BIOTROPICA 36(4): 447-473 2004 Why Do Some Tropical Forests Have So Many Species of Trees?^ Egbert Giles Leigh Jr. Smithsonian Tropical Researcli Institute, Unit 0948, APO AA 34002-0948, U.S.A. Priya Davidar Salim AN School of Ecology and Environmental Science, Pondicherry University, Kalapet, 605014 Pondicherry, India Christopher W. Dicl< Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002-0948, U.S.A Jean-Philippe Puyravaud Dept. of Ecology and Evolution, State University of New York at Stony Brook, Stony Brook, New York 11794, U.S.A. John Terborgh Center for Tropical Conservation, Duke University, Box 90381, Durham, North Carolina 27708, U.S.A Hans ter Steege National Herbarium of the Netherlands, Utrecht University Branch, Heidelberglaan 2, 3584 CS Utrecht, The Netherlands and Stuart Joseph Wright Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002-0948, U.S.A. ABSTRACT Understanding why there are so many kinds of tropical trees requires learning, not only how tree species coexist, but what factors drive tree speciation and what governs a tree clade's diversification rate. Many report that hybrid sterility evolves very slowly between separated tree populations. If so, tree species rarely originate by splitting of large popu- lations. Instead, they begin with few trees. The few studies available suggest that reproductive isolation between plant populations usually results from selection driven by lowered fitness of hybrids: speciation is usually a response to a "niche opportunity." Using Hubbell's neutral theory of forest dynamics as a null hypothesis, we show that if new tree species begin as small populations, species that are now common must have spread more quickly than chance allows. Therefore, most tree species have some setting in which they can increase when rare. Trees face trade-offs in suitability for different microhabitats, difiFerent-sized clearings, different soils and climates, and resistance to different pests. These trade-offs underlie the mechanisms maintaining a-diversity and species turnover. Disturbance and microhabitat spe- cialization appear insufficient to maintain a-diversity of tropical trees, although they may maintain tree diversity north of Mexico or in northern Europe. Many studies show that where trees grow readily, tree diversity is higher and temperature and rainfall are less seasonal. The few data available suggest that pest pressure is higher, maintaining higher tree diversity, where winter is absent. Tree a-diversity is also higher in regions with more tree species, which tend to be larger, free for a longer time from major shifts of climate, or in the tropics, where there are more opportunities for local coexistence. RESUMEN Comprender por qu? hay tantos tipos de ?rboles tropicales, se requiere aprender no s?lo c?mo las especies de ?rboles coexisten, sino tambi?n, cu?les factores conducen a su especiaci?n, y qu? determina la velocidad de diversificaci?n de un ciado de ?rboles. Muchos reportan que la esterilidad h?brida evoluciona muy lentamente entre poblaciones separadas de ?rboles. De ser as?, las especies de ?rboles raramente se originar?an por la separaci?n de grandes poblaciones; m?s bien empezar?an con pocos ?rboles. Los pocos estudios disponibles sugieren que el aislamiento reproductivo entre las poblaciones vegetales usualmente resulta de selecci?n derivada del bajo ?xito de los h?bridos: la especiaci?n general- mente responde a una "oportunidad de nicho". Usando la teor?a neutral de Hubbell de din?mica de bosques como hip?tesis nula, nosotros mostramos que si las nuevas especies de ?rboles comienzan como poblaciones peque?as, ^ Received 29 March 2003; revision accepted 10 June, 2004. 447 448 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit especies que ahora son comunes deber?an haberse expandido m?s r?pido que lo que el azar permite. Por lo tanto, la mayor?a de las especies de ?rboles tendr?an alguna condici?n donde sus poblaciones podr?an crecer cuando son raras. Los ?rboles enfrentan compromisos en su adecuaci?n por diferentes microhabitats, claros de diferentes tama?os, diferentes suelos y climas, y resistencia a diferentes plagas. Estos compromisos sirven de base para los mecanismos que mantienen la diversidad a y al reemplazo espacial de especies. Los disturbios y la especializaci?n de microhabitats parecen ser insuficiente para mantener la diversidad o: de ?rboles tropicales, sin embargo ellos pueden mantener diversidad de ?rboles al norte de M?xico o en Europa del norte. Muchos estudios muestran que en lugares donde los ?rboles crecen f?cilmente, la diversidad de ?rboles es mayor donde la temperatura y la lluvia son menos estacionales. Los pocos estudios disponibles sugieren que la presi?n de las plagas es mayor, manteniendo as? la diversidad de ?rboles en lugares donde no hay invierno. La diversidad a. de ?rboles tambi?n es m?s alta en regiones con m?s especies de ?rboles, las cuales tienden a ser m?s largas, exentas por un largo periodo de tiempo de grandes cambios clim?ticos, o en los tr?picos donde hay m?s oportunidades de coexistir localmente. Key words: a-diversity; ?-diversity; Janzen?Connell; neutral theory; pest pressure; regional diversity; stability-time; sped- ation. WHY ARE THERE SO MANY KINDS OF TROPICAL TREES? In particular, why do some tropical hectares con- tain over 250 species of trees 10 cm DBH or great- er, while in the eastern United States, a hectare rarely contains more than 20 species (Leigh 1999)? How can 0.5 km^ of rain forest in Borneo or Ama- zonia contain as many tree species as the 4.2 mil- lion km^ of temperate zone forest in Europe, North America, and Asia combined (Wright 2002)? This question has attracted increasing attention ever since it was reopened by Dobzhansky (1950); yet, the explanations proposed for this phenomenon differ as much as ever (Huston 1994, Givnish 1999, Wright 2002, Hawkins et al. 2003). Some explanations focus exclusively on the ecological mechanisms that allow so many species to coexist locally, while others focus on latitudinal differences in rates of speciation and extinction, without show- ing how these species coexist. Moreover, new hy- potheses to explain the latitudinal gradient in tree diversity keep appearing (Willig et al. 2003). This paper grew out of a symposium with the same title at the 2002 meeting of the Association for Tropical Biology. The task of both the sym- posium and this paper was to set out what w^e know about the factors that promote diversity among tropical trees, and identify what we need to learn in order to understand why some tropical for- ests contain so many kinds of tropical trees. We focus on four aspects of our question: (1) What factors promote the sympatric coexistence of tree species?; (2) What factors govern the change in tree species composition (?-diversity) along a series of plots spanning a great expanse of forest such as Amazonia?; (3) How do different features of the environment influence the diversity that factors promoting coexistence can maintain?; and (4) What factors influence a biogeographic realm's tree diversity? We first argue that there must be factors that tend to stabilize species composition, even in trop- ical forests. Then, we consider possible factors and what we must learn to assess their relative impor- tance in maintaining tree diversity. TREE DIVERSITY REFLECTS DIFFERENCES AMONG TREE SPECIES To learn why there are so many kinds of tropical trees, we first ask whether, as a rule, tree species increase in abundance if they are temporarily made rarer. In other words, do tree species coexist be- cause one or more factors tend to stabilize species composition (Chesson 2000)? We begin our answer by using Hubbell's (1979, 1997, 2001) neutral theory of tropical forest dy- namics as a null hypothesis by which to judge whether widespread species began with an advan- tage over their competitors. Hubbell's theory as- sumes that what species a tree belongs to is irrele- vant to its prospects of mortality or reproduction. In Hubbell's theory, no process stabilizes species composition, and species diversity expresses a bal- ance between speciation and random extinction. Hubbell derived predictions from his theory in analogy with the neutral theory of population ge- netics for a multi-allelic locus (Ewens 1979) in a haploid population: the population geneticist's re- productive adults correspond to Hubbell's repro- ductive trees, and the geneticist's allelic types to Hubbell's species. Here, we use a prediction from the neutral the- ory of population genetics that follows equally from Hubbell's theory, to judge whether wide- spread species began with an advantage over their fellows. Fisher (1930) showed that for a neutral al?ele in a large population, all copies of which de- scend from a single mutant, the number of adults now carrying this al?ele cannot greatly exceed the Why Are There So Many Kinds of Tropical Trees? 