BIOTROPICA 31 (4): 570-578 1999 Responses of Tropical Understory Plants to a Severe Drought: Tolerance and Avoidance of Water Stress1 Michael F. Tobin2 Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Panama Omar R. Lopez3 Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Panama and Thomas A. Kursar Department of Biology, University of Utah, Salt Lake City, Utah 84112, U.S.A. ABSTRACT Shade-tolerant understory shrubs and subcanopy trees constitute most of the woody species in Neotropical moist forest, but studies demonstrating physiological differences among these species are few. Shade-tolerant species that coexist in the forest understory exhibit differences in leaf life span that have been associated with variation in physi- ological traits. We hypothesized that water relations of understory species with widely divergent leaf life spans differ in response to drought. Although severe drought is infrequent in Neotropical moist forest, we studied the water relations of shade-tolerant understory species with short or long leaf life spans during the severe 1991-1992 dry season on Barro Colorado Island, Panama. The predawn leaf watet potential declined to -2.8 and -3.6 MPa during the dry season in Hybanthus prunifolius and Psycbotria horizontalis, respectively, two species with short leaf life spans, but remained above -1.3 MPa in two species with long leaf life spans, Swartzia simplex and Ouratea lucens. The midday leaf water potential dropped as low as -3.4 and -4.5 MPa for H. prunifolius and P. horizontalis, respectively. The osmotic potential of H. prunifolius and P. horizontalis and another species with short leaf life span, Alseis bkckiana, decreased early in the dry season, a period during which all three had substantially negative predawn water potential. In contrast, the osmotic potential of S. Simplex, O. lucens, and Licania platypus, a third species with long leaf life span, declined late in the dry season, even though we observed little change in predawn watet potential for 5. simplex and O. lucens. We conclude that the variable and potentially severe dry season in Neotropical moist forest can be sufficiently intense to severely limit soil moisture availability for understoty plants. H. prunifolius and P. horizontalis tolerated dehydration, whereas 5. simplex and O. lucens postponed dehydration. relations. IN NEOTROPICAL MOIST FORESTS, PLANTS EXPERIENCE studies have examined the physiological differences MODERATE WATER STRESS REGULARLY and severe water among these species. Tropical shade-tolerant species deficits every few years, even though soil moisture differ considerably in physiology, including is abundant throughout most of the year (Robi- drought acclimation (Wright et al. 1992), light in- chaux et al 1984). Soil moisture can decrease dra- duction of photosynthesis (Kursar & Coley 1993), matically during the annual dry season in the trop- and a number of other physiological traits (Mulkey ical moist forest (Becker et al. 1988). Understory et al. 1993). plants have been observed to experience substantial The leaf life span of tropical shade-tolerant spe- water stress and even wilt during the dry season cies also varies considerably, ranging from nine (Robichaux et al. 1984, Wright 1991). months to five years (Coley 1981). A long leaf life Even though shade-tolerant shrubs and subcan- span has been suggested to improve carbon balance opy trees constitute most of the woody species in and nutrient-use efficiency, thereby providing an tropical moist forest (Hubbell & Foster 1992), few advantage in environments poor in nutrients, light, and water (Chapin 1980, Chabot & Hicks 1982). ' Received 2 May 1997; revision accepted 23 April 1998. ^ variation in leaf lib span among tropical tree % Present address: Department of Forest Resources, Uni- speacst? been corfelated with leaf physiology and versity of Minnesota, St. Paul, Minnesota 55108, U.S.A. morphology (Reich et al. 1991). Additionally, leaf 3 Present address: Department of Biology, University of kre span has been related to leaf, plant, and stand Utah, Salt Lake City, Utah 84112, U.S.A. traits across ecosystems and biomes (Reich et al. 570 Responses of Understory Plants to Severe Drought 571 Nov I Dec I Jan I Feb I Mar I Apr I May ' Jim FIGURE 1. Rainfall and soil moisture in a tropical moist forest on Barro Colorado Island for the 1991-1992 study period. The bars represent rainfall and the open circles represent gravimetric soil water content at 30 to 40 cm. Error bars represent SE (data from Smithsonian Environmental Studies Program). 