Mortality Rates of 205 Neotropical Tree and Shrub Species and the Impact of a Severe Drought Author(s): Richard Condit, Stephen P. Hubbell and Robin B. Foster Reviewed work(s): Source: Ecological Monographs, Vol. 65, No. 4 (Nov., 1995), pp. 419-439 Published by: Ecological Society of America Stable URL: http://www.jstor.org/stable/2963497 . Accessed: 10/12/2012 15:52 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. . Ecological Society of America is collaborating with JSTOR to digitize, preserve and extend access to Ecological Monographs. http://www.jstor.org This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions Ecological Monographs, 65(4), 1995, pp. 419-439 ? 1995 by the Ecological Society of America MORTALITY RATES OF 205 NEOTROPICAL TREE AND SHRUB SPECIES AND THE IMPACT OF A SEVERE DROUGHT' RICHARD CONDIT Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002-0948, USA; or Apartado 2072, Balboa, Repliblica de Panamd STEPHEN P. HUBBELL Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002-0948, USA; or Apartado 2072, Balboa, Reptiblica de Panamd and Department of Ecology, Evolution, and Behavior, Princeton University, Princeton, New Jersey 08544 USA ROBIN B. FOSTER Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002-0948, USA; or Apartado 2072, Balboa, Republica de Panamd and Department of Botany, Field Museum of Natural History, Chicago, Illinois 60605 USA Abstract. Mortality rates of 205 tree and shrub species were estimated during two intervals, 1982-1985 and 1985-1990, in two size classes, 1-10 and ?10 cm in diameter, in a 50-ha census plot in tropical moist forest on Barro Colorado Island in Panama. The severe dry season of 1983 was the focus of the study, since prior observations had dem- onstrated that it caused mortality in the forest. Here we document that forest-wide mortality was z3%/yr during the drought interval but only 2%/yr during the period afterwards, and that excess mortality during the first interval amounted to 2% of stems in the larger size class and 1% in the smaller. Overall, just under 70% of all species had higher mortality during the first census interval, but not all species were equally affected. Canopy trees had significantly higher mean mortality rates during 1982-1985 than during 1985-1990, but treelets and shrubs showed no or slight differences. This was counter to our prediction that species with short root systems would suffer more from a long drought. Shrubs did, however, have higher mortality rates than trees and treelets during both census intervals. We also evaluated mortality rates for subgroups of species that specialized on different microhabitats in the forest. As we predicted, colonist species (those associated with light gaps) had higher mortality rates than generalist species, 7-10%/yr compared to 2-4%/yr, but only in the smaller size class. Unexpectedly, colonizers had similar mortality rates as non-colonizers in the larger size class. Gap colonizers and generalist species were similarly affected by the drought-both had elevated mortality during 1982-1985. Species whose distributions were associated with moister soils (on the slopes around the island's plateau or in a swamp in the midst of the 50-ha plot) also had elevated mortality during the drought period, but no more so than generalist species. This was counter to our prediction that species from moist microhabitats would suffer more during an extended drought han generalists. Understory treelets that were slope specialists had higher mortality than generalists during both census intervals, but not large trees that were slope specialists. Our conclusions emphasize diversity as well as pattern. Every trend we illustrated had well-documented exceptions: large trees with lower mortality during the drought period, for example. Clearly, accurate predictions about how tropical forests will respond to climatic perturbations will require much detailed information from many species. Key words: demography; drought; El Niho; forest; mortality; neotropics; trees; tropical. INTRODUCTION Turnover rates of tropical forests are often said to be high, with mortality rates >1% and sometimes >2%/yr (Lieberman et al. 1985, Manokaran and Ko- chummen 1987, Swaine et al. 1987a, b, Proctor et al. 1989, Phillips et al. 1994). These are forest-wide mor- tality rates, however, and tropical forests are exceed- ingly diverse communities; we should be emphasizing diversity of mortality patterns and diversity in response I Manuscript received 20 December 1994; accepted 12 Feb- ruary 1995; final version received 14 March 1995. to environmental perturbations, not uniform estimates for entire communities. Studies reporting demographic parameters of a wide variety of species from a single community are needed to properly evaluate life history paradigms of tropical trees (Hubbell and Foster 1986a, b, Swaine and Whitmore 1988, Whitmore 1989, Al- varez-Buylla and Martinez-Ramos 1992, Clark and Clark 1992, Zimmerman et al. 1994). Unfortunately, there have been few studies giving mortality rates for individual tree species in the tropics; indeed, rates based on reasonable sample sizes have been reported in no more than -50 cases (Lang and Knight 1983, Primack et al. 1985, Hay and Barreto 1988, Martinez- This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions 420 RICHARD CONDIT ET AL. Ecological Monographs Vol. 65, No. 4 Ramos et al. 1988, Primack and Lee 1991, Alvarez- Buylla and Martinez-Ramos 1992, Bullock 1992, Clark and Clark 1992, Milton et al. 1994, Zimmerman et al. 1994). The very reason that diversity has been ignored in mortality studies is diversity. Individual species are rare, and small plots have too few stems from any one species to provide a reliable mortality estimate. To overcome this problem, we began a study of 50 ha of forest on Barro Colorado Island (BCI) in Panama in 1980 (Hubbell and Foster 1983, Condit 1995). With three complete censuses of the plot finished over the last decade, we now have good estimates of mortality rate for >200 species-two-thirds of those in the plot- and here we present the first community-wide survey of species-specific mortality rates in tropical trees. Our data allow an evaluation of basic hypotheses about for- est mortality patterns by examining individual species: how many fit predictions and how many do not. Adding interest o this survey was an unusually se- vere dry season that struck Barro Colorado in 1983 (Leigh et al. 1990). In Central Panama, the dry season typically lasts from mid-December to mid-April, when ;215 mm of rain falls, but in 1982-1983, just 88 mm of rain fell during this period (Leigh et al. 1990, Wind- sor 1990). During 12 wk from late January to late April, 1983, just 3 mm of rain fell, and during March and April, temperatures were 20C higher than normal (Leigh et al. 1990). Unusual wilting was already evi- dent on Barro Colorado as early as March, and became more severe during the next several weeks (Leigh et al. 1990). This unusually severe dry season was as- sociated with the strong El Nifio event of 1982-1983. The El Nifio drought fell in the midst of the first census interval in the 50-ha plot, and mortality rates during this period thus include any impact caused by the drought. As of 1990, when the third census of the BCI plot was completed, we at last had the opportunity to compare mortality during the drought period with the period afterwards, evaluating changes in mortality for a large number of species. Did shrubs suffer more from the drought han trees, as the shorter oot systems of shrubs might lead one to believe (Wright 1992)? Did species associated with moist microhabitats uffer more from the drought than more generally distributed spe- cies? Did species restricted to the swamp suffer less because the swamp remained moist? Did gap-coloniz- ing species, which ought to have fairly broad tolerances for drought, suffer less? Answers to these questions as well as an overall description of mortality patterns are presented here. MATERIALS AND METHODS Study site The study was carried out in tropical moist forest on Barro Colorado Island (BCI) in central Panama. De- tailed descriptions of the climate, flora, and fauna of BCI can be found in Croat (1978) and Leigh et al. (1982). Censuses of 50 ha of forest were carried out in 1981-1983, 1985, and 1990 (Hubbell and Foster 1983, 1986a, b, c, 1990a, b, 1992, Condit et al. 1992a, b; we refer to the first census, which lasted two years, as the 1982 census). About 48 ha in the plot are in old- growth forest (>600 yr); the remaining two are in 90- yr-old forest. All free-standing, woody stems ?10 mm diameter at breast height (dbh) were identified, tagged, and mapped. The diameter of each stem was measured at breast height (1.3 m) unless there were irregularities in the trunk there, in which case the measurement was taken at the nearest lower point where the stem was cylindrical. Diameter at breast height of buttressed trees were taken above the buttresses. Species analyzed We included species in the analysis only if there were -20 live stems of the species at the beginning of a census interval for the particular period under consid- eration. This was an arbitrary cutoff, but we wanted to remove very small samples in which mortality rate could be greatly altered by one or two unusual events. There was no indication that our major results were affected by one cutoff as opposed to another: we re- peated analyses using only species with N 2 50 stems and found identical patterns to those presented here. A total of 205 species qualified with N ' 20 stems for both census intervals: 194 species for the 10-99 mm size class and 128 species for the ?100 mm size. Species names follow those given in D'Arcy (1987), except for eight cases where we use more recent names. Appendix 1 lists these eight species and their synonyms in D'Arcy (1987), plus an additional 20 species whose names have changed since our early publications (Hub- bell and Foster 1983, 1986a, 1990b, Welden et al. 1991). Appendix 1 allows any species listed in this study to be matched with a species mentioned in our earlier publications, in D'Arcy (1987), or in Croat (1978). Species characteristics We considered mortality rate as a function of three species characteristics-growth form, moisture pref- erence, and tendency to recruit into light gaps. Species were divided into four growth forms-large trees, mid- sized trees, treelets (or understory trees), and shrubs- based on the maximum size attained (the sizes are given in Hubbell and Foster 1986c). Moisture regime was defined using the slopes in the 50-ha plot, which have higher soil moisture content during the dry season than the flat regions (Becker et al. 1988), and the swamp, which is flooded throughout he wet season (Hubbell and Foster 1986c). The slopes are moist because of a perched water table below the plateau that drains around its edge onto the slopes. Many species have distributions clearly demarcated by the slopes and the swamp (Hubbell and Foster 1986c), and K. E. Harms This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions November 1995 MORTALITY IN 205 TROPICAL TREE SPECIES 421 (personal communication) calculated the density of each species in the different habitats. We used the ratio of density on the slopes to density on the low-lying flat areas as an index of "slope-specialization," and the ratio of density in the swamp to density on the low- lying flat areas as an index for "swamp-specializa- tion." We considered "slope specialists" and "swamp specialists" as those species with ratios ?1.5 (Condit et al., in press). Finally, we used the fraction of recruits found in light gaps, given in Welden et al. (1991), as a "colonizing index" for each species (Hubbell and Foster 1986b used a similar but not identical "index of heliophily"). Colonizers were defined as those