449 number of generations since this mutant occurred. Indeed, the average time required for an al?ele orig- inally represented by y copies in a huge population to spread by chance alone until there are n ? y copies is likewise n generations or more (Kimura & Ohta 1973, equations 13 and 17). Fisher's pre- diction applies to trees if no one adult leaves many mature offspring, and is irrelevant if a few trees have multitudes of successful young and most have none at all. Many species of tropical trees range from Cen- tral America through all Amazonia as far as Bolivia, each represented by trees 10 cm DBH or greater in a third or more of this region's 1 ha rain forest plots. Such species must each be represented by 10 million or more mature trees. For most of these species, a generation is over 50 years (Gentry 1989: 123). If as Willis (1922) and Stebbins (1982) have suggested, these species began with very small pop- ulations, the neutral theory implies that they ap- peared well before the origin of angiosperms 140 million years ago. This cannot be true. The tree species Symphonia glohulifera, for example, reached the Neotropics from Africa in three separate trans- atlantic dispersal events, all less than 25 million years ago. This species now ranges from Belize and Dominica all through Amazonia as far as Bolivia (Dick et al. 2003). The simplest explanation why they have spread so quickly and widely is that such species began with some advantage: they must have exploited a "niche opportunity" (Shea & Chesson 2002) by using some resource or resisting herbi- vores and pathogens more effectively or economi- cally than their competitors. Although tree species can become common and Wdespread only through an advantage over their competitors, no tree species replaces all of its com- petitors?indeed, in most tropical forests, no spe- cies even comes close to doing so. Symphonia glo- hulifera, for example, has spread throughout nearly all Neotropical forests. This species no^v grow^s on good soils of Western Amazonia and central Pan- ama and on poor soils in Central and Eastern Ama- zonia, in everwet forests and forests with severe dry seasons (Dick et al. 2003), but in almost every for- est it accounts for less than 2 percent of the trees 10 cm DBH or greater. Thus rarer, less widespread species also must exploit niche opportunities (less lucrative, to be sure, than those exploited by com- mon species) that have protected them from being replaced by rapidly spreading competitors. There- fore, most tree species must have some setting in w^hich they increase when rare. This circumstance stabilizes tree species composition sufficiently to al- low many species to coexist, but it does not ensure each species an equilibrium abundance (Chesson 2000). These are indeed sweeping conclusions to draw from a single test of Hubbell's null model. After all, to make definite predictions about forests with trees that face no trade-offs, this model assumes, along Avith many other simplifications, that species always begin from small populations. Later in this paper, w^e propose and apply more direct tests of the proposition that most new species form in re- sponse to niche opportunities. Meanwhile, this first test of the null model will guide our next steps. CLASSIFYING AND ORDERING EXPLANATIONS FOR TREE DIVERSITY Abstractly speaking, niche opportunities and the diversity they favor reflect trade-offs. No one spe- cies can do everything w^ell; the inevitability of trade-offs allow^s different species to coexist (Mac- Arthur 1961). We now ask, what are the relevant trade-offs? What circumstances influence the diver- sity that different trade-offs maintain? A plethora of factors can influence biotic di- versity (Pianka 1994). To understand tree diversity, w^e must ask questions on at least two scales (Rick- lefs & Schl?ter 1993). First, what governs the tree diversity of a biogeographic region, the number of species available for stocking its various habitats? A region's diversity represents a balance between ex- tinction, and speciation -I- immigration (MacAr- thur Sc Wilson 1967, Terborgh 1973), so we must reckon with mechanisms of speciation in trees and Wth how this region's climate and topography have varied in the geologic past. Second, what factors control local (O?) diversity of trees and the turnover of tree species from one locale to another within the region (?-diversity)? Here, we deal with the properties or circumstances that allow different spe- cies to coexist. The two scales of diversity are close- ly related. Local diversity tends to be higher in re- gions with higher total diversity (Ricklefs 2004), w^hereas a species survives only if it is superior to its competitors in some respect, in some setting (Chesson 2000). In this paper, we start from the bottom up. We begin with the factors that main- tain local diversity because no species can enter the regional pool unless it spreads initially in some lo- cale by exploiting a local niche opportunity. To evaluate the relative impact of the various factors influencing species diversity, we order our questions as follows: 450 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit ( 1 ) What factors favor the sympatric coexistence of tree species? (A) DISTURBANCE.?The fall of individual trees, and the opening of larger clearings by windthrows, allow light-demanding and shade-tolerant tree spe- cies to coexist (Skellam 1951); in some regions at least, occasional windstorms may be needed to keep a slow-growing, competitive "supertree" from tak- ing over the forest (Connell 1978, ter Steege & Hammond 2001). (B) SPECIALIZATION.?Specialization to various aspects of a complex habitat can allow different species to coexist. Different tree species may coexist because they specialize to different strata of the for- est (Terborgh 1985), because their seedlings spe- cialize to different microsites on the forest floor (Molofsky & Augspurger 1992), or perhaps be- cause year-to-year variation in climate and abun- dances of different types of pollinators, seed-dis- persers, and seedling-browsers cause temporal sort- ing in their recruitment (Chesson & Warner 1981). (C) PEST PRESSURE.?The presence of a diversity of pests and pathogens able to penetrate different types of plant defense and capable of preventing any one species from taking over the forest can also allow different species to coexist (Gillett 1962). (2) What factors influence species turnover? (A) LIMITED DISPERSAL.?Different species evolve in different places and spread only a limited dis- tance from their points of origin (Willis 1922, Condit et al. 2002). (B) SPATIAL HETEROGENEITY.?Different species specialize to different habitats within a region? different elevations along a mountainside (Brown 1919, Grubb et al. 1963) or different edaphic con- ditions such as floodplain versus uplands or poor versus better soil (Richards 1952, Phillips et al. 2003). (3) How do different features of the environment influence the diversity that these mechanisms of coexistence can maintain? (A) PRODUCTIVITY AND CLIMATE.?How do cli- mate, soil quality, and ecosystem productivity in- fluence tree diversity (Ricklefs 1977, D. H. Wright 1983, Terborgh 1985)? (B) STABILITY OF PRODUCTIVITY AND CLIMATE.? How does winter or a severe dry season affect the ability of tree species to specialize to different as- pects of their habitat (Janzen 1967) or the ability of specialized pests and pathogens to maintain tree diversity (Janzen 1970)? (C) DIVERSITY AND BIOGEOGRAPHY.?How much is local tree diversity influenced by the tree diversity in its biogeographic region (Ricklefs 2004)? (4) What factors influence a biogeographic region's tree diversity? (A) MECHANISMS.?^What are the mechanisms of speciation among trees? Do new species usually form in response to niche opportunities? (B) SPECIATION/EXTINCTION BALANCE.?How is a region's tree diversity influenced by the time avail- able for its trees to diversify in response to currently prevailing environmental conditions (Fischer 1960, Morley 2000)? More generally, how does specia- tion/extinction balance vary in different regions, and why? (C) MUSEUM OR CRADLE?.?^Are the tropics pri- marily a museum of tree diversity or a cradle of tree speciation? We now proceed to discuss these questions in the order that they were outlined. FACTORS MAINTAINING LOCAL (a) DIVERSITY AMONG TREES Diversity reflects the inability of any one species to do all things well: enhancing one ability usually entails diminishing others. Organisms therefore face trade-offs that allow different species to coexist by exploiting different niche opportunities. What trade-offs play a role in the coexistence of tree spe- cies? Most trade-offs involve more than tw^o char- acteristics. The relationship between any two of these characteristics will show scatter created by the variation of other characteristics involved in the trade-off. If, for example, birds face a triple trade- off among producing many young, avoiding pred- ators, and competing effectively with other guild members (Cody 1966), a tw^o-factor plot of clutch size against success in avoiding predators will show scatter created by differences in competitive ability. The tw^o-factor trade-offs we discuss are all influ- Why Are There So Many Kinds of Tropical Trees? 451 enced by other unspecified factors, but they still contribute to the maintenance of tree diversity. (A) DISTURBANCE AS A GENERATOR OF NICHE OP- PORTUNITIES.?Skellam (1951) showed that a "pi- oneer" tree species can coexist with a mature forest species even if a mature forest seedling under a pioneer's crown can always grow up to shade and replace that pioneer. To do so, pioneers must col- onize light gaps opened by the fall of mature forest trees quickly enough, grow fast enough, and ma- ture soon enough that they can produce enough young to replace themselves before their inevitable displacement. Here, w^e show what trade-offs allow light-demanding species to coexist with shade-tol- erant species. Then we show that, even though tree species are distributed along a continuum from light-demanding to shade-tolerant, there is reason to believe that adaptation to different levels of dis- turbance is not what allows hundreds of tree species to coexist in the tropics. The trade-off between competitive ability and the ability to disperse seeds quickly to clearings was long considered to be the primary factor allowing the coexistence of pioneers with mature forest spe- cies (Grubb 1977). Indeed, Tilman (1994) argued that the competition?colonization trade-off can al- low an indefinite number of species to coexist if these species can be ranked so that any species j is competitively inferior to, but a better colonizer than, species y ? 