1992). These studies suggest that the leaf life span of a species is only one of a suite of interdependent physiological ttaits that consistently co-occur. Var- iation in leaf life span among shade-tolerant species may reflect their physiology and constrain their re- sponses to changes in resource availability in con- trasting ways. We evaluated the response of shade-tolerant understory plants to a severe dry season in a trop- ical moist forest on Barro Colorado Island (BCI) in Panama. We hypothesized that shade-tolerant species with divergent leaf life spans differ in their response to drought. To test our hypothesis, we compared species with short leaf life spans (1-2 yr) to those having long leaf life spans (3-6 yr). Since we predicted that species with similar leaf life spans would be physiologically convergent, unrelated spe- cies were compared to avoid possible phylogenetic constraints on physiology. We evaluated the devel- opment of water stress through the dry season by monitoring seasonal changes in predawn and mid- day leaf water potential (^J. Acclimation to drought was assessed by measuring changes in leaf water relations through the dry season using the pressure-volume technique. MATERIALS AND METHODS SITE AND SPECIES.?The plants monitored in this study occur in the understory of the tropical moist forest on Barro Colorado Island, Republic of Pan- ama (9?9'N, 79051'W). The island receives an av- erage annual precipitation of 2600 mm and expe- riences a pronounced dry season from mid-Decem- ber to April or May (Fig. 1; Windsor 1990). Soil water potentials decline steadily during the dry sea- son, increasing only after heavy rains (Wright 1991), and have been measured as low as ?2.0 to ? 2.3 MPa at a depth of 20 cm in the understory (Becker et al. 1988). The following six shade-tolerant species were chosen for this study because of the large variation in the longevity of their leaves (species after Croat 1978): the shrub Hybanthus prunifolius (Schult.) Schultze (Violaceae), the shrub Psychotria horizon- talis Sw. (Rubiaceae), the tree Alseis blackiana Hemsl. (Rubiaceae), the treelet Swartzia simplex (Sw.) Spreng. (Caesalpinioideae), the treelet Oura- tea lucens (H.B.K.) Engler in Mart. (Ochnaceae), and the tree Licania platypus (Hemsl.) Fritsch (Chrysobalanaceae). Although leaf life spans vary continuously and do not fall into two clear cate- gories, we grouped species with similar leaf life spans into categories of short or long leaf life span for this study (Table 1). TABLE I. Z^af/(# jyam caAgpr^ 6ofYi/& {p#?, jpfo/k /aaf ana* (5/^/1), #%/ 6a^Wry pmzm.mwgz*/ ?%za MAoy&r ?% fAddb- during the early dry season. Leaf dry mass.turgid Leaf life span Leaf life span SLA mass ratio Species category (yr) (cm2/g) (mass/mass) Hybanthus prunifolius short 0.8" 308 ? 13 0.227 ? 0.004 Psychotria horizontalis short <2.Qb 231 ? 8 0.234 ? 0.008 Alseis blackiana short 0.8= 350 ? 16 0.251 ? 0.003 Swartzia simplex long 4.0^ 157 ? 13 0.388 ? 0.013 Ouratea lucens long 4.7'' 123 ? 5 0.412 ? 0.020 Licania platypus long ~3.0d 147 ? 8 0.359 ? 0.012 "Aide (1989). ? C. Sagers (pers. comm.), = Coley (1988). d Kursar and Coley (pers. comm.). 572 Tobin, Lopez, and Kursar LEAF WATER RELATIONS.?Predawn and midday Mi^ were determined for four of the six study species during the 1991-1992 study period. All of the plants sampled were 1- to 3-m tall and located in the understory of a level 2-ha area. Individuals with different leaf life spans were sometimes located within a meter of each other. Leaf age was stan- dardized within species by selecting mature leaves at the same distance from the branch tip and hav- ing similar epiphyll cover (Coley et al. 1993). Sam- pled leaves were estimated to have completed 40? 80 percent of their lifetime. A leaf from each of five replicate plants was sampled at predawn and midday on four measurement dates: 22 November 1991, and 24 February, 16 April, and 3 July 1992. The leaves were placed in a plastic bag immediately after excision. The ^P^ was determined in the field with a pressure chamber (PMS Inc., Corvallis, Oregon) and completed within 5 min of excision. Additionally, predawn and midday ^L were measured for five individuals of P. horizontalis and O. lucens on three measurement dates (3 February, 5 March, and 8 April 1993) through the 1992- 1993 dry season at two ridge sites. Aluminum foil was placed over the leaf and enclosed in a plastic bag the preceding night to