spe- cies with an index ?30 (Condit et al., in press); col- onizers by our definition probably correspond with "pioneers" as defined by Swaine and Whitmore (1988), although they emphasized seed germination characteristics, which we do not consider here. Species missing from Welden's or Harms' calculations were omitted from the corresponding analyses here, that is, a species was considered a "non-colonizer" in our study only if it had a recruit index and the index was <30 (likewise for slope and swamp indices). Mortality rate Mortality was defined as death or disappearance. We recorded four different states of death: a standing stem, a fallen or broken stem, no stem at all but with the tree's tag located, and finally, neither stem nor tag. Many trunks were never found, as even large trees often died and completely rotted away during five years. Stems that snapped but resprouted were considered alive (Condit et al. 1993a). Mortality rate was calculated in two different census intervals, 1982-1985 and 1985-1990, and in two size classes, 10-99 mm dbh and ?100 mm dbh (based on the dbh at the beginning of each interval). Thus, every species had four mortality rates-in two size classes and two census intervals. Mortality rate m was cal- culated as = ln(NO) - ln(N,) (1) t where No is the number of initial stems, N, the number remaining alive at year t, and ln(N) is the natural log- arithm of N. This m is an approximation of the instan- taneous mortality rate, or the derivative of the popu- lation trajectory, but since t is small relative to the mortality rates, the approximation should be very good. Eq. 1 is the most commonly used formula for calcu- lating mortality in tropical forests (Swaine et al. 1987b, Clark and Clark 1992, Phillips et al. 1994), but an alternative formulation that yields an estimate equal to [1 - e-n] is sometimes used (Primack et al. 1985, Gil- bert et al. 1994). The two estimates are nearly identical when m is small. The time interval, t, used in Eq. 1 required close consideration because the census intervals for different 20 X 20 m subquadrats in the 50-ha plot were different. (The census interval was defined as the time elapsed between censuses for each 20 X 20 m subquadrat, which was accurate to ?2 wk, since individual sub- quadrats took <2 wk to complete). For all the sub- quadrats in the 50 ha, the first census interval (1982- 1985) varied from 1.9 to 4.5 yr and the second (1985- 1990) from 4.9 to 5.6 yr. For t, we used the arithmetic mean of the census intervals for individual stems of any one species. Using the arithmetic mean is not strict- ly accurate, and it yields only an approximation of the true instantaneous mortality; however, as we demon- strate in Appendix 2, the bias is slight. Given the actual variation in census intervals during 1982-1985, we cal- culated an upper bound of 0.5% for the bias given a true mortality rate <0.06/yr, and 5% for the highest mortality rate observed, or 0.50/yr. Thus, even with census intervals varying as widely as they did during 1982-1985, our estimates based on the arithmetic mean time interval are accurate. Statistical tests for individual species To assess statistical significance of differences in mortality rates for individual species, confidence limits for the mortality rate of each species in each size class and census interval were calculated using the normal approximation to the binomial variance, as long as there were more than five dead stems (D > 5). This is the recommendation given by Dixon and Massey (1969); Sokal and Rohlf (1973) give looser restrictions for use of the normal approximation. For D ' 5 and No ' 500, we calculated exact confidence limits using binomial probabilities. We created a table of 95% con- fidence limits for every pair of D and N (stopping at 500 because no species with No > 500 had D ' 5) by searching for a population mean D for which the bi- nomial probability of observing D or less would be <0.025; this was the upper 95% confidence limit (Dix- on and Massey 1969, Sokal and Rohlf 1973). The lower confidence limit was found analogously. Confidence limits were converted into annual mortality rates using Eq. 1. Statistical tests for groups of species Analyses of mortality patterns across groups of spe- cies were designed to determine whether mortality var- ied between census intervals, between size classes, and between growth forms; and as a function of colonizing, slope, or swamp status. Because this was an analysis of individual species, we did all tests on unweighted mortality rates of individual species; that is, the mor- tality rate of a species with 40 stems counted just as much as that of a species with 40,000 stems. A standard analysis of variance did not work on this dataset be- cause it was extremely unbalanced, with several empty cells. Therefore, we tested each of the various factors separately, as much as possible testing each in isolation from the others. This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions 422 RICHARD CONDIT ET AL. Ecological Monographs Vol. 65, No. 4 First, we tested for differences in mean mortality rate between census intervals, with all growth forms com- bined, then with the four growth forms separated, using the Wilcoxon paired-sample (signed rank) test. The two size classes were always tested separately. Second, we tested for differences between size classes using ex- actly the same approach and statistical test, always evaluating the two census intervals separately. Thirdly, we tested for differences in mortality rates among the four growth forms, using the Kruskal-Wallis test, sep- arately testing the two census intervals and two size classes. In each statistical test, a species was included only if N ' 20 in all intervals and size classes relevant to the test. This meant that different ests had slightly different numbers of species included. Just one shrub species, Sorocea affinis, had >20 stems in the large size class, and so shrubs of this size class were never included in statistical tests. We also considered the effect of swamp, slope, and colonization status on mortality. To do so, we used the Mann-Whitney U-test, comparing slope versus non- slope, swamp versus non-swamp, and colonizer versus non-colonizer species. The two size classes and census intervals were always considered separately. Initially, all growth forms were combined, then each test was repeated with the growth forms separated. Since there was no association between swamp status and either colonizer or slope status, the swamp category was al- ways tested by itself, ignoring the other two categories. But slope and colonizer status were associated: there were only four slope specialist/colonizer species in the plot (significantly fewer than expected by chance, based on a chi-square test with a 2 X 2 contingency table). Thus, we did all slope tests considering only non-colonizers, and all colonizer tests considering only non-slope species; in both cases, swamp status was ignored. The only assumption about underlying distributions that must be met for use of these non-parametric sta- tistics is that when multiple samples are compared, they have similar distributions (Siegel 1956, Ghent 1973). This assumption seems warranted here, since the sam- ples being compared were always different sets of mor- tality rates calculated in the same way. Results are illustrated as unweighted mean mortality rates of various species groups. In addition, we cal- culated forest-wide mortality rate, based on all stems combined, with species identity ignored. The latter was mainly for comparison with the many other studies which report just this statistic from a forest. Earlier publications.-A number of earlier papers have given some mortality information based on the same dataset used here (Hubbell and Foster 1990b, Leigh et al. 1990, Welden et al. 1991, Condit et al. 1992b, 1993b, 1994, Gilbert et al. 1994). Discrepancies between figures in earlier papers and those given here should be minor, and are due to on-going corrections to the dataset. RESULTS Distribution of mortality rates The modal mortality rate was 0.5-2%/yr in all four growth forms (Figs. 1 and 2). Nearly all species had rates below 6-8%/yr, with only a few >10% (Figs. 1 and 2). Mortality rates for all 205 species in each size class and census interval are given in Appendix 3, in- cluding estimates of 95% confidence limits. The highest mortality rate observed was 48%/yr, in Cecropia obtusifolia (Moraceae) in the small size class and the first census interval, during which 18 of 23 stems died. The lower confidence limit in this instance was 26%, which was the highest of all lower confidence limits calculated (Table 1); there were a number of other cases where lower confidence limits were >10% (Table 1). Species with the highest mortality rates were all colonizing species. The lowest mortality rate was zero, observed in a number of instances, the most ex- treme being Chamguava schippii (Myrtaceae), in which no stems died of 194 (small size class, first interval). The upper confidence limit in this case was 0.66%, but other upper limits were lower, reaching 0.35%/yr in Malmea sp. (Annonaceae) and Swartzia simplex var. grandifolia (Leguminosae). In all four growth forms, there were species with upper confidence limits <1 %/ yr (Table 1). Comparison of census intervals Mean mortality rates were higher during 1982-1985 than during 1985-1990 (Table 2). For all growth forms combined, the difference between means was statisti- cally significant in both size classes (P < 0.0001; Wil- coxon test). For each growth form separately, the dif- ference in mean mortality was much more pronounced in large and mid-sized tree species than in treelets; shrubs were intermediate (Table 2). Indeed, there was no statistically significant difference in treelets in either size class. About 70% of all species had higher mortality in the early census interval, but the percentage was lower in treelets and shrubs than in larger trees (Table 3). Nine- teen species had significantly higher mortality during the early interval, fifteen in the smaller size class, three in the larger size class, and one species, Poulsenia ar- mata (Moraceae), in both size classes (Table 3). In contrast, only two species had significantly higher mor- tality during the later census interval, both in the small- er size class (Table 3). Mortality rates for individual species were fairly consistent between census intervals. Regressions of mortality during the first interval against mortality dur- ing the second interval were highly significant with positive slopes (data not shown); r2 values for different growth forms and size classes were between 0.46 and 0.83. There were, however, exceptional species with very different mortality rates in the two intervals. In small stems of the mid-sized tree, Garcinia madruno This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions November 1995 MORTALITY IN 205 TROPICAL TREE SPECIES 423 20- Large trees (71 species) 0 15- (n E3 1982-1985 ()5i L a 0 !