1, and if certain other conditions are met. Tilman's result is invalid for two reasons. First, it applies only if a seedling of species j under the crown of a competitively inferior species will replace that inferior equally quickly whatever that inferior's competitive rank. If, however, a seedling under a tree of a competitively inferior species re- places it more slowly when the competitive abilities of the tw^o species are closer, as seems biologically reasonable, this trade-off allow^s only a few species to coexist (Adler & Mosquera 2000). Second, ef- fective seed dispersal is not the primary factor that enables a species to coexist with superior compet- itors. For example, seedlings of the tree species that dominate the first few centuries of succession on a growing floodplain in western Amazonia all grow together on this floodplain's new beachfront; dif- ferential growth, not differential colonization, or- ders these stages of succession (Foster et al. 1986). Even a large, multi-tree windthrow clearing is not a blank slate; pioneers dominate large clearings by outgrowing the other plants present, not by getting there first. Pioneers must be good colonizers, to be sure, but they must also grow fast. The primary trade-off allowing pioneer and mature forest tree species to coexist is that between fast growth of saplings in bright light and survival in shade. On Barro Colorado Island, Panama, sap- ling growth rates in large clearings of three pioneer tree species are inversely related to the minimum size of clearing that will allow new saplings to sur- vive for at least nine years (Brokaw 1987). In gen- eral, the more light-demanding canopy tree species on Barro Colorado have shorter-lived and more poorly defended leaves and fewer saplings per adult; their saplings suffer higher mortality and grow faster when they are well lit, and a higher proportion of their germinating seedlings occur in gaps. More shade-tolerant species have more seed- lings that germinate in shade, slower-growing sap- lings with denser wood, lower mortality, longer- lived, better defended leaves, and more saplings per adult (Coley et al. 1985; Coley 1987, 1988; King 1994; Wright et al. 2003). Judging by these crite- ria, the canopy tree species of Barro Colorado Is- land's 50 ha Forest Dynamics Plot are distributed along a continuum from light-demanding to shade- tolerant (Wright et al 2003). Simulations suggest that the trade-off betw^een fast sapling growth in bright light and survival in shade allows six tree species to coexist in a Con- necticut forest if they are subject to a regime of "wide-scale disturbance typical of natural stands" (Pacala et al 1996: 36). On the other hand, fre- quent gaps of many sizes do not appear to be es- sential for high tree diversity. Tree diversity in the 50 ha Forest Dynamics Plot in Pasoh Reserve, Ma- laysia, is nearly three times higher and its most common species much rarer than on Barro Colo- rado's 50 ha plot. This disparity in diversity occurs even though at Pasoh, tree cro\vns are narro\ver and trees usually die standing (Putz & Appanah 1987); so gaps are smaller and less frequent and pioneer trees far less common and diverse at Pasoh than on Barro Colorado (Condit et al 1999). Sixteen of the 141 canopy tree species on Barro Colorado's plot are pioneers with saplings' diameter growth averaging more than 4 mm/yr, compared to 3 of the 422 canopy tree species on the Pasoh plot (Condit et al 1999). Within Sarawak, mixed dip- terocarp forest on fertile soil is less diverse than dipterocarp forest on poorer soil (Ashton 1989), even though gaps are larger and more frequent among dipterocarps on more fertile soil (van Schaik & Mirmanto 1985, Ashton & Hall 1992). Here, too, a-diversity is greater where there are fewer, smaller, gaps. In some regions, such as the Budongo forest in 452 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit Uganda, or in Guyana, occasional wide-scale dis- turbance may be needed to keep a slow-growing, dense-w^ooded, competitive "supertree" from taking over the forest (Eggeling 1947, Connell 1978, ter Steege &C Hammond 2001). These takeovers, how- ever, are rarely complete. In the Congo basin's Ituri Forest, large tracts are dominated by the canopy tree species Gilbertiodendron dewevrei, where more than 70 percent of the trees 30 cm DBH or greater belong to this species. Yet, in a 10 ha plot of mo- nodominant forest, the trees 30 cm DBH or great- er that belong to other species contain as many species as an equivalent number of such trees on a mixed forest plot (Makana et al. 2004). Is the dom- inant "supertree" Gilbertiodendron an intruder in a forest where diversity among non-Gilbertiodendron tree species is maintained by factors other than dis- turbance? In short, although disturbance can contribute to the maintenance of tree diversity, there is no clear evidence that disturbance is the only factor allowing hundreds of tree species to coexist locally. Moreover, it appears that high tree diversity is maintained in some areas without frequent or widespread disturbance. To understand better what maintains tropical tree diversity, however, we must learn what properties allow a species such as Pen- taclethra macroloba (Lieberman et al. 1985) or G. dewevrei (Makana et al. 2004) to dominate an oth- erwise diverse forest. (B) SPECIALIZATION TO DIFFERENT MICROHABITATS.? A forest's trees create a complex habitat that offers many niche opportunities. The trees create a light gradient from the well-lit canopy to the understory; thus, different tree species can coexist if they spe- cialize to different strata of the forest (Terborgh 1985). In Borneo, even hemiepiphytic figs, which start as epiphytes and later drop roots to the forest floor, also specialize to different strata of the forest, even though their growth form is far rarer than trees (Shanahan & Compton 2001, Harrison et al. 2003). Trees also create a mosaic of microsites of different types on the forest floor, each of which favors germination and establishment of a different set of tree species (Grubb 1977). How much tree diversity can the complexity of a forest habitat maintain? To what extent is a tree's regeneration niche governed by its way of life as an adult, that is to say, the adult's size, successional status, or po- sition in the canopy? Do regeneration niches pro- vide different, additional mechanisms of coexis- tence, or are regeneration niches governed by adult niches (Nakashizuka 2001)? Specialization to forest stratum is driven in part by the trade-off between fast sapling growth in bright light and survival in shade, the same trade- off that allows coexistence between pioneer and shade-tolerant tree species. In Malaysia's Pasoh Re- serve, leaves of understory species have lower pho- tosynthetic capacity per unit area, or per gram of nitrogen, than leaves of young mid-story congeners in the same light environment, which in turn have lower photosynthetic capacity than leaves of young canopy congeners in that light environment (Thomas & Bazzaz 1999). In Borneo, diameter growth decelerates faster once diameter exceeds 11 cm in trees of understory species than in their can- opy counterparts, but understory species recruit more saplings 6 cm DBH or greater per unit basal area of conspecific adults than do their canopy counterparts (Kohyama et al. 2003), presumably because the more shade-tolerant seedlings and sap- lings of understory species survive much better than do their more light-demanding canopy coun- terparts (^cf. Wright et al. 2003). A species' degree of shade tolerance, however, is not the only factor influencing its stratum in the forest. Shade-tolerant and light-demanding species coexist in forest can- opies (Grubb 1977). Indeed, the 73 most common species of canopy tree on Barro Colorado's 50 ha plot span the spectrum from extremely light-de- manding to extremely shade-tolerant (Wright et al. 2003). Rain forests, however, have no more than five strata of trees (Terborgh 1985). Does the complex- ity of the forest habitat offer other opportunities for coexistence? Specialization to soil type contrib- utes greatly to species turnover (Richards 1952, Ashton 1964). Tilman (1982) proposed that nu- trient availability is heterogeneous at all scales, and that small-scale nutrient heterogeneity enhances tree a-diversity to the greatest degree on moderate- ly infertile soils. Ashton (1989) uses Tilman's (1982) theory to explain why the most diverse for- ests in Sarawak occur on moderately infertile soils. No one, however, has yet been able to associate seedlings of particular tree species on a "uniform" forest hectare with specific soil qualities or to pro- vide independent criteria for predicting the precise relationship between soil fertility and tree diversity. Moreover, given the prevalence of dispersal limita- tion (Harms et aL 2000, Dalling et al 2002), es- pecially in rare species (MuUer-Landau et al. 2002), is it reasonable to expect seedlings of particular spe- cies to establish on specific microsites? The spe- cialization of plants to microsites with particular Why Are There So Many Kinds of Tropical Trees? 