Journal of Tropical Ecology (2006) 22:11?24. Copyright ? 2006 Cambridge University Press doi:10.1017/S0266467405002774 Printed in the United Kingdom The contribution of interspecific variation in maximum tree height to tropical and temperate diversity David A. King*1, S. Joseph Wright? and Joseph H. Connell? * Center for Tropical Forest Science ? Arnold Arboretum Asia Program, Harvard University, 22 Divinity Ave., Cambridge, MA 02138, USA ? Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002-0948, USA ? Department of Ecology, Evolution, and Marine Biology, University of California, Santa Barbara, California 93106 USA (Accepted 9 June 2005) Abstract: Maximum height was assessed for tree species from seven temperate deciduous forests, one subtropical forest and one tropical forest and combined with published tree heights for three other tropical forests. The temperate deciduous forests showed a strong concentration of canopy species and a dearth of subcanopy species. In contrast, the four tropical forests showed more uniform distributions of maximum heights, while the subtropical forest had an intermediate distribution. The tropical and subtropical sites had greater densities of small trees than did the temperate sites and most of these small trees were members of small- to medium-sized species. Sapling recruitment per unit stem basal area increased with declining maximum height in Panama, which is consistent with the criterion for coexistence of species of differing stature derived from Kohyama?s forest architecture hypothesis. Greater penetration of light into gaps and favourable conditions for growth over most of the year may allow more smaller-statured species to coexist with canopy trees in tropical vs. temperate forests. Key Words: Australia, biodiversity, forest architecture hypothesis, forest stratification, Panama, temperate forest, tree height, tropical forest, USA INTRODUCTION Tropical forest trees showgreat diversity in size (Kohyama 1996, Richards 1996, Turner 2001), with adult heights ranging from 1 to 40?70 m on sites without severe waterornutrient limitations.Understoreyandsubcanopy trees appear to be quite diverse among wet lowland forests of the neotropics (Pitman et al. 2002, Popma et al. 1988, Valencia et al. 2004). In contrast, the deciduous forests of eastern North America include some shrubs and understorey trees, but show greater concentrations of large-statured canopy trees and a dearth of subcanopy species (Pacala et al. 1996). The higher diversity of tropical vs. temperate forests has been associated with this greater diversity in adult stature, which may re?ect reduced exclusion of smaller-statured species by canopy species at lower latitudes (Kohyama 1993, 1996). Despite the importance of tree height in forest commu- nities, there have been few quantitative studies of within- forest variation in the adult height of the constituent 1Corresponding author. Email: dkingaz@yahoo.com species. Earlier studies, summarized by Richards (1952) andSmith (1973),havedescribedvariation in tree stature in terms of forest strati?cation, reporting three to ?ve strata of trees in wet tropical forests, as compared to two strata (understorey and overstorey) in temperate deciduous forests. Distinct concentrations of foliage at different heights may often be observed along any single vertical transect (Ashton & Hall 1992, Koike & Syahbuddin 1993). However, such distinctions are blurred when averaged over multiple vertical transects (Kira 1978, Parker & Brown 2000, Popma et al. 1988), and the co-occurrence of trees in all ontogenetic stages obscures the effects of interspeci?c differences in stature on forest strata (Richards 1996, Whitmore 1998). More recent classi?cations of species into four or ?ve adult size categories on large-scale plots also indicate broad variation in stature within tropical forests (Condit et al. 1996, Kochummen et al. 1990), but these categories are approximations and include no measured heights of the constituent species. Thus, temperate and tropical forests differ in their distributions of adult stature, but the exact nature of these differences and their relation to diversity remain uncertain. 12 DAVID A. KING, S. JOSEPH WRIGHT AND JOSEPH H. CONNELL Comprehensive measurements of tree height in seven temperate deciduous forests, one subtropical and one tropical forest were combined with reported height distributions for three other tropical forests to address the following questions: (1) What are the quantitative differences in the dis- tributions of maximum tree heights (hmax) among species between tall temperate deciduous forests of the USA and tropical forests? (2) Do the relative abundances of adult understorey trees vs. juveniles of taller species differ between temperate and tropical forests? (3) To what extent are differences in overall tree diversity among the sites associated with differences in the distribution of hmax? In addition, we evaluated recruitment ef?ciency (de?ned as saplings recruited per unit basal area of larger conspeci?cs) and seed and seedling survival to assess possible mechanisms of coexistence of tree species of different adult stature. This de?nition of recruitment ef?ciency was chosen to test the forest architecture hypothesis of Kohyama (1993, 1996), which identi?es sapling recruitment per unit basal area as a key variable in the coexistence of different-sized species. METHODS Sites and tree selection Trees of temperate deciduous forests were measured in seven forestsacross theeastern, centralandsouth-eastern sections of the USA (Table 1). The study sites included rare remnants of once widespread, old-growth forests, usually on moist, fertile soils of protected hollows, north- facing slopes or more level ground (Kershner & Leverett 2004, Martin 1975, McClain et al. 2001, Schmeltz et al. 1974, White & White 1996). The forests were all in the deciduous forest formation of Braun (1950), but varied greatly in species composition. The forests were Table 1. Site descriptions based on United States Geological Survey (1970), Martin (1975), Schmeltz et al. (1974), Muller (1982), Connell et al. (1984) and Leigh (1999). Coordinates for Brown Co. and Great Smoky Mountains National Park are means for multiple stands. Temperate forest types (in USA) after Ku?chler (1964). Annual Site Forest type Location Elevation (m) precip. (m) Mohawk Trail State Forest, MA, USA Northern hardwoods 42?39?N, 72?58?W 200?300 1.3 Brown, Co., IN, USA Beech-maple 39?9?N, 86?16?W 200?250 1.1 Donaldson?s Woods, IN, USA Oak-hickory 38?44?N, 86?24?W 200 1.1 Beall Woods Nature Pres., IL, USA Southern ?oodplain 38?23?N, 87?53?W 120 1.2 Lilley Cornett Woods, KY, USA Mixed mesophytic 37?5?N, 83?0?W 320?500 1.3 Big