- U 1985-1 990 0 2 4 6 8 10 12 14 16 18 20 22 24 20- g 15- Mid-sized trees (54 species) a)- 10- 3 FIG. 1. Frequency distribution of mortal- oX ity rates for 10-99 mm dbh stems in four a) growth forms and two census intervals. An- C/) nual percentage mortality is plotted in 0.5% 0 l intervals: 0.0%-0.499%, 0.5%-0.999%, 1.0- 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 20 22 24 1.499%, etc. Only species with -20 stems in 20 the 10-99 mm size class in both 1982 and Treelets (41 species) 1985 are included. Two species of medium- 6 1 5- sized trees had mortality estimates >25% <,, (30.2% and 48.0%) in 1982-1985, based on ) 1 0- >20 stems, but both had <20 stems in 1985 and so were omitted from the plot. (C 5 - 0 2 4 6 8 10 12 14 16 18 20 22 24 1 5- Shrubs (28 species) 10-0 - ?11111 1tI 1 I l I ., ,,,,.. (1) 5 U kfl' 0 2 4 6 8 10 12 14 16 18 20 22 24 annual mortality for 10-99 mm dbh stems (%) (Guttiferae), mortality was 0.7%/yr over the first in- terval and 6.6% over the second (N > 600); 15 stems died over the first three years, but 193 over the next five (Appendix 3). A reverse example was the shrub, Anaxagorea panamensis, with mortality of 5.3% during the first interval then only 1.2% during the second (N > 400). Condit et al. (1992b) gave other examples of extreme changes in mortality. Comparison of size classes With all growth forms combined, mean mortality rate was significantly higher in the smaller size class than in the larger (P < 0.05; Wilcoxon test) during both census intervals (Table 2). This overall trend, however, masked a sharp difference between growth forms. In large trees, the difference was highly significant (P - 0.001) during both intervals, but in mid-sized trees and treelets, there were no differences. In fact, in mid-sized trees, smaller stems had lower mortality than larger in 1985-1990. About 70% of large tree species had higher mortality in the small size class, but in mid-sized trees and treelets, species were equally divided (Table 4). A total of 40 species had significant differences in mor- tality between size classes; 15 of these species were significantly different during both census intervals. The significant differences were equally divided between higher mortality at the small size versus higher mor- tality at the large size class (Table 4). Most with higher mortality at smaller size were large trees, while most with higher mortality at the larger size were mid-sized trees (Table 4). Two extremes illustrate the variation in how mor- tality changed with size. In 1982-1985, Pterocarpus rohrii (Leguminosae) had 2% mortality in the small size class but 8% in the large (with sample sizes >100); it also had higher mortality in the large size class in 1985-1990, but the difference was less extreme (Ap- pendix 3). Conversely, Ocotea whitei (Lauraceae) in 1982-1985 had 7% mortality at the small size class but 2% at the large size (sample size > 100); again, the pattern was repeated in 1985-1990 (Appendix 3). Comparison of growth forms The mean mortality of shrub species was about dou- ble that of trees or treelets in the smaller size class (Table 2); this difference was statistically significant in both census intervals (P < 0.001; Kruskal-Wallis test). After removing shrubs from the analysis, there was no This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions 424 RICHARD CONDIT ET AL. Ecological Monographs Vol. 65, No. 4 25 Large trees (63 species) 20- 0 Cl- 1 53 1982-1985 0 2 4 6 8 10 12 14 16 18 20 22 24 20- 1 5- ~ 15 Mid-sized trees (481species) 0 ~~~~~~~~~~~~~~~~~~FIG. 2. Frequency distribution of mortal- C/) .a) 10- ~~~~~~~~~~~~~ity rates for stems ?100 mm dbh in two dif- cJ ] ~~~~~~~~~~~~~~~ferent census int rvals. Annual percentage n . l g r ~~~~~~~~~~~~~mortality is plotted in 0.5% intervals: 0.0%- c) 5 0.499%, 0.5-%-0.999%, 1.0-1.499%, etc. 0 dbh in both 1982 and 1985 are included. 0 2 4 6 8 10 12 14 16 18 20 22 24 30- 25zTreelets (16 species) ,0~ Cl 1 0 0 2 4 6 8 10 12 14 16 18 20 22 24 annual mortality for stems ? 100 mm dbh (%i) significant variation among the remaining three groups. In the larger size class, there was no significant vari- ation during 1982-1985, but in 1985-1990 large trees had significantly lower mortality than mid-sized trees and treelets (P < 0.05, Table 2). Colonizers Overall mortality.-Colonizers had higher mean mortality rates than non-colonizers in all growth forms and both census intervals (Fig. 3), but only in the small size class (for the small size class, P < 0.01 in seven tests and P < 0.05 in the eighth; Mann-Whitney test). The mean mortality rate of colonizers in various growth forms was two to three times higher than that of non- colonizers. In the large size class, there were no sig- nificant differences between colonizers and non-colo- nizers. Colonizers and the inter-census comparison.-Col- TABLE 1. Maximum lower and minimum upper confidence limits (CL) on mortality rates in annualized percentages, for each growth form and size class. 1982-1985 1985-1990 Maximum lower CL Minimum upper CL Maximum lower CL Minimum upper CL Growth form Species* Rate Species Rate Species Rate Species Rate Small size class: 10-99 mm dbh stems Large trees Prioria 0.71 Cecropia i. 19.3 Brosimum 0.71 Cecropia i. 12.9 Mid-sized trees Heisteria c. 0.80 Cecropia o. 26.0 Malmea 0.35 Solanum 14.7 Treelets Swartzia s. g. 0.38 Croton 17.8 Swartzia s. g. 0.35 Croton 16.0 Shrubs Ouratea 1.05 Piper cu. 15.7 Capparis 0.96 Miconia n. 14.2 Large size class: ?100 mm dbh stems Large trees Alseis 0.63 Ocotea o. 3.9 Drypetes 0.74 Inga m. 8.7 Mid-sized trees Oenocarpus 0.38 Solanum 11.5 Oenocarpus 0.88 Solanum 13.1 Treelets Swartzia s. g. 0.97 Croton 16.0 Swartzia s. o. 1.71 Croton 7.6 * Species are designated by their genus, plus initials of the species name if necessary for locating in Appendix 3. This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions November 1995 MORTALITY IN 205 TROPICAL TREE SPECIES 425 TABLE 2. Mean ? 1 SE of mortality rate for each growth form, with the number of species given in parentheses. Means are unweighted (see Methods). For each entry, all species in the appropriate category with 220 stems at the start of the census were included in the mean; however, different statistical tests used different subsets of these totals (see Methods). 10-99 mm dbh 100 mm dbh Growth form 1982-1985 1985-1990 1982-1985 1985-1990 Mortality rate Large trees 4.6 + 0.5 (73) ** 3.2 + 0.4 (71) 3.2 + 0.4 (64) ** 1.9 + 0.2 (63) Mid-sized trees 4.5 + 1.0 (58) ** 2.7 + 0.4 (54) 4.0 + 0.6 (50) * 3.3 + 0.2 (49) Treelets 3.3 + 0.6 (41) 2.9 + 0.6 (41) 2.9 + 0.7 (16) 2.9 + 0.8 (16) ** ** Shrubs 7.3 + 1.1 (28) * 6.3 + 0.9 (28) 6.8 + 0.0 (1) 7.5 + 0.0 (1) Total 4.7 + 0.4 (200) ** 3.5 ? 0.3 (194) 3.4 + 0.3 (131) ** 2.6 + 0.2 (129) *,** Asterisks between the 1982-1985 and 1985-1990 data denote statistically significant differences between census intervals, with * for P < 0.05 and ** for P < 0.01 (Wilcoxon test). For the 10-99 mm dbh size class, the row of asterisks between shrubs and the other groups indicates a statistically significant difference (Kruskal-Wallis t test) among growth forms due to shrubs, and likewise for large trees ?100 mm dbh in 1985-1990. Statistically significant differences between size classes are not indicated, but are given in the text. onizers and non-colonizers had similar changes in mor- tality between census intervals, as can be seen by com- paring 1982-1985 rates with 1985-1990 (Fig. 3). With all growth forms combined, mortality was higher dur- ing 1982-1985 than during 1985-1990 for both colo- nizers and non-colonizers in both size classes (P < 0.01 in three of four cases and P < 0.05 in the fourth; Wilcoxon test). When separating growth forms, there were no significant differences among treelets nor shrubs, but all comparisons were significant for large tree species, for both colonizers and non-colonizers (P < 0.01 for the small size class, P < 0.05 for the large). In mid-sized trees, there was one significant difference in mortality between intervals: non-colonizers in the small size class had higher mortality during the early period (P < 0.01). Colonizers and the size comparison.-Colonizers behaved much differently from non-colonizers in terms of size differences in mortality. Colonizers had higher mean mortality rate in the small size class in both cen- sus intervals (P < 0.01; Wilcoxon test) when all growth forms were combined. With growth forms separated, only large colonizing trees had significantly higher mortality in the small size class compared to the large (P < 0.01, both census intervals; Fig. 3). In contrast, non-colonizers of all growth forms showed no signif- icant differences in mortality between size classes (Fig. 3). Colonizers and growth form.-In the small size class, shrubs had higher mean mortality than treelets and trees of both colonizer and non-colonizer species (P < 0.05 in both groups and both census intervals; Kruskal-Wallis test; Fig. 3). There were no significant effects of growth form on mortality for colonizers nor non-colonizers in the large size class. Slope specialists Overall mortality.-In the small size class, slope- specialists had higher mean mortality rate than gen- eralist species (Fig. 3; with all growth forms combined, P < 0.05 for 1985-1990 and P = 0.06 for 1982-1985; Mann-Whitney test). With growth forms separated, treelets in the small size class showed a significant difference (P < 0.01) in both census intervals (Fig. 3). There were no significant differences between slope and non-slope species in the larger size class, whether growth forms were combined or separated. Recall that all comparisons between slope and non-slope species excluded colonizers. Slope and the inter-census comparison.-Slope and TABLE 3. Differences in mortality rate between census intervals. Entries in the table are the number of species with higher mortality rate in either the early or late census interval; numbers in parentheses are the number of species with significant differences. In the large size class, four species of large trees had equal mortality (all 0.00) in the two intervals and are not included in the tallies. Size class 10-99 mm dbh 100 mm dbh Fraction higher in 1982-1985 Higher in Higher in Higher in Higher in 10-99 2100 Growth form 1982-1985 1985-1990 1982-1985 1985-1990 mm dbh mm dbh No. species No. species Large trees 54 (6) 17 (0) 42 (2) 17 (0) 0.76 0.71 Mid-sized trees 38 (4) 16 (1) 31 (2) 17 (0) 0.70 0.65 Treelets 25 (2) 16 (0) 10 (0) 6 (0) 0.61 0.63 Shrubs 18 (4) 10 (1) 0 (0) 1 (0) 0.64 0.00 Total 135 (16) 59 (2) 83 (4) 41 (0) 0.70 0.67 This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions 426 RICHARD CONDIT ET AL. Ecological Monographs Vol. 65, No. 4 TABLE 4. Differences in mortality rate between size classes. Entries in the table are the number of species with higher mortality rate in either the small size class or the large size class; numbers in parentheses are the number of species with significant differences. In 1985-1990, one large tree species had equal mortality (at 0.00%) in the two size classes and is not included in the tallies. Census interval 1982-1985 1985-1990 Fraction higher at small size Higher at Higher at Higher at Higher at Growth form 10-99 mm dbh 100 mm dbh 10-99 mm ?100 mm 1982-1985 1985-1990 No. species No. species Large trees 44 (9) 18 (5) 39 (8) 19 (3) 0.71 0.67 Mid-sized trees 18 (5) 29 (7) 23 (3) 19 (7) 0.38 0.55 Treelets 8 (1) 8 (2) 8 (1) 8 (2) 0.50 0.50 Shrubs 0 (0) 1 (1) 0 (0) 1 (1) 0.00 0.00 Total 70 (15) 56 (15) 70 (12) 47 (13) 0.56 0.60 non-slope species had similar changes in mortality be- tween census intervals (Fig. 3). With all growth forms combined, mortality was significantly higher in 1982- 1985 than in 1985-1990 for slope and non-slope spe- cies alike, in both size classes (P < 0.05; Wilcoxon test). Separating growth forms, the effect held only for large trees (P < 0.05 for slope and non-slope in the small size class but just for non-slope in the large size class). None of the other three growth forms showed significant differences between census intervals: nei- ther slope nor non-slope species and neither size class (Fig. 3). Slope and the size comparison.-Both slope and non-slope specialists showed no differences in mor- tality between size classes (Fig. 3). Recall again that the slope comparison excluded colonizers, and only large colonizing tree species had a significant mortality difference between size classes. Slope and growth form.