prevent declines in ^ due to nocturnal stomatal conductance. PRESSURE?VOLUME CURVES.?The water relations of bulk leaf tissue were analyzed with the pressure- chamber technique (Tyree &: Jarvis 1982). Leaves of all study species were collected during four cen- sus periods: late wet season (November?Decem- ber), early dry season (February?March), late dry season (April), and early wet season (June-July) in the 1991-1992 study. During each census period, we attempted to complete pressure?volume (P-V) curves for five leaves from different individuals of each species. We applied the same criteria for choosing plants and leaves as those used in sam- pling for water potential measurements, except that individual plants were different for each census. Leaves were collected over a 3-h period (1 h apart) beginning at 0530 h. Immediately after ex- cising the leaf, the petiole was recut underwater and sealed inside a plastic bag with the petiole im- mersed in distilled water for a 3-h rehydration pe- riod. During the first census, the petiole was not recut underwater and the collection time of the leaves was not staggered; thus the time of rehydra- tion varied from 3 to 6 h. During the early-dry- season census (second census), two adjacent leaves were collected from each plant. The additional leaf was sealed in a bag immediately after excision and measurements were begun within 15 min. This procedure allowed a comparison of water relations for rehydrated versus non-rehydrated leaves. A P-V curve was generated by measuring leaf mass and balance pressure at periodic intervals; the leaf was allowed to dry on the laboratory bench between measurements (Hinckley et at 1980). The mass of the leaf was determined immediately before and after the balance-pressure measurement. The average of these two values was used to calculate relative water content [(fresh mass ?dry mass)/(sat- urated mass?dry mass)]. These measurements were continued until a relative water content of ca 0.5 had been reached. The number of points taken af- ter turgor loss was kept to fewer than seven to min- imize loss of leaf structural integrity due to repeated pressurization. After measurements had been com- pleted, the leaves were dried at 60?C for 2 to 3 d to determine dry mass. The saturated mass of each leaf was estimated by extrapolating a plot of fresh mass against balance pressure to zero after excluding plateau points (Ku- biske & Abrams 1990, 1991a, b). The estimated saturated mass of non-rehydrated samples was not considered reliable unless the highest balance pres- sure was >0.4 MPa. For each leaf, we plotted the reciprocal of the balance pressure against the rela- tive water deficit (1-relative water content). The osmotic potential at full turgor (^J00) was ob- tained from the y-intercept of a least squares linear regression fit to the linear portion of this plot. The osmotic potential at zero turgor {"^^) was derived from the y-coordinate of the turgor loss point. The leaf dry mass: turgid mass ratio was calculated as the dry mass of a leaf divided by its estimated sat- urated mass. During the early dry season, we col- lected an additional five leaves of each species and measured the area of each using a leaf area meter (LI-3000 and LI-3050A, LI-COR, Lincoln, Ne- braska), determined its dry mass, and used these to calculate specific leaf area (SLA; leaf area/leaf mass). STATISTICAL ANALYSIS.?The analysis of predawn and midday ^L measured at different sites during the 1992-1993 dry season was treated as a mixed model ANOVA using the JMP statistical package (Version 2.05, SAS Institute, Cary, North Caroli- na). Site, species, date, and all interactions were treated as fixed effects and repeatedly measured in- dividual plants were treated as random effects. Site, species, and their interaction were tested against individuals nested within site and species. Since measures were repeated on individual plants, we Responses of Understory Plants to Severe Drought 573 tested date effects and all interactions with date against the individual plant by date error term. The analysis of the effect of rehydration on M*/- was treated as a mixed model ANOVA. Leaf life span and rehydration treatment were considered fixed effects. Species was nested within leaf life span, and species and interactions with species were considered random effects. The effect of an indi- vidual plant was nested within species, and this plant effect and its interactions were considered random effects. Hypothesis tests were