Oak Woods, NC, USA Oak-hickory-pine (bottomland) 35?53?N, 79?1?W 80 1.15 Great Smoky Mtns N.P., TN, USA Mixed mesophytic (cove) 35?43?N, 83?24?W 800?1100 1.5+ Lamington N.P., Australia Subtropical evergreen 28?14?S, 153?10?E 900 1.9 Gigante, Panama Tropical lowland 9?6?N, 79?51?W 60 2.6 entirely or predominantly deciduous, with the inclusion of one or two coniferous overstorey species at three sites and the large evergreen shrub, Rhododendron maximum L., at two sites. There were ?ve upland forests and two bottomland forests (Illinois and North Carolina). Two stands of similar (north-facing) slope and species composition, separated by 7 km, were combined as a single site that included Ogle Hollow in Brown County State Park and the hollow to the south of Crooked Lake in Yellowwood State Forest, Indiana. Six mid-elevation cove forests (hollows or sheltered north- or east-facing slopes) located within 35 km of each other were combined as a single site in the Great Smoky Mountains National Park, Tennessee. The north-facing slope north-west of Todd Mountain was searched at Mohawk Trail State Forest, Massachusetts. At each site, 2?5-d searches were made for the tallest individuals of every free-standing species attaining a height ? 4 m. All species were included for which relatively large, old individuals could be found. On the siteswithsubstantialvariation inelevation, searcheswere restricted to hollows and low to midslope positions, as trees grow taller in such sheltered positions thanonupper slopesandridges (McNab1989).Only the?oodplainareas were searched at Beall Woods, Illinois, as the adjacent upland foresthadadifferent species composition (McClain et al. 2001). Broad-leavedevergreen treesof a subtropical rain forest were measured on a 1.94-ha plot (Connell et al. 1984), plus additional trees were measured in a small adjacent area, inLamingtonNationalPark,Queensland,Australia. A list of tree coordinates was used to ?nd the largest diameter trees of themore common species.All 18 species with ? 15 tagged trees ? 10 cm dbh were measured, as were 15 of the 30 species with 4?14 tagged trees. For species seldom or never exceeding 10 cm dbh, we measured four of the eight species with ? 5 stems among the 2.5?10 cm dbh trees in the 0.59-ha area censused for saplings (Connell et al. 1984). The largest trees of each selected species generally had rounded crowns and relatively thick limbs indicating maturity or old age. Adult tree size in tropical vs. temperate forest 13 A set of 6256 height measurements was used to characterize thedistributionof adultheights for a lowland tropical forest located on the Gigante Peninsula of the BarroColoradoNatureMonument,Panama.Treeheights were measured for individuals from a wide size range for 95 of the more common species. For each of these species, all individuals ? 10 and ? 1 cm dbh were measured within 36 nested subplots of 0.16 and 0.06 ha each, respectively. The nested subplots were spread uniformly over a 38.4-ha plot where all individuals ? 20 cm dbh were mapped and identi?ed. Additional large trees were measured for the larger species over the entire 38.4-ha plot. The forest at this site is more than 200 y old. The chosen sites were all relatively favourable for tree growth. The deciduous forest stands all receive substantial rain during the summer growing season, with a mean annual precipitation of 1.0 to 1.5 + m y?1 (Table 1) and a frost-free growing season of 140?210 d (Martin 1975, United States Geological Survey 1970). The subtropical forest site receives about 1.9 m y?1 of precipitation, with a summer rainfall maximum (Connell et al. 1984). Mean daily maximum and minimum temperatures for the warmest and coolest months for the nearest weather station at Mt Tamborine (525 m elevation) are respectively, 25.7 and 17.1 ?C in January and 17.1 and 8.0 ?C in July (http://www.bom.gov.au/climate/averages/tables/ca qld names.shtml). The tropical site receives about 2.6 m y?1 of rain with a 4-mo dry season, during which time some of the canopy trees lose their leaves, as described by Leigh (1999) for the nearby forest on Barro Colorado Island. Mean monthly temperatures are 27 ?C in April and 26 ?C in all other months (Leigh 1999). Species diversity was characterized by Fisher?s ?, a measure which is relatively insensitive to plot size and shape (Leigh 1999). The measure was calculated for the four enumerated sites (Donaldson?s Woods, Indiana; Lilley Cornett Woods, Kentucky; and the tropical and subtropical sites). Height measurements Tree heights on the subtropical and temperate sites were determined by ?rst measuring the eye-to-leaf distance of the highest twigs or leaves with a calibrated optical or laser range?nder and then multiplying that distance by the sine of the sighting angle to the horizontal (measured by clinometer). Similar methods were used to calculate the height of the sighting point above the tree base, which was then added to the eye-to-top height to yield the total height above the base. This technique is judged superior to the traditional method of estimating heights fromhorizontaldistancesandsightingangles,becausethe usual assumption that the highest visible point is directly above the base, albeit reasonable for young, conical trees, is questionable for spreading old trees. At the tropical site, heights to 15 m were measured with a telescoping pole. Taller trees were measured from the tree base with a laser range?nder, with tree height taken as observer height plus the greatest of multiple vertical re?ections from the upper crown of the tree. The resulting plots of height vs. dbh were inspected to see if the tallest trees were outliers for each species. The more obvious outliers were remeasured, from which it was inferred that for a few of the hundreds of measurements of mature trees, heights were likely misrecorded by 10 m, or a branch of a taller, overarching tree was measured. Based on this inference, a few additional outlying heights were discarded. The number of species was also reduced to 85 by omitting 10 species with few trees in the upper halves of their diameter distributions. Maximum height estimation A number of methods have been used to estimate maximum height per species including asymptotic approaches (Thomas 1996) and the choice of the 95th percentile in height among all trees greater than 5 or 10 cm dbh (Kohyama et al. 2003, Poorter et al. 2003). As we lacked the necessary height distributions for such methods at most sites, we simply took the heights of the tallest trees per species, as direct measures of maximum height. However, a potential problem with this approach is that trees of greater heights are likely to be encountered among the commonest species. This bias was reduced by averaging the heights of the three tallest measured trees for the most common species, averaging the heights of the two tallest trees for species of intermediate abundance and using the greatest height recorded for less-abundant species. As the resulting maximum heights differed little from those derived by using the maximum measured height for every species, the approach was judged adequate for comparing different forest types. To better assess height patterns in tropical forests, we included maximum heights of the 27 species measured by Kohyama et al. (2003) on two 1-ha plots in a lowland mixed dipterocarp forest of western Borneo and the 53 species measured by Poorter et al. (2003) on 20 1-ha plots in lowland evergreen moist forest of Liberia. The species includedhad?20 individuals? 