453 soil qualities needs to be documented before we can assess its importance. Grubb (1977) considered that different tree species could coexist because their seedlings were favored by different microhabitats. Fallen logs, tip- up mounds from uprooted trees, and variation in thickness of litter and in amount of light reaching the ground offer a variety of different microsites, each favoring different species of seedlings. In southern Chile, litter thickness seems to be a mi- crosite's decisive attribute; small-seeded species fa- vor fallen logs and tip-up mounds because they are free of litter (Christie &C Armesto 2003). Seeds of some light-demanding pioneers persist in the soil and grow when exposed by an uprooted tree, an event that usually opens a large gap (Putz 1983), but in other cases, there is no obvious relationship between regeneration niche and way of life as an adult. At Yasuni in Amazonian Ecuador, small ju- veniles of Oenocarpus bataua are distributed inde- pendently of light level, while those of Iriartea del- toidea tend to occur in the better lit understory sites?yet most adult Oenocarpus occur in major canopy gaps, while the majority of 15 rn tall Iriar- tea are shaded by closed canopy (Svenning 2001: 14). Can different "regeneration niches" allow tree species with the same "adult niche" to coexist? Greenhouse experiments in central Panama suggest that in gaps, the large seeds of the successional tree Gustavia superha germinate and survive better un- der thick litter, whereas the small seeds of other successional species do better where litter is thin or absent (Molofsky & Augspurger 1992). If thick- and thin-litter microsites are close together in the same gap, one sapling will reproduce after cro'wd- ing out the others on these various microsites. On the other hand, if whole gaps differ in litter thick- ness, different successional species could coexist be- cause their seedlings are favored by different thick- nesses of litter. Different saplings also differ in their ability to survive damage by branches fallen from above (Guariguata 1998). If some sites escape branch-falls and others not, a more susceptible, faster-growing species might coexist with a slower- growing, more damage-resistant species of the same forest stratum. We have much to learn about how the complexity of forest habitats contributes to maintenance of tree diversity. Do the year-to-year fluctuations in climate and in the abundance of different pollinators, seed dis- persers, and seedling browsers create a temporally sorted array of regeneration niches that allow dif- ferent species to coexist because they recruit in dif- ferent years (Grubb 1977, Chesson & Warner 1981)? There is some temporal sorting in fruit pro- duction among a forest's tree species (Grubb 1977, Connell & Green 2000). Coexistence, however, de- mands temporal sorting in recruitment, not just reproduction; coexistence requires, for example, that each species occupy a disproportionately high fraction of the gaps that occur in those years when it fruits heavily. This can rarely be true. Seedlings of most mature forest tree species grow slowly (Hubbell 1998, Connell & Green 2000), so slowly that differences in their growth rates will prevent temporal sorting in reproduction from assuring temporal sorting in tree recruitment. Kelly and Bowler (2002) argued that in dry forest at Chamela, Mexico, coexistence within a ge- nus reflects temporal sorting in recruitment, where- by the faster-growing species only recruits in good years (with abundant rain or low herbivory?). Their evidence consists of five pairs of sympatric conge- ners in which the rarer, faster-growing congener has a less regular size distribution, presumably reflect- ing more pulsed recruitment (Kelly et al. 2001, Kelly ?? Bowler 2002). Their evidence, however, suggests that each pair of congeners coexists simply because one is more gap-demanding and the other more shade-tolerant; is a more subtle explanation needed? Their argument shares another problem \vith other explanations of tree diversity by habitat spe- cialization: Kelly and Bowler (2002) fail to explain how different genera coexist. The coexistence of 34 palm species at Yasuni (Svenning 1999, 2001), 27 species of hemiephytic figs at Lambir, Sarawak (Harrison et al. 2003), or 11 species of pioneer Macaranga at Lambir (Davies 1998, Davies et al. 1998) are all three explained in terms of differences in shade tolerance and different preferences for for- est stratum, soil quality, and topographic position. The figs, being hemiepiphytes, also differ according to what parts of a tree they colonize (Harrison et al. 2003). Otherwise, species in each group coexist by the same types of difference in habitat prefer- ence. Large, varied families such as Euphorbiaceae have each evolved species suitable for all tropical forest microhabitats. In the Solomon Islands, Cor- ner (1967: 32) describes a forest with canopy trees, understory trees, treelets, and lianas all belonging to the genus Ficus. Neither disturbance nor the complexity and heterogeneity of forest habitats ex- plains why 30 or more plant families coexist on a hectare of tropical forest. Does diversity of defenses against pests promote the coexistence of plant fam- ilies? 454 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit (C) PESTS, PATHOGENS, AND TREE DIVERSITY.? Pathogens and insect herbivores face trade-offs in what plants they can eat. In those tropical forests where neither winter nor prolonged dry season de- presses insect populations, pest pressure is intense (Ridley 1930, GiUett 1962, Janzen 1970, Connell 1971, Coley & Barone 1996). Young, tender leaves are easiest to eat (Coley 1983), so the defenses of young, expanding tropical leaves are particularly strong (Coley et al. 2003). Therefore, trade-offs be- tween eating young leaves of species with different defenses should be particularly acute, so that the most damaging pathogens and insect pests should be specialists. If specialist pests are the attackers, w^e expect: (1) that these pests will find and kill most quickly those young plants nearest conspecific adults, making space for plants of other species be- tween adults and their surviving young; and (2) as adult hosts become rarer, specialist pests are less likely to spread to other adults, and the adults' load of specialist pests should decrease, until this plant species can maintain its density. As a result of (1) and (2), wide-scale seedling survival in a species should decrease as the abundance of its adults in- creases, and seedlings of this species should survive less w^ell in patches where they are more abundant or closer to conspecific adults. All three proposi- tions are true for rain forest trees in Borneo (Webb & Peart 1999). Can pest pressure maintain the diversity of tropical trees? This can only be so if the most dam- aging tropical pests are specialists. In most forests, caterpillars (larvae of Lepidoptera) "consume more living leaves than all other animals combined" (Jan- zen 1988: 120). Tropical caterpillars appear to face a trade-off betw^een fast growth and a generalized diet (Janzen 1984, Bernays & Janzen 1988). In any one tropical forest, the most damaging caterpillars usually consume leaves of a particular species (Jan- zen 1988) or genus (Novotny et al. 2002) of plants. The same is true of seed-eating weevils (Janzen 1980, 1981). Specialized pests can maintain tree diversity by causing mutual repulsion among conspecifics. Some pests are known to do so. In Central Amer- ica, seedlings or saplings of Casearia corymhosa (Howe 1977), Platypodium elegans (Augspurger 1983), Dipteryxpanamensis (Clark & Clark 1985), Quararibea asterolepis (Sffon^ et al. 1990), and Oco- tea whitei (Gilbert et al. 1994, 2001) suffer more from pathogens and/or herbivores when they are closer together or closer to adult conspecifics. In other tree species, such as Attalea butyracea (Scheelea zonensis: S. J. Wright 1983) and Virola surinamensis (Howe et al. 1985) on Barro Colora- do, Astrocaryum muruniuru in western Amazonia (Terborgh et al. 1993), and Maximiliana maripa in northern Amazonia (Fragoso et al. 2003), insects cause mutual repulsion among conspecifics by in- flicting more damage on seeds near conspecific adults. The mutual repulsion acts on very different scales; a toucan greatly improves the survival pros- pects of a Virola seed by carrying it 40 m from its parent's crown before dropping it, whereas Maxi- miliana palm seeds may have a future only if a tapir defecates them a kilometer or more from existing palm clumps. In general, many insects and some pathogens, but very few mammals, preferentially attack seeds or seedlings near conspecific adults (Hammond & Brown 1998, Gilbert 2002). A meta-analysis by Hyatt et al. (2003) suggests that seed pr?dation is as likely to be heavier farther from, as nearer to, parent plants. If so, mutual re- pulsion may be governed mostly by heavier mor- tality of seedlings or saplings, rather than seeds, near conspecific adults. Is mutual repulsion among conspecifics the rule for tropical trees? In the tw^o commonest canopy tree species on Barro Colorado's 50 ha Forest Dy- namics Plot, Trichilia tuberculata and Alseis blac- kiana, the number of juveniles per adult on a hect- are is sufficiently depressed by increased numbers of conspecific adults on that hectare to regulate the populations of their species (Hubbell et al. 1990). Other populations may be regulated on more local scales. Per capita recruitment of stems 1 cm DBH or greater of a species onto 10 x 10 or 20 x 20 m quadrats of the 50 ha plots on Barro Colorado, Panama, or Pasoh, Malaysia, is lower on plots in which this species has higher density or basal area (Wills etal 1997, Wills & Condit 1999). On Ba- rro Colorado's 50 ha plot, the probability that a stem 1 cm DBH or greater in 1983 survived to 1995 was lower if there were more conspecifics among its 20 nearest neighbors (Ahumada et al. 2004). For shade tolerant plants, the probability of surviving from 1983 to 1995 was diminished by an average of 1 percent for each extra conspecific among its 20 nearest neighbors (Fig. 8D in Ahu- mada et al 2004). In another analysis, Peters (2003) divided stems on the 50-ha plots at Pasoh and Barro Colorado into three diameter classes: DBH < 5 cm, 5 cm ? DBH < 10 cm, and DBH > 10 cm. For each species with 30 or more stems in some diameter class, he calculated the partial correlation between the prospects of a stem of that size class surviving from the first census (1983 on Barro Colorado, Why Are There So Many Kinds of Tropical Trees? 