constructed accordingly to account for paired sampling of leaves from individual plants. The analysis of the water relations parameter "^J00 was treated as a mixed model nested AN- OVA. Date and leaf life span were treated as fixed effects, and species and interactions with species were treated as random effects (Sokal & Rohlf 1981). Hypothesis tests were as shown in Table 3. The fit model contains random effects and unequal cell sizes; thus the denominator in some hypothesis tests was synthesized by making a linear combina- tion of the mean squares of more than one effect (JMP, SAS Institute). The denominator used in the actual hypothesis test may therefore deviate slightly from the approximate one given in Table 3. The denominator degrees of freedom given in the text also deviate slightly from the actual values. In all analyses, we examined the residuals for normality and homoscedasticity and found no need for trans- formation. Nov Dec Jan Feb Mar Apr May Jun Jul Month FIGURE 2. Predawn (A) and midday (B) leaf water potential for O. lucens (open squares), S. simplex (open circles), H. prunifollus (closed circles), and P. horizontalis (closed squares) for dates in the late wet season, early dry season, late dry season, and early wet season (1991- 1992). Open and closed symbols represent species with long and short leaf life spans, respectively. The dashed lines represent leaf turnover. Points are means (?SE) of five observations. RESULTS The understory plant community experienced se- vere water stress during the 1991-1992 dry season. The dry season was not unusually long, but was one of the driest, and the following wet season was one of the wettest in the last 30 years (Paton 1993). The only substantial amount of rainfall during the dry season occurred in early January (Fig. 1). Soil moisture was partially recharged, but overall gravi- metric soil water content declined steadily through- out the dry season (Fig. 1). WATER POTENTIAL.?The predawn and midday "VL did not differ among species in the late wet season on 22 November (Fig. 2). The similarity among species in midday ^L was probably the result of constant rain during the ^L determination. In the early dry season on 24 February, and especially in the late dry season on 16 April, S. simplex and O. lucens had substantially higher predawn ^ than H. prunifollus and P. horizontalis. By the late dry season census, H. prunifolius and P. horizontalis had a predawn ^; of ?2.8 and ?3.6 MPa, respectively, below their turgor loss points. Both predawn and midday ^ increased in all species during the rainy season and did not differ among species on 3 July. The February predawn ^ of the two species not included in the seasonal measurements, A. blacki- ana and L. platypus, were ?1.8 and ?0.3 MPa, respectively, for five haphazardly chosen individu- als. The predawn M*L for P. horizontalis and O. lu- cens at two ridge sites during the relatively mild 1992-1993 dry season showed a pattern similar to the 1991-1992 data, but the magnitude of the changes was smaller (Fig. 3). The seasonal change in #i differed between the two species (^2,28 = 64.30, P < 0.0001), with the %, of P. horizontalis decreasing substantially as the dry season pro- gressed while the ^P^ of O. lucens remained rela- tively constant (Fig. 3). The effect of site also dif- fered for the two species (i*2,28 = 10.71, P < 0.001). O. lucens showed little difference in ^L be- 574 Tobin, Lopez, and Kursar a. TABLE 3. Analysis of variance for osmotic potential at full -0.5 -1.0 Feb Mar Month FIGURE 3. Predawn leaf water potential for O. lucens (open symbols), a species with a long leaf life span, and P. horizontalis (closed symbols), a species with a short leaf life span, for dates in the early, mid, and late dry season (1992?1993). The circle and square symbols represent different sampling sites. Points are means (?SE) of five observations. tween sites on any date, but P. horizontalis had sub- stantially different ty~L (-1.0 and -1.7 MPa) be- tween sites in the late dry season on 8 April (Fig. 3). OSMOTIC POTENTIAL.?Although plant material used for generating P-V curves is rehydrated rou- tinely to allow observations to begin near full tur- gor, some researchers have found that rehydration can alter parameters derived from P-V curves, thereby masking seasonal change (Kubiske & Abrams 1991a) or response to water deficits (Ea- rn us & Narayan 1990). We evaluated the effect of rehydrating the excised leaves for 3 h on ty^ during the first census period of the dry season. Limited pressure-volume data for non-rehydrated leaves did not allow comparisons of additional parameters. The comparison of ^P,/ for rehydrated and non- rehydrated leaves was significant across species (7V5 = 6.93, P < 0.05). Increases in ?