5 cm dbh on the Bornean plots and ? 10 individuals of 10? 20 cm dbh and a maximum diameter ? 15 cm on the Liberian plots. We also included the inferred distribution of asymptotic maximum heights for the 50-ha plot at Pasoh Forest, Malaysia derived by Thomas (2003) from the diameter distributions of all species and a relation between asymptotic height and the 97th percentile of diameter of trees ?1 cm dbh determined for 42 of these 14 DAVID A. KING, S. JOSEPH WRIGHT AND JOSEPH H. CONNELL species.Of thespeciesmeasuredbyKohyama et al. (2003), 12/27 also occurred on the Pasoh plot. There was no overlap among the species sampled for Africa, South-East Asia, subtropical Australia, Central America and North America. Sapling recruitment Kohyama (1993) hypothesized that small-statured spe- cies require greater recruitment ef?ciency than do larger species to persist in tall forests, i.e. that recruitment per unit of basal area must be higher in small-statured species. This hypothesis was assessed at the tropical site, where sapling recruitment could be determined over a substantial, 2.16-ha area (the 36 0.06-ha plots where all trees ? 1 cm dbh were measured). For each species, the rateof recruitmentperunitbasalareawasestimatedusing themethodofKohyama&Takada (1998).Theyestimated recruitment from sapling density in the size class centred on the recruitment size threshold and the mean growth rate for this size class. We used a sapling size class of 1? 1.9 cm and a recruitment size threshold of 1.5 cm dbh. Thus, the sapling recruitment rate was estimated as R = G ? f , where G is the mean stem diameter growth rate (cm y?1), f is the number of saplings per unit area and diameter class width (no. ha?1 cm?1) for saplings of 1?1.9 cm dbh, and R is the estimated recruitment rate per unit area and time (no. ha?1 y?1) into the ? 1.5 cm dbh size class. The mean value of G of 0.06 cm y?1 reportedbyConditet al. (1999) for1?1.9-cm-dbhsaplings at Barro Colorado (BCI), Panama was used here with the observed species-speci?c values of f to calculate R. This recruitment ratewas thendivided by the basal area perha of all conspeci?c stems ? 1.5 cm dbh to yield recruitment per unit basal area. Seed and ?rst-year seedling survival were assessed for small- vs. large-statured species using 200 stations located in a strati?ed random fashion within the 50-ha plot on BCI, which is 4 km from the Gigante plot. Each station included one 0.5-m2 seed trap, constructed of 1-mm mesh screen, and three adjacent 1-m2 seedling plots (see Wright et al. 2003 for detailed methodology). All seeds were identi?ed and counted in weekly censuses conducted from 1 January 1987 through 31 December 2001. All woody plants ?50 cm tall were tagged and identi?ed between January and March 1994. Survivors were re-measured and new recruits were tagged and identi?ed between January and March each year from 1995 to 2002. Seed survival (seedlings per seed) was estimated as the density of seedling recruits divided by the density of conspeci?c seeds for species with 10 or more seeds captured in the appropriate years. First-year seedling survival (% y?1) was estimated as the proportion of new recruits that survived until the next annual census for species with 10 or more seedling recruits. Seed and seedling survival and seed mass were transformed logarithmically to normalize the residuals. The effect of tree stature on seed and seedling survival was then assessed using an analysis of covariance (ANCOVA), which treated survival as the dependent variable, tree stature as the grouping factor, and seed mass as the covariate. The grouping factor had two levels, which were understorey trees including shrubs and treelets vs. canopy trees includingmid-sizedand large treesasde?ned by Condit et al. (1996). Mean dry seed mass was for endosperm plus embryo for up to four seeds chosen at random from up to ?ve randomly chosen fruits from up to ?ve randomly chosen individuals of each species. RESULTS The temperate sites all showed strong concentrations of maximum height over a narrow range. Therefore we present numbers of species per quarter octave height range (Figure 1). Quarter octaves result in four geo- metrically or logarithmically equal divisions for every doubling in height. Geometric height classes were used because relative height may be more important than absolute height in determining the crown overlap of trees of differing heights. Of the 69 temperate species measured across all sites, 14 to 25 were measured on any one site. The temperate speciesweremeasuredat anaverageof 2.0 of the seven temperate sites. Thus, our seven temperate forests are not repeated measures of the same forest, but rather are closer to being independent measures of temperate deciduous forest structure. The across-site coef?cient of variation of conspeci?c hmax averaged 6.7% for the 31 species occurring at multiple sites. This small coef?cient of variation of approximately one-third of the quarter octave height class range represents an upper bound on the random error in hmax, as it also re?ects genetic and environmental differences between sites. Striking differences in the distributions of maximum height among species were observed between the temperate deciduous sites and the subtropical and tropical sites (Figures 1 and 2). For each of the seven temperate sites, themajority of themeasured specieswere in the largest height classes (hmax ? 28.3 m), with a peak in the 33.6 6) and was therefore excluded from this analysis only. The interaction between tree stature (understorey vs. canopy species) and seed mass was insigni?cant for both seed survival (F1,54 =1.98, P=0.17) and seedling survival (F1,45 =1.25, P=0.27). Survival increased with seed mass for both seeds (F1,55 =21.2, P<0.001) and seed- lings (F1,46 =10.3,P<0.01); andsurvivalwasgreater for understorey trees than for canopy trees after controlling Figure 3. Per cent of trees ? 