455 1987 at Pasoh) to 1995 and the density of con- specifics within its size class ^5, ^10, ?15, or ?20 m away fi'om it, holding the density of he- terospecifics in the neighborhood constant. He tested significance relative to the correlation from 100 data sets in each of which the survival fate of each tree was randomly reassigned to another tree of the same species and size class. For over half the species in each plot, he found a correlation for at least one diameter class and neighborhood size that he considered significant or nearly so. Such corre- lations w^ere most prevalent among small stems, but they occurred among trees 10 cm DBH or greater in 40 percent of the species tested. On each plot, over 75 percent of these correlations w^ere negative; a stem was less likely to survive if there ^vere more conspecifics nearby (Peters 2003). Unfortunately, local density of conspecific trees and local mortality rates among stems of all species are both spatially autocorrelated on BCI's 50 ha plot (Condit et al. 2000, Hubbell et al. 2001: 861). The simulations underlying the significance tests of both Peters (2003) and Wills et al (1997), however, removed the spatial autocorrelations in mortality, invalidat- ing the tests. This topic needs more w^ork. Nonetheless, what might these results mean? First, many species are increasing in numbers even though their stems die faster ^vhen more conspe- cifics are nearby. Species having stems that survive better with fewer conspecific neighbors include some that can increase quite rapidly ^vhen rare (Condit et al 1996, Ahumada et al 2004). Secondly, some species ^vhose stems survive better w^hen more conspecifics are nearby are hab- itat specialists. On Barro Colorado, stems of the habitat specialists Erythrina costaricana, which grows along streams, and V. surinamensis, which grows on slopes where soil moisture content is higher during the dry season, survived better w^hen more conspecifics were nearby (Condit et al 1996, Ahumada et al. 2004); however, this w^as not true for all habitat specialists. Also, there is other evidence that specialized pests affect forest dynamics in a way that enhances tree diversity. Rausher (1981) found that specialist s\vallowtail caterpillars located Aristolochia vines more readily if surrounding plants of different spe- cies virithin 50 cm of an Aristolochia were cleared away. Similarly, Wills and Green (1995) proposed a herd immunity hypothesis ^vhich, applied to trees, predicts that a stem survives better when more stems of other species are nearby, because they "hide" that stem from pests specialized to its species. At Pasoh, the average stem survives better w^hen the density of stems of other species Wthin 15 m or less is higher, even though it survives less \vell when the basal area of stems of other species is higher (Peters 2003). The discrepancy occurs be- cause most stems are small, and small stems suffer less from the competitive impact of other stems their size than from the shade and root competition of bigger trees nearby. On Barro Colorado, the av- erage stem survives worse when the density of near- by stems of other species is higher. Nonetheless, for 69 of the 188 species on the plot that w^ere com- mon enough to test, the "herd immunity" effects of extra heterospecifics nearby overrode the com- petitive impact of these heterospecifics enough to significantly enhance survival (Peters 2003). It is not obvious what, besides pest pressure, would cause neighboring stems of other species to en- hance a plant's survival. Finally, if the pressure of specialized pests is responsible for the mutual repulsion among con- specifics documented by Peters (2003), can the ob- served degree of mutual repulsion among conspe- cifics (and the observed degree of herd immunity) maintain observed levels of tree diversity? Tree seeds disperse only a limited distance from their parents (Harms et al 2000), while a seed's recruit- ment prospects, and a plant's survival prospects, de- pend on nearness to stems of different sizes and on w^hat species these stems belong to. Learning howr repulsion among conspecifics governs tree diversity demands a dynamics of spatial pattern, which is not an easy task (Molofsky et al 1999) even in the neutral case (Nagylaki 1976). (D) CONCLUSIONS.?Testing Hubbell's (2001) neu- tral theory suggested that if species begin as small populations, common species must have become so thanks to some advantage, and other species must avoid replacement by new competitors spreading through their region by increasing when rare enough, at least in some habitat (Chesson 2000). Therefore, some mechanisms must promote the co- existence of tree species. Disturbance, habitat com- plexity, and pressure of specialized pests all provide opportunities for different species to coexist. At the moment, it is easier to visualize how pest pressure, rather than microhabitat specialization, maintains many tree species in a single forest stratum. To assess the relative contributions of different factors to the a-diversity of trees, (1) we need a more com- prehensive understanding of the various possible modes of habitat specialization; (2), we must learn w^hether pest pressure is the prime cause of mutual repulsion among conspecifics; and (3) we must 456 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit learn ^vhether or not observed levels of mutual re- pulsion among conspecifics can maintain observed levels of tree diversity. Now, however, we turn to causes of species turnover. CAUSES OF SPECIES TURNOVER (?-DIVERSITY) Differences in tree species composition among bio- geographic realms separated by oceans or by re- gions of very different climate have long been known (Good 1964). Changes in tree species com- position (?-diversity) along elevational gradients (Brown 1919) or along gradients of climate and soil (Gleason &C Cronquist 1964: chapter 22) have likewise long been known, especially in the north temperate zone. Data on the rate at which tree spe- cies composition changes with location in a tropical region such as Amazonia have become available more recently (Ashton 1964, Schulze &C Whitacre 1999, Pyke etal. 2001, Pitman et ai. 2001, Pitman, Terborgh, Silman, Nunez et al. 2002). What fac- tors influence species turnover among tropical trees? (A) LOCAL SPECIATION AND DISPERSAL LIMITATION.? One cause of species turnover is that different spe- cies originate in different places and spread only a limited distance from their points of origin. Dif- ferences in tree species composition among differ- ent biogeographic realms arise largely from this cause. To learn how local speciation and dispersal lim- itation might influence species turnover within a biogeographic realm. Chave and Leigh (2002) modified Hubbell's (2001) neutral theory by in- corporating limited dispersal. They imagined a for- est in an endless, uniform habitat in which the distribution of seeds about their parents is the same for all reproductive trees regardless of their species, and each tree has the same minute probability of producing a young of an entirely new species. They calculated the probability F{r) that two trees r km apart belong to the same species, as a function of tree density, speciation rate, and mean square dis- persal distance of seeds from their parents, when speciation is in balance with random extinction. Condit et al. (2002) used this theory to fit data from censuses of trees 10 cm DBH or greater on 1 ha plots separated by 200 to 100,000 m in Am- azonian Peru near Manu, Amazonian Ecuador near Yasuni, and central Panama near Barro Colorado. They used observed values of tree density, assumed that mean square dispersal distance was near that observed on Barro Colorado, and adjusted specia- tion rate to fit the data. They fit the trend of the data, and they found more scatter about the trend in central Panama, which has a strong rainfall gra- dient and marked variation in soil type, than in the more uniform expanses of upland forest in western Amazonia. To fit these data, however, Condit et al. (2002) had to assume that speciation rate was a thousand-fold lower near Yasuni than in central Panama, and a thousand-fold lower near Manu than near Yasuni. The speciation rate fitting the data near Manu is so low that it would take several times the age of the universe for random extinction to balance speciation (Chave & Leigh 2002). In later attempts to compare the relative con- tributions of dispersal limitation and habitat het- erogeneity to species turnover, the partial correla- tion of distance or the logarithm of distance be- tween plots with the divergence in their species composition, holding topography and soil con- stant, was used as a surrogate for the impact of dispersal limitation (Potts et al. 2002; Phillips et al. 2003; Tuomisto, Ruokolainen, Aguilar et al. 2003; Tuomisto, Ruokolainen, & Yli-Halla 2003). All these authors, especially Tuomisto, Ruokolainen, and Yli-Halla (2003), found that distance explains some differences after the effects of soil and topog- raphy are accounted for. They ascribe this distance effect to the limited dispersal of species that origi- nate in different places. They all, how^ever, consider that differences in soil and topography contribute far more to species turnover. We now turn to this contribution. (B) HABITAT HETEROGENEITY AND SPECIES TURN- OVER.?Plants face a fundamental trade-off be- tween competing for light and competing for water and nutrients (Tilman 1982, King 1993); resources expended on roots and mycorrhizae are not avail- able for making leaves or lifting them above neigh- boring crowns. Forests on drier (Murphy & Lugo 1986) or less fertile soil (Medina & Cuevas 1989) must expend more energy procuring water or nu- trients and have larger proportions of their total biomass underground (Table d.G in Leigh 1999). Moreover, in poorer soils, trees conserve their nu- trients by making tougher, better-defended, longer- lived leaves (Janzen 1974, Reich et al 1992). Fi- nally, trees on waterlogged soils and those subject to seasonal floods, have special features which, among other things, assure their roots an adequate supply of oxygen (Junk 1989). These features pre- sumably render these species less competitive in up- land forests. Why Are There So Many Kinds of Tropical Trees? 