variation in sunfleck activity can account for differences in seedling establishment, growth, and distribution of plants in understorey microsites. Atkins et al. (1937) concluded that sunflecks were of relatively little concern for plant growth and distribution in deciduous forests because they were thought to contribute a relatively small percentage of total radiation at any particular site. Recent studies, however, indicate that, at least in some forest types, sunfleck activity is an excellent predictor of plant growth (Pearcy, 1983). Responses to irradiance at the whole-plant level often do not correspond with predicted responses based on light responses of individual leaves. These relationships are complicated because of changes in leafsize, structure, and duration in response to changes in irradiance (Hughes, 1959; Blackman and Wilson, 1951, 1954). In some cases, increased production of leafarea may compensate for lower unit leaf rate in low-light conditions, suchthat relative growth rates remain constant or decrease only slightly (Blackman and Wilson, 1954; Hughes, 1966). Moreover, plant growth ratesare also affected by self-shading and by plant size in ways that may be only indirectly related to light conditions. The light compensation point for an individual leaf, for example, may be significantly lower than that for an entire plant. forest understorey habitats. Sunflecks, despite their relatively low frequency, often contribute over 50% of the daily photon flux (Table 5). The ability to In this section, I review studies of the influence of sunfleck activityon seed germination, early establishment, and growth of forest understorey plants. take advantage of these sunflecks, however brief and unpredictable, may prove to be at least as important for long-term carbon gain as maintaining positive carbon balance under diffuse light conditions. variation in light availability in natural habitats, not all of them considerhow VI. SEED GERMINATION, ESTABLISHMENT AND GROWTH IN RELATION TO SUNFLECK ACTIVITY All phases of a plant's life-cycle may be influenced by light variation in understorey habitats. Sunfleck activity affects both the quantity and quality of light available within a microsite (Holmes and Smith, 1977a; Chazdon and Fetcher, 1984b; Lee, 1987). Light quality has well-characterized morpho- Although these studies describe plant responses in relation to microsite these responses are specifically affected by differences in sunfleck activity among microsites. Despite a long history of research on light relations of forest understorey plants, the influence of sunflecks on plant establishment and growth in natural forest understorey sites remains a relatively unex plored area of research. A. Seed Germination and Establishment in Understorey Habitats Differences in light conditions between understorey and gap environments r 44 R. L. CHAZDON SUNFLECKS AND tJNDERSTOREY PLANTS 45 are known to affect seed germination in several tropical pioneer species (see cense in a Hawaiian evergreen forest understorey (Fig. 6). Based on hemi review by Vazquez-Yanes and Orozco-Segovia, 1984). Far less is known sphericalphotographs, the potential minutes of sunfleck activityfor an entire about the extent to which smaller-scale light differences, such as those subsequent establishment in a Costa Rican wet forest showed no significant year were estimated for each of 15 plants. The mean potential minutes of sunflecks per day was closely correlated with the relative growth rate of plants over the year. In contrast, the diffuse site factor estimated from the same photographs was not significantly correlated with growth. Growth difference in seed germination between high- and low-cover plots (Marquis et rates of the two species were similar under similar sunfleck regimes. This al., 1986). In this study, understorey vegetation cover was removed in half of the plots, which decreased total cover from 90 to 85%. Sunfleck activity in a Mexican rainforest was found to aflfect seed study provides the strongest evidence available that sunfleck activity directly created by sunfleck activity, can affect seed germination of species that establish in the forest understorey. Comparisons of seed germination and germination in the photoblastic pioneer species Piper auritum and P umbellatum (Orozco-Segovia, 1986). Among seeds placed in three understorey microsites, percentage germination was significantly higher in the microsite that received longer sunflecks. In the understorey species P. aequah, percent age germination after one month was also significantly higher in the microsite with longer sunflecks, but after six months no significant dif affects growth of understorey plants over an entire season. Oberbauer et al. (1988) investigated growth and crown light environments of saplings of Dipteryx panamensis and Lecythis ampla, two rainforest canopy tree species. Height growth of Lecythis, but not of Dipteryx, was significantly correlated with the proportion of daily PFD contributed by sunflecks (instantaneous PFD above SOpmoI m~^ s"')- In both species, height growth over a year was correlated with measurements of weekly total ferences were found among the three microsites (Orozco-Segovia, 1986) Not all species of Piper require red light for germination; some forest species exhibit high germination rates in darkness (Vazquez-Yanes, 1976) Many temperate forest species produce seeds that are capable ofgermination under Euphorbia • y =-55.2 + 3-35x o y =-37.8 + 3.1 1 x Claoxylon r = 0.88 = 0.95 low-light conditions (Angevine and Chabot, 1979). D 0) B. Growth of Understorey Plants >. CT> 1. Tree Seedlings and Saplings O) Seedling growth ofthree dipterocarp species has been studied in relation to microsite variation in light availability. Within the study forest, Sasaki and Mori (1981) found that the frequency and intensity of sunflecks was <U "o correlated with measurements of difl'use irradiance. Growth of seedlings was closely correlated with difl'use light levels in a range below 20% offull sun Within a given level of steady-state difl'use light, however, dipterocarp seedlings showed uniform growth, even though sunfleck incidence may have 0) varied (Sasaki et al, 1981). Although sunfleck activity was not measured in these microsites, these results suggest that growth was more dependent on 0 difl'use light levels than on sunfleck activity. Similar results were obtained in a study of seedlings of the dipterocarp Hopea pedicellata (Gong, 1981). not differ significantly, although leaf length ofshaded seedlings was signifi Astriking relationship between sunfleck activity andgrowth was described by Pearcy (1983) for seedlings of Euphorbia forbesH and Claoxylon sandwi- 40 Potential minutes of sunflecks per day Survival and height growth under green mesh or under natural sunflecks did cantly less than seedlings grown under sunflecks. 20 Fig. 6. Relative growth rate of Euphorbia forbesii and Claoxylon sandwicense as a function of average duration of potential sunflecks per day (in minutes), estimated from hemispherical photographs. From Pearcy (1983), with permission of the publisher. 46 R. L. CHAZDON 47 SUNFLECKS AND UNDERSTOREY PLANTS PFD and the weekly percentage of full sun received, whereas diameter growth was only weakly correlated with light conditions. 