10 cm dbh belonging to species whose maximum height falls within each quarter octave height class. Heights listed along the abscissa are lower limits of quarter octave height classes. Table 3. The overall densities of trees and the percentages of those trees belonging to canopy species (hmax ? 28.3 m) for two diameter classes. Density refers to all species; percentages to those species for which hmax was measured (making up >62% of the stems in all cases). Data for Donaldson?s Woods, Lilley Cornett Woods, and the subtropical and tropical (Gigante) sites are from Schmeltz et al. (1974), Muller (1982) and this study, respectively. Trees 2.5?10 cm dbh Trees 10?20 cm dbh % stems of % stems of Site Trees ha?1 canopy spp. Trees ha?1 canopy spp. Donaldson?s 112 95 Lilley Cornett? 784 79 Subtropical 2140 15 301 30 Tropical 1847 17 331 27 ?Species abundances at this site were given for 2.5?10 cm dbh and ?10 dbh trees, but not 10?20 cm dbh trees. for the in?uence of seed mass for both seeds (F1,55 =30.3, P<0.001) and seedlings (F1,46 =6.21, P<0.05). DISCUSSION General patterns Our results suggest that temperate deciduous forests of North America differ markedly from tropical Adult tree size in tropical vs. temperate forest 17 Table 4. Diversity of canopy species (hmax ? 28.3 m) and all species attaining 10 cm dbh for the enumerated sites. Fisher?s ? is de?ned by S = ? ln(1 + N/?), where S and N represent the number of species and individuals enumerated, respectively (Leigh 1999). At Lamington and Gigante, the numbers of canopy species and stems of these species (among all stems ?10 cm dbh) were estimated assuming the same proportions of species and trees of this stature range as measured among the more common species (Table 2). Number of trees Number of species Fisher?s ? Sample Site area (ha) Canopy All trees Canopy All trees Canopy All trees Donaldson?s Woods, USA 7.9 2141 2200 20 25 3.1 4.0 Lilley Cornett Woods, USA 1.92 707 788 22 28 4.3 5.7 Lamington, N.P., Australia 1.94 656 1349 30 75 6.5 17.1 Gigante, Panama 5.76 828 2158 45 155 10.2 38.3 Figure 4. Species stature and shade tolerance for the seven deciduous forest sites and the 50-ha plot on Barro Colorado Island, Panama. Shade tolerance classes 1?5 are, respectively, very tolerant, tolerant, intermediate, intolerant and very intolerant of shade, as rated by foresters (Baker 1949, Burns & Honkala 1990), or in the case of the smallest-statured species, by the authors, based on co-occurrence in shade with saplings of shade-tolerant trees and other studies of these species (e.g. Lei & Lechowicz 1990). Shaded sapling mortalities for BCI are for saplings under canopies >10 m tall (Welden et al. 1991); adult size classes 1?4 are, respectively, shrubs, treelets, mid-sized trees and large trees, as de?ned by Condit et al. (1996). Some points were jiggled slightly to the right or left to reduce overlap in the upper panel. forests in the distribution of adult statures of their constituent species. The deciduous forests showed greater dominance by canopy trees (Figure 3) and much higher proportions of large species than did the tropical forests (Figure 2). Extreme dominance by overstorey species was also found for the tall (60 to 100+ m) old- growth coniferous forests of the North American west coast (VanPelt & Franklin2000).A relatively continuous Figure 5. Estimated sapling recruitment per unit stem basal area vs. maximum height per species at Gigante, Panama. The 70 species had at least three 1?1.9-cm-dbh saplings on the sapling subplots. Figure 6. Seed survival is greater for understorey tree species (open symbols) than forcanopytree species (closedsymbols)onBarroColorado Island, Panama. Understorey trees include shrubs and treelets and canopy trees include mid-sized and large trees, as de?ned by Condit et al. (1996). 18 DAVID A. KING, S. JOSEPH WRIGHT AND JOSEPH H. CONNELL Figure 7. First-year seedling survival is greater for understorey tree species (open symbols) than for canopy tree species (closed symbols) on Barro Colorado Island, Panama. Understorey trees include shrubs and treelets and canopy trees include mid-sized and large trees, as de?ned by Condit et al. (1996). distribution of maximum tree heights was observed for both a warm-temperate evergreen angiosperm forest and a montane temperate deciduous forest in southern Japan (Aiba & Kohyama 1996, Koike & Hotta 1996). The Japanese forests were, however, considerably shorter (22- and 29-m maximum height, respectively) than the tall (?40 m) North American forests studied here, and most species in the montane forest were less than half the height of the tallest species. The marked difference in the proportions of large vs. smaller species among the study forests was associated with corresponding differences in tree diversity. For the Panamanian and subtropical sites, a sample containing a given number of stems will include many more species if drawn from all trees than if it is restricted to large species, as indicated by contrasting values of Fisher?s ? (Table 4). This pattern is consistent with the observation of higher? values for small-statured species than for canopy species in a wet tropical forest in Sri Lanka (Gunatilleke et al. 2004). Thus, diversi?cation in tree stature contributes greatly to tropical tree diversity, but is less important in tall temperate forests of North America. Within diverse tropical forests there is substantial diversi?cation in maximum height among co-occuring species of the more speciose genera (Davies et al. 1998, Thomas 1996). At Gigante, Panama, nine of the 14 genera with two or more measured species showed a within-genus range in hmax of more than 1.5 fold (Appendix 3). In contrast, only three of the 12 temperate generawithtwoormoremeasuredspeciesshowedarange this large (Appendix 1). These three are Cornus, Magnolia and especially, Acer, which shows similar patterns in Eurasia (Ackerly & Donoghue 1998). Small-statured species do occur among some of the other large-statured genera of this study (e.g. Quercus), but they are usually restricted to short forests on unfavourable sites. Tree size distributions and the proportions of smaller stems that are juveniles of large-statured species also appear to differ with latitude. As observed by Hartshorn (1978), tropical forests have higher densities of small individuals (2.5?20 cm dbh) than do old-growth temperate deciduous forests. This greater density was associatedwith thedominanceof smaller-statured species among stems of 2.5?20 cm dbh at both the tropical and subtropical sites (Table 3). Coexistence of species of differing stature Kohyama (1993, 1996) used a forest dynamics model to demonstrate that the number of species that can coexist increaseswith the range in adult height