-The higher mortality of shrubs in the smaller size class held in slope and non- slope species alike (Fig. 3; P < 0.05 in all comparisons in both census intervals; Kruskal-Wallis test). The ef- fect of growth form on mortality in the larger size class (lower mortality among large trees, 1985-1990 only) did not hold in slope nor non-slope species. This did not appear to be just a sample size problem, but was instead due to the fact that the slope comparison was made only among non-colonizers, and it was colonizer ..I ** .Total .......... C) .. shrubs. ) I*( ..I.**.. treelets ........ l I) ** mid-sized trees.. _O (0 _ slope . large trees. * generalist El colonizer ..** . Total ........ LO ....... shrubs ...... (O C) 1** ...C CM I ** ........ treelets. I . c\j (0 CX) 00) C) _..mid-sized trees.. ...... large trees.... 0 2 4 6 8 10 12 14 0 2 4 6 mortality of stems mortality of stems of 10-99 mm dbh (%) of 2 100 mm dbh (%) FIG. 3. Mean mortality rates for slope specialists and colonizing species compared to generalist species, of four growth forms and separated into two census intervals (above and below) and two size classes (left and right). The total category is the combination of the four growth forms that follow. Asterisks at the end of each bar indicate a statistically significant difference between the marked column and the generalist column (*, P < 0.05; **, P < 0.01; Mann-Whitney U-test). All the main results of this study-mortality comparisons between species groups-are summarized in this figure (except for swamp specialists). This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions November 1995 MORTALITY IN 205 TROPICAL TREE SPECIES 427 TABLE 5. Unweighted mean mortality rates of swamp specialists. Entries in parentheses are the number of species on which the mean was based. 1982-1985 1985-1990 Swamp Non-swamp Swamp Non-swamp Small size class: 10-99 mm dbh Mortality rate Large trees 4.3 (14) 4.6 (59) 3.0 (13) 3.2 (58) Mid-sized trees 8.0 (11) 3.7 (47) 3.2 (9) 2.6 (45) Treelets 5.8 (9) 2.6 (32) 4.9 (9) 2.4 (32) ** ** Shrubs 10.3 (8) 6.4 (20) 7.6 (8) 5.8 (28) Total 6.7 (42) * 4.2 (158) 4.4 (39) 3.2 (155) Large size class: '100 mm dbh Mortality rate Large trees 1.6 (13) ** 3.6 (51) 0.9 (13) ** 2.2 (50) Mid-sized trees 3.2 (12) * 4.2 (38) 2.6 (12) 3.5 (37) Treelets 12.4 (1) 2.3 (15) 11.4 (1) 2.4 (15) Total 2.7 (26) ** 3.6 (104) 2.1 (26) * 2.7 (102) *,** Asterisks between two columns denote a statistically significant difference between the mortality rates given in the two columns, * for P < 0.05 and ** for P < 0.01 (Mann-Whitney test). Among non-swamp species, the row of asterisks between shrubs and the other groups indicates statistically significant differences among growth forms due to shrubs only, and likewise for large trees in 1985-1990. species only that showed the contrast in mortality be- tween large trees and mid-sized trees or treelets at the large size. Swamp specialists Overall mortality.-In the larger size class, swamp specialists had lower mortality rates than non-swamp (Table 5B). The difference was statistically significant (Mann-Whitney test) when all growth forms were com- bined, and for large tree species (both census intervals) and mid-sized tree species (1982-1985 only) separately (Table 5B). The trend did not hold in the smaller size class; here, swamp specialists had higher mortality than non-swamp, but the difference was significant only in 1982-1985 and only when all growth forms were com- bined (Table 5A). Swamp and the inter-census comparison.-Swamp and non-swamp species had similar changes in mor- tality between census intervals (Table 5). The differ- ence was significant in the small size class, with P < 0.05 in swamp species and P < 0.01 in non-swamp; in the larger size class, P = 0.09 among swamp species and P < 0.01 for non-swamp (Mann-Whitney test). The difference remained statistically significant in some cases when growth forms were separated-mid-sized and large trees, both size classes (P < 0.05)-but only for non-swamp species. The lack of significant results among swamp specialists when growth forms were iso- lated was probably due to small sample size, since even among swamp species, every growth form had higher mean mortality during the early census interval (Table 5). Swamp and the size comparison.-The difference in mean mortality between size classes, seen only in large tree species (Table 2), was maintained even when swamp and non-swamp species were separated (P < 0.05 for both groups in both census periods; Wilcoxon test). This can be seen by comparing mortality rates in Table 5A with those in Table 5B, for swamp and non- swamp species separately. The other growth forms did not show significant differences. Swamp and growth form.-Shrubs had higher mor- tality rates than trees among non-swamp species (Table 5; P < 0.01 for both intervals; Kruskal-Wallis test), but in swamp species, significance was not achieved. The lack of significance in the latter case appeared to be due to low sample size, since swamp species showed the same trend as non-swamp, and P < 0.10 in both census periods. The effect of growth form in the larger size class (lower mortality among large trees, 1985- 1990 only) was upheld at P < 0.05 only among swamp specialists; in non-swamp species, patterns were sim- ilar but 0.05 < P < 0.10 (Table 5). Forest-wide mortality With all stems in the forest combined, annual mor- tality was 2.66% in 1982-1985 and 2.26% in 1985- 1990 in the small size class, 2.75% and 1.98% in the large size class. In the small size class, there were 214,530 total stems with 18,142 dead over 1982-1985; there were 221,284 total stems with 24,864 dead over 1985-1990. In the large size class, there were 20,891 total stems with 1843 dead over 1982-1985, and 20,727 total stems with 2069 dead over 1985-1990. The dif- ferences between census periods were highly signifi- cant for both sizes, whereas the differences between size classes were significant in 1985-1990 but not in 1982-1985. These rates are lower than the averages shown in Table 2, which are unweighted means of in- dividual species' mortality rates. This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions 428 RICHARD CONDIT ET AL. Ecological Monographs Vol. 65, No. 4 0.05- 0.04 1982-1985 0.03 0.01- 0.000- 0 10 20 30 40 50 60 70 80 90 100 110 size class (dbh in cm) FIG. 4. Forest-wide annual mortality rates by size class. Points are plotted above the mid-point of each size class: 10- 19, 20-49, 50-99, 100-199, 200-299, 300-399, etc., and fi- nally, all stems ?1000 mm dbh. Mortality was higher during 1982-1985 in every size class (Fig. 4). The differences were significant in all size classes <500 mm dbh, in the 600-699 mm class, and in the -1000 mm class. The greatest differences in mortality between census intervals were in stems -200 mm dbh (Fig. 4) for which mortality was 3.37% over 1982-1985 and 1.99% over 1985-1990. DIscusSION How long do tropical trees live? Mortality rates give the answer. Prioria copaifera had annual mortality rates of ?0.6% in both saplings and large trees, and at this rate, a cohort of 1000 trees would last > 1100 years (until the fraction alive fell to <0.001). Swartzia sim- plex var. ochnacea had mean mortality rates <0.35%/ yr in 1982-1985, so a cohort of this species would last almost 2000 years. If large trees senesce, these would be over-estimates of life span, but our detailed study of mortality in Prioria (Condit et al. 1993b) did not show evidence for senescence: even larger trees had mortality no more than -1%/yr. Thus, it seems rea- sonable to conclude that some trees of the BCI forest are >1000 yr old. On the other hand, a substantial number of species had annual mortality rates >2%/yr, including some abundant canopy dominants. Trichilia tuberculata (Meliaceae), the most abundant tree in the plot (Hub- bell and Foster 1983, 1987), had mortality rates >2.25%/yr. At this rate, a cohort of 1000 trees would last only 300 years. More extreme, there were several colonizing species with annual mortality >10% (Cro- ton billbergianus) or even 20% (Solanum hayesii), rates which would eliminate 1000 stems in 35-70 years. The lowest mortality estimate, even using the upper end of 95% confidence limits, was 0.35%/yr in Swart- zia. The highest value for the lower end of 95% limits was 26%/yr in Cecropia. This is almost a 100-fold range in mortality rates, quite a comment on the di- versity of mortality rates in a tropical forest. Our estimate of forest-wide mortality over 1982- 1985 was 2.75%/yr in the large size class, which is high but not unprecedented for tropical forests (Swaine et al. 1987b)-Phillips et al. (1994) gave a figure of 2.8% for two plots in Peru. Our high figure represents the drought period, though, and the 1985-1990 rate of 1.98%/yr represents our best estimate of normal tree mortality at BCI. This is close to the annual rates of 2.2% (Lang and Knight 1983) and 2.0% (Milton et al. 1994) derived from smaller plots at BCI, and 2% for the La Selva forest in Costa Rica (Lieberman et al. 1985). In a few species, we can compare mortality rates from the 50-ha plot with reports elsewhere. Lang and Knight (1983) and Milton et al. (1994) gave mortality data for 30 species on BCI, nearly all of which appear in Appendix 3 here; most species showed close matches in annualized mortality rates. Clark and Clark (1992) reported mortality estimates for Dipteryx panamensis and Hyeronima alchornoides in Costa Rica, and Al- varez-Buylla and Martinez-Ramos (1992) for Cecropia obtusifolia in Mexico; their figures corresponded close- - ly to what we found at BCI. Mortality rates in temperate forests appear to be sim- ilar. In old growth forest in Indiana, annual forest-wide mortality was 1.23% over 10 years (Abrell and Jackson 1977). For individual species, some hardwoods such as beech (Fagus grandifolia) and sugar maple (Acer saccharam) have very low mortality, often well below 1 %/yr, but various species of Betula, Quercus, and Os- trya have rates of 3-10%/yr (Leak 1970, Monserud 1976, Harcombe and Marks 1983). Conifers tend to have fairly high mortality: 1.5-6%/yr (Yarranton and Yarranton 1975, Knowles and Grant 1983, Johnson and Fryer 1989). A consistent problem comparing studies of mortality in temperate and tropical forests is that temperate studies are often based on reconstruction of a population's history based on aging with tree rings, whereas all tropical studies have been based on per- manent plots. Nevertheless, our overall conclusion would be that there are no obvious tropical-temperate differences in tree mortality. The drought of 1983 led to general increases in mor- tality among BCI trees. If mortality in 10-99 mm stems had been the same in 1982-1985 as it was in 1985- 1990, then 15,867 trees would have died over the 3.4- yr census interval instead of the 18,142 that did die. Thus, an additional 2275 trees died as a result of the drought, or 1.1% of the 10-99 mm stems. For stems ?100 mm dbh, the drought killed an additional 483 trees, or 2.3% of the total, and for stems -200 mm dbh, an additional 338 stems, or 4.3% of the total, died. These calculations assume that mortality during 1985- 1990 was not elevated by the earlier drought, so they are minimum estimates of excess mortality. The greater impact of the drought on larger stems was borne out by our observations of individual species: treelet and shrub species were less affected than larger trees. These patterns ran counter to our prediction that larger trees This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions November 1995 MORTALITY IN 205 TROPICAL TREE SPECIES 429 with deeper roots would be buffered against long dry seasons compared to shrubs and small trees (Wright 1992). The dry period afflicted colonizers, non-colonizers, slope, non-slope, swamp, and non-swamp species equally. We predicted that slope specialists would be affected more than non-slope specialists, but we could not demonstrate this. Nor was our prediction that col- onizing species would be less affected than non-col- onizers borne out. The drought had