*/? after Source of error Hypoth- esis test df MS A. Date A/E 3 1.11 26.85** B. Leaf lifetime B/C 1 034 0.54 C. Species C/E 3 0.70 17.01** D. Leaf life span X Date D/E 3 0.24 5.82* E. Species X Date E/F 9 0.04 1.41 F. Error 88 0.03 * P < 0.05; **P< 0.001. rehydration ranged from 0.02 to 0.51 MPa (Table 2). The effect of rehydration did not vary among species (F44 = 2.06, P > 0.25) or vary between the species with short and long leaf life spans (f, ..4 = 1.10, P > 0.35). Although rehydration increased "tyj- in leaves, there was no significant difference among species or between leaf life span categories. Such a rehydration-induced increase in ^P/ has been observed in other species (Saliendra & Mein- zer 1991), and may result from loss of symplastic solutes. Regardless of the origin, we conclude that the rehydration-induced artifacts for our study spe- cies were consistent in direction and magnitude and did not distort our analysis. We were unable to complete P-V curves for all species during the late-dry-season census period (April). P. horizontalis leaves partially lost structural integrity, so P-V curves were not possible. Also, the wet season rains began before all April P-V curves had been completed for H. prunifolius, A. blacki- ana, and L. platypus. The additional P-V curves were completed for all four species, but the 'Pir100 of P. horizontalis, excluded from the statistical analysis, decreased in the early dry season, as did those of H. prunifolius and A. black- iana (Fig. 4B). The seasonal changes in %/? were similar among S. simplex, O. lucens, and L. platypus, even though the tyj00 was substantially higher in L. platypus than the other two species. DISCUSSION The seasonal patterns of predawn ^Pr, during the 1991-1992 study period suggest that S. simplex and O. lucens, species with long leaf life spans, ex- perienced water stress later in the dry season than species with short leaf life spans, H. prunifolius and P. horizontalis (Fig. 2). Differences in predawn M^L among co-occurring species during drought have been related to rooting depth (Hinckley et al. 1981, 1983; Davis & Mooney 1986). Similarly, in a tropical dry forest, the decrease in predawn ^Pr of deeper rooted evergreens during the dry season lagged several months behind that of the more shal- lowly rooted deciduous species (Sobrado & Cuenca 1979, Sobrado 1986). On BCI, greater rooting depth among understory species has been related to maintenance of a higher water potential during the dry season (Rundel & Becker 1987, Wright 1992). P. horizontalis and H. prunifolius have shal- low rooting depths (0.3 and 0.6 m, respectively; Becker & Castillo 1990) and extract moisture from surface soil layers that can develop soil water po- tentials of ?2.3 MPa in the understory during even mild dry seasons on BCI (Becker et al. 1988). Our results suggest that S. simplex and O. lucens have a greater rooting depth, enabling them to maintain a high ^PL as the soil dries out. During the 1992-1993 dry season, O. lucens maintained a high predawn ^, while the predawn ^L of P. horizontalis declined steadily (Fig. 3). Al- though this dry season was less severe, differences in predawn ^ between P. horizontalis and O. lu- cens were apparent at the beginning of March and were substantial by April, late in the dry season. The two populations of O. lucens we compared did not show differences in predawn ^P^. In contrast, the two populations of P. horizontalis had different predawn \PL by the end of the dry season (Fig. 3). There may be considerable spatial heterogeneity in the water stress experienced by shallowly rooted species such as P. horizontalis. The seasonal changes in ^P^100 were correlated 576 Tobin, Lopez, and Kursar with the two categories of leaf life span. The M^100 of H. prunifolius, P. horizontal, and A. blackiana decreased in the early dry season concomitantly with decreases in predawn tp% (Figs. 2 and 4). The Tjr^ioo Qf ? simplex, O. lucens, and L. platypus de- creased in the late dry season at a point when the midday ^L indicated that at least one of these spe- cies, S. simplex, had experienced substantial midday water stress (Figs. 2 and 4). Substantial decreases in 1P?100 for O. lucens between the early and late dry season without