of the constituent species. However, coexistence was only predicted for limited ranges in the establishment, growthandmortality rates of the constituent species and these allowable ranges decreased with increasing species packing along the height gradient. Anecessarycondition for thecoexistenceof specieswith identicalmortalityandgrowth functions (relatinggrowth to current size,maximumsize and standbasal area) is that sapling recruitment per unit basal area must increase as maximum size decreases (Kohyama 1993, 1996). This condition was met by the species of the warm temperate rain forest to which the model was applied, a South-East Asian forest (Kohyama et al. 2003) and the Panamanian site considered here (Figure 5). On Barro Colorado Island, near this site,understorey species showedgreater seedling establishment per seed and greater seedling survival rates than did larger-statured species (Figures 6 and 7). At Pasoh Forest, Malaysia, saplings of understorey species tend to have lower photosynthetic light compensation points and lower photosynthetic capacities than do saplings of larger-statured species (Thomas & Bazzaz 1999). Thus, adaptations for rapid growth under high light in canopy species may result in lower seedling shade tolerance than in understorey species, despite considerable plasticity and interspeci?c differenceswithin both groups (Thomas 2003, Thomas & Bazzaz 1999, Turner2001,Valladares et al.2000,Welden et al.1991). As only about 1 of 100,000 seeds produced on Barro Colorado grow into saplings ? 1 cm dbh (Harms et al. 2000), small differences in mortality rates may have substantial effects on sapling recruitment and hence species coexistence. Adult tree size in tropical vs. temperate forest 19 Density-dependentdeathof seedlingsmaybeassociated with the build-up of host-speci?c pests and pathogens on conspeci?c leaf litter and seedlings on the forest ?oor (Leigh1999).Becausesmallunderstoreytreesproduce far less leaf litter and seed mass per tree than do canopy trees, they are effectively rarer with respect to such effects. This inference is supported by the observation of decreasing density-dependent effects on sapling recruitment with decreasing adult size on Barro Colorado Island, with the exception of the most common treelet species, Faramea occidentalis (Condit et al. 1992, Leigh 1999). Thus, density-dependent processes thought to favour rare species may also contribute to the observed higher recruitment ef?ciency of understorey species. Possible causes of latitudinal shifts in tree stature distributions The current distribution of adult tree statures within forests is in?uenced by the regional species pool and the relative abilities of species of differing size to compete or coexist. Factors that may in?uence relative abilities to compete and that differ between tropical and temperate regions includegrowingseason length, sunangleand tree crown shape. A year-round growing season and leaf life spans of 1 ? 4 + y among understorey plants (Coley 1988) make it possible to recover leaf construction costs over longer periods and may therefore lower whole-plant light compensation points in the tropics (Givnish 1988, Terborgh 1985). Somewhat less light penetrates to the forest ?oor in tropical vs. temperate deciduous forests (Brown & Parker 1994, Leigh 1999, Richards 1996), consistent with lower whole-plant light compensation points and hence greater utilization of light by tropical understorey trees. As a result, wet tropical forests support greater leaf area per unit ground area (LAI) than do temperate deciduous forests (Leigh 1999) and vertical transects in tropical forests commonly encounter crowns of several trees or saplings, onebeneath theother (Koike& Syahbuddin 1993). High midday sun angles enhance penetration of incident light into canopy gaps and to the vegetation below in equatorial forests (Canham et al. 1990). How- ever, gaps providing direct illumination of the ground are ephemeral in tropical forests, due to the germination of fast-growing pioneer species and the vigorous growth response of existing vegetation (Denslow 1987). Gaps between canopy trees are ubiquitous, and gaps that reach down to the forest ?oor are usually wider at higher levels within forests (Connell et al. 1997, Hubbell & Foster 1986). This extra light associated with upper-level gaps is essential for thematurationofmost canopy species (Canham 1985) and increases the reproductive output of understorey trees (Levey 1988). Thus, the bene?cial in?uence of upper canopy gaps on adult understorey and midstorey trees may have larger effects on diversity than the small, shifting area of ephemeral gaps reaching down to the forest ?oor. Furthermore, the abundance of these smaller-statured species, along with lianas (Putz 1984), may also aid in maintaining upper level openings by retarding their closure by large-statured species. In contrast, the tendency for crown depth in canopy trees to increase with increasing latitude (Kuuluvainen 1992), coupled with changes in sun angle, reduces the penetration of light into gaps in high-latitude forests (Canham et al. 1990), thereby increasing the competitive advantage of overstorey vs. understorey trees with increasing latitude. This reduced light penetration may also explain the lack of small-statured shade-intolerant species at the temperate sites (Figure 4); these species are typically associated with persistent openings, such as stream banks, or widespread disturbances, such as stand- replacing ?res (Fernald 1950). Thus, the combined effects of sun angle, overstorey crown geometry and a year- roundgrowingseasonmayact synergistically toprovidea relative advantage to understorey and midstorey trees at lowlatitudesby increasing lightpenetration to thecrowns of these trees and by increasing their capacity to subsist on limited light (Terborgh 1985). ACKNOWLEDGEMENTS We thank Carmi Korine, Ru?no Gonzalez and Omar Hernandez for measuring tree heights at Gigante, Panama, and the Great Smoky Mountains National Park, Lilley Cornett Woods Appalachian Ecological Research Station, Eastern Kentucky University and the North Carolina Botanical Garden for logistical support and permission to conduct research and Robert Leverett for locations of old-growth stands at Mohawk Trail State Forest, Massachusetts. Helpful reviews were provided by Peter Ashton, Takashi Kohyama, Egbert Leigh Jr., Helene Muller-Landau and two anonymous reviewers. The Scholarly Studies and Environmental Sciences Programs of the Smithsonian Institution provided support for measurements made in Panama. LITERATURE CITED ACKERLY, D. D. & DONOGHUE, M. J. 1998. Leaf size, sapling allometry, and Corner?s rules: phylogeny and correlated evolution in maples (Acer). American Naturalist 152:767?791. AIBA, S. & KOHYAMA, T. 1996. Tree species strati?cation in relation to allometry and demography in a warm-temperate rain forest. Journal of Ecology 84:207?218. 