consistent effects on large and mid-sized tree species, whatever their soil or light gap preferences. A few species ran counter to the overall trend and had much lower mortality during the drought han after. In one species, we know why: peccaries chewed the base of a large number of Garcinia madruno trees in 1989 and presumably were responsible for the greatly elevated mortality rate observed during 1985-1990. In addition, there were many species whose mortality rates were similar in the two periods, such as Prioria copaifera. Many species were apparently quite tolerant of the long drought. Droughts have long been recognized as important disturbances in temperate forests, causing slight or sub- stantial increases in tree mortality (Hursh and Haasis 1931, Yarranton and Yarranton 1975). In the tropics, the recognition of the importance of drought in forest dynamics is more recent, arising following the 1982- 1983 El Nifio, which affected forests on BCI, and in Borneo where large tracts of moist forest burned fol- lowing a long dry period. Although the fires were partly the result of logging damage (Woods 1989), regular drought probably does play a role in structuring east Bornean forests (P. Ashton, personal communication). Although the drought at BCI affected species of dif- ferent microhabitats equally, there were consistent dif- ferences in baseline mortality among groups. Coloniz- ing species of all growth forms had much higher mor- tality rates than non-colonizers, both during and after the drought. Although this is what everyone would ex- pect (Swaine and Whitmore 1988), there were unan- ticipated results. The difference only showed up in the small size class, not the large: this is illustrated clearly by Cecropia insignis, with 15.8% mortality among sap- lings during 1985-1990 but only 3.2% mortality in the larger size class, and by Zanthoxylum belizense, which had 14.4% and 3.7% mortality in the small and large size classes, respectively. A thorough analysis of mor- tality in Cecropia obtusifolia in Mexico revealed the same pattern (Alvarez-Buylla and Martinez-Ramos 1992). Also contrary to the predicted pattern, some colonizers had low mortality even as saplings: Gustavia superba (Lecythidaceae) is an abundant roadside tree and a colonizer by Welden's index, yet its mortality in 1985-1990 was 1.9%/yr at the small size and 0.3%/yr at the large; Macrocnemum glabrescens (Rubiaceae) had a high colonizing index, but had mortality 1%/ yr or less in the small size class; and Jacaranda copaia had the highest colonizing index in the plot and high growth rates (Condit et al. 1993a), yet had mortality <3.5%/yr in both size classes during 1985-1990. Discussions of tropical tree life history have focused on the dichotomy between colonizers and non-colo- nizers (or pioneers and non-pioneers in some termi- nology): colonizers are species with small seeds that require high light levels to germinate, have high growth rates and mortality rates, and tend to recruit in light gaps. Shade-tolerant species possess the opposite suite of characters. In the current analysis, we evaluated just two of these features: the tendency to recruit in light gaps and mortality rate. Although we did find the ex- pected correlation, there were many exceptions, and theories must account for the correlations as well as the exceptions. We must be wary that the dichotomy of life history traits is really a continuum (Whitmore 1989, Alvarez-Buylla and Martinez-Ramos 1992, Zim- merman et al. 1994), as illustrated by the range of mortality rates we found. The other large difference among species groups was that shrubs had higher mortality than treelets and trees; this was consistent during and after the drought. The few non-colonizing species with very high mortality rates were shrubs, such as Psychotria deflexa (Rubi- aceae), Piper aequale (Piperaceae), and Conostegia cinnamomea (Melastomataceae), which had mortality rates >10%/yr during both census periods. Mean mor- tality for shrubs was >6%/yr, and the four colonist shrub species had even higher mortality. Shrubs, how- ever, had mortality rates less affected by the drought than larger trees, and treelets were like shrubs in this regard. The excess mortality of slope specialists was less pronounced than that of colonizers or shrubs, but slope species did have higher mortality than generalists dur- ing both census intervals, although the difference was due solely to treelets and perhaps shrubs, not larger trees. Our prediction was that slope specialists would suffer more during the drought than generalists, be- cause they are moisture-demanding species. Instead, we observed differences consistent across censuses: slope specialists of small stature had higher mortality rates during the drought and afterwards as well. Similar effects showed up in population trends: nearly all shrubs and treelets that are slope specialists in the plot declined in abundance over 1982-1985 and 1985-1990 (Condit et al., in press). Our working hypothesis for these observations is that small-stature species (shrubs and treelets) that require moister soils (such as on the slopes) are uniformly suf- fering unusually high mortality rates and population declines on BCI because of the post-1966 drying trend, during which total rainfall has been 14% lower and severe dry seasons twice as frequent as before 1966 (Windsor 1990, Windsor et al. 1990, Condit et al., in press). Some larger trees of the moist microhabitats are also suffering high mortality and population declines- This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions 430 RICHARD CONDIT ET AL. Ecological Monographs Vol. 65, No. 4 Poulsenia armata and Ocotea whitei are examples- but other tree species affiliated with slopes, such as Guatteria dumetorum and Calophyllum longifolium, are doing just fine. We hypothesize that the division among large trees is caused by differences in seedling biology. Trees are associated with the moist slopes be- cause their seedlings are drought-sensitive (Howe 1990), but some are also drought-sensitive as adults (Poulsenia and Ocotea) while others have longer roots and are drought tolerant as adults (Guatteria and Cal- ophyllum). The latter group had normal mortality rates among the stems we measured, but the former had high mortality and appear to be dying back as a result of drought. On the other hand, we hypothesize that shrubs and treelets that are drought-intolerant as seedlings are also intolerant as adults, because their roots are short throughout development. (Among all groups, there are species with drought-resistance mechanisms not based on deep roots [see Mulkey et al. 1994], and these spe- cies are not restricted to the slopes.) These hypotheses lead to clear predictions about developmental changes in drought sensitivity that can be tested by evaluating physiological condition during drought (Wright and van Schaik 1994). The most important aspect of our hypothesis from the perspective of forest dynamics is that the effects of climate change have not been caused solely by the 1983 El Nifio, but are due to a longer term pattern of drought. This hypothesis leads to a clear prediction about future censuses of the 50-ha plot: the drought- sensitive species should continue to suffer very high mortality and population declines as long as the drier weather continues. If wetter conditions return, their populations should stabilize as mortality rates decline. Regardless, we see climatic shifts driving continuous shifts in demography and composition in the BCI for- est, and we suggest that they probably always have and always will (Condit et al. 1992b, in press). Ongoing change is probably typical of tropical forests (Bush and Colinvaux 1990, Bush et al. 1990, Hart et al., in press). ACKNOWLEDGMENTS The Smithsonian Tropical Research Institute in Panama provided generous logistical and financial support for the cen- suses. We also thank the field workers who contributed tothe censuses on BCI, more than 100 people from 10 countries, R. P6rez and S. Loo de Lao for their persistent work main- taining the plot and its database, and P. Coley, T. Kursar, J. Wright, and E. Leigh for many useful discussions, and Ira Rubinoff for his long-term support. This project has been supported by grants from the National Science Foundation, the Smithsonian Scholarly Studies Program, the Smithsonian Tropical Research Institute, the John D. and Catherine T. Mac- Arthur Foundation, the World Wildlife Fund, the Earthwatch Center for Field Studies, the Geraldine R. Dodge Foundation, and the Alton Jones Foundation. This publication is a sci- entific contribution from the Center for Tropical Forest Sci- ence, which is supported by the John D. and Catherine T. MacArthur Foundation. LITERATURE CITED Abrell, D. B., and M. T. Jackson. 1977. A decade of change in an old-growth beech-maple forest in Indiana. American Midland Naturalist 98:22-32. Alvarez-Buylla, E. R., and M. Martinez-Ramos. 1992. De- mography and allometry of Cecropia obtusifolia, a neo- tropical pioneer tree-an evaluation of the climax-pioneer paradigm for tropical rain forests. Journal of Ecology 80: 275-290. Becker, P., P. E. Rabenold, J. R. Idol, and A. P. Smith. 1988. Water potential gradients for gaps and slopes in a Pana- manian tropical moist forest's dry season. Journal of Trop- ical Ecology 4:173-184. Bullock, S. H. 1992. Effects of sex, size and substrate on growth and mortality of trees in tropical wet forest. Oec- ologia 91:52-55. Bush, M. B., and P. A. Colinvaux. 1990. A pollen record of a complete glacial cycle from lowland Panama. Journal of Vegetation Science 1:105-118. Bush, M. B., P. A. Colinvaux, M. C. Wiemann, D. R. Piperno, and K. Liu. 1990. Late pleistocene temperature depression and vegetation change in Ecuadorian Amazonia. Quater- nary Research 34:330-345. Christensen, N. L. 1977. Changes in structure, pattern and diversity associated with climax forest maturation in Pied- mont, North Carolina. American Midland Naturalist 97: 176-188. Clark, D. A., and D. B. Clark. 1992. Life history diversity of canopy and emergent trees in a neotropical rain forest. Ecological Monographs 62:315-344. Condit, R. 1995. Research in large, long-term tropical forest plots. Trends in Ecology and Evolution 10:18-22. Condit, R., S. P. Hubbell, and R. B. Foster. 1992a. Recruit- ment near conspecific adults and the maintenance of tree and shrub diversity in a neotropical forest. American Nat- uralist 140:261-286. Condit, R., S. P. Hubbell, and R. B. Foster. 1992b. Stability and change of a neotropical moist forest over a decade. Bioscience 42:822-828. Condit, R., S. P. Hubbell, and R. B. Foster. 1993a. Identifying fast-growing native trees from the neotropics using data from a large, permanent census plot. Forest Ecology and Management 62:123-143. Condit, R., S. P. Hubbell, and R. B. Foster. 1993b. Mortality and growth of a commercial hardwood, "El Cativo," Prior- ia copaifera, in Panama. Forest Ecology and Management 62:107-122. Condit, R., S. P. Hubbell, and R. B. Foster. 1994. Density dependence in two understory tree species in a neotropical forest. Ecology 75:671-705. Condit, R., S. P. Hubbell, and R. B. Foster. In press. Changes in a tropical forest with a shifting climate: results from a 50 ha permanent census plot in Panama. Journal of Tropical Ecology. Croat, T. R. 1978. Flora of Barro Colorado Island. Stanford University Press, Stanford, California, USA. D'Arcy, W. G. 1987. Flora of Panama. Part I: Introduction and checklist. Missouri Botanical Garden, Saint Louis, Missouri, USA. Dixon, W. J., and E J. Massey. 1969. Introduction to statis- tical analysis. Third edition. McGraw-Hill, New York, New York, USA. Ghent, A. W. 1973. Theory and application of some non- parametric statistics I. Exact solutions (extended by number triangles) to the Wilcoxon two-sample and paired-sample tests. Biologist 55:149-177. Gilbert, G. S., S. P. Hubbell, and R. B. Foster. 