2. Understorey Species At light levels below 20% of full sun, irradiance usually limits growth of vegetation below forest canopies (Shirley, 1929). In temperate deciduous forests, growth and distribution of the understorey herb Hyacinthoides (Scilla) non-scripta was more dependent on the amount of light received during the high-light phase of spring than on that received during the lowlight conditions following canopy closure (Blackman and Rutter, 1946). Thus, even if microsites varied significantly in sunfleck activity, the effects on plant growth would be relatively small compared to differences in growth duringearly spring. Growth studies of the forest annual Impatiensparviflora showed that when the diffuse site factor was constant, large increases in O) O) I c Q. c (0 0) direct sunlight produced only a small increase in unit leaf rate, and very little change in leaf weight ratio (Coombe, 1966). Specific leaf area, however, showed a relatively large decrease. The influence of light availability on growth of Aster acuminatus, a rhizomatous perennial of eastern deciduous forests in the USA, was also affected by seasonal distribution (Pitelka et ai, 1985). When light levels were initially high, as in the temperate forest in early spring, ramet growth increased more than when high-light levels occurred later in the growing season. Early season exposures to high light also appeared to enhance phenological development. Regardless of timing, high-light periods led to increased ramet height, weight, and rhizome production. According to Pitelkaet al. (1985), it is likely that in at leastsome Aster patches there can be substantial seasonal variation in light availability because of different temporal patterns in sunfleck activity. Long-term studies of Aster acuminatus haveshown that light levels within understorey microsites were significantly correlated with average plant size within patches, and with thelocations of patches (Pitelka et al, 1980; Fig. 7). Although this species isoften found in relatively open sites, such as treefalls, it exhibits extensive phenotypic plasticity and can occupy a wide range of microsites with different degrees of light availability (Pitelka et al., 1980; Ashmun et al., 1980). Microsite variation in light availability did not significantly affect patterns of biomass allocation to vegetative parts (Pitelka et al., 1980). Transplant experiments in deciduous forest sites showed that mean ramet size increased with light level over a three-year period (Ashmun and Pitelka, 1984). Measurements of PFD in eight transplant gardens were positively correlated with survival of ramets, the number of new ramets produced, and the total number of ramets at the end of the experiment. 20 40 Patch Light Level (jjmol 60 ) Fig. 7. Mean plant biomass (g) ofAster acuminatus as a function of patch light level (PFD; ^mol m"^ s"')- From Pitelka et al., (1980), with permission of the publisher. Moreover, ramets were correlated in size from one year to the next. In a different field site, similar relationships were observed between ramet size andlight availability (Ashmun et al., 1985). In four different growth seasons, multiple regression analysis using direct and diffuse site factors as indepen dent variables showed a highly significant dependence ofmean ramet weight, ramet density, and standing crop on light availability (Ashmun et al., 1985). Plants growing in the understorey of tropical evergreen forests are subjected to low daily PFD on a year-round basis, unless they are located jxear canopy gaps. Along the edge of gaps, seasonal variation in sunfleck activity was evident (Chazdon, 1986). Mean daily PFD at the northern edge of a gap was more than double that measured at the southern edge of the same gap in February, but was similar in March, when the sun was almost directly overhead (Chazdon, 1986). Further research is needed to determine whether differences in sunfleck activity associated with gap location signifi cantlyaffect plant growth. Studies of understorey herbsin a seasonal tropical forest in Panama indicate that, for many species, growth and establishment are highly dependent on canopy gaps (Smith, 1987). 48 R. L. CHAZDON vn. THE INFLUENCE OF SUNFLECKS ON REPRODUCTIVE BEHAVIOR AND DISTRIBUTIONS OF UNDERSTOREY SPECIES Although it is often assumed that reproduction and distribution of understorey species are limited by light availability, relatively little quantitative data has been gathered to support or refute this claim. Resource allocation to reproduction is an elusive quantity to measure (Bazzaz and Reekie, 1985; Bazzaz et al., 1987). Attributing reproductive effort to measured light conditions during flowering or fruiting is also problematic. In some species, reproductive buds may be initiated at least a year before actual flowering, when Hght conditions may have been quite different. For insect-pollinated plants, seed set may be pollinator-limited, whereas resource allocation to flowering structures and fruit maturation may or may not be light-limited (Bierzychudek, 1981). Furthermore, the costs of reproduction may impose substantial constraints on vegetative growth in light-limited understorey species (Clark and Clark, 1987).Reduced vegetative growth may then lead to periods of little or no reproduction, despite relatively unchanged light conditions. Clearly, long-term studies are required to gain insights into the relationships between reproductive effort and light availability in under storey species. Light availability is only one of the many possible causal factors in plant distribution; othersinclude dispersal, herbivory, pathogens, disturbance, and both local and regional history (Augspurger, 1984; Augspurger and Kelly, 1984). Many understorey species are capable of vegetative reproduction, and individual clones may persist in a site for decades or longer. Therefore, what may, at first glance, appear as patches of current regeneration, may in fact represent the remnants of a persistent, long-lived clone that at some previous time had proliferated in response to increased light availability (GomezPompa and Vazquez-Yanes, 1985; Smith, 1987). Studies of the distributions of long-lived perennials in relation to microenvironmental conditions must recognize this historical dimension. In this section, I review studies of the reproductive behavior and distribution of understorey species in relation to the patchiness of light availability in understorey habitats. Although the patchy nature of light availability is almost certainly correlated with microsite variation in sunfleck activity, more research is needed to determine the extent to which patterns of reproduction and distribution are affected by sunfleck activity rather than by other environmental factors. SUNFLECKS AND UNDERSTOREY PLANTS 49 A. Light Availability, Size Variation and Reproductive Behavior Many demographic studies of forest understorey species have shown that plant size is often correlated with the frequency and amount of reproduction (Sohn and Policansky, 1977; Solbrig, 1981; Pitelka et al, 1980; Sarukhan et al., 1984). In the understorey palm Astrocaryum mexicanum, patterns of reproduction were associated with differences in leaf number and crown size (Pinero and Sarukhan, 1982). Moreover, plants with above- and belowaverage flowering frequency were clumped in their distributions. Individuals growing in gaps produced move leaves and fruits during a 2 yr period than plants growing in other areas (Pinero and Sarukhan, 1982). Inflorescence size and number were significantly higher in reproductive