20 DAVID A. KING, S. JOSEPH WRIGHT AND JOSEPH H. CONNELL ASHTON, P. S. & HALL, P. 1992. Comparisons of structure among mixed dipterocarp forests of north-western Borneo. Journal of Ecology 80:459?481. BAKER, F. S. 1949. A revised tolerance table. Journal of Forestry 47: 179?181. BRAUN, E. L. 1950. Deciduous forests of Eastern North America. Blakiston, Philadelphia. 596 pp. BROWN, M. J. & PARKER, G. G. 1994. Canopy light transmittance in a chronosequence of mixed-species deciduous forests. Canadian Journal of Forest Research 24:1694?1703. BURNS, R. M. & HONKALA, B. H. 1990. Summary of tree character- istics. Pp. 846?849 in Burns, R. M. & Honkala, B. H. (Technical coordinators). Silvics of North America Vol. 2. U.S. Department of Agriculture, Agricultural Handbook No. 654. CANHAM, C. D. 1985. Suppression and release during canopy recruitment in Acer saccharum. Bulletin of the Torrey Botanical Club 112:134?145. CANHAM, C. D., DENSLOW, J. S., PLATT, W. J., RUNKLE, J. R., SPIES, T. A. & WHITE, P. S. 1990. Light regimes beneath closed canopies and tree-fall gaps in temperate and tropical forests. Canadian Journal of Forest Research 20:620?631. COLEY,P.D.1988.Effect of plantgrowthand leaf lifetimeon theamount and type of anti-herbivore defense. Oecologia 74:531?536. CONDIT, R., HUBBELL, S. P. & FOSTER, R. B. 1992. Recruitment near conspeci?c adults and the maintenance of tree and shrub diversity in a Neotropical forest. American Naturalist 140:261?286. CONDIT, R., HUBBELL, S. P. & FOSTER, R. B. 1996. Changes in a tropical forest with a shifting climate: results from a 50 ha permanent census plot in Panama. Journal of Tropical Ecology 12: 231?256. CONDIT, R., ASHTON, P. S., MANOKARAN, N., LAFRANKIE, J. V., HUBBELL, S. P. & FOSTER, R. B. 1999. Dynamics of the forest communities at Pasoh and Barro Colorado: comparing two 50-ha plots. Philisophical Transactions of the Royal Society of London B 334:1739?1748. CONNELL, J. H., TRACEY, J. G. & WEBB, L. J. 1984. Compensatory recruitment, growth, andmortality as factorsmaintaining rain forest tree diversity. Ecological Monographs 54:141?164. CONNELL, J. H., LOWMAN, M. D. & NOBLE, I. R. 1997. Subcanopy gaps in temperate and tropical forests. Australian Journal of Ecology 22:163?168. D?ARCY, W. G. 1987. Flora of Panama. Missouri Botanical Garden, St. Louis. 328 pp. DAVIES, S. J., PALMIOTTO, P. A., ASHTON, P. S., LEE, H. S. & LAFRANKIE, J.V.1998.Comparativeecologyof11sympatric species of Macaranga in Borneo: tree distribution in relation to horizontal and vertical resource heterogeneity. Journal of Ecology 86:662?673. DENSLOW, J. S. 1987.Tropical rainforest gaps and tree species diversity. Annual Review of Ecology and Systematics 18:431?451. FERNALD, M. L. 1950. Gray?s Manual of Botany. (Eighth (centennial) edition). Dioscorides Press, Portland. 1632 pp. GIVNISH, T. J. 1988. Adaptations to sun and shade: a whole plant perspective. Australian Journal of Plant Physiology 15:63?92. GUNATILLEKE,C.V.S.,GUNATILLEKE, I.A.U.N., ETHUGALA,A.U.K., WEERASEKARA, N. S., ESUFALI, S., ASHTON, P. S., ASHTON, P. M. S. & WIJESUNDARA, D. S. A. 2004. Community ecology in an everwet forest in Sri Lanka. Pp. 119?144 in Losos, E. C. & Leigh, E. G. (eds). Tropical forest diversity and dynamism: ?ndings from a large-scale plot network. University of Chicago Press, Chicago. 645 pp. HARMS, K. E., WRIGHT, S. J., CALDERON, O., HERNANDEZ, A. & HERRE, E. A. 2000. Pervasive density-dependent recruitment enhances seedling diversity in a tropical forest. Nature 404:493? 495. HARTSHORN, G. S. 1978. Tree falls and tropical forest dynamics. Pp.617?638 inTomlinson,P.B.&Zimmermann,M.H. (eds).Tropical trees as living systems. Cambridge University Press, Cambridge. HUBBELL, S. P. & FOSTER,R. B. 1986. Canopy gaps and the dynamics of a Neotropical forest. Pp. 77?96 in Crawley, M. J. (ed.). Plant ecology. Blackwell Scienti?c Publications, Oxford. KERSHNER, B. & LEVERETT, R. T. 2004. The Sierra Club guide to ancient forests of the Northeast. Sierra Club Books, San Francisco. 276 pp. KIRA, T. 1978. Community architecture and organic matter dynamics in tropical lowland rain forests of Southeast Asia with special ref- erence to Pasoh Forest, West Malaysia. Pp. 561?590 in Tomlinson, P. B. & Zimmermann, M. H. (eds). Tropical trees as living systems. Cambridge University Press, Cambridge. KOCHUMMEN, K. M., LAFRANKIE, J. V. & MANOKARAN, N. 1990. Floristic composition of Pasoh Forest Reserve, a lowland rain forest in Peninsular Malaysia. Journal of Tropical Forest Science 3: 1?13. KOHYAMA, T. 1993. Size-structured tree populations in gap-dynamic forest ? the forest architecture hypothesis for the stable coexistence of species. Journal of Ecology 81:131?143. KOHYAMA,T.1996.The roleof architecture inenhancingplant species diversity. Pp. 21?33 in Abe, T., Levin, S. A. & Higashi, M. (eds). Biodiversity: an ecological perspective. Springer-Verlag, New York. KOHYAMA,T.&TAKADA,T.1998.Recruitment rates in forest plots:Gf estimates using growth rates and size distributions. Journal of Ecology 86:633?639. KOHYAMA, T., SUZUKI, E., PARTOMIHARDJO, T., YAMADA, T. & KUBO, T. 2003. Tree species differentiation in growth, recruitment and allometry in relation to maximum height in a Bornean mixed dipterocarp forest. Journal of Ecology 91:797?806. KOIKE, F. & HOTTA, M. 1996. Foliage-canopy structure and height distribution of woody species in climax forests. Journal of Plant Research 109:53?60. KOIKE, F. & SYAHBUDDIN 1993. Canopy structure of a tropical rain forest and thenatureof anunstrati?edupper layer.FunctionalEcology 7:230?235. KU?CHLER, A. W. 1964. Potential natural vegetation of the Coterminous United States: manual to accompany the map. American Geographical Society, New York. 156 pp. KUULUVAINEN, T. 1992. Tree architectures adapted to ef?cient light utilization: is there a basis for latitudinal gradients? Oikos 65:275? 284. LEI, T. T. & LECHOWICZ, M. J. 1990. Shade adaptation and shade tolerance in saplings of three Acer species from eastern North America. Oecologia 84:224?228. LEIGH, E. G. 1999. Tropical forest ecology: a view from Barro Colorado Island. Oxford University Press, Oxford. 