1994. Density and distance-to-adult effects of a canker disease in a moist tropical forest. Oecologia 98:100-108. Harcombe, P. A., and P. L. Marks. 1983. Five years of tree death in a Fagus-Magnolia forest, southeast Texas (USA). Oecologia 57:49-54. Hart, T. B., J. A. Hart, R. Dechamps, M. Fournier, and M. Ataholo. In press. Changes in forest composition over the This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions November 1995 MORTALITY IN 205 TROPICAL TREE SPECIES 431 last 4000 years in the Ituri Basin, Zaire. Proceedings of the Fourteenth AETFAT (Association pour l'Etude Taxono- mique de la Flore d'Afrique Tropicale) Congress, Wag- eningen Agricultural University, Wageningen, The Neth- erlands. Hay, J. D., and E. J. M. Barreto. 1988. Natural mortality of Vochysia thyrsoidea in an unburnt cerrado ecosystem near Brasilia. Biotropica 20:274-279. Howe, H. E 1990. Survival and growth of juvenile Virola surinamensis in Panama: effects of herbivory and canopy closure. Journal of Tropical Ecology 6:259-280. Hubbell, S. P., and R. B. Foster. 1983. Diversity of canopy trees in a neotropical forest and implications for conser- vation. Pages 25-41 in S. L. Sutton, T. C. Whitmore, and A. C. Chadwick, editors. Tropical rain forest: ecology and management. Blackwell Scientific, Oxford, England. Hubbell, S. P., and R. B. Foster. 1986a. Biology, chance, and the history and structure of tropical rain forest tree communities. Pages 314-329 in J. Diamond and T. J. Case, editors. Community ecology. Harper and Row, New York, New York, USA. Hubbell, S. P., and R. B. Foster. 1986b. Canopy gaps and the dynamics of a neotropical forest. Pages 77-96 in M. J. Crawley, editor. Plant ecology. Blackwell Scientific, Ox- ford, England. Hubbell, S. P., and R. B. Foster. 1986c. Commonness and rarity in a neotropical forest: implications for tropical tree conservation. Pages 205-231 in M. Soul6, editor. Conser- vation biology: the science of scarcity and diversity. Sin- auer Associates, Sunderland, Massachusetts. Hubbell, S. P., and R. B. Foster. 1987. The spatial context of regeneration in a neotropical forest. Pages 395-412 in M. Crawley, P. J. Edwards, and A. Gray, editors. Coloni- zation, succession, and stability. Blackwell Scientific, Ox- ford, England. Hubbell, S. P., and R. B. Foster. 1990a. Structure, dynamics, and equilibrium status of old-growth forest on Barro Col- orado Island. Pages 522-541 in A. Gentry, editor. Four neotropical rain forests. Yale University Press, New Haven, Connecticut, USA. Hubbell, S. P., and R. B. Foster. 1990b. The fate of juvenile trees in a neotropical forest: implications for the natural maintenance of tropical tree diversity. Pages 317-341 in M. Hadley and K. S. Bawa, editors. Reproductive ecology of tropical forest plants. Parthenon Publishing, New Jersey, USA. Hubbell, S. P., and R. B. Foster. 1992. Short-term population dynamics of a neotropical forest: why ecological research matters to tropical conservation and management. Oikos 63:48-61. Hursh, C. R., and F. W. Haasis. 1931. Effects of 1925 summer drought on southern Appalachian hardwoods. Ecology 12: 380-386. Johnson, E. A., and G. I. Fryer. 1989. Population dynamics in lodgepole pine-Engelmann spruce forests. Ecology 70: 1335-1345. Knowles, P., and M. C. Grant. 1983. Age and size structure analyses of Engelmann spruce, ponderosa pine, lodgepole pine, and limber pine in Colorado. Ecology 64:1-9. Lang, G. E., and D. H. Knight. 1983. Tree growth, mortality, recruitment, and canopy gap formation during a 10-year period in a tropical moist forest. Ecology 64:1075-1080. Leak, W. B. 1970. Successional change in northern hard- woods predicted by birth and death simulation. Ecology 51:794-801. Leigh, E. G., Jr., S. A. Rand, and D. M. Windsor, editors. 1982. The ecology of a tropical forest: seasonal rhythms and long-term changes. Smithsonian Institution Press, Washington, D.C., USA. Leigh, E. G., Jr., D. M. Windsor, S. A. Rand, and R. B. Foster. 1990. The impact of the "El Nifio" drought of 1982-1983 on a Panamanian semideciduous forest. Pages 473-486 in P. W. Glynn, editor. Global ecological consequences of the 1982-1983 El Nifio-southern oscillation. Elsevier, Am- sterdam, The Netherlands. Lieberman, D., M. Lieberman, R. Peralta, and G. Hartshorn. 1985. Mortality patterns and stand turnover rates in a wet tropical forest in Costa Rica. Journal of Ecology 73:915- 924. Lorimer, C. G. 1980. Age structure and disturbance history of a southern Appalachian virgin forest. Ecology 6:1169- 1184. Manokaran, N., and K. M. Kochummen. 1987. Recruitment, growth and mortality of tree species in a lowland dipter- ocarp forest in Peninsular Malaysia. Journal of Tropical Ecology 3:315-330. Martfnez-Ramos, M., J. Sarukhdn, and D. Pifiero. 1988. The demography of tropical trees in the context of fore$t gap dynamics: the case of Astrocaryum mexicanum at Los Tux- tlas tropical rain forest. Pages 293-313 in A. J. Davy, M. J. Hutchings, and A. R. Watkinson, editors. Plant popula- tion ecology. Blackwell Scientific, Oxford, England. Milton, K., E. A. Laca, and M. W. Demment. 1994. Suc- cessional patterns of mortality and growth of large trees in a Panamanian lowland forest. Journal of Ecology 82:79- 87. Monserud, R. A. 1976. Simulation of forest tree mortality. Forest Science 22:438-444. Mulkey, S. S., A. P. Smith, S. J. Wright, J. L. Machado, and R. Dudley. 1994. Contrasting leaf phenotypes control sea- sonal variation in water loss in a tropical forest shrub. Pro- ceedings of the National Academy of Sciences 89:9084- 9088. Phillips, 0. L., P. Hall, A. H. Gentry, S. A. Sawyer, and R. Vasquez. 1994. Dynamics and species richness of tropical rain forests. Proceedings of the National Academy of Sci- ence 91:2805-2809. Primack, R. B., P. S. Ashton, P. Chai, and H. S. Lee. 1985. Growth rates and population structure of Moraceae trees in Sarawak, East Malaysia. Ecology 66:577-588. Primack, R. B., and H. S. Lee. 1991. Population dynamics of pioneer (Macaranga) trees and understorey (Mallotus) trees (Euphorbiaceae) in primary and selectively logged Bornean rain forests. Journal of Tropical Ecology 7:439- 458. Proctor, J., C. Phillips, G. K. Duff, A. Heaney, and F. M. Robertson. 1989. Ecological studies on Gunung Silam, a small ultrabasic mountain in Sabah, Malaysia. II. Some forest processes. Journal of Ecology 77:317-331. Putz, E E., and K. Milton. 1990. Tasas de mortalidad de los drboles en la isla de Barro Colorado. Pages 157-162 in E. G. Leigh, A. S. Rand, and D. M. Windsor, editors. Ecologia de un bosque tropical: ciclos estacionales y cambios a largo plazo. Smithsonian Tropical Research Institute, Balboa, Republica de Panama. Siegel, S. 1956. Non-parametric statistics for the behavioral sciences. McGraw Hill, New York, New York, USA. Sokal, R. R., and F. J. Rohlf. 1973. Introduction to biosta- tistics. W. H. Freeman, San Francisco, California, USA. Swaine, M. D., J. B. Hall, and I. J. Alexander. 1987a. Tree population dynamics at Kade, Ghana. Journal of Tropical Ecology 3:331-345. Swaine, M. D., D. Lieberman, and F E. Putz. 1987b. The dynamics of tree populations in tropical forest: a review. Journal of Tropical Ecology 3:359-366. Swaine, M. D., and T. C. Whitmore. 1988. On the definition of ecological species groups in tropical rain forests. Ve- getatio 75:81-86. Welden, C. W., S. W. Hewett, S. P. Hubbell, and R. B. Foster. 1991. Survival, growth, and recruitment of saplings in can- This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions 432 RICHARD CONDIT ET AL. Ecological Monographs Vol. 65, No. 4 opy gaps and forest understory on Barro Colorado Island, Panama. Ecology 72:35-50. Whitmore, T. C. 1989. Canopy gaps and the two major groups of forest trees. Ecology 70:536-538. Windsor, D. M. 1990. Climate and moisture variability in a tropical forest: long-term records from Barro Colorado Is- land, Panama. Smithsonian Contribution to the Earth Sci- ences, Number 29. Smithsonian Institution Press, Wash- ington, D.C., USA. Windsor, D. M., A. S. Rand, and W. M. Rand. 1990. Cara- teristicas de la precipitaci6n en la isla de Barro Colorado. Pages 53-71 in E. G. Leigh, A. S. Rand, and D. M. Windsor, editors. Ecologia de un bosque tropical: ciclos estacionales y cambios a largo plazo. Smithsonian Tropical Research Institute, Balboa, Repiiblica de Panama. Woods, P. 1989. Effects of logging, drought, and fire on structure and composition of tropical forests in Sabah, Ma- laysia. Biotropica 21:290-298. Wright, S. J. 1992. Seasonal drought, soil fertility and the species diversity of tropical forest plant communities. Trends in Ecology and Evolution 7:260-263. Wright, S. J., and C. P. van Schaik. 1994. Light and the phenology of tropical trees. American Naturalist 143:192- 199. Yarranton, M., and G. A. Yarranton. 1975. Demography of a jack pine stand. Canadian Journal of Botany 53: 310-314. Zimmerman, J. K., E. M. Everham, III, R. B. Waide, D. J. Lodge, C. M. Taylor, and N. V. L. Brokaw. 1994. Re- sponses of tree species to hurricane winds in subtropical wet forest in Puerto Rico: implications for tropical tree life histories. Journal of Ecology 82:911-922. APPENDIX 1 The 28 species in this study that have been renamed since the 50-ha plot was initiated in 1981, or that do not appear in Croat (1978). The current name is the one appearing in Appendix 3 of this paper. Following it are synonyms from D'Arcy (1987) and from Croat (1978). The first publications from the 50-ha plot project used Croat's names, except for those species found in the plot but not listed in Croat (1978). Current name Name in D'Arcy Name in Croat Chamaedorea tepejilote Chamaedorea tepejilote Chamaedorea wenlandiana Chamguava schippii Psidium anglohondurensis Psidium anglohondurensis Chrysochlamys eclipes Tovomitopsis nicaraguensis Tovomitopsis nicaraguensis Chrysophyllum argenteum Cynodendron panamense Cynodendron panamense Garcinia intermedia Garcinia intermedia Rheedia edulis Garcinia madruno Garcinia madruno Rheedia acuminata Guarea sp. nov. none none Guarea grandifolia Guarea grandifolia Guarea multiflora Heisteria acuminata Heisteria acuminata Heisteria longipes Hyeronima alcheornoides Hyeronima laxiflora Hyeronima laxiflora Inga sp. nov. none none Lonchocarpus latifolia Lonchocarpus latifolia Lonchocarpus pentaphyllus Malmea sp. nov. none Crematosperma sp. Nectandra purpurea Nectandra purpurea Nectandra purpurescens Ocotea puberula Ocotea puberula Ocotea pyramidata Ocotea whitei Ocotea whitei Ocotea skutchii Oenocarpus mapoura Oenocarpus mapoura Oenocarpus panamanus Ormosia croatii Ormosia coccinea Ormosia coccinea Phoebe cinnamomifolia Phoebe cinnamomifolia Phoebe mexicana Pourouma bicolor Pourouma guianensis Pourouma guianensis Pouteria reticulata Pouteria unilocularis Pouteria unilocularis Sapium aucuparium Sapium caudatum both (now considered synonyms) Senna dariensis Senna dariensis Cassia fruticosa Socratea exorrhiza Socratea exorrhiza Socratea durissima Terminalia oblonga Terminalia oblonga Terminalia chiriquensis Trichilia pallida Trichilia pallida Trichilia montana Trichilia tuberculata Trichilia tuberculata Trichilia cipo Virola sp. nov. none none This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions November 1995 MORTALITY IN 205 TROPICAL TREE SPECIES 433 APPENDIX 2 The purpose here is to determine what m, the estimated mortality rate, would be if a population suffering idealized instantaneous mortality at a rate M were censused over a range of time intervals. Let No be the initial population, with all individuals first censused at time 0. Each plant is re-censused again at time t, which varies for different plants. Let F, be the fraction of the original population re-censused at time t. For example, consider 1000 plants censused on day 0, with 100 of the plants recensused after 0.5 years, 200 more