individuals of two Costa Rican understorey palm species growing in gap-edge plots compared to closed-canopy understorey (Chazdon, 1984). The frequency of reproduc tion in the understorey cycad Zamia skinneri was correlated with both plant size and an index of light availability (Clark and Clark, 1987). These studies provide anecdotal evidence that patterns of reproduction in understorey species are related to heterogeneity in light availability. To the extent that plant size is correlated with sexual expression in dioecious understorey species, differences inlight availability may also be associated with changes in sex-ratio within plant populations (Bierzychudek, 1982). Flowering of Aster acuminatus was highly dependent on both individual plant size and light availability (Pitelka et al., 1980; Ashmun et al, 1985). In all patches containing flowering plants, non-flowering plants were always smaller than flowering plants. In contrast, the proportion of total biomass allocated to vegetative reproduction (production of new rhizomes) remained constant (Pitelka etal., 1980). Further studies oftransplanted ramets showed that the percentage of ramets flowering increased with garden light level in each ofthree successive years (Ashmun and Pitelka, 1984). Variation inlight availability alone was the principal factor that explained the large differences in reproduction observed among gardens. Plant size, however, was not the only determinant of flowering behavior. Reduction in light levels after deciduous canopy closure strongly affected sexual reproductive behavior, whereas ramet size was not significantly affected (Pitelka etal., 1985). Inthis case, phases of vegetative growth and sexual reproduction became uncou pled, allowing plants to respond to seasonal changes in the availability of limiting resources. In a study of understorey herbs of a seasonally dry tropical forest in Panama, Smith (1987) found that only a few species reproduce regularly in closed forest. Most species remain in a suppressed, vegetative state in the understorey until light availability increases following the creation of a treefall gap in the vicinity. 50 SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON B. Vegetative and Sexual Reproductive Effort Many, if not most, forest understorey species reproduce vegetatively as well as sexually. Allocation of resources (carbon or biomass) to both vegetative and sexual reproductive functions may vary according to light availability within the understorey. In Aster acuminatus, vegetative reproductive effort remained constant over a wide range of light levels, whereas sexual repro ductive effort increased with greater light availability (Pitelka et al., 1980). Plants transplanted to deeply-shaded sites did not flower, and few produced clonal offspring (Ashmun and Pitelka, 1984). In sites with higher irradiance, however. Aster acuminatus was capable of rapid clonal spread and a high incidence of sexual reproduction. When biomass allocation to the perennat- ing rhizome was separated from allocation to clonal growth, Ashmun et aL (1985) found that vegetative reproductive allocation increased with light availability. 51 resembled understorey sites on north-facing slopes, where seedlings were not patchily distributed. Ustin et al. (1984) suggest that aggregations of seedlings are the result of differential germination and seedlingsurvival, which may be susceptible to water and thermal stresses during prolonged sunfleck exposures. The small-scale distributions of Arnica cordifolia and A. latifolia in subalpine coniferous forests may also be strongly linked to the influence of sunfleck patterns on water relations (Young and Smith, 1979). Computer simulations of carbon gain and water-use efficiency of A. cordifolia indicate that water-use efficiency may be more important in microsite distributions than carbon gain (Young and Smith, 1982). Frequent cloud-cover during the summer growth season may mitigate the effects of sunflecks on water-use efficiency, however (Knapp and Smith, 1987). Studies of the distribution of understorey herbs in a Panamanian forest showed that patterns of abundanceand distribution can, to a largeextent, be explained by temporal and spatial variation in canopy gaps (Smith, 1987). Most understorey herbs are capable of rapid growth and recruitmentin gaps, Other studies of reproduction in forest understorey plants indicate that sexual reproduction is not always more sensitive to light availability than is vegetative (clonal) reproduction, although both may be affected. Two strawberry species, Fragaria virginiana and F. vesca, both showed decreasing but are also able to persist in closed-canopy forest, usually in a vegetative state. For many species, clumped spatial patterns are closely linked to the sexual reproductive effort in relatively shadier environments (Jurik, 1983, 1985). In this case, however, vegetative reproductive effort decreased more than sexual reproductive effort under shadier conditions, so that allocation between species distributions and canopy gaps were described for two understorey species in a wet lowland tropical forest in Costa Rica (Richards was shifted in favor of sexual reproduction. These studies show that the relationship between vegetative reproduction, sexual reproduction, and lightavailability iscomplex, and that patterns often differ among species and among forest types. Most studies agree, however, that seedling establishment of long-lived perennials is very rare in shaded forest understorey. Therefore, the ability of individuals to persist, may ultimately bedetermined by patternsof clonal growthand vegetative spread. previous occurrence ofa canopy gap in that location. Similar relationships and Williamson, 1975). Preliminary studies of the distributional patterns of understorey shrubs in the genus Piper within the understorey ofprimary and secondary rainforest suggest that some species are distributed differentially with regard to light availability and sunfleck activity (C. B. Field and R. L. Chazdon, unpub lished data). Piper hispidum occurs in both early successional and primary forest habitats; two-thirds of the plants sampled received from 70 to 90 potential minutes of sunflecks per day (yearly average, based on hemispheri cal photographs). In contrast, all ofthe individuals ofthe understorey species C. Sunflecks, Canopy Gaps and Species Distributions Forest understorey species exhibit many different patterns of spatial distribu tion, ranging from random to highly clumped patterns. Relationships between distributional patterns and sunfleck activity have been studied in detail foronly a few species. Seedings of redfir (Abies magnified) showpatchy distributions on south-facing slopes over much of their range. In a study by Ustin et al. (1984), low levels of sunfleck activity were the environmental factor that best accounted for the clumped distribution of these seedlings. Areas with low seedling density received long sunflecks at midday, with PFD at full-sun intensity. Incontrast, areas with high seedling density had smaller canopy openings with shorter, less intense sunflecks. These areas closely p aequale and P. amalago received less than 70 min ofpotential sunflecks per day- p. Vertical Distribution ofUnderstorey Species ^ siinplc theoretical model of the three-dimensional distribution of light vvrithin forests shows that, immediately below the canopy, light levels exhibit extremely high