457 The long-recognized impact on tree species composition of habitats as different as floodplain forests, swamp forests, white sand forests, and "nor- mal" upland forests (Richards 1952, Brunig 1983, Proctor et al. 1983, Balslev et al. 1987, Medina & Cuevas 1989, Dumont et al. 1990, Terborgh et al. 1996, Schulze & Whitacre 1999), reflect these trade-offs, as does the impact of dry season length on tree species composition (Richards 1952, Pyke et al. 2001, Condit et al 2004). Within upland forests, differences in soil quality influence tree spe- cies composition. In the 52 ha Forest Dynamics Plot at Lambir, none of the six most common can- opy species, or the seven most common subcanopy species in a 4 ha subplot on fertile soil w^ere among their most common counterparts in a 4 ha subplot on very poor soil (Lee et al 2002). Nearby hectares on very different soils can dif- fer greatly in tree diversity as well as species com- position. In a national park in Sarawak, a hectare of dipterocarp forest contains 214 species of tree 10 cm DBH or greater; a hectare of heath forest, 123; and a hectare of forest on limestone, 73 (Proc- tor et al. 1983). In Pakitza, along the Rio Manu in southw^estern Amazonia, a hectare of upland for- est averages 128 tree species, whereas a hectare of swamp forest contains 61 (Appendix in Pitman et al. 1999). If they are distinctive enough, rare hab- itats, like small regions, have fewer tree species. Gradients in climate also have a marked impact on species turnover. In a network of 22 small plots spanning 6.5? of latitude in south India's Western Ghats, at elevations ranging from 400 to 1400 m, species turnover is most rapid along axes where the length and severity of the dry season varies most rapidly (Davidar ?c Puyravaud 2002). Heteroge- neity in soil and climate is clearly a major contrib- utor to species turnover. (C) WHAT FACTORS LIMIT SPECIES TURNOVER?.? Some tree species, in some places, are much less influenced by habitat heterogeneity than others. In the rain forests of Sarawak, species composition of trees 10 cm DBH or greater depends much less on soil type when soil fertility exceeds a certain thresh- old (Potts et al 2002). On Barro Colorado, Pan- ama, where soil is relatively fertile (Leigh & Wright 1990), 26 of the 41 most common species on the 50 ha Forest Dynamics Plot are equally common on the flat plateau, on slopes exceeding 10 percent in ravines and along streambeds, and in a 2 ha seasonal swamp (Hubbell & Foster 1986), even though in the dry season, soil moisture is more readily available on the slopes (Becker et al. 1988) or along streambeds than on the plateau. On the relatively fertile soils of western Amazonia, com- mon species tend to grow in many habitats (Pitman et al 1999). In a network of 15 1 ha upland forest plots near Yasuni, 15 percent (150) of their species averaged at least one tree 10 cm DBH or greater per hectare. These 150 species accounted for 63 percent of these plots' trees (Pitman et al 2001) and for 32 percent of the trees on floodplain and swamp forest plots nearby (Pitman, Terborgh, Sil- man, Thompson et al 2002). Where soils are fer- tile, many tree species live in a variety of habitats. Smaller plants can be more sensitive to habitat type. Indeed, common species of ferns and melas- tomes of Western Amazonia are rather more sen- sitive to habitat type than rare ones (Tuomisto, Ruokolainen, & Yli-Halla 2003). Ferns and melas- tomes include mostly herbs and shrubs, but tree species composition also changes with habitat type, even in Western Amazonia (Phillips et al 2003; Tuomisto, Ruokolainen, Aguilar et al. 2003). Tuo- misto, Ruokolainen, Aguilar et al (2003: 754), however, noticed a tendency "at least among Am- azonian trees and palms, that large-statured trees are more wide-spread both geographically and eco- logically than small-statured trees," a result quan- tified for geographical range by Ruokolainen et al (2002). Pitman et al (2001) likewise found that common species on their plots averaged greater maximum height than their rarer counterparts. Likewise, canopy or emergent tree species that oc- cur in at least half of lowland or mid-elevation plots were more common than species occurring in fewer plots in south India's Western Ghats (Davidar ?C Puyravaud 2002). Although common plants of smaller size are often habitat specialists, canopy trees of common species, at least those that grow on more fertile soils, are more likely to be habitat generalists. Bigger trees probably disperse their seeds far- ther than shrubs or treelets. This circumstance may hinder speciation among bigger trees (Davidar & Puyravaud 2002). Does some other factor allow common species to override the trade-offs that in- hibit growing successfully in different habitats? Is effective defense against specialized pests what al- low^s a tree species to grow in many habitats? There is some evidence that invasive plants that have es- caped the largest proportion of the fungal patho- gens plaguing them in their native habitats are the most widespread in North America (Mitchell & Power 2003). More detailed study of invasive plants may well reveal the extent to which freedom 458 Leigh, Davidar, Dick, Puyravaud, Terborgh, ter Steege, and Wright TABLE 1. Annual rainfall, P, and ?vapotranspiration, AET^ (mm); annual wood production, WP; total fine litter fall, LF; and total abovegroundproduction, TP; measured as ^V + LF, (tons dry weight/ha-yr); the total num.ber N of trees ^10 cm DBH in a plot sam.ple and the num.ber S of species among them at selected sites. Site Lat. P AET WP LF TP N 5 1. New Hampshire 44?N 1295 ASA 5.7 5.7 11.4 156 8 2. Western Oregon 44?N 2370 825 2.7 4.3 7.0 2825'' 12 3. Eastern Maryland 39?N 1080 758 6.5 6.9 13.4 435 15 4. W. North Carohna 35?N 1813 858 2.9 5.5 8.4 579 22 5. Chamela, Mexico 20?N 707 <700 2.4 3.6 6.0 451' 75 6. Costa Rica 10?N 3600 3.0 8.7 11.7 529 102 7. Panama 9?N 2600 1600 5.5 12.9 18.4 409 91 8. French Guiana 6?N 3357 1492 3.1 7.8 10.9 654 175 9. Amazonia, Brazil 2?S 2609 1319 4.4 8.4 12.8 618 285 ^ A?"?" measured as rainfall minus runoff except at sites 3 and 9. ''Trees >15 cm DBH in a 10.24 ha catchment. *= Freestanding woody stems a2.5 cm DBH in ten 100 m^ plots scattered over 5 ha. Data Sources: Site 1, Hubbard Brook: P and AET, Likens et al. (1977); WP, Whittaker et al. (1974); LF, Gosz et al (1972). N and 5 are from a forest in central Vermont (Bormann & Buell 1964). Site 2, Andrews Experimental Forest, WS 10: P and AET, SoUins et ai (1980); other data from Grier & Logan (1977). Site 3, SERC, Edgewater, MD: G. G. Parker, pers. comm. Site 4, Coweeta WS 18: P and AET, Johnson & Swank (1973); WP and LF, Monk & Day (1985); A^and S, J. A. Yeakley, pers. comm. Site 5, dry deciduous forest; P, WP, LF, Martinez-Yrizar et al (1996); A^and S, Lott et al. (1987). Site 6, everwet forest. La Selva; P WP, Lieberman et al (1990); LF, Parker (1994; Fig. 5.1); A^and 5, Lieberman et al (1985). Site 7, seasonal forest. Barro Colorado Island: data from Leigh et al. 2004. Site 8, rain forest, ECEREX site. Piste de St. Elie: _/? A?Tand LF, Sarrailh (1989); WP'\& standing crop, 318 tons/ha (Sarrailh 1989) times the average proportional increase in basal area from recruitment and growth, Pelissier & Riera (1993); 5, Sabatier & Pr?vost (1989); A^, average number of trees >10 cm DBH/ha in ECEREX plots of Puig & Lescure (1981). Site 9, Biological Dynamics of Forest Fragments Project; P, W. Laurance, pers. comm. AET from nearby Ducke Reserve where P = 2648 mm (Shuttleworth 1988); WP and LT (mm Fazenda Dimona (Clark et al 2001); A^and S, de Oliveira & Mori (1999). from pests allows plants to invade a wider range of in plant diversity over distances exceeding 800 km habitats. (Hawkins et al. 2003). And indeed, aboveground productivity is higher in tropical forests than in the ENVIRONMENTAL INFLUENCES "average" natural forest of the temperate zone. In f\-Ki TRP'p' F?TVTi'W?T'TV '^^ tropics, moreover, tree diversity tends to be low on the least fertile and least productive soils, such How do different aspects of the environment influ- as white sands (Bruenig 1996). ence the diversity that different modes of coexis- On the other hand, gradients in productivity tence among trees can support? Disentangling caus- ?o not predict gradients of tree diversity within a es from correlates is not easy because different as- region. Over gradients of 500 km or less, plant pects of the environment often vary concurrently. diversity peaks at intermediate productivity (Til- For example, when average annual temperature is ^^^ ^ p^^^j^ 1993^ Waring et al. 2002). More higher, winter is usually less severe; when rainfall is generally, tree diversity is not closely related to ei- higher, the dry season is usually shorter and wetter. ?