major changes in predawn and midday ^ may be a response to drying roots in the upper soil layers (Davies & Zhang 1991). The range of changes we observed in tp^ and ,??100 and the differentiation of these between spe- cies with short and long leaf life spans may not be discernable during less severe dry seasons. For ex- ample, P. horizontalis had a predawn ^ of ?3.6 MPa in April 1992, but did not have a predawn ^L below ?1.0 MPa when measured in February and April 1989 (Wright ft d6 1992). Without irrigating plants, it is not possible to differentiate between osmotic adjustment in re- sponse to limited soil water availability, atmospher- ic changes, or one of internal origin {e.g., ontogeny or seasonal endogenous rhythms). Nevertheless, the observed dry season decreases in "V^100 are likely, in part, a response to water limitations. Irrigation on BCI of five species of understory shrubs in the genus Psychotria throughout a mild dry season did not diminish the seasonal decrease in ^ relative to control plants (Wright et al. 1992). In a study conducted during an earlier dry season, however, one of these species, Psychotria limonensis, did have lower ^ in nonirrigated plants relative to irrigated plants (Mulkey et al. 1991). Decrease in #? due to water limitation also can be inferred for P. horizon- talis. Individuals growing in sites with lower water availability, as indicated by lower predawn ^P^, had lower WJ00, suggesting an osmotic response to greater levels of water stress (Becker et al. 1988). Finally, effects of leaf aging alone cannot explain the late-dry-season decrease in ^p^100 of S. simplex, O. lucens, and L. platypus. In these species, 9%^o rose in the early wet season to pre-dry-season levels in leaves that had been retained through the dry season (Fig. 4). It appears that decreases in ^ in response to drought are superimposed onto internal or atmosphere-related changes, and are discernible in understory plants only during severe dry seasons. The 1991-1992 dry season was severe. In the early dry season, H. prunifolius and P. horizontalis both had substantially negative predawn "^L and lost turgor during midday. In the late dry season, they had a predawn ^L below their turgor loss point. Only one of the two species that experienced less water stress, S. simplex, lost turgor at midday late in the dry season. Thus, although all species returned to pre-dry-season water status when the wet season began in April, H. prunifolius and P. horizontalis dehydrated during this severe dry sea- son, whereas S. simplex and O. lucens postponed dehydration. Drought occurs on BCI with a frequency that suggests understory plants will experience multiple periods of severe water limitation during their life- times. Dry seasons with lower than normal rainfall on BCI have been related to El Nino Southern Oscillation (ENSO) events (Windsor 1990). ENSO is an aperiodic, unstable interaction be- tween the Pacific Ocean and the atmosphere that causes widespread climatic perturbations on an in- terannual 2- to 5-year cycle (Enfield 1992). ENSO has existed over at least the last 5000 years (Enfield 1992), suggesting that the fauna and flora of af- fected regions may be adapted to patterns of cli- mate produced by ENSO (Nicholls 1992). Water stress associated with severe drought may have ma- jor impacts on the local abundance and regional distribution of shrubs and trees in the tropical moist forest (Condit et al. 1995). The populations of H. prunifolius, P. horizontalis, A. blackiana, S. simplex, and O. lucens in a 50-ha upland forest site increased between 1982 and 1985 (Hubbell & Fos- ter 1990), a time period including the most severe drought recorded for BCI (Leigh et al. 1990). This suggests that these species have high survivorship through periods of drought, although increased light availability and reduced competition may play important roles in the survival of drought by un- derstory species (Aide &c Zimmerman 1990). Al- though the understory species studied responded differently to drought, they appear to have high survivorship during severe dry seasons. We hypothesized that shade-tolerant understo- ry species may differ considerably in physiology, even though these species are often lumped into a single group or guild. Our results demonstrate con- siderable differentiation in drought response among shade-tolerant species with divergent leaf life spans. Specifically, some species avoided water stress while other species tolerated water stress. Al- though