variance. At this level, a point is either directly beneath a crown or directly in a gap (Terborgh, 1985). At greater depths below the canopy, however, a higher fraction of the space along a horizontal plane receives at least some direct radiation. The horizontal variance decreases, because the expanding cones of light beneath alternate canopy openings 52 R- L. CHAZDON eventually intersect. Based on this "sunfleck" model of vertical canopy structure, Terborgh (1985) hypothesized that understorey (midlayer) trees with their crowns in this intersection plane will maximize photosynthetic production and reproductive output because of improved light conditions. Understorey trees, such as Cornusflorida^ should therefore grow in height up to, but not exceeding this level. The height of the intersection plane can be predicted based on measured angular distributions of canopy openings. Terborgh (1985) compared heights of the understorey tree stratum with the predicted height of the intersection plane, and found close agreement. An implication of this model is that the vertical distribution of sunflecks is an important controller of the vertical distribution of plant species. The sunfleck model is also useful for predicting the extent of stratification within forests at different latitudes. In boreal forests, an intersection plane is not predicted above ground level; no woody substratum is observed in these forests. In contrast, tropical forests are predicted to haveat leastone additional canopylayerbecause light is able to penetrate the canopy at relatively shallow angles (Terborgh, 1985). Vin. CONCLUSIONS A. The Importance of Sunflecks: Scaling Up From Leaves to Whole Plants Responses of understorey plants to sunflecks can be found at many different levels (Tables 1 and 2). Leaves show increased photosynthetic rates and stomatal conductance, plants often gain more biomass and produce more propagules, and some plant populations become restricted in their distribu tion within the forest. Although sunfleck activity may be highly correlated with these biological processes, these correlations do not necessarily imply that sunfleck activity plays a causal role. Despite greater logistical difficul ties, the causal effects of direct radiation during sunflecks are far easier to interpret for individual leaves than for whole plants and populations. Because of the non-linearity of photosynthetic responses, different pat terns of light variation can yield different photosynthetic outcomes even when total PFD remains constant (Chazdon, 1986). High PFD during sunflecks may have detrimental effects on photosynthesis and water-use that could not be predicted from a knowledge of daily total PFD (Young and Smith, 1979; Powles and Bjorkman, 1981). For other biological processes, however, it appears that the incidence of direct radiation during sunflecks may be important in a strictly quantitative sense. Changes in photosynthetic capacity in leaves grown in different regimes appear to SUNFLECKS AND UNDERSTOREY PLANTS 53 depend on integrated PFD rather than on instantaneous values (Chabot et al., 1979). Recent studies have shown that relative growth rate may be linearly related to sunfleck activity (Pearcy, 1983), and that sexual repro ductive allocation is linearly related to patch light level (Pitelka et al., 1980). Thus, even though photosynthetic responses to light are non-linear, scaling up from leaves to whole plants may effectively linearize many of these relationships. Whatever the basis, the apparent linearization and integration of organis- mal responses to changing light conditions make it exceedingly difficult to quantify the influence of sunflecks as opposed to other components of the light environment (diffuse irradiance, light quality) on plant responses. In this regard, computer simulations of whole-plant carbon balance andgrowth may be the most useful technique for elucidating the mechanisms by which whole plants respond dynamically to spatial and temporal light fluctuations. B. Directions for Future Research As this review amply demonstrates, many gaps remain in our understanding of sunfleck utilization by leaves, whole plants, and populations. Below, I discuss the subject areas that, in my view, are most in need of investigation. We still know relatively little about sunfleck frequency, duration, and intensity within coniferous and deciduous temperate forests and tropical wet and dry forests. In particular, no sunfleck data have been published on tropical dry forests. Moreover, few studies have addressed the extent of seasonal variation in sunfleck activity in these different forest types. Light measurements need to be sufficiently frequent to account for light variation during even the shortest sunflecks. These studies should also incorporate analyses ofthe spatial scale ofsunfleck activity. Except for investigations ofleaf movements in Oxalis (Powles and Bjork man, 1980' morphological responses of leaves and whole plants to sunflecks jiave not been investigated in natural populations. These responses range from modifications of leaf structure to changes in leaf orientation and (janopy structure. These studies should also take into account changes in spectral quality associated with sunfleck activity. The effect ofplant canopy structure on variation in sunfleck activity among leaf microsites is yet another relatively unexplored area. Studies of light acclimation in forest understorey plants have traditionally been concerned with steady-state photosynthetic responses and with plants grown under steady-state conditions. Acclimatory responses to constant light conditions may well affect photosynthetic dynamics, as suggested by laboratory studies (Chazdon and Pearcy, 1986b; Sharkey et a/., 1986). Moreover, we do not know whether growth under fluctuating light conditions 54 SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON may lead to changes in dynamic or steady-state photosynthetic responses. Relatively little is known about constraints on sunfleck utilization in natural populations. These constraints may operate seasonally, such as water stress during the dry season in tropical forests, or on a shorter time-scale, such as photoinhibition and induction loss. Studies of daily courses of light, photosynthesis, stomatal conductance, and leaf temperature within a par ticular forest are needed during different times of the year as well as under different weather conditions. Extensive field studies in conjunction with laboratory investigations will provide a detailed understanding of potential and actual constraints on sunfleck utilization. 