^^^ aboveground productivity or actual evapo- (A) CLIMATE AND PRODUCTIVITY.?Here, we assume transpiration (Table 1). The forest at Hubbard that a forest's productivity reflects its "average" cli- Brook, New Hampshire, has higher productivity mate (Scheiner & Rey-Benayas 1994). Productivi- than a forest at Coweeta in western North Caro- ty, or an environmental factor such as rainfall or Una, but the Coweeta forest is more diverse than ?vapotranspiration that probably governs produc- forests of the Hubbard Brook region (Table 1). In- tivity for the region in question, is the best predic- deed, aboveground productivity is higher in eastern tor of plant diversity in 20 of 21 studies of changes Maryland than in French Guyana, La Selva, Costa Why Are There So Many Kinds of Tropical Trees? 459 Rica or north of Manaus, Brazil, which have forests that are much more diverse, and actual ?vapotrans- piration is higher in western Oregon and Coweeta than at Chamela, although tree diversity is much higher at Chamela (Table 1). In the tropics, net forest productivity levels off at an annual rainfall of 2000 mm, and declines from 3000 mm onward (Clark et al. 2001; Fig. 1 in Schuur 2003). Diversity of woody stems 2.5 cm DBH or greater on 0.1 ha plots increases with an- nual rainfall up to 3500 mm, and is no lower in wetter places (Gentry 1988a). Here, rainfall, not productivity, appears to be the best predictor of diversity. Productivity is necessary for tree diversity, but within the tropics, other factors such as total annual rainfall and brevity of dry season exert greater influence on tree diversity. How might productivity constrain tree diver- sity? There is no consistent variation of tree density from boreal forest to the tropics (Tilman ?c Pacala 1993); the increased productivity of tropical climes does not allow finer "niche partitioning" merely by supporting more trees per hectare. Higher produc- tivity can increase tree diversity by providing enough resources to support viable populations of energetic animal pollinators able to travel long dis- tances in search of appropriate pollen and such pol- linators can maintain adequate genetic variation in low-density tree populations (Regal 1977, Nason et al. 1998). Higher productivity also supports more trophic levels, multiplying the number of ways dif- ferent tree species can coexist (Paine 1966, Oksa- nen et al. 1981). Both these effects, however, are more pronounced where productivity is less season- al. (B) TREE DIVERSITY AND ENVIRONMENTAL STABILI- TY.?The most obvious contrast between the trop- ics and the temperate zone is winter. The frost-free parts of South Florida shelter a diverse array of tree species that do not grow farther north (Gleason & Cronquist 1964). Adapting to frost is costly. Wide xylem vessels transport water far more rapidly, but freezing embolizes wider vessels (Zimmermann & Brown 1971). Temperate zone trees must somehow parry the effects of frost. Ring-porous trees, such as oaks and elms, build a new set of wide vessels each year after the danger of frost is over and before leafing out: they opt for a shorter growing season in return for rapid water transport during this sea- son. Other kinds of trees have short, narrow vessels, as do diffuse-porous maples, or even narrower tra- cheids, as do conifers, sacrificing rapid water trans- port in return for a longer growing season and re- duced danger of embolism. There should accord- ingly be rapid species turnover near the southern- most boundary of frosts. But does frost limit tree diversity? Why cannot many frost-adapted tree spe- cies coexist? Winter has other effects. At higher latitudes, the low and variable sun angle reduces the number of strata a forest can support and diminishes the ability of understory trees to specialize to particular light environments (Terborgh 1985). A low and variable sun angle also reduces the contrast between light gaps and shaded understory (Ricklefs 1977). Finally, while trees at high latitudes must face a great variety of temperatures during their growing season, tropical trees can specialize more closely to particular temperature conditions; so species turn- over should be more rapid on the slopes of tropical mountains than on mountains at higher latitudes (Janzen 1967). Can such factors, however, increase the number of tree species on a hectare from 16 in Maryland to 91 in Panama and more than 280 in Amazonia (Leigh 1999)? Many believe that the latitudinal gradient in tree diversity occurs because the absence of winter enhances pest pressure. Is this true? Many types of organisms suffer heavier pr?dation in the tropics (Paine 1966, Moles & Westoby 2003). Specialist seed-eaters are considered a major cause of mutual repulsion among conspecifics (Hammond & Brown 1998). Moles and Westoby (2003) were therefore surprised to discover that there was no latitudinal gradient in the proportion of seeds de- stroyed prior to dispersal, or the proportion of seeds removed after dispersal, among the 122 and 205 plant species of different latitudes in which these types of seed pr?dation were measured. If, however, the most damaging tropical seed predators are specialists Qanzen 1980), their ability to destroy the seeds of rare tropical plant species as effectively as their counterparts destroy seeds of the more common plant species of the temperate zone sug- gests that if a plant species in a tropical setting became much more common, it would suffer in- tolerable seed pr?dation. To test this proposition, we must learn whether most seed pr?dation is the work of specialists and if pr?dation on the seeds of a tree species is heavier where that species is more common. Leaf-chewing and seedling-eating insects also cause mutual repulsion among conspecifics. Ab- sence of winter lengthens their activity season (Wolda 1983), intensifying their pressure on plants. Because young, expanding leaves should be equally tender everywhere, a comparative measure 460 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit of an area's pest pressure should assess the rate at which young leaves are eaten in relation to their toxicity and the area's plant diversity. Young trop- ical leaves are eaten much more rapidly even though they are far more poisonous and are pro- duced by much rarer tree species than young dicot leaves in the north temperate zone (Coley & Ba- rone 1996). Perhaps because young tropical leaves are more poisonous, butterfly caterpillars are more specialized in the tropics than in the temperate zone (Scriber 1973, 1995; Marquis & Braker 1994); this is presumably true of moth caterpillars as well, the main chewers of plant leaves (Janzen 1988). Phloem-sucking treehoppers (Membraci- dae) are less specialized in the tropics (Marquis & Braker 1994), presumably because phloem is not poisonous. Even so, the enormous pressure of con- sumers on seeds, seedlings, and young leaves of tropical tree species, despite their low density, sug- gests that increased pest pressure entailed by ab- sence of winter could account for the latitudinal gradient in tree diversity. On the other hand, there must be a "niche op- portunity" in the coniferous forests of eastern Ca- nada for a tree species resistant to spruce bud- worms. Is that habitat more difficult to adapt to than others? Are cold winters so difficult to adapt to, and habitats with cold winters so temporary, geologically speaking, that there has not been time enough for a tree species resistant to spruce bud- w^orms to evolve in this habitat? Within the tropics, tree diversity tends to be higher where dry season is shorter (Givnish 1999, Leigh 1999), that is to say, where climate and pro- ductivity vary less from season to season. In India's Western Ghats, annual rainfall is not correlated with the length of the dry season; there, tree di- versity on small plots is not correlated with total annual rainfall, but it is higher where dry season is shorter (Davidar & Puyravaud, pers. obs.). In Amazonia, the average tree diversity on 1 ha plots within a 1? latitude x 1? longitude block is higher where dry season, as measured by the number of months averaging less than 100 mm of rain, is shorter (i? = 0.35, P< < 0.001; ter Steege et al. 2003). If all the plots in this study with one dry month or less are assigned to one group, those with two dry months to another, those with three dry months to a third, and so forth, then the average diversity of the most diverse 10 percent of the plots in each group is quite tightly correlated with dry season length {R^ = 0.91, P? 0.001; ter Steege et al. 2003); here, dry season length is a decisive constraint on tree diversity. More generally, tree diversity is highest where environmental conditions vary least. "Hyperdiver- se" forests with over 250 tree species per hectare are restricted, not only to everwet climates, but to latitudes between 4?40'N (Davies & Becker 1996) and 5?S (de Lima Filho et al. 2001), where sun angle at zenith varies least (Table 3). Are light en- vironments partitioned more finely in these "hyper- diverse" forests? There is little evidence to support this proposition. Everwet climates usually allow year-round her- bivory. The abundance of insect herbivores varies far less with the season in a forest of central Guyana with only one dry month a year (Basset 2000) than on Barro Colorado, with its four-month dry season (Wolda 1983). In a deciduous dry forest of south India, insect herbivores