our study species fell into only two groups, an analysis at the community level would likely demonstrate a continuum of responses to drought in shade-tolerant species. Substantial differences among shade-tolerant understory species are a gen- Responses of Understory Plants to Severe Drought 577 eral phenomenon, observed in the rates of photo- ACKNOWLEDGMENTS synthetic induction, epiphyll colonization of leaves, We thank David S. Ellsworth, S. Joseph Wright, Frederick survivorship in low light, maximum growth rate, Meinzer, and Phyllis Coley for valuable discussions and susceptibility to pathogens, and other traits (Aug- Gregory S. Gilbert for assistance with computer program- .??/ TT i i 11 o T- i nm ,vr- L ming. We thank Mark G. Tioelker, Jose-Luis Andrade, spureer 1984, Hubbell & Foster 1992, Wright et ,5 , ? <- ., r o ' ' b and several anonymous reviewers tor comments on the al. 1992, Coley et al. 1993, Kursar & Coley 1993, manuscript. We are grateful to the Smithsonian Tropical Mulkey et al. 1993, Kitajima 1994). Overall, these Research Institute for logistical support. The research was results suggest conside^lephWogWdi&rcn. j^%%^U%g%J%%%^S:3%^:%%:^3:^ nation among shade-tolerant understory plants in ship awarded by the Smithsonian Tropical Research In- the Neotropical moist forest. stitute (to Omar R. Lopez). LITERATURE CITED AIDE, T. M. 1989. The relationship between leaf phenology and herbivory in a tropical understory community. Ph.D. dissertation, University of Utah, Salt Lake City, Utah. , AND J. K. ZIMMERMAN. 1990. Patterns of insect herbivory, growth, and survivorship in juveniles of a Neo- tropical liana. Ecology 71: 1412-1421. AUGSPURGER, C. K. 1984. Light requirements of Neotropical tree seedlings: a comparative study of growth and survival. J. Ecol. 72: 777-795. BECKER, P., P. E. RABENOI.D, J. R. IDOL, AND A. P. SMITH. 1988. Water potential gradients for gaps and slopes in a Panamanian tropical moist forest's dry season. J. Trap. Ecol. 4: 173-184. , AND A. CASTILLO. 1990. Root architecture of shrubs and saplings in the understory of a tropical moist forest in lowland Panama. Biotropica 22: 242-249. CHABOT, B. E, AND D. J. HICKS. 1982. The ecology of leaf life spans. Annu. Rev. Ecol. Syst. 13: 229-259. CHAI>IN, F. S., III. 1980. The mineral nutrition of wild plants. Annu. Rev. Ecol. Syst. 11: 233-260. COLEY, P. D. 1981. Ecological and evolutionary responses of tropical trees to herbivory: a quantitative analysis of grazing damage, plant defenses, and growth rates. Ph.D. dissertation, University of Chicago, Chicago, Illinois. . 1988. Effects of plant growth rate and leaf lifetime on the amount and type of anti-herbivore defense. Oecologia (Berl) 74: 531-536. , T. A. KURSAR, AND J. L. MACHADO. 1993. Colonization of tropical rain forest leaves by epiphylls: effects of site and host plant leaf lifetime. Ecology 74: 619-623. CONDIT, R., S. P. HUBBELL, AND R. B. FOSTER. 1995. Mortality rates of 205 Neotropical tree and shrub species and the impact of a severe drought. Ecol. Monogr. 65: 419?439. CROAT, T. B. 1978. Flora of Barro Colorado Island. Stanford University Press, Stanford, California. DAVIPS, W. J., AND J. ZHANG. 1991. Root signals and the regulation of growth and development of plants in drying soil. Annu. Rev. Plant Physiol. Plant Mol. Biol. 42: 55-76. DAVIS, S. 0., AND H. A. MOONEY. 1986. Water use patterns of four co-occurring chaparral shrubs. Oecologia (Berl) 70: 172-177. EAMUS, D., AND A. NARAYAN. 1990. A pressure?volume analysis of Solatium melongena leaves. J. Exp. Bot. 41: 661? 668. ENPIELD, D. B. 1992. Historical and prehistorical overview of El Nino/Southern Oscillation. In H. F. Diaz and V. Markgraf (Eds.). El Nifio: historical and paleoclimatic aspects of the Southern Oscillation, pp. 95?117. Cam- bridge University Press, Cambridge, England. HINCKLEY, T. M., F. DUHME, A. R. HINCKLEY, AND H. RICHTER. 1980. Water relations of drought hardy shrubs: osmotic potential and stomatal reactivity. Plant Cell Environ. 3: 131?140. , , , and . 1983. Drought relations of shrub species: assessment of the mechanisms of drought resistance. Oecologia (Berl) 59: 344-350. , R. O. TESKEY, F. DUHME, AND H. RICHTER. 1981. Temperate hardwood forests. In T. T. Kozlowski (Ed.). Woody plant communities, pp. 153?208. Academic Press, New York, New York. HUBBELL, S. P., AND R. B. FOSTER. 1990. Structure, dynamics, and equilibrium status of old-growth forest on Barro Colorado Island. In A. H. Gentry (Ed.). Four Neotropical Rainforests, pp. 522-541. Yale University Press, New Haven, Connecticut. , AND . 