after 1 year, 300 after 2 years, 200 after 3 years, and 200 after 4 years. Then F05 = 0.1 (100 out of 1000), F1 = 0.2 (200 out of 1000), F2 = 0.3, F3 = 0.2, and F4 = 0.2. Under idealized mortality, the fraction of plants still alive at time t is e-M', so the number of plants recorded alive at time t would be N = NOFteM. (Al) In the example above, with M = 0.02/yr, at the 6-mo census there would be 99 of 100 plants still alive, at the 1-yr census there would be 196 of 200 alive, after 2 years, 288 of 300, after 3 years, 188 of 200, and after 4 years, 185 of 200. The total number of stems found alive throughout he survey, N, is the sum of N, over all t, or N = | Ndt = | NOFte Mtdt. (A2)- In the example, the total alive from all censuses would be 956 out of 1000. Since the mean time interval for the 1000 stems would be 2.25 years, our crude estimate of mortality (based on Eq. 1) would be m = 0.019999, only infinitesimally different from the true value of M = 0.02. We consider more generally the case where the function F, is constant between time t1 and t2, that is, where equal numbers of stems are censused at each time over the interval. We chose this case because it is mathematically simple, and because it is worse than the actual situation in our study, in which F, was somewhat bell-shaped. Thus, F, = k over the specified interval, with k = [1/(t2- t1)] and F, = 0 outside the interval. Substituting this for F, in Eq. A2 and integrating from t1 to t2 yields: e-Mt2 e-Mt, M(t2 -t (t) Given M, Eq. A3 allowed us to calculate N and thus m (using Eq. 1). We calculated N and m for a set of values of M, using t1 = 4.6 and t2 = 1.9 (actual bounds for the 1982- 1985 census interval) and found that the estimated m < M in all cases, but the discrepancy was below 0.05-M for all M < 0.5 and below 0.01 M for all M < 10%. The vast majority of species in the plot had mortality rates <6%/yr, where the bias was only 0.005-M. Since we took a worst-case scenario (the uniform distribution for Ft), the actual bias would be even less. This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions 434 RICHARD CONDIT ET AL. Ecological Monographs Vol. 65, No. 4 APPENDIX 3 Mortality rates of 205 tree and shrub species in the BCI 50-ha plot that had ?20 individuals in one size class for one census interval. A code for the growth form of each species is given immediately after its name (T = large tree, M = medium- sized tree, U = treelet, and S = shrub), followed by a dash, then zero to three codes indicating slope specialists (S), swamp specialists (W), and colonizers (C). For each species, the first row of data is for stems 10-99 mm in dbh and the second row for stems ?100 mm dbh; if there is no second row, then N < 20 in both censuses for the larger size class. 1982-1985 1985-1990 Mor- 95% Mor- 95% tality confidence tality confidence Genus and species N* D* rate* limits N D rate limits Acalypha diversifolia S-WC 1566 449 9.65 10.57 8.77 1201 417 8.07 8.87 7.30 Acalypha macrostachya U-C 78 19 9.89 14.65 5.69 65 33 13.64 19.10 9.40 Adelia triloba U-C 230 33 4.38 5.92 2.93 199 52 5.71 7.34 4.22 115 6 1.49 2.71 0.32 114 11 1.91 3.07 0.81 Aegiphila panamensis M- 111 22 7.32 10.54 4.39 103 32 7.08 9.72 4.76 23 3 4.14 12.10 0.83 21 2 1.89 6.85 0.22 Alchornea costaricensis T-C 224 73 11.94 14.83 9.30 154 55 8.40 10.79 6.29 160 10 1.89 3.08 0.74 158 19 2.39 3.50 1.34 Alibertia edulis U-W 303 11 1.13 1.80 0.47 340 19 1.10 1.59 0.61 Allophyllus psilospermus M- 145 26 5.81 8.13 3.65 139 23 3.42 4.88 2.07 30 2 2.00 7.24 0.24 32 7 4.67 8.50 1.49 Alseis blackiana T- 6748 307 1.42 1.58 1.26 7194 453 1.23 1.35 1.12 847 26 0.91 1.27 0.56 857 20 0.44 0.63 0.25 Amaioua corymbosa U- 29 1 0.88 4.91 0.02 29 0 0.00 2.41 0.00 Anacardium excelsum T-SW 5 1 5.54 31.30 0.13 3 1 7.63 44.45 0.15 23 0 0.00 4.47 0.00 23 2 1.67 6.03 0.19 Anaxagorea panamensis S-SC 472 59 5.33 6.72 3.99 472 28 1.18 1.62 0.75 Andira inermis T- 276 9 0.94 1.56 0.33 270 7 0.50 0.87 0.13 42 6 4.48 8.31 1.10 36 6 3.44 6.42 0.87 Annona acuminata S-W 509 42 2.65 3.46 1.86 525 47 1.78 2.30 1.28 Annona spraguei M-C 39 13 12.84 20.78 6.49 53 14 5.83 9.17 2.99 Apeiba membranacea T- 151 22 5.14 7.37 3.06 116 12 2.08 3.30 0.94 238 23 3.06 4.34 1.83 226 14 1.19 1.83 0.58 Apeiba tibourbou M- 20 4 6.93 17.80 1.84 16 4 5.49 14.14 1.44 26 3 3.88 11.36 0.80 23 9 9.36 16.85 4.02 Ardisia fendleri U-S 76 2 0.72 2.60 0.08 79 3 0.73 2.14 0.15 Aspidosperma cruenta T-S 403 14 0.97 1.48 0.47 418 15 0.69 1.04 0.34 48 1 0.57 3.14 0.01 53 3 1.09 3.20 0.22 Astrocaryum standleyanum M-W 15 7 17.28 34.91 6.63 7 1 2.91 16.33 0.07 233 12 1.55 2.44 0.69 225 20 1.76 2.54 1.00 Astronium graveolens T- 30 5 5.49 12.84 1.75 24 1 0.81 4.52 0.02 35 0 0.00 3.32 0.00 35 1 0.54 3.02 0.01 Beilschmiedia pendula T- 2068 136 2.17 2.54 1.81 2366 171 1.43 1.64 1.22 308 20 1.95 2.81 1.11 303 19 1.21 1.77 0.68 Brosimum alicastrum T- 682 13 0.57 0.88 0.26 717 18 0.48 0.71 0.26 183 11 1.85 2.96 0.78 179 9 0.96 1.60 0.34 Calophyllum longifolium T-S 594 48 2.53 3.26 1.83 668 74 2.24 2.75 1.73 55 5 2.88 6.74 0.93 54 8 3.05 5.29 1.04 Capparis frondosa S- 3536 109 0.89 1.06 0.73 3669 157 0.83 0.96 0.70 Casearia aculeata U- 443 41 2.82 3.70 1.97 449 44 1.95 2.54 1.38 24 3 3.95 11.56 0.81 26 1 0.74 4.13 0.02 Casearia arborea T-C 120 29 9.09 12.59 5.93 98 31 7.24 9.99 4.83 151 27 6.13 8.53 3.89 128 21 3.38 4.88 1.98 Casearia sylvestris M- 176 27 5.36 7.45 3.40 164 26 3.29 4.60 2.06 72 10 4.31 7.12 1.75 67 9 2.73 4.60 1.02 Cassipourea elliptica M-W 698 17 0.80 1.18 0.42 778 31 0.77 1.05 0.50 67 3 1.43 4.17 0.29 72 10 2.84 4.69 1.16 Cavanillesia platanifolia T-S 1 0 0.00 100.00 0.00 0 0 21 1 1.30 7.22 0.03 21 0 0.00 3.22 0.00 Cecropia insignis T-C 237 127 23.17 27.61 19.30 194 109 15.77 19.09 12.94 280 43 5.28 6.90 3.74 249 40 3.22 4.24 2.25 Cecropia obtusifolia M-WC 23 18 48.01 81.78 26.04 13 7 14.66 31.44 5.90 38 15 16.12 25.65 8.78 24 9 8.83 15.80 3.76 Ceiba pentandra T- 30 9 11.58 20.25 4.75 27 8 6.73 12.11 2.54 42 3 2.27 6.65 0.46 40 6 2.90 5.39 0.72 Celtis schippii M-S 140 12 2.58 4.08 1.16 128 16 2.53 3.82 1.33 42 5 3.75 8.75 1.19 38 9 5.13 8.83 2.03 Cestrum megalophyllum S-SW 309 95 10.69 12.94 8.61 236 105 11.18 13.48 9.12 Chamaedorea tepejilote S-S 32 14 17.40 28.42 9.33 22 11 13.21 23.52 6.55 Chamguava schippii U- 194 0 0.00 0.66 0.00 239 13 1.07 1.66 0.50 Chrysochiamys eclipes 5-5 458 34 2.23 2.99 1.49 432 32 1.46 1.97 0.96 This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions November 1995 MORTALITY IN 205 TROPICAL TREE SPECIES 435 APPENDIX 3. Continued. 1982-1985 1985-1990 Mor- 95% Mor- 95% tality confidence tality confidence Genus and species N* D* rate* limits N D rate limits Chrysophyllum argenteum T-C 348 12 1.07 1.67 0.47 393 12 0.59 0.93 0.26 75 1 0.41 2.29 0.01 84 4 0.92 2.37 0.25 Chrysophyllum cainito T-WC 49 2 1.27 4.58 0.15 57 3 1.02 3.00 0.21 21 0 0.00 5.10 0.00 23 0 0.00 2.90 0.00 Coccoloba coronata M- 159 13 2.60 4.04 1.22 169 9 1.04 1.73 0.37 22 0 0.00 5.38 0.00 21 1 0.92 5.10 0.02 Coccoloba manzanillensis U- 437 17 1.24 1.84 0.66 439 18 0.80 1.17 0.43 Conostegia cinnamomea S-SW 391 138 12.27 14.41 10.28 280 116 10.15 12.11 8.36 Cordia alliodora T-C 50 11 7.99 13.11 3.57 48 7 3.02 5.41 0.90 61 6 3.16 5.79 0.73 60 7 2.35 4.18 0.68 Cordia bicolorM-C 460 52 4.11 5.25 3.01 500 112 4.87 5.79 3.98 255 18 2.38 3.50 1.30 258 16 1.22 1.83 0.63 Cordia lasiocalyx M-S 1278 96 2.29 2.75 1.84 1222 122 2.00 2.36 1.65 420 34 2.39 3.21 1.60 442 47 2.13 2.75 1.53 Coussarea curvigemmia U- 1462 38 0.86 1.13 0.59 1616 63 0.76 0.95 0.57 40 1 0.82 4.53 0.01 46 6 2.66 4.92 0.64 Croton billbergianus U-WC 553 255 20.20 22.82 17.78 552 336 17.88 19.97 15.99 67 22 12.35 18.03 7.55 67 30 11.37 16.01 7.63 Cupania latifolia T-S 46 6 4.14 7.66 1.00 39 4 2.06 5.27 0.55 Cupania rufescens T-W 50 1 0.61 3.38 0.01 65 2 0.60 2.16 0.07 Cupania sylvatica U- 934 20 0.66 0.95 0.37 1008 16 0.30 0.45 0.16 28 1 1.11 6.16 0.02 30 0 0.00 2.36 0.00 Dendropanax stenodontus T- 63 11 6.87 11.19 3.02 51 9 3.73 6.34 1.43 96 6 2.25 4.10 0.49 93 12 2.65 4.21 1.20 Desmopsis panamensis U- 11 718 1126 3.04 3.22 2.87 12119 1639 2.76 2.89 2.62 Diospyros artanthifolia M- 39 3 2.61 7.65 0.52 42 0 0.00 1.70 0.00 Dipteryx panamnensis T- 23 4 5.79 14.88 1.53 19 2 2.12 7.66 0.24 33 0 0.00 3.40 0.00 33 0 0.00 2.00 0.00 Drypetes standleyi T-S 1977 51 0.70 0.89 0.51 2037 99 0.94 1.13 0.76 196 1 0.13 0.74 0.01 227 7 0.59 1.03 0.16 Elaeis oleifera M-W 0 0 0 0 22 1 1.62 9.04 0.04 21 0 0.00 3.31 0.00 Erythrina costaricensis U-S 242 49 7.00 9.03 5.10 190 45 5.14 6.71 3.69 47 5 3.31 7.71 1.06 46 12 5.73 9.31 2.73 Erythroxylum ultifiorum M- 309 35 3.58 4.79 2.42 288 45 3.23 4.20 2.31 18 1 1.66 9.23 0.04 20 5 5.41 12.71 1.70 Erythroxylum panamense U-W 104 7 2.19 3.86 0.61 105 10 1.91 3.13 0.76 Eugenia coloradensis T- 725 33 1.39 1.87 0.92 761 84 2.22 2.70 1.75 78 9 3.47 5.83 1.29 79 14 3.69 5.73 1.85 Eugenia galalonensis U- 940 29 0.94 1.28 0.60 1131 53 0.91 1.16 0.67 21 1 1.48 8.25 0.03 25 9 8.48 15.10 3.58 Eugenia nesiotica M- 461 14 0.91 1.39 0.44 481 9 0.36 0.59 0.13 48 3 1.92 5.62 0.40 48 1 0.40 2.19 0.01 Eugenia oerstedeana M- 1955 138 2.23 2.61 1.86 2068 198 1.91 2.18 1.65 133 15 3.44 5.24 1.75 140 25 3.71 5.23 2.31 Faramea occidentalis U- 22232 743 1.05 1.12 0.97 23742 1091 0.89 0.95 0.84 1228 88 2.22 2.68 1.76 1402 146 2.08 2.42 1.75 Ficus tonduzii M-S 24 1 1.19 6.66 0.03 24 8 7.70 14.01 2.97 42 5 3.85 9.00 1.22 37 8 4.61 8.12 1.65 Garcinia intermedia M- 3577 91 0.74 0.89 0.59 3948 164 0.80 0.93 0.68 77 12 4.88 7.79 2.24 75 10 2.72 4.49 1.10 Garcinia madruno M- 629 15 0.73 1.11 0.36 655 193 6.61 7.57 5.70 23 2 2.97 10.73 0.34 26 7 5.94 10.97 1.97 Genipa americana T-W 71 5 2.11 4.94 0.68 68 4 1.15 2.95 0.31 20 0 0.00 5.19 0.00 20 0 0.00 3.52 0.00 Guapira standleyanum T- 140 21 4.94 7.14 2.90 117 15 2.61 3.98 1.33 90 3 1.00 2.93 0.21 91 3 0.62 1.80 0.13 Guarea sp. nov. M- 1453 155 3.34 3.88 2.82 1397 200 2.93 3.35 2.53 105 13 3.78 5.92 1.80 95 19 4.23 6.24 2.41 Guarea grandifolia T- 45 1 0.75 4.17 0.01 47 2 0.83 3.01 0.10 Guarea guidonia M-SW 1411 86 1.77 2.15 1.40 1465 99 1.32 1.59 1.07 370 23 1.76 2.49 1.05 363 16 0.85 1.27 0.44 Guatteria dumetorum T-S 1302 122 3.05 3.59 2.51 1248 130 2.09 2.46 1.74 285 31 3.46 4.70 2.26 280 34 2.43 3.26 1.63 Guazuma ulmifolia T-W 27 4 4.52 11.59 1.21 19 7 8.65 16.57 3.09 28 0 0.00 3.83 0.00 30 1 0.62 3.48 0.01 This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions 436 RICHARD CONDIT ET AL. Ecological Monographs Vol. 65, No. 4 APPENDIX 3. Continued. 