245 pp. Adult tree size in tropical vs. temperate forest 21 LEVEY, D. J. 1988. Tropical wet forest treefall gaps and distributions of understory birds and plants. Ecology 69:1076?1089. MARTIN, W. H. 1975. The Lilley Cornett Woods: a stable mixed mesophytic forest in Kentucky. Botanical Gazette 136:171?183. MCCLAIN, W. E., LARIMORE, R. L. & EBINGER, J. E. 2001. Woody vegetation survey of Beall Woods Nature Preserve, Wabash, County, Illinois. Proceedings of the Indiana Academy of Science 110:41?50. MCNAB, W. H. 1989. Terrain shape index: quantifying the effect of minor landforms on tree height. Forest Science 35:91?104. MULLER,R.N.1982.Vegetationpatterns in themixedmesophytic forest of Eastern Kentucky. Ecology 63:1901?1917. PACALA, S. W., CANHAM, C. D., SAPONARA, J., SILANDER, J. A., KOBE, R. K. & RIBBENS, E. 1996. Forest models de?ned by ?eld measurements: estimation, error analysis and dynamics. Ecological Monographs 66:1?43. PARKER, G. G. & BROWN, M. J. 2000. Forest canopy strati?cation ? is it useful? American Naturalist 155:473?484. PITMAN, N. C. A., TERBORGH, J. W., SILMAN, M. R., NUNEZ, V. P., NEILL, D. A., CERON, C. E., PALACIOS, W. A. & AULESTIA, M. 2002.Acomparisonof tree species diversity in twoupperAmazonian forests. Ecology 83:3210?3224. POORTER, L., BONGERS, F., STERCK, F. J. & HANNSJO?RG, W. 2003. Architecture of 53 rain forest tree species differing in adult stature and shade tolerance. Ecology 84:602?608. POPMA, J., BONGERS, F. & DEL CASTILLO, J. M. 1988. Patterns in the vertical structure of the tropical lowland rain forest of Los Tuxtlas, Mexico. Vegetatio 74:81?91. PUTZ, F. E. 1984.Thenatural historyof lianas onBarroColorado Island, Panama. Ecology 65:1713?1724. RICHARDS, P. W. 1952. The tropical rain forest: an ecological study. (First edition). Cambridge University Press, Cambridge. 450 pp. RICHARDS, P. W. 1996. The tropical rain forest: an ecological study. (Second edition). Cambridge University Press, Cambridge. 575 pp. SCHMELTZ, D. V., BARTON, J. D. & LINDSEY, A. A. 1974. Donaldson?s Woods: two decades of change. Proceedings of the Indiana Academy of Science 84:234?242. SMITH, A. P. 1973. Strati?cation of temperate and tropical forests. American Naturalist 671?683. TERBORGH, J. 1985. The vertical component of plant species diversity in temperate and tropical forests. American Naturalist 126:760?776. THOMAS, S. C. 1996. Asymptotic height as a predictor of growth and allometric characteristics in Malaysian rain forest trees. American Journal of Botany 83:556?566. THOMAS,S.C.2003.Comparativebiologyof tropical trees:aperspective from Pasoh. Pp. 171?194 in Okuda, T., Manokaran, N., Matsumoto, Y., Niiyama, K., Thomas, S. C. & Ashton, P. S. (eds). Pasoh: ecology of a lowland rain forest in Southeast Asia. Springer, London. 628 pp. THOMAS, S. C. & BAZZAZ, F. A. 1999. Asymptotic height as a predictor of photosynthetic characteristics in Malaysian rain forest trees. Ecology 80:1607?1622. TURNER, I. M. 2001. The ecology of trees in the tropical rain forest. Cambridge University Press, Cambridge. 298 pp. UNITED STATES GEOLOGICAL SURVEY 1970. The national atlas of the United States of America. U.S. Dept. of the Interior Geological Survey, Washington, DC. 417 pp. VALENCIA,R., CONDIT,R., ROMOLEROUX,K., FOSTER,R. B.,MUNOZ, G. V., LOSOS, E. C., BALSLEV, H., SVENNING, J.-C. & MAGARD, E. 2004. Tree species diversity and distribution in a plot at Yasuni National Park, Amazonian Equador. Pp. 107?118 in Losos, E. C. & Leigh, E. G. (eds). Tropical forest diversity and dynamism: ?ndings from a large-scale plot network. University of Chicago Press, Chicago. 645 pp. VALLADARES, F., WRIGHT, S. J., LASSO, E., KITAJIMA, K. & PEARCY, R. W. 2000. Plastic phenotypic response to light of 16 congeneric shrubs from a Panamanian rainforest. Ecology 81:1925? 1936. VAN PELT, R. & FRANKLIN, J. F. 2000. In?uence of canopy structure on the understory environment in tall, old-growth, conifer forests. Canadian Journal of Forest Research 30:1231?1245. WELDEN, C. W., HEWETT, S. W., HUBBELL, S. P. & FOSTER, R. B. 1991. Sapling survival, growth, and recruitment: relationship to canopy height in a neotropical forest. Ecology 72:35?50. WHITE, P. S. & WHITE, R. D. 1996. Old-growth oak and oak-hickory forests. Pp. 178?198 in Davis, M. B. (ed.). Eastern old-growth forests: prospects for rediscovery and recovery. Island Press, Washington DC. WHITMORE, T. C. 1998. An introduction to tropical rain forests. (Second edition). Oxford University Press, Oxford. 282 pp. WRIGHT, S. J., MULLER-LANDAU, H. C., CONDIT, R. & HUBBELL, S. P. 2003. Gap-dependent recruitment, realized vital rates, and size distributions of tropical trees. Ecology 84:3174?3185. 22 DAVID A. KING, S. JOSEPH WRIGHT AND JOSEPH H. CONNELL Appendix 1. Maximum heights (m) of species of temperate deciduous forests of theUSA. Locations are: 1.Mohawk Trail State Forest,Massachussetts, 2. Brown County, Indiana, 3. Donaldson?s Woods, Indiana, 4. Beall Woods, Illinois (?oodplain section), 5. Lilley Cornett Woods, Kentucky, 6. Big Oak Woods, North Carolina and 7. selected cove forests of Great Smoky Mountains National Park, Tennessee. Maximum heights are means of the one to three tallest trees encountered in limited search times (see Methods). hmax (m) for site no. Species Family 1 2 3 4 5 6 7 Acer negundo L. Aceraceae 24.5 Acer pensylvanicum L. Aceraceae 16.85 17.9 Acer rubrum L. Aceraceae 34.2 34 35 27.5 37.2 Acer sacharinum L. Aceraceae 34.8 Acer saccharum Marsh. Aceraceae 35.6 31.8 32 34 36 Acer spicatum Lam. Aceraceae 8.5 9.1 Aesculus ?ava Ait. Hippocastanaceae 39.7 Asimina triloba (L.) Dunal Annonaceae 11.7 11.8 11 11.2 Betula alleghaniensis Britton Betulaceae 29.4 29 Betula lenta L. Betulaceae 31.2 28.2 32.6 Betula papyrifera Marsh. Betulaceae 28.7 Carpinus caroliniana Walt. Betulaceae 9.1 9.2 10.7 Carya cordiformis (Wangenh.) K. Koch Juglandaceae 35.5 31 33 35.5 39 Carya glabra (Mill.) Sweet Juglandaceae 36 Carya illinoensis (Wangenh.) K. Koch Juglandaceae 36.2 Carya lacinosa (Michx. f.) Loud. Juglandaceae 33.6 Carya ovata (Mill.) K. Koch Juglandaceae 33 35.2 38 C. ovata and/or Carya carolinae-septentrionalis Juglandaceae 35.7 (Ashe) Engl. & Graeb. Carya tomentosa (Poir.) Nutt. Juglandaceae 39.8 Celtis laevigata Willd. Ulmaceae 28 Celtis occidentalis L. Ulmaceae 31.3 Cercis canadensis L. Fabaceae 13.3 11.2 Cladrastis kentukea (Dum.