55 Anderson, M.C. (1966). Some problems of simple characterization of the light climate in plant communities. In: Light as an Ecological Factor (Ed. by R. Bainbridge, G. C. Evans and O. Rackham), pp. 77-90. Oxford: Blackwell. Anderson, M. C. and Miller, E. E. (1974). Forest cover as a solar camera: penumbral effects in plant canopies. J. appl. Ecol. 11, 691-697. Angevine, M. W. and Chabot, B. F. (1979). Seed germination syndromes in higher plants. In: Topics in Plant Population Biology (Ed. by O. T. Solbrig, S. Jain, G. B. Johnson and P. H. Raven), pp. 188-206. New York: Columbia Univ. Press. Ashmun, J. W. and Pitelka, L. F. (1984). Light-induced variation in the growth and dynamics of transplanted ramets of the understory herb. Aster acuminatus. Oecologia (Berlin) 64, 255-262. Ashmun, J. W., Brown, R. L. and Pitelka, L. F. (1985). Biomass allocation in Aster Although many studies have shown that increases in light availability are acuminatus: variation within and among populations over 5 years. Can. J. Bot. 63, often associated with increases in plant growth and reproductive output, most of these studies have not specifically focused on the role of sunflecks. Long-term field studies are needed to assess the ecological significance of Ashton, P. S. (1958). Light intensity measurements in a rain forest near Santarem, sunflecks as opposed to other components of the environment. Complicating circumstances such as storage, time-lags, and interactions with other envir onmental factors may require long-term studies, experimental approaches, and computer simulations in these investigations. In this era of unprecedented deforestation in the tropics, studies of regeneration of secondary and primary forest trees are greatly needed. During tropical forest succession, rapidly growing trees, such as Ochroma and Cecropia^ quickly form a thin canopy. Within a few years, however, a relatively tall, dense canopy is formed, producing a heavily-shaded understorey. It is under these heavily-shaded conditions that longer-lived trees of secondary and primary forests initially become established. There is a great need for comparative studies of sunfleck activity in successional forests of different ageand composition and of physiological responses of regenerating tree seedlings to sunflecks. Studies in this area would greatly contribute to our understanding of tropical forest regeneration, forest management, and reforestation efforts. 2035-2043. Brazil. J. Ecol. 46, 65—70. Atkins, W. R. G. and Poole, H. H. (1926). Photo-electric measurements of illumina tion in relation to plant distribution. Part I. Sci. Proc. Roy. Dublin Soc. N.S. 18, 277—298 Atkins W. R. G., Poole, H. H. and Stanbury, F. A. (1937). The measurement of the intensity and the colour of the light in woods by means ofemission and rectifer photoelectric cells. Proc. Roy. Soc. B. 121, 427^50. Augspurger C K. (1984). Seedling survival of tropical tree species: Interactions of dispersal'distance, light-gaps and pathogens. Ecology 65, 1705-1712. Augspurger, C. K. and Kelly, C.K. (1984). Pathogen mortality of tropical tree seedlings' Experimental studies of the effects of dispersal distance, seedling density and light conditions. Oecologia (Berlin) 61, 211-217. Baldocchi,'D. D., Hutchison, A., Matt, D. R. and McMillen, R. T. (1984). Seasonal variations in the radiation regime within an oak-hickory forest. Agric. For. Meteorol. 33, 177—191. Bazzaz, F.A. and Carlson, R. W. (1982). Photosynthetic acclimation to variability in the light environment ofearly and late successional plants. Oecologia (Berlin) 54, Ba^az, F. A. and Reekie, E.G. (1985). The meaning and measurement of repro ductive effort in plants. In: Studies on Plant Demography (Ed. by J. White), 373-387. London: Academic Press. Ba^az, F. A., Chiariello, N. R., Coley, P. D. and Pitelka, L. F. (1987). Allocating resources to reproduction and defense. Bioscience 37, 58-67. ACKNOWLEDGEMENTS I thank A. H. Fitter and an anonymous reviewer for their helpful comments on the manuscript. REFERENCES Anderson, M. C. (1964a). Studies ofthe woodland light climate. I. The photographic computation oflight conditions. 7. Ecol. 52, 27-41. Anderson, M.C. (1964b). Studies of the woodland climate. II. Seasonal variation in theseasonal light climate. J. Ecol. 52, 643-663. Bierzychudek, P. (1981). Pollinator limitation of plant reproductive effort. Am. Naturalist 117, 838-840. Bierzychudek, P. (1982). The demography of Jack-in-the-pulpit, a forest perennial that changes sex. Ecol. Monogr. 52, 335-351. piggs, W.W., Edison, A. R., Easton, J.D., Brown, K.W., Maranville, J.W. and Clegg, M. C. (1971). Photosynthesis light sensor and meter. Ecology 52, 125-131. Bjorkman, O. (1968). Carboxydismutase activity in shade-adapted and sun-adapted species of higher plants. Carnegie Inst. Wash. Yearbook 67, 487-488. Bjorkman, O. (1981). Responses todifferent quantum flux densities. In: Physiological Plant Ecology I. Encyclopedia of Plant Physiology (Ed. by O. L. Lange, P. S. Nobel, C. B. Osmond, H. Ziegler), New Series, Vol. 12A, pp. 57-107. New York: Springer-Verlag. 56 SUNFLECKS AND UNDERSTOREY PLANTS R. L. CHAZDON Bjorkman, O. and Holmgren, P. (1963). Adaptability of the photosynthetic apparatus to hght intensity in ecotypes from exposedand shaded habitats. Physiol. Plant. 16, 889-914. Bjorkman, O. and Ludlow, M. M. (1972). Characterization of the light climate on the floor of a Queensland rainforest. Carnegie Inst. Wash. Yearbook 71, 85-94. Bjorkman, O., Boardman, N. K., Anderson, J. M. and Thorne, S. W. (1972a). Effect of light intensity during growth of Atriplex iriangularis on the capacity of photosynthetic reactions, chloroplast components and structure. Carnegie Inst. Wash. Yearbook 1\, 115-135. Bjorkman, O., Ludlow, M.M. and Morrow, P. S. (1972b). Photosynthetic perform ance of two rainforest species in their native habitat and analysis of their gas exchange. Carnegie Inst. Wash. Yearbook 71, 94—102. Bjorkman, O. and Powles, S. B. (1981). Leaf movement in the shade species Oxalis oregana. I. Response to light level and light quality. Carnegie Inst. Wash. Yearbook 80, 59-62. Blackman, G. E. and Rutter, A.J. (1946). Physiological and ecological studies in the analysis of plant environment. I. The light factor and the distribution of the bluebell (Scilla non-scripta) in woodlandcommunities. Ann. Bot. n.s. 10, 361-390. Blackman, G. E. and Wilson, G. L. (1951). Physiological and ecological studiesin the analysis of plant environment. VIL An analysis of the differential eff'ects of light intensity on the net assimilation rate, leaf-area ratio, and relative growth rate of different species. Ann. Bot. n.s. 15, 373-408. Blackman, G. E. and Wilson, G. L. (1954). Physiological and ecological studies in the analysis of plant environment. IX. Adaptive changes in the vegetative growth and development of Helianthus annuus induced by an alteration in light level. Ann. Bot. n.s. 18, 71-94. Boardman, N. K. (1977). Comparative photosynthesis of sun and shade plants. Ann. Rev. Plant Phys. 28, 355-377. Bohning, R. H. and Burnside, C. A. (1956). The effect of light intensity on rate of apparent photosynthesis in leaves of sun and shade plants. Am. J. Bot. 43, 557561. Brun, W. A. (1972). Rhythmic stomatal opening response in banana leaves. Physiol. Plantarum 15, 623-630. Caemmerer, S. von and Farquhar, G. D. (1981). Some relationships between the biochemistry of photosynthesis and the gas exchange of leaves. Planta 153, 376387. Chabot, B. F., Jurik, T. W. and Chabot, J. F. (1979). Influence of instantaneous and integrated light flux density on leaf anatomy and photosynthesis. Am. J. Bot. 86, 940-945, Chan, S. S., McCreight, R. W., Walstad, J. D. and Spies, T. A. (1986). Evaluating forest vegetative cover with computerized analysis of fisheye photographs, forest Sci. 32, 1085-1091. Chazdon, R. L. (1984). Ecophysiology and architecture of three rain forest understory palm species, Ph.D. Thesis, Cornell University, Ithaca, New York. Chazdon, R. L. (1986). Light variation and carbon gain in rain forest understorey palms, J- Ecol. 74, 995-1012. Chazdon, R. L. (1987). Aspectos importantes para el estudio de los regimenes de luz en bosques tropicales. Rev. Biol. Trop. 35 (Suppl. 