1992. Short-term dynamics of a Neotropical forest: why ecological research matters to tropical conservation and management. Oikos 63: 48-61. KITAJIMA, K. 1994. Relative importance of photosynthetic traits and allocation patterns as correlates of seedling shade tolerance of 13 tropical trees. Oecologia (Berl.) 98: 419-428. KUBJSKE, M. E., AND M. D. ABRAMS. 1990. Pressure?volume relationships in non-rehydrated tissue at various water deficits. Plant Cell Environ. 13: 995-1000. , AND . 1991a. Rehydration effects on pressure?volume relationships in four temperate woody species: variability with site, time of season and drought conditions. Oecologia (Berl) 85: 537-542. 578 Tobin, Lopez, and Kursar , AND . 1991b. Seasonal, diurnal and rehydration-induced variation of pressure?volume relationships in Pseudotsuga menziesii. Physiol. Plant. 83: 107?116. KURSAR, T. A., AND P. D. COLEY. 1993. Phocosynthetic induction times in shade-tolerant species with long- and short- lived leaves. Oecologia (Berl) 93: 165-170. LEIGH, E. G., JR., D. M. WINDSOR, A. S. RAND, AND R. B. FOSTER. 1990. The impact of the "El Nino" drought of 1982?83 on a Panamanian semideciduous forest. In P. W. Glynn (Ed.). Global ecological consequences of the 1982-83 El Nino-Southern Oscillation, pp. 473-486. Elsevier, New York, New York. Mui.KEY, S. S., S. J. WRIGHT, AND A. P. SMITH. 1991. Drought acclimation of an understory shrub {Psychotria limonensis; Rubiaceae) in a seasonally dry tropical forest in Panama. Am. J. Bot. 78: 579?587. , , AND . 1993. Comparative physiology and demography of three Neotropical forest shrubs: alternative shade-adaptive character syndromes. Oecologia (Berl) 96: 526-536. NIO TOLLS, N. 1992. Historical El Nino/Southern Oscillation variability in the Australasian region. In H. F. Diaz and V. Markgraf (Eds.). El Nino: historical and paleoclimatic aspects of the Southern Oscillation, pp. 151?173- Cambridge University Press, Cambridge, England. PAION, S. 1993. How wet is it? Smithsonian Trop. Res. Ins. Newsl. 20: 3. REICH, P. B., M. B. WAITERS, AND D. S. ELLSWORTH. 1992. Leaf lifespan in relation to leaf, plant, and stand char- acteristics among diverse ecosystems. Ecol. Monogr. 62: 365-392. , C. UIII., M. B. WAITERS, AND D. S. ELLSWORTH. 1991. Leaf life-span as a determinant of leaf structure and function among 23 Amazonian tree species. Oecologia (Berl) 86: 16?24. ROBICIIAUX, R. H., P. W. RuNDEi, L. STEMMERMANN, J. E. CANFIELD, S. R. MORSE, AND W. E. FRIEDMAN. 1984. Tissue water deficits and plant growth in wet tropical environments. In E. Medina, H. A. Mooney, and C. Vasquez- Yanes (Eds.). Physiological ecology of plants in the wet tropics, pp. 99-112. Dr. W Junk Publishers, The Hague, The Netherlands. RUNDEL, P. W, AND P. F. BECKER. 1987. Cambios estacionales en las relaciones hidricas y en la fenologia vegetativa de plantas del estrato bajo del bosque tropical de la Isla de Barro Colorado, Panama. Rev. Biol. Trop. 35: 71- 84. SALIENDRA, N. Z., AND F. C. MEINZER. 1991. Symplast volume, turgor, stomatal conductance and growth in relation to osmotic and elastic adjustment in droughted sugarcane. J. Exp. Bot. 42: 1251-1259. SOBRADO, M. A. 1986. Aspects of tissue water relations and seasonal changes of leaf watet potential components of evergreen and deciduous species coexisting in tropical dry forests. Oecologia (Berl) 68: 413-416. , AND G. CUENCA. 1979. Aspectos del uso de agua de especies deciduas y siempreverdes en un bosque seco tropical de Venezuela. Acta Cient. Venez. 30: 302-308. SOKAL, R. R., AND F. J. ROHLE. 1981. Biometry. Freeman and Company, New York, New York. TYREE, M. T., AND P. G. JARVIS. 1982. Water in tissues and cells. In O. L. Lange, P. S. Nobel, C. B. Osmond, and H. Zeigler (Eds.). Physiological plant ecology. II. Encyclopedia in plant physiology, NS, vol. 12B, pp. 36?77. Springer-Verlag, Berlin, Germany. WINDSOR, D. M. 1990. Climate and moisture variability in a tropical forest: long-term records from Barro Colorado Island, Panama. Smithsonian Contrib. Earth Sci. 29: 1?145. WRIGHT, S. J. 1991. Seasonal drought and the phenology of understory shrubs in a tropical moist forest. Fxology 72: 1643-1657. . 1992. Seasonal drought, soil fertility and the species density of tropical forest plant communities. Trends Ecol. Evol. 7: 260-263. , J. L. MACHADO, S. S. MULKEY, AND A. P. SMITH. 1992. Drought acclimation among tropical forest shrubs {Psychotria, Rubiaceae). Oecologia (Berl) 89: 457-463.