-Cours.) Rudd Fabaceae 25.2 Cornus alternifolia L. Cornaceae 6.8 Cornus ?orida L. Cornaceae 11.5 11.4 11.4 9.4 Fagus grandifolia Ehrh. Fagaceae 35.6 33.7 35.5 33 34.3 Fraxinus americana L. Oleaceae 40.6 37 36.2 33.5 36.5 37.4 Fraxinus pennsylvanica Marsh. Oleaceae 36.5 Gymnocladus dioicus (L.) K. Koch Fabaceae 34.4 Halesia carolina L. Styracaceae 32.6 Hamamelis virginiana L. Hamamelidaceae 5.7 4.8 6.9 Ilex decidua Walt. Aquifoliaceae 6.6 Juniperus virginianab L. Cupressaceae 18.5 Juglans nigra L. Juglandaceae 35 34 34.2 36.5 Lindera benzoin (L.) Blume Lauraceae 4.4 4.4 4.6 Liquidambar styraci?ua L. Hamamelidaceae 38.5 34.5 Liriodendron tulipifera L. Magnoliaceae 40 40 45.1 44.8 Magnolia acuminata L. Magnoliaceae 33.5 36.4 Magnolia fraseri Walt. Magnoliaceae 33.6 Magnolia macrophylla Michx. Magnoliaceae 26.2 Magnolia tripetala L. Magnoliaceae 18.7 Nyssa sylvatica Marsh. Nyssaceae 30 34 Ostrya virginiana (Mill.) K. Koch Betulaceae 19 16 14.4 Oxydendrum arboreum (L.) DC. Ericaceae 24 Pinus strobusb L. Pinaceae 46.5 Pinus taedab L. Pinaceae 36 Platanus occidentalis L. Platanaceae 39.4 Populus deltoides Bartr. ex Marsh. Salicaceae 39.5 Prunus serotina Ehrh. Rosaceae 34.2 37 38 Quercus alba L. Fagaceae 34.5 35.4 39 32.8 Quercus falcata var. pagodifolia Ell. Fagaceae 35.8 Quercus macrocarpa Michx. Fagaceae 37.1 Quercus michauxii Nutt. Fagaceae 35.5 Quercus phellos L. Fagaceae 36.7 Quercus prinus L. Fagaceae 33.7 Quercus rubra L. Fagaceae 37.5 34.8 34.7 38 38 Quercus shumardii Buckl. Fagaceae 38.3 Adult tree size in tropical vs. temperate forest 23 Appendix 1. Continued. hmax (m) for site no. Species Family 1 2 3 4 5 6 7 Quercus velutina Lam. Fagaceae 30 33 33.5 Rhododendron maximuma L. Ericaceae 5.1 6.5 Robinia pseudoacacia L. Fabaceae 36 Sassafras albidum (Nutt.) Nees Lauraceae 29.2 Tilia americana L. Tiliaceae 34 32 Tilia heterophylla Vent. Tiliaceae 35.6 35.2 Tsuga canadensisb (L.) Carr. Pinaceae 34 38 42.2 Ulmus alata Michx. Ulmaceae 25 Ulmus americana L. Ulmaceae 28 Ulmus rubra Muhl. Ulmaceae 31.8 32.8 Viburnum alnifolium L. Caprifoliaceae 4.2 Viburnum prunifolium L. Caprifoliaceae 4.5 a Evergreen angiosperm. b Evergreen conifer. Appendix 2. Maximum heights of species in the plot of Connell et al. (1984) in Lamington, National Park, Queensland, Australia. Species Family hmax (m) Actephila lindleyi (Steudel) Airy Shaw Euphorbiaceae 9 Acronychia pubescens (F. Muell.) C. T. White Rutaceae 18 Acronychia suberosa C. T. White Rutaceae 21.6 Argyrodendron actinophyllum (F. M. Bailey) Edlin Sterculiaceae 33.2 Argyrodendron trifoliolatum F. Muell. Sterculiaceae 37 Baloghia inophylla (Forst. F.) P. S. Green Euphorbiaceae 24 Baurella simplicifolia (Endl.) T. Hartley Rutaceae 23.5 Caldcluvia paniculosa (F. Muell.) Hoogland Cunoniaceae 28.2 Clerodendrum ?oribundum R. Br. Verbenaceae 13 Cinnamomum virens R. T. Baker Lauraceae 31.5 Denhamia pittosporoides F. Muell. Celastraceae 16 Diospyros pentamera Woods & F. Muell. ex F. Muell. Ebenaceae 27.8 Diploglottis cunninghamii Hook. F. Sapindaceae 28.5 Doryphora sassafras Endl. Monimiaceae 31 Ellatostachys nervosa (F. Muell.) Radlk. Sapindaceae 23 Emmenosperma alphitonioides F. Muell. Rhamnaceae 34.1 Euodia micrococca F. Muell. Rutaceae 24.8 Eupomatia laurina R. Br. Eupomatiaceae 10.5 Ficus watkinsiana F. M. Bailey Moraceae 40 Geissois benthamii F. Muell. Cunoniaceae 29 Halfordia kendack (Montr.) Guillaumin Rutaceae 26.5 Litsea reticulata (Meisn.) F. Muell. Lauraceae 31 Melicope octandra (F. Muell.) Druce Rutaceae 30.8 Mischocarpus pyriformis (F. Muell.) Radlk. Sapindaceae 19 Orites excelsa R. Br. Proteaceae 28.8 Polyscias elegans (C. Moore & F. Muell.) Harms Araliaceae 25.7 Premna lignum-vitae (Cunn. ex Schauer) Pieper Verbenaceae 32.5 Pseudoweinmannia lachnocarpa (F. Muell.) Endl. Cunoniaceae 35.2 Psychotria simmondsiana F. M. Bailey Rubiaceae 4.9 Quintinia sieberi A. DC. Escalloniaceae 16.2 Randia benthamiana F. Muell. Rubiaceae 14.8 Sarcopteryx stipitata (F. Muell.) Radlk. Sapindaceae 25 Synoum glandulosum (Sm.) A. Juss. Meliaceae 17.5 Syzygium crebrenerve (C. T. White) L. Johnson Myrtaceae 31 Tasmannia insipida R. Br. ex DC. Winteraceae 6.2 Wilkiea austroqueenslandica Domi. Monimiaceae 9.4 Wilkiea huegeliana (Tul.) A. DC. Monimiaceae 5.4 24 DAVID A. KING, S. JOSEPH WRIGHT AND JOSEPH H. CONNELL Appendix 3. Maximum heights of species in the 38.4 ha plot on the Gigante peninsula of the Barro Colorado Nature Monument, Panama. Nomenclature of D?Arcy (1987), as updated by Condit et al. (1996). Species Family hmax (m) Species Family hmax (m) Alseis blackiana Rubiaceae 30.2 Miconia argentea Melastomaceae 19.1 Amaioua corymbosa Rubiaceae 17.8 Miconia minuti?ora Melastomaceae 25.8 Apeiba aspera Tiliaceae 32.7 Mouriri myrtilloides Melastomaceae 8.8 Ardisia fendleri Myrsinaceae 13.0 Myrcia sp. Myrtaceae 9.6 Aspidospermum cruentum Apocynaceae 39.9 Myrcia gatunensis Myrtaceae 13.9 Beilschmiedia pendula Lauraceae 30.1 Myrcia zetekeniana Myrtaceae 4.5 Brosimum alicastrum Moraceae 35.0 Nectandra purpurea Lauraceae 21.4 Brosimum guianensis Moraceae 29.6 Ouratea lucens Ochnaceae 8.4 Calophyllum longifolium Clusiaceae 34.8 Oxandra panamensis Annonaceae 17.6 Cassipourea elliptica Rhizophoraceae 22.8 Perebea xanthochyma Moraceae 17.0 Copaifera aromatica Fabaceae 35.0 Phoebe cinnamomifolia Lauraceae 24.7 Cordia bicolor Boraginaceae 25.5 Pachira sessilis Bombacaceae 38.1 Cordia lasiocalyx Boraginaceae 12.0 Poulsenia armata Moraceae 26.6 Coussarea curvigemmia Rubiaceae 7.3 Pourouma bicolor Cecropiaceae 25.5 Couratari guianensis Lecythidaceae 31.5 Pouteria reticulata Sapotaceae 29.5 Desmopsis panamensis Annonaceae 9.0 Prioria copaifera Fabaceae 41.5 Dialium guianense Fabaceae 30.2 Protium sp. Burseraceae 17.5 Diospyros artanthifolia Ebenaceae 20.3 Protium panamense Burseraceae 23.1 Dipteryx panamensis Fabacaceae 43.4 Protium correae Burseraceae 15.3 Drypetes standleyi Euphorbiaceae 22.4 Protium tenuifolium Burseraceae 25.5 Eugenia coloradoensis Myrtaceae 21.9 Quassia amara Simaroubaceae 11.9 Faramea luteovirens Rubiaceae 5.9 Rinorea crenata Violaceae 8.2 Faramea occidentalis Rubiaceae 16.1 Rinorea squamata Violaceae 12.1 Garcinia intermedia Clusiaceae 20.5 Rinorea sylvatica Violaceae 10.7 Garcinia madruno Clusiaceae 20.1 Simarouba amara Simaroubaceae 27.8 Guatteria dumetorum Annonaceae 33.3 Sloanea zulianensis Elaeocarpaceae 18.3 Heisteria acuminata Olacaceae 12.5 Sorocea af?nis Moraceae 8.9 Heisteria concinna Olacaceae 20.4 Sterculia recordiana Sterculiaceae 26.6 Hirtella americana Chrysobalanaceae 20.9 Swartzia panamensis Fabaceae 21.5 Hirtella racemosa Chrysobalanaceae 7.8 Swartzia simplex var. ochnacea Fabacaceae 8.7 Hirtella triandra Chrysobalanaceae 23.4 Tabebuia guayacan Bignoniaceae 39.4 Hybanthus prunifolius Violaceae 6.3 Tachigali versicolor Fabaceae 33.7 Inga cocleensis Fabaceae 21.9 Talisia nervosa Sapindaceae 9.6 Inga umbellifera Fabaceae 12.5 Tetragastris panamensis Burseraceae 33.5 Jacaranda copaia Bignoniaceae 31.9 Tovomita stylosa Clusiaceae 10.4 Lacistema aggregatum Flacourtiaceae 9.0 Trattinnickia aspera Burseraceae 30.1 Lacmellea panamensis Apocynaceae 23.4 Trichilia tuberculata Meliaceae 24.9 Laetia procera Flacourtiaceae 27.3 Vatairea lundellii Fabacaceae 41.2 Laetia thamnia Flacourtiaceae 14.7 Virola sebifera Myristicaceae 27.8 Licania hypoleuca Chrysobalanaceae 22.0 Virola multi?ora Myristicaceae 34.1 Lonchocarpus latifolius Fabaceae 32.0 Virola surinamensis Myristicaceae 37.0 Mabea occidentalis Euphorbiaceae 13.3 Xylopia macrantha Annonaceae 20.1 Mosannona garwoodii Annonaceae 15.1