1), 191-196. Chazdon, R. L. and Field, C. B. (1987a). Determinants of photosyntheticcapacity in six rainforest Piper species. Oecologia (Berlin) 73, 222-230. 57 Chazdon, R. L. and Field, C. B. (1987b). Photographic estimation of photosynthetically active radiation: Evaluation of a computerized technique. Oecologia (Berlin) 73, 525-532. Chazdon, R. L. and Fetcher, N. (1984a). Photosynthetic light environments in a lowland tropical rainforest in Costa Rica. J. Ecol. 72, 553-564. Chazdon, R. L. and Fetcher, N. (1984b). Light environments of tropical forests. In; Physiological Ecology of Plants of the Wet Tropics (Ed. by E. Medina, H. A. Mooney and C. Vazquez-Yanes), pp. 27-36. The Hague: Junk. Chazdon, R. L. and Pearcy, R. W. (1986a). Photosynthetic responses to light variation in rain forest species. I. Induction under constant and fluctuating light conditions. Oecologia (Berlin) 69, 517-523. Chazdon, R. L. and Pearcy, R. W. (1986b). Photosynthetic responses to hght variation in rainforest species. II. Carbon gain and photosynthetic eflSciency during lightflecks. Oecologia (Berlin) 69, 524-531. Chazdon, R. L., Williams, K. and Field, C. B. (1988). Interactions between crown structure and light environment in five rainforest Piper species. Am. J. Bot., in press. Chiariello, N. (1984). Leaf energy balance in the wet lowland tropics. In: Physiologi calEcology of Plants of the Wet Tropics (Ed. by E. Medina, H. A. Mooney and C. Vazquez-Yanes), pp. 85—98. The Hague: Junk. Clark D B and Clark, D. A. (1987). Leaf production and the cost of reproduction in a tropical rain forest cycad, Zamia skinneri. J. Ecol., in press. Coombe D E (1957). Thespectral composition of shade light in woodlands./. £'co/. 45, 823-830. Coombe D. E. (1966). Theseasonal light climate and plant growth in a Cambndgeshire wood In: Light as an Ecological Factor (Ed. by R. Bainbridge, G. C. Evans, and O. Rackman), pp. 148-166. Oxford: Blackwell. Corre W J. (1983). Growth and morphogenesis of sun and shade plants. II. The influence of light quality. Acta. Bot. Neerlandica 32, 185-202. rurtis W. F. and Kincaid, D. T. (1984). Leafconductanceresponses of Viola species from sun and shade habitats. Can. J. Bot. 62, 1268-1272. navies W. J- and Kozlowski, T. T. (1975). Stomatal responses to changes in light intensity as influenced by plant water stress. For. Sci. 21, 129-133. P (1983). Water relations pattern of understorey species influenced by sunflecks. Biol. Plant. 25, 68-74. Ellenberg, H. (1963). Vegetation Mitteleuropas mit den Alpen. Stuttgart: Eugen Ulmer. 989 pp. pvans, G. C. (1939). Ecological studies on the rain forest ofsouthern Nigeria II. The atmospheric environmental conditions. J. Ecol. 27, 436-462. gvans, G. C. (1956). An area survey method ofinvestigating the distribution oflight intensity in woodlands, with particular reference to sunflecks. J. Ecol. 44, 391-428. gvans, G. C. (1966). Model and measurement in the study ofwoodland light climates. In: Light as an Ecological Factor (Ed. by R. Bainbridge, G.C. Evans, and O, Rackman), pp. 53-76. Oxford: Blackwell. Evans, G.C. and Coombe, D. E. (1959), Hemispherical and woodland canopy photography and the light climate. J. Ecol. 47, 103-113. gvans, G. C., Whitmore, T. C. and Wong, T. K. (1960). The distribution of light reaching the ground vegetation in a tropical rainforest. J. Ecol. 48, 193-204. Evans, G. C., Freeman, P. and Rackham, O. (1975), Developments in hemispherical photography. In: Light as an Ecological Factor. II (Ed. by G. C. Evans, R. Bainbridge and O. Rackham), pp. 549-556. Oxford: Blackwell. R. L. CHAZDON 58 Federer, C. A. and Tanner, C. B. (1966). Spectral distribution of light in the forest. Ecology 4n^ 555-560. Fetcher, N., Oberbauer, S. F. and Strain, B. R. (1985). Vegetation effects on microclimate in lowland tropical forest in Costa Rica. Int. J. Biometeorol. 29, 145155. Fetcher, N., Oberbauer, S. F., Rojas, G. and Strain, B. R. (1987). Efectos del regimen de luz sobre la fotosintesis y el crecimiento en plantulas de arboles de un bosque lluvioso tropical de Costa Rica. Rev. Biol. Trop. 35 (Suppl. 1), 97-110. Fonteno, W. C. and McWilliams, E. L. (1978). Light compensation points and acclimatization of four tropical foliage plants. J. Am. Soc. hort. Sci. 103, 52-56, Gomez-Pompa, A. and Yazquez-Yanes (1985). Estudios sobre la regeneracion de selvas en regiones calido-humedas de Mexico. In: Investigaciones sobre la regenera cion de selvas alias en Veracruz, Mexico. II (Ed. by A. Gomez-Pompa and S. del Amo R.), pp. 1-25. Mexico: Editorial Alhambra Mexicana. Gong, W. K. (1981). Studies on the natural regeneration of a hill Dipterocarp species, Hopea pedicellata. Malaysian Forester 44, 357-369. Goodchild, D.J., Bjorkman, O. and Pyliotis, N. A. (1972). Chloroplast ultrastructure, leaf anatomy, and content of chlorophyll and soluble protein in rainforest species. Carnegie Inst. Wash. Yearbook 71, 102-107. Gregory, F. G. and Pearse, H. L. (1937). The effects on the behavior of stomata of alternative periods of light and darkness of short duration. Ann. Bat. n.s. 1, 3-10. SUNFLECKS AND UNDERSTOREY PLANTS 59 Holmes, M. G. and Smith, H. (1977b). The function of phytochrome in the natural environment. IV. Light quality and plant development. Photochem. Photobiol. 25, 551-557. Horowitz, J. L. (1969). An easily constructed shadow-band for separating direct and diffuse solar radiation. Solar Energy 12, 543-545, Hughes, A, P, (1959). Effects of the environment on leaf development in Impatiens parviflora D.C. J. Linn. Soc. (Bot.) 56, 161-165. Hughes, A. P. (1966). The importance of light compared with other factors affecting plant growth. In: Light as an Ecological Factor (Ed. by R. Bainbridge, G. C. Evans and O. Rackham), pp. 121-147. Oxford: Blackwell. Hutchison, B. A. and Matt, D. R. (1976). Beam enrichment of diffuse radiation in a deciduous forest. Agric. Meteorol. 17, 93-110. Hutchison, B. A. and Matt, D. R, (1977), The distribution of solar radiation within a deciduous forest, Ecol. Monogr. 47, 185-207. Jupp, D, L. B., Anderson, M. C., Adomeit, G. M. and Witts, S. J. (1980). Pisces—a computer program for analysing hemispherical canopy photographs. CSIRO Technical Memorandum 80123. Canberra. Jurik, T. W, (1983). Reproductive effort and COj dynamics of wild strawberry populations. Ecology 1329-1342. Jurik, T. W. (1985). Differential costs of sexual and vegetative reproduction in wild strawberry populations, Oecologia (Berlin) 66, 394-403. Gross, L. J. (1982). Photosynthetic dynamics in varying light environments: a model and its application to whole leaf carbon gain. Ecology 63, 84-93. Gross, L. J. (1984). Reply to McCree and Loomis. Ecology 65, 1018-1019. Jurik, T. W., Chabot, J, F, and Chabot, B, F. (1979), Ontogeny of photosynthetic performance in Fragaria virginiana under changing light regimes. Plant Physiol. Gross, L. J. (1986). Photosynthetic dynamics and plant adaptation to environmental Kawano, S., Takasu, H, and Nagai, Y. (1978). The productive and reproductive biology of flowering plants. IV. Assimilation behavior of some temperate wood variability. Lectures Math. Life Sci. 18, 135-169. Gross, L. J. and Chabot, B. F. (1979). Time course of photosynthetic response to changes in incident light energy. Plant Physiol. 63, 1033-1038. Grubb, P. J. and Whitmore, T. C. (1967). A comparison of montane and lowland forest in Ecuador. III. The light reaching the ground vegetation, J. Ecol. 55,33-57. Gutschick, V. P,, Barron, M. H., Waechter, D. A. and Wolf, M. A. (1985). Portable monitor for solar radiation that accumulates irradiance histograms for 32 leafmounted sensors. Agric. For. Meteorol. 