are sufficiently less com- mon during the dry season that trees reduce their "herbivore tax" by leafing out before the rains come (Murali & Sukumar 1993). In a Brazilian Cerrado, however, trees and shrubs avoid pathogen damage, not herbivore damage, by flushing leaves before the rainy season; leaves flushed after the rains come suffer more pathogen damage than those flushed beforehand (Marquis et al. 2001). Is insect herbivory more intense in everwet for- ests where it is less seasonal? We summarize the few available data. Young leaves are eaten much more rapidly in the seasonal forest of Barro Colorado than in drier Brazilian Cerrado or Mexican dry for- est (Table 2), despite Barro Colorado's higher tree diversity; insect pressure is higher on Barro Colo- rado than at these drier sites. Mutual repulsion among conspecifics is as prevalent on the 50 ha Forest Dynamics Plot at Pasoh, Malaysia, as on Ba- rro Colorado's plot, even though tree diversity is far higher and tree species correspondingly rarer at Pasoh (Peters 2003). Were Pasoh less diverse, its plants would be more damaged by specialist pests than Barro Colorado's. On the other hand, the everwet forest at La Selva, Costa Rica, has hardly greater tree diversity than Barro Colorado (Table 1); yet, on average, both young and mature leaves of both pioneer and mature forest species are eaten much more rapidly on Barro Colorado than at La Selva (Table 2). Indeed, young leaves appear to be consumed most rapidly in forests with dry seasons of intermediate length (Marquis et al. 2002). Are anti-herbivore defenses so much stronger at La Sel- va than on Barro Colorado? A contrast within cen- tral Panama suggests that this may be true. In dry forest near Panama City, trees need birds to help defend them against insect herbivores (Van Bael et al. 2003), whereas in forest near Col?n, where an- Why Are There So Many Kinds of Tropical Trees? 461 TABLE 2. Consumption rate by chewing herbivores, percent area per day, of young and mature leaves at selected sites. Site Lat. Young Leaves Mature Leaves L Deciduous dry forest 2. Cerrado 3. Seasonal forest 4. Everwet forest 19.5?N 15.9?S 9.rN 10.4?N 0.352 0.097 0.254 + 0.113= 0.007 o.g?'' 0.04^ O.MS'' 0.012^ 0.060" " Damage rate from pathogens. ? Average for mature forest species. Data Sources: Site 1, Chamela, Mexico: Filip et al, (1995). Site 2, Fazenda Agua Limpa, Brazil: Marquis et al. (2001). Site 3, Barro Colorado Island, Panama: Coley (1983). Site 4, La Selva, Costa Rica: Marquis & Braker (1994). nual rainfall is 50 percent higher and dry season shorter and wetter, trees suffer no more insect dam- age when birds are excluded (S. Van Bael, pers. comm.), as if the leaves of wetter forests have far better anti-herbivore defenses. To learn if pest pres- sure is what maintains higher tree diversity in more nearly everwet forests, the relation between season- ality, anti-herbivore defenses, pr?dation pressure on herbivores, and the resulting damage from pests needs far more exact study, at many more sites. On the other hand, it appears that trees 10 cm DBH or greater are no more likely to be habitat specialists in western Amazonia than in South Car- olina (Pitman, Terborgh, Silman, Thompson et al. 2002). We also need to learn whether species turn- over contributes to the latitudinal gradient in tree diversity (Pitman, Terborgh, Silman, Thompson et al. 2002), and whether this answer depends on the size of the plants considered. In summary, even within the tropics, less sea- sonal climates allow higher tree diversity. More work is needed to learn why this is so. The mystery of how 43 species of Inga can coexist on a 25 ha plot in Amazonian Ecuador (Bermingham Sc Dick 2001) demands attention. Moreover, some forests in everwet climates, such as that at San Carlos de Rio Negro, have very low tree diversity (Table 3). This circumstance reflects the effects of past his- tory, which we will consider later. (C) THE RELATION BETWEEN LOCAL AND GLOBAL TREE DIVERSITY.?Local tree diversity tends to be higher in regions with high regional tree diversity (Ricklefs 2004). Does this mean that in regions with more tree species, local diversity is enhanced by species unfit for the locale, which would inevi- tably be replaced by superior competitors were it not for immigration from outside (Shmida & Wil- son 1985)? Would this circumstance imply that mechanisms promoting stable coexistence of tree species are irrelevant to gradients in tree species diversity? If we have interpreted our test of HubbelFs (1997, 2001) neutral theory correctly, each tree species in the region started off with some advan- tage over its competitors, and a tree species usually has some setting in which it can increase when rare. A region's tree diversity cannot exceed the diversity of that region's local niche opportunities. To be sure, a larger proportion of such opportunities may remain unexploited in rarer habitats or smaller re- gions such as oceanic islands. We know too little of the ecology of most trop- ical tree species to even think of assessing what proportion of the species on a hectare, or in a 50 ha plot, is unsuited to that plot's environment. It is reasonable to believe, however, that more species will spill over from habitats offering more niche opportunities than from those offering fewer (Mac- Arthur & Wilson 1967). Immigration into unsuit- able habitats may render diversity gradients less steep, but immigration cannot reverse these gradi- ents (Rohde 1992). These gradients require expla- nation in terms of the relative abundance of niche opportunities. WHAT FACTORS INFLUENCE A REGION'S TREE DIVERSITY? A region's tree diversity represents a balance be- tween speciation (and some immigration) and ex- tinction (MacArthur & Wilson 1967, Terborgh 1973). We must first review mechanisms of speci- ation among woody plants to learn what conditions are necessary for speciation, whether or not species usually begin from small populations, and if, as our test of Hubbell's (2001) neutral theory has sug- gested, speciation is usually a response to a niche 462 Leigh, Davidar, Dick, Puyravaud, Terborgh, ter Steege, and Wright TABLE 3. Latitude, annual rainfall V (mm), rainfall during year's driest quarter V 2, (mm), number^ of trees ^10 cm DBH, number S of species among them, and Fisher's a^ in selected everwet forests, and in selected pairs of sites with similar climates. In each pair of sites, the upper site was much less disrupted by Pleistocene climate change. Site Lat. P Pi N S a. Everwet Sites 1. Papua New Guinea 6.7?S 6400 1350 693 228 119 2. Amazonia, Brazil 5?S 3404 <300 769 322 208 3. Amazonia, Brazil 5?S 2715 365 779 271 147 4. Amazonia, Peru 4?S 2845 513 608 295 226 5. Amazonia, Brazil 2.4?S 2609 344 618 285 205 6. Lambir, Sarawak 4.3?N 2664 498 637 247 148 7. Andulau, Brunei 4.7?N 3000 500 572 256 178 8. San Carlos, Venezuela 1.9?N 3500 600 744 83 23.4 Paired Sites 9. Manu, SW Amazonia, Peru 12?S 2028 186 586 174 83.6 10. Korup, Cameroon 5?N 5272 172 492 87 30.7 11. E. Amazonia, Brazil 5?S 1900 87 530 98 35.4 12. W. Ghats, S. India 8.9?N 2499 90 482 57 16.8 ^ Fisher's a is calculated recursively from the equation S = CL ln(l + NICL). Fishers alpha depends less on sample size //than does S (Leigh 1999). Data sources: Site 1, Crater Mt: Wright et al. (1997). Site 2, Ha 3, Rio Urucu: de Lima Filho et al. (2001). Site 3, Mungaba plot, Rio Juru?; Lima da Silva et al. (1992). Site 4, Allpahuayo, plot 2: V?squez-Mart?nez & Phillips (2000); rainfall for nearby Iquitos from M?ller 1982. Site 5, Ha C, 90 km N of Manaus: de Oliveira & Mori (1999); rainfall from W. Laurance, pers. comm. Site 6, Lee et al. 2004. Site 7, Davies & Becker (1996); P3 from nearby Site 6. Site 8, San Carlos de Rio Negro, forest on Oxisol: Uhl and Murphy (1981). Site 9, Average Manu ha from Pitman et al (2002); rainfall from Gentry (1990, facing p. 1). Site 10, Chuyong et al 2004. Site 11, Salom?o (1991). Site 12, Parthasarathy & Karthikeyan (1997). opportunity. Next, we ask what factors affect the diversity different regions support, and how long it takes speciation and immigration to build up the tree diversity of a region newly opened, say, by re- treating glaciers or a major, favorable change in cli- mate. Finally, w^e ask whether the tropics are pri- marily a cradle of tree speciation or a museum of tree diversity. (A) MECHANISMS OF SPECIATION AMONG WOODY PLANTS.?This paper's argument hinges on the propositions that speciation in woody plants is nor- mally a response to a niche opportunity and that new species arise from small populations. Thus, w^e inquire into mechanisms of speciation among woody plants, especially tropical ones. The first question is, do tropical woody plants usually sp?ciale allopatrically? Sympatric speciation can occur in woody plants. Polyploidy seems a log- ical means of sympatric speciation, but only 2?4 percent of today's species of flowering plants evolved by polyploidy, and this proportion is far lower among w^oody plants (Otto & Whitton 2000). Sympatric speciation sometimes occurs among tropical trees when a sexual species gives rise to an obligately asexual clade (Ehrendorfer 1982, Gentry 1989), but asexual species can hardly last long in competitive, pest-ridden tropical forests. Gentry (1989) proposed that sympatric speciation is contributing to the "explosive " diversification of smaller woody plants in the topographically het- erogeneous low^er slopes of the northern Andes, but he provided no evidence for how it happened. In groups such as Bignoniaceae and Dipterocarpaceae, however, the most closely related species have ad- jacent rather than overlapping or distantly separat- ed ranges (Ehrendorfer 1982, Ashton 1