1982-1985 1985-1990 Mor- 95% Mor- 95% tality confidence tality confidence Genus and species N* D* rate* limits N D rate limits Guettarda foliacea U- 304 12 1.20 1.89 0.53 299 26 1.72 2.40 1.07 78 2 0.76 2.77 0.08 83 5 1.18 2.75 0.39 Gustavia superba M-C 244 31 3.94 5.36 2.58 179 17 1.87 2.78 1.00 637 15 0.67 1.00 0.33 642 11 0.32 0.52 0.13 Hamelia axillaris S-W 113 38 12.20 16.39 8.53 92 36 9.43 12.82 6.55 Hampea appendiculata M- 35 22 30.21 47.40 19.28 16 9 15.79 31.28 7.35 41 12 10.84 17.71 5.22 33 14 10.51 17.11 5.62 Hasseltia foribunda M-W 885 130 4.90 5.76 4.07 762 148 4.10 4.78 3.45 262 25 3.06 4.28 1.88 255 33 2.63 3.55 1.75 Heisteria acuminata U- 100 8 2.44 4.18 0.80 101 2 0.38 1.38 0.04 Heisteria concinna M- 642 11 0.50 0.80 0.21 708 39 1.08 1.42 0.74 246 6 0.72 1.30 0.15 256 13 0.99 1.53 0.46 Herrania purpurea U-C 522 31 1.87 2.53 1.22 531 30 1.11 1.51 0.72 Hirtella americana T- 42 2 1.26 4.55 0.14 39 3 1.52 4.44 0.30 Hirtella triandra M-S 3628 140 1.16 1.35 0.96 4102 189 0.90 1.02 0.77 516 30 1.70 2.31 1.10 554 33 1.16 1.56 0.77 Hura crepitans T-W 27 3 3.11 9.09 0.64 22 0 0.00 3.17 0.00 100 4 1.08 2.77 0.29 97 2 0.39 1.40 0.04 Hybanthus prunifolius S- 39 869 3648 2.92 3.01 2.83 41 107 4996 2.46 2.53 2.39 Hyeronima alcheornoides T-WC 57 7 3.98 7.08 1.16 58 15 5.70 8.83 3.00 44 3 2.07 6.07 0.42 42 1 0.44 2.45 0.01 Inga cocleensis M-C 180 17 3.44 5.11 1.84 179 8 0.88 1.50 0.28 39 6 5.70 10.61 1.42 39 9 5.07 8.71 2.00 Inga fagifolia T-W 50 1 0.65 3.63 0.01 53 4 1.50 3.84 0.40 Inga goldmanii T- 436 30 2.13 2.91 1.38 417 45 2.17 2.81 1.55 62 16 8.78 13.44 4.75 51 13 5.57 8.89 2.75 Inga marginata T-S 832 194 7.56 8.65 6.52 734 222 6.82 7.75 5.94 81 29 13.18 18.46 8.70 74 16 4.60 7.01 2.47 Inga pezizifera T-S 183 25 3.57 5.01 2.21 187 25 2.70 3.79 1.67 25 9 10.56 18.80 4.46 15 7 11.84 23.93 4.54 Inga quaternata M-S 701 41 1.70 2.23 1.19 701 58 1.64 2.06 1.22 34 5 4.38 10.23 1.40 40 5 2.53 5.91 0.80 Inga ruiziana T- 70 6 2.95 5.40 0.67 64 6 1.88 3.44 0.43 Inga sapindoides M-C 324 24 2.37 3.33 1.44 287 31 2.17 2.95 1.42 67 9 4.32 7.28 1.62 67 7 2.08 3.69 0.60 Inga sp. nov. U- 196 5 0.75 1.74 0.24 236 14 1.16 1.78 0.56 Inga umbellifera M- 921 48 1.64 2.11 1.18 979 78 1.58 1.94 1.23 21 6 10.01 19.40 2.89 17 5 6.59 15.53 2.06 Jacaranda copaia T-C 118 26 8.19 11.51 5.18 89 15 3.52 5.39 1.81 224 7 1.02 1.78 0.27 230 18 1.55 2.28 0.85 Lacistema aggregatum U- 1514 142 3.15 3.67 2.63 1620 202 2.54 2.89 2.19 43 8 6.29 11.01 2.21 40 5 2.54 5.93 0.81 Lacmellea panamensis M-W 55 0 0.00 2.10 0.00 54 1 0.36 1.99 0.01 36 2 1.75 6.34 0.20 37 0 0.00 1.90 0.00 Laetia thamnia U- 499 39 2.58 3.40 1.78 502 64 2.60 3.25 1.97 Licania hypoleuca M- 101 3 0.92 2.69 0.18 105 4 0.73 1.87 0.20 Licania platypus T-S 244 11 1.20 1.93 0.50 284 26 1.82 2.53 1.13 Lindackeria laurina M-W 24 7 10.43 19.40 3.52 19 6 7.18 14.07 2.14 85 4 1.44 3.70 0.39 78 9 2.30 3.86 0.85 Lonchocarpus latifolia T-C 695 46 2.20 2.84 1.57 707 62 1.75 2.20 1.32 147 17 3.89 5.79 2.09 137 24 3.67 5.20 2.25 Luehea seemannii T-WC 101 27 10.13 14.21 6.50 97 32 7.63 10.50 5.14 87 1 0.34 1.91 0.01 93 0 0.00 0.71 0.00 Macrocnemum glabrescens M-SC 72 0 0.00 1.33 0.00 76 4 1.02 2.60 0.27 24 2 2.26 8.19 0.26 25 1 0.75 4.19 0.02 Malmea sp. nov. M- 262 5 0.59 1.37 0.20 304 1 0.06 0.35 0.01 Malpighia romeroana S- 54 3 1.81 5.28 0.37 63 8 2.58 4.46 0.87 Maquira costaricana M- 1195 62 1.58 1.97 1.19 1245 58 0.91 1.14 0.67 223 35 5.18 6.95 3.51 200 39 4.13 5.47 2.87 Maytenus schippii M- 62 0 0.00 1.66 0.00 63 1 0.30 1.70 0.01 20 3 4.73 13.86 0.94 19 2 2.12 7.65 0.24 Miconia affinis U-C 367 32 3.15 4.27 2.08 391 76 4.15 5.10 3.23 Miconia argentea M-WC 486 125 9.30 10.98 7.70 628 273 10.84 12.19 9.58 45 4 2.80 7.17 0.75 50 15 6.71 10.47 3.57 Miconia elata U- 32 5 6.13 14.34 1.96 28 6 4.65 8.79 1.24 Miconia hondurensis U-W 23 4 6.71 17.25 1.78 24 .4 3.48 8.92 0.92 This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions November 1995 MORTALITY IN 205 TROPICAL TREE SPECIES 437 APPENDIX 3. Continued. 1982-1985 1985-1990 Mor- 95% Mor- 95% tality confidence tality confidence Genus and species N* D* rate* limits N D rate limits Miconia nervosa S-C 359 146 18.55 21.73 15.63 293 170 16.60 19.37 14.19 Mouriri myrtilloides S- 6948 600 2.87 3.10 2.64 7707 1077 2.87 3.04 2.70 Myrcia gatunensis U- 38 2 1.74 6.26 0.20 47 3 1.26 3.70 0.26 Nectandra cissifiora T- 326 40 4.71 6.20 3.28 319 25 1.57 2.20 0.96 23 2 3.05 11.02 0.35 20 0 0.00 3.54 0.00 Nectandra globosa M-SWC 105 19 5.54 8.16 3.15 103 16 3.20 4.83 1.69 Nectandra purpurea M- 76 7 3.03 5.37 0.86 75 6 1.60 2.92 0.36 Neea amplifolia S- 62 11 6.19 10.08 2.72 71 11 3.20 5.19 1.39 Ocotea cernua M- 322 33 3.31 4.46 2.20 309 24 1.54 2.16 0.93 24 0 0.00 4.89 0.00 28 2 1.40 5.08 0.16 Ocotea oblonga T- 182 49 9.55 12.36 6.97 152 47 7.05 9.19 5.12 33 5 4.53 10.61 1.45 31 11 8.22 13.89 3.87 Ocotea puberula T- 260 50 7.09 9.12 5.18 233 34 3.00 4.04 2.02 Ocotea whitei T-S 963 227 7.10 8.04 6.19 770 202 5.74 6.55 4.96 167 14 2.47 3.80 1.20 170 14 1.60 2.45 0.78 Oenocarpus mapoura M- 1038 74 2.33 2.86 1.80 966 19 0.38 0.55 0.21 752 14 0.57 0.88 0.27 746 9 0.23 0.38 0.08 Olmedia aspera U-S 392 71 5.35 6.62 4.13 334 82 5.35 6.54 4.22 50 7 4.05 7.24 1.20 42 8 4.00 7.01 1.41 Ormosia croatii T-S 48 1 0.58 3.21 0.01 53 0 0.00 1.31 0.00 Ouratea lucens S- 1122 29 0.77 1.05 0.49 1240 60 0.94 1.18 0.71 Palicourea guianensis S-WC 377 113 11.79 14.05 9.67 659 322 12.82 14.30 11.44 Pentagonia macrophylla U-S 566 64 3.54 4.42 2.68 510 54 2.13 2.70 1.57 Perebea xanthochyma M- 255 23 3.53 5.01 2.12 255 25 1.98 2.77 1.22 Phoebe cinnamomifolia T-C 70 12 5.95 9.52 2.75 64 10 3.24 5.37 1.33 Picramnia latifolia U-S 1131 97 2.51 3.01 2.01 1137 145 2.58 3.00 2.16 38 4 2.96 7.59 0.79 39 5 2.60 6.07 0.83 Piper aequale S-S 219 74 12.59 15.62 9.83 158 68 10.73 13.50 8.30 Piper arboreum U-S 107 24 7.98 11.35 4.93 82 22 5.97 8.66 3.62 Piper cordulatum S- 3147 400 4.48 4.92 4.04 3708 1273 8.05 8.50 7.61 Piper culebranum S-SW 120 56 20.87 26.95 15.73 65 17 5.81 8.81 3.22 Piper perlasense S-S 110 23 6.14 8.79 3.74 117 39 7.67 10.27 5.39 Piper reticulatum U-SW 171 19 3.55 5.19 1.99 160 28 3.66 5.06 2.35 Platymiscium pinnatum T- 185 18 3.16 4.66 1.73 179 28 3.24 4.48 2.07 71 5 2.26 5.29 0.73 69 4 1.11 2.85 0.30 Platypodium elegans T-C 112 18 5.36 7.95 2.98 108 18 3.48 5.16 1.93 58 9 5.22 8.83 1.98 49 5 1.96 4.59 0.63 Posoqueria latifolia M- 63 1 0.48 2.67 0.01 65 5 1.53 3.56 0.49 Poulsenia armata T- 2507 593 7.77 8.40 7.15 1822 418 4.95 5.43 4.48 922 158 5.43 6.29 4.59 857 161 3.93 4.54 3.33 Pourouma bicolor T-S 29 1 1.24 6.88 0.03 31 3 1.97 5.77 0.41 Pouteria reticulata T- 1495 98 2.03 2.44 1.63 1548 117 1.49 1.76 1.22 157 9 1.80 3.00 0.65 170 8 0.90 1.53 0.29 Pouteria stipitata M- 28 1 1.19 6.62 0.03 30 0 0.00 2.34 0.00 33 3 2.93 8.58 0.58 31 2 1.26 4.56 0.15 Prioria copaifera T- 1077 15 0.47 0.71 0.23 1099 33 0.59 0.79 0.39 279 3 0.36 1.07 0.08 309 8 0.50 0.85 0.16 Protium costaricense M-S 803 75 2.98 3.66 2.31 758 91 2.43 2.94 1.94 110 24 7.42 10.56 4.58 101 20 4.18 6.10 2.43 Protium panamense M- 2650 149 1.87 2.17 1.57 2790 206 1.47 1.67 1.27 65 17 9.00 13.64 4.99 51 6 2.37 4.37 0.56 Protium tenuifolium M- 2310 86 1.05 1.27 0.83 2552 169 1.30 1.50 1.10 354 26 2.11 2.94 1.31 353 39 2.21 2.91 1.53 Psidium 35 0 0.00 2.83 0.00 37 1 0.52 2.88 0.01 Psychotria defiexa S- 88 27 12.75 17.95 8.22 77 40 14.13 19.23 10.11 Psychotria grandis U-SW 102 18 5.21 7.73 2.90 94 35 8.78 11.96 6.05 Psychotria horizontalis S- 6167 829 4.22 4.51 3.93 6437 1431 4.74 4.99 4.50 Psychotria marginata S- 582 92 5.47 6.61 4.37 691 183 5.84 6.71 5.01 Pterocarpus rohrii T- 1441 90 2.04 2.46 1.62 1521 138 1.81 2.12 1.51 136 35 8.52 11.50 5.82 103 22 4.52 6.52 2.72 Quararibea asterolepis T- 1691 88 1.55 1.87 1.23 1684 109 1.26 1.50 1.03 703 43 1.81 2.35 1.27 694 38 1.03 1.36 0.71 Quassia amara U- 148 5 0.92 2.14 0.29 143 7 0.95 1.67 0.26 Randia armata U- 900 43 1.49 1.94 1.05 920 73 1.57 1.93 1.21 228 15 2.04 3.09 1.02 234 23 1.96 2.78 1.18 Rinorea sylvatica 5- 2570 168 2.01 2.31 1.70 2612 201 1.52 1.73 1.31 This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions 438 RICHARD CONDIT ET AL. Ecological Monographs Vol. 65, No. 4 APPENDIX 3. Continued. 1982-1985 1985-1990 Mor- 95% Mor- 95% tality confidence tality confidence Genus and species N* D* rate* limits N D rate limits Sapium aucuparium T-W 24 8 11.99 21.82 4.62 21 10 12.21 22.11 5.75 23 5 7.16 16.80 2.26 20 4 4.07 10.46 1.08 Scheelea zonensis M-W 0 0 0 0 44 3 2.10 6.15 0.42 40 6 3.08 5.72 0.76 Senna dariensis S-C 204 79 14.53 17.96 11.46 136 71 14.02 17.68 10.94 Simarouba amara T- 993 165 5.92 6.83 5.02 995 241 5.31 5.99 4.65 247 25 3.44 4.82 2.12 255 51 4.27 5.48 3.13 Siparuna paucifiora U-S 407 54 4.07 5.18 3.00 328 31 1.88 2.56 1.23 24 2 2.61 9.43 0.30 26 1 0.74 4.15 0.02 Sloanea ternifiora T- 516 20 1.16 1.68 0.66 505 17 0.65 0.97 0.35 85 1 0.33 1.82 0.01 85 5 1.13 2.65 0.37 Socratea exorrhiza M- 438 74 6.39 7.88 4.96 380 92 5.33 6.46 4.27 374 45 4.69 6.09 3.34 357 64 3.75 4.69 2.85 Solanum hayesii M-SC 85 43 23.25 31.23 16.82 64 42 20.24 28.07 14.71 40 19 21.90 33.77 13.12 25 16 19.54 33.68 11.50 Sorocea affinis S- 3255 227 2.15 2.43 1.87 3326 267 1.59 1.78 1.40 47 10 6.79 11.35 2.86 44 14 7.27 11.55 3.78 Spondias mombin T-WC 39 10 10.09 17.02 4.33 42 9 4.66 7.98 1.83 24 3 4.36 12.78 0.89 23 1 0.84 4.67 0.02 Spondias radlkoferi T-C 137 36 9.49 12.77 6.52 108 28 5.71 7.97 3.70 55 4 2.29 5.84 0.62 56 2 0.69 2.47 0.08 Sterculia apetala T-W 43 7 5.45 9.79 1.65 29 0 0.00 2.44 0.00 25 0 0.00 4.39 0.00 24 0 0.00 2.79 0.00 Stylogyne standleyi S-W 712 51 2.26 2.89 1.65 732 54 1.45 1.85 1.07 Swartzia simplex var. grandifolia U- 2057 18 0.26 0.38 0.14 2219 30 0.26 0.35 0.17 198 8 1.24 2.11 0.39 203 4 0.38 0.97 0.10 Swartzia simplex var. ochnacea U- 2597 28 0.32 0.44 0.20 2708 50 0.35 0.45 0.26 104 1 0.31 1.71 0.01 112 4 0.69 1.77 0.19 Symphonia globulifera T-S 142 11 2.32 3.72 0.98 140 18 2.61 3.86 1.44 46 14 10.39 16.46 5.38 38 8 4.45 7.83 1.59 Tabebuia guayacan T- 46 3 2.13 6.23 0.43 46 2 0.85 3.08 0.10 30 2 2.08 7.54 0.25 28 0 0.00 2.45 0.00 Tabebuia rosea T-W 235 26 3.55 4.95 2.22 224 24 2.15 3.03 1.31 81 9 3.28 5.52 1.22 75 9 2.38 4.01 0.89 Tabernaemontana arborea T-W 994 53 1.75 2.22 1.28 1026 66 1.27 1.57 0.96 293 10 1.15 1.86 0.44 302 24 1.53 2.15 0.93 Tachigalia versicolor T- 2837 263 3.03 3.40 2.67 2895 442 3.16 3.46 2.87 86 12 4.55 7.24 2.08 82 15 3.77 5.78 1.95 Talisia nervosa U- 813 23 0.75 1.06 0.45 819 35 0.83 1.10 0.55 Talisia princeps M- 616 19 0.87 1.27 0.48 629 14 0.42 0.65 0.20 Terminalia amazonica T-W 34 1 0.96 5.32 0.02 32 2 1.23 4.44 0.14 28 1 1.18 6.57 0.03 28 0 0.00 2.43 0.00 Terminalia oblonga T- 49 1 0.58 3.25 0.01 48 3 1.22 3.58 0.25 43 2 1.27 4.56 0.15 42 0 0.00 1.61 0.00 Tetragastris panamensis T- 2935 94 0.99 1.19 0.79 3375 201 1.17 1.33 1.01 318 15 1.47 2.22 0.73 323 12 0.71 1.12 0.31 Thevetia ahouai U-SW 105 14 4.52 6.98 2.24 96 11 2.33 3.76 1.00 Trattinickinia spera T- 62 12 6.71 10.76 3.12 49 8 3.38 5.88 1.17 50 8 5.41 9.41 1.87 47 8 3.56 6.21 1.24 Trichilia pallida M-W 491 41 2.68 3.51 1.87 497 43 1.72 2.24 1.21 76 5 2.16 5.03 0.70 75 5 1.31 3.06 0.42 Trichilia tuberculata T. 10905 912 2.47 2.63 2.31 11252 1268 2.26 2.38 2.13 2022 205 3.25 3.70 2.81 1901 239 2.54 2.87 2.22 Triplaris cumingiana M-W 246 17 1.96 2.91 1.05 215 22 2.03 2.90 1.20 125 9 2.11 3.53 0.76 127 5 0.75 1.75 0.24 Trophis racemosa M-S 258 8 0.89 1.51 0.28 275 17 1.21 1.80 0.64 54 8 4.64 8.06 1.59 48 13 5.98 9.57 2.97 Turpinia occidentalis T- 84 23 10.04 14.46 6.17 57 19 7.72 11.58 4.51 69 18 9.19 13.78 5.20 56 9 3.29 5.58 1.25 Unonopsis pittieri M-S 635 49 2.36 3.02 1.70 636 63 1.98 2.48 1.50 136 11 2.56 4.11 1.08 147 6 0.79 1.43 0.17 Virola sp. nov. T-S 37 3 2.39 6.99 0.48 33 3 1.79 5.24 0.36 21 3 4.40 12.90 0.88 20 0 0.00 3.40 0.00 Virola sebifera M- 1799 152 2.74 3.18 2.31 1688 188 2.25 2.57 1.93 607 62 3.27 4.10 2.47 588 46 1.55 2.00 1.10 This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions November 1995 MORTALITY IN 205 TROPICAL TREE SPECIES 439 APPENDIX 3. Continued. 1982-1985 1985-1990 Mor- 95% Mor- 95% tality confidence tality confidence Genus and species N* D* rate* limits N D rate limits Virola surinamensis T-S 126 17 4.35 6.49 2.35 96 12 2.53 4.03 1.15 174 20 3.50 5.08 2.01 163 13 1.53 2.38 0.72 Vismia baccifera U-W 74 17 8.72 13.16 4.80 74 16 4.64 7.06 2.49 Xylopia macrantha M-S 738 16 0.60 0.90 0.31 816 27 0.63 0.88 0.40 79 4 1.44 3.69 0.39 99 4 0.78 1.99 0.21 Xylosma oligandrum S- 182 22 3.91 5.59 2.32 167 23 2.83 4.02 1.70 Zanthoxylum belizense T-C 117 42 14.86 19.73 10.61 144 76 14.39 18.03 11.34 103 13 4.13 6.46 1.96 108 19 3.68 5.41 2.09 Zanthoxylum panamense T-C 214 51 8.37 10.76 6.15 168 45 5.94 7.76 4.27 83 17 7.02 10.55 3.84 70 13 3.92 6.18 1.89 Zanthoxylum procerum M-C 181 16 2.88 4.33 1.50 184 45 5.33 6.96 3.83 26 5 6.38 14.96 2.03 28 6 4.56 8.63 1.22 Zuelania guidonia M-W 27 2 2.13 7.68 0.25 27 6 4.75 9.00 1.28 * N is sample size, or the total number of stems alive at the start of a census interval, and D is the number of stems that died by the end of the interval. Mortality rates are given as annualized percentages along with upper and lower 95% confidence limits. This content downloaded by the authorized user from 192.168.52.67 on Mon, 10 Dec 2012 15:52:07 PM All use subject to JSTOR Terms and Conditions