33, 281-290. Harbinson, J. and Woodward, F. I, (1984). Field measurements of the gas exchange of woody plant species in simulated sunflecks. Ann. Bat. 53, 841-851. Hariri, M, and Prioul, J, L. (1978), Light-induced adaptive responses under green house and controlled conditions in the fern Pteris cretica var ouvardii. II. Photosynthetic capacities. Physiol. Plantarum42, 97-102. Harvey, G. W. (1980), Seasonal alteration of photosynthetic unit sizes in three herb layer components of a deciduous forest community. Am. J. Bot. 67, 293-299. Herbert, T. J. (1988). Area projections of fisheye photographic lenses. Agric. For. Meteorol., in press. Hicks, D. J. and Chabot, B, F, (1985). Deciduous forest. In: Physiological Ecology of North America Plant Communities (Ed. by B, F. Chabot and H. A, Mooney), pp, 257-277. London: Chapman and Hall. Hill, R. (1924). A lens for whole sky photographs. Q. J. Roy. met. Soc. 50, 227235. Holmes, M. G. and Smith, H. (1977a), The function of phytochrome in the natural environment. II. The influence of vegetation canopies on the spectral energy distribution of natural daylight. Photochem. Photobiol. 25, 539-545. 63, 542-547, land herbs. J. Coll. Lib. Arts Toyama Univ. (Nat. Sci.) 11, 33-60. Kawano, S., Masuda, J., Takasu, H. and Yoshie, F. (1983). The productive and reproductive biology of flowering plants. XI. Assimilation behavior of several evergreen temperate woodland plants and its evolutionary-ecological impli cations. J. Coll. Lib. Arts Toyama Univ. (Nat. Sci.) 16, 31-65, Kirschbaum, M, and Pearcy, R. W. (1988), Gas exchange analysis of the relative importance of stomatal and biochemical factors in photosynthetic induction in Alocasia macrorrhiza. Plant Physiol. 86, 782-785, Knapp, A, K, and Smith, W,K. (1987). Stomatal and photosynthetic responses during sun/shade transitions in subalpine plants: influence on water useefficiency. Oecologia (Berlin) 74, 62-67. Kozlowski, T.T. (1957). Effect of continuous high light intensity on photosynthesis of forest tree seedlings. For. Sci. 3, 221-224. Kriedemann, P., Torokfalvy, S. and Smart, R.E. (1973). Natural occurrence and photosynthetic utilization of sunflecks in grapevine leaves. Photosynthetica 7, 18-27. Langenheim, J. H., Osmond, C, B,, Brooks, A, and Ferrar, P, J, (1984). Photosynthe tic responses to light in seedlings of selected Amazonian and Australian rainforest tree species. Oecologia (Berlin) 63, 215-224. Lawrence, W.T, (1984), Photosynthetic response of a tropical understory species to naturally occurring sunflecks. Plant Physiol. (Suppl, 1) 5, Lee, D. W, (1987). The spectral distribution of radiation in two neotropical rainfor ests. Biotropica 19, 161-166. Lundegardh, H. (1922). Ziir Physiologic und Okologie der Kohnlensaurassimilation. Biol. Zbl. 42, 337-358. 60 R. L. CHAZDON McCree, K. J. and Loomis, R. S. (1969). Photosynthesis in fluctuating light. Ecology 50, 422-428. SUNFLECKS AND UNDERSTOREY PLANTS 61 Pearcy, R. W. and Calkin, H. (1983). Carbon dioxide exchange of C3 and C4 tree species in the understory of a Hawaiian forest. Oecologia (Berlin) 58, 26-32. McCree, K.J. and Loomis, R. S. (1984). Photosynthetic dynamics—a comment. Ecology 65, 1016-1018. Marks, T. C. and Taylor, K. (1978).The carbon economy of Rubus chamaemorus L. I. Photosynthesis. Ann. Bot. 42, 165-179. Marquis, R. J., Young, H. J. and Braker, H. E, (1986). The influence of understory Pearcy, R. W., Osteryoung, K. and Calkin, H. (1985). Photosynthetic responses to dynamic light environments by Hawaiian trees. Plant Physiol. 79, 896-902. Pearcy, R. W., Chazdon, R. L. and Kirschbaum, M. U. F. (1987a), Photosynthetic utilization of lightflecks by tropical forest plants. In: Progress in Photosynthesis Research, Vol. IV (Ed. by J. Biggens), pp. 257-260. Dordrecht: Martinus Nijhoff. vegetation cover on germination and seedling establishment in a tropical lowland Pearcy, R. W., Bjorkman, O., Caldwell, M. M., Keeley, J. E., Monson, R. K. and Strain, B. R. (1987b). Carbon gain by plants in natural environments. BioSci. 37, wet forest. Biotropica 18, 273-278. Masarovicova, E. and Elias, P. (1986). Photosynthetic rate and water relations in 21-29. some forest herbs in spring and summer. Photosynthetica 20, 187-195. Meinzer, F. C. (1982). Models of steady-state and dynamic gas exchange responses to Pinero, D. and Sarukhan, J. (1982). Reproductive behavior and its individual vapor pressure and light in Douglas Fir (Psudotsugamenziesii) saplings. Oecologia Pitelka, L. F. and Curtis, W. F. (1986). Photosynthetic responses to light in an (Berlin) 55, 403-408. Miller, E. E. and Norman, J. M. (1971). A sunfleck theory for plant canopies. I. Lengths of sunlit segments along a transect. Agron. J. 63, 735-738. Mooney, H. A., Field, C. B., Vazquez-Yanes, C. and Chu, C. (1983). Environmental controls on stomatalconductance in a shrub of the humid tropics. Proc. nat. Acad. 5d. 80, 1295-1297. Morgan, D. C. and Smith, H. (1978). Simulated sunflecks have large, rapid effects on plant stem extension. Nature ITi, 534-536. Nobel, P. S. (1976). Photosynthetic rates of shade versus shade leaves of Hyptis emoryi Tow. Plant Physiol. 58, 218-223. Nobel, P. S. and Hartsock, T. L. (1981). Development of leaf thickness for Plectran- thus parviflorus—ln^ntnce of photosynthetically active radiation. Physiol. Plant. 51, 163-166. Norman, J. M., Miller, E. E. and Tanner, C. B. (1971). Light intensity and sunflecksize distributions in plant canopies. Agron. J. 63, 743-748. Oberbauer, S. F., Clark, D. B.,Clark, D. A. and Quesada, M. A. (1988). Crown light environments of saplings of two species of rain forest emergent trees. Oecologia (Berlin) 75, 207-212. Oker-Blom, P. (1984). Penumbral effects of within-plant shading on radiation distribution and leaf photosynthesis: a Monte-Carlo simulation. Photosynthetica 18, 522-528. Orozco-Segovia, A.D. L. (1986). Fisiologia ecologia delphotoblastismo en semillas de cuatro especiesdelgenero Piper L. Ph.D. Thesis, Universidad Nacional Autonoma de Mexico, Mexico, 123 pp. Osmond, C. B. (1983). Interactions between irradiance, nitrogennutrition, and water stress in the sun-shade responses of Solanum dulcamara. Oecclogia (Berlin) 57, 316-321. Pearce, R. B.and Lee, D. R. (1969). Photosynthetic and morphological adaptation of alfalfaleaves to lightintensity at different stagesof maturity. Crop Sci. 9,791-794. Pearcy, R. W. (1983). The lightenvironment and growth of Cj and C4 species in the understory of a Hawaiian forest. Oecologia (Berlin) 58, 26-32. Pearcy, R. W. (1987a). Photosynthetic gas exchange responses of Australian tropical forest trees in canopy, gap and understory microenvironments. Functional Ecology 1, 169-178. Pearcy, R. W. (1988a). Photosynthetic utilization of lightflecks by understory species. Austr. J. Plant Physiol. 15, in press. Pearcy, R. W. (1988b). Radiation and light measurements. In: Physiological Plant Ecology: Field Methods and Instrumentation. (Ed. by R. W. Pearcy, J. R. Ehleringer, and P. W. Rundel). London: Chapman and Hall, in press. variability in a tropical palm, Astrocaryum mexicanum. J. Ecol. 70, 461-472. understory herb. Aster acuminatus. Am. J. Bot. 73, 535-540. Pitelka, L. F., Stanton, D. S. and Peckenham, D.O. (1980). Effects of light and density on resource allocation in a forest herb. Aster acuminatus (Compositae). Am. J. Bot. 67, 942-948. Pitelka, L. F., Ashmun, J. W. and Brown, R. L. (1985). The relationships between seasonalvariation in lightintensity, rametsize, and sexual reproductionin natural and experimental populations of Aster acuminatus (Compositae). Am. J. Bot. 72, 311-319. Pollard, D.F.W. (1970). The effect of rapidly changing light on the rate of photosyn thesis in largetooth aspen {Populus grandidentata). Can. J. Bot. 48, 823-829. powles, S. B. and Bjorkman, O. (1981). Leaf movement in the shade species Oxalis oregana. II.Role inprotection against injury by intense light. Carnegie Inst. Wash. Yearbook 80, 63-66. powles, S.B. and Thome, S.W. (1981). Effect of high light treatments in inducing photoinhibition of photosynthesis in intact leaves of low-light grown Phaseolus vulgaris and Lastreopsis microsora. Planta 152, 471-477. Rabinowitch, E. I. (1956). Photosynthesis and Related Processes, Vol. II Part 2. New York: Interscience. Rackham, O.(1975). Temperatures ofplant communities asmeasured bypyrometric and other methods. In: Light as an Ecological Factor. II (Ed. by G.C. Evans, R. Bainbridge and O. Rackham), pp. 423-450. Oxford: Blackwell. Reifsnyder, W. E., Furnival, G. M. and Horovitz, J.L. (1971). Spatial and temporal distribution ofsolar radiation beneath forest canopies. Agric. Meteorol. 9,21-37. Rich, P.M., Clark, D. B., Clark, D. A. and Oberbauer, S.F. (1987). Canopy photography for assessment oflocal light environment oftropical forest trees and palms. Bull. Ecol. Soc. Am. 68, 397. Richards, P. and Williamson, G. B. (1975). Treefalls and patterns of understory species in a wet lowland tropical forest. Ecology 56, 1226-1229. Robichaux, R. H. and Pearcy, R. W. (1980). Photosynthetic responses ofCj and C4 species from cool shaded habitats in Hawaii. Oecologia (Berlin) 47, 106-109. Salisbury, E. J. (1916). The oak-hornbeam woods of Hertfordshire. Parts I and II. J. Ecol. 4, 83-117. Salminen, R., Nilson, T., Hari, P., Kaipiainen, L. and Ross, J. (1983). A comparison ofdifferent methods formeasuring thecanopy light regime. J. appl. Ecol. 20,897904. Sarukhan, J., Martinez-Ramos, M. and Pinero, D. (1984). The analysis of demo graphic variability at the individual level and its population consequences. In: Perspectives on Plant Population Ecology (Ed. by R. Dirzo and J. Sarukhan), pp. 83-106. Sinauer: Sunderland, MA. 62 R. L. CHAZDON Sasaki, S. and Mori, T. (1981). Growth responses of dipterocarp seedlings to light. Malaysian Forester 44, 319-345. Sasaki, S., Mori, T and Ng, F. S. P. (1981). Seedling growth under various light conditions in the tropical rain forest. In: Proc. XVII lUFRO World Congress. Kyoto, Japan, 1981, 79-85. Sharkey, T. D., Seemann, J. R. and Pearcy, R. W. (1986). Contribution of metabo lites of photosynthesis to postillumination CO2 assimilation in response to lightflecks. Plant Physiol. 82, 1063-1068. Shirley, H. L. (1929). The influence of light intensity and light quality upon the growth of plants. Am. J. Bot. 16, 354. Smith, A. P. (1987). Respuestas de hierbas del sotobosque tropical a claros ocasionados por la caida de arboles. Rev. Biol. Tropical 35 (Suppl. 1), 111-119. Smith, W. K. (1981). Temperature and water relation patterns in subalpine, understory plants. Oecologia (Berlin) 48, 353-359. Smith, W. K. (1985). Western montane forests. In: Physiological Ecology of North American Plant Communities (Ed. by B. F. Chabot and H. A. Mooney), pp. 95126. London: Chapman and Hall. Sohn, J. J. and Policansky, D. (1977). The costs of reproduction in the mayapple Podophyllum peltatum (Berberidaceae). Ecology 58, 1366-1374. Solbrig, O. T. (1981). Studies on the population biology of the genus Viola. II. The effect of plant size on fitness in Viola sororia. Evolution 35, 1080-1093. Sparling, J. H. (1967). Assimilation rates of some woodland herbs in Ontario. Bot. Gaz. 128, 160-168. Taylor, R. J. and Pearcy, R. W. (1976). Seasonal patterns of the CO2 exchange characteristics of understory plants from a deciduous forest. Can. J. Bot. 54,10941103. Terborgh, J. (1985). The vertical component of plant species diversity in temperate and tropical forests. Am. Nat. 126, 760-776. Ustin, S. L., Woodward, R. A., Barbour, M. G. and Hatfield, J. L. (1984). Relation ships between sunfleck dynamics and red fir seedling distribution. Ecology 65, 1420-1428. Vazquez-Yanes, C. (1976). Estudios sobre ecofisiologia de la germinacion en una zona calido-humeda de Mexico. In: Regeneracion de Selvas (Ed. by A. GomezPompa, S. del Amo and A. Butanda), pp. 279-387. Mexico: Editorial Continental. Vazquez-Yanes, C. and Orozco-Segovia, A, (1984). Ecophysiology of seed germina tion in thetropical humid forests of theworld: a review. In: Physiological Ecology of Plants of the Wet Tropics (Ed. by E. Medina, H. A. Mooney and C. VazquezYanes), pp. 37-50. The Hague: Junk. Walker, D. A. (1981). Photosynthetic induction. In: Proc 5th Int. Cong. Photosyn, Vol. IV (Ed. by G. Akoyonoglou), pp. 189-202. Philadelphia; Balaban Int. Sci. Series. Walters, M. B. and Field, C. B. (1987). Photosynthetic light acclimation in two rainforest Piper species with different ecological amplitudes. Oecologia (Berlin) 72, 449^56. Weber, J. A., Jurik, T. W., Tenhunen, J. D. and Gates, D. M. (1985). Analysis of gas exchange in seedlings of Acer saccharum: integration of field and laboratory studies. Oecologia (Berlin) 65, 338-347. Whitmore, T. C. and Wong, T. K. (1959). Patterns of sunfleck and shade in tropical rain forest. Malayan Forester 22, 50-62. View publication stats SUNFLECKS AND UNDERSTOREY PLANTS 63 Willis, A. J. and Balasubramaniam, S. (1968). Stomatal behavior in relation to rate of photosynthesis and transpiration in Pelargonium. New Phytol. 67, 265-285. Woods, D. R. and Turner, N. C. (1971).Stomatal responses to changing light by four tree species of varying shade tolerance. New Phytol. 70, 77-84. Woodward, F. I. (1981). Shoot extension and water relations of Circaea lutetiana in sunflecks. In: Plants and Their AtmosphericEnvironment(Ed. J. Grace), pp. 83-91. Oxford: Blackwell. Woodward, F.I. and Yaqub, M. (1979). Integrator and sensors for measuring photosynthetically active radiation and temperature in the field. J. appl. Ecol. 16, 545-552. Yoshie, F. and Kawano, S. (1986). Seasonal changes in photosynthetic characteristics oT Pachysandra terminalis (Buxaceae), an evergreen woodlandchamaephyte, in the cool temperate regions of Japan. Oecologia (Berlin) 71, 6-11. Yoshie, F. and Yoshida, S. (1987). Seasonal changes in photosynthetic characteristics of Anemone raddeana, a spring-active geophyte in the temperate region of Japan. Oecologia (Berlin) 72, 202-206. Young, D. R.(1985). Microclimatic effects onwater relations, leaf temperatures, and thedistribution of Heracleum lanatum at high elevations. Am. J. Bot. 72,357-364. Young, D. R. and Smith, W. K. (1979). Influence ofsunflecks on the temperature and water relations of two subalpine understory congeners. Oecologia (Berlin) 43, 195-205. Young, D.R. and Smith, W. K. (1980). Influence of sunlight on photosynthesis, water relations and leaf structure m the understory species Arnica cordifolia. Ecology 61, 1380-1390. Young, D.R- and Smith, W.K. (1982). Simulation studies on the influence of understory location on the water and photosynthetic relations of Arnica cordifolia Hook. Ecology 63, n6\-\m. Young, D. R. and Smith, W. K. (1983). Effect ofcloudcover on photosynthesis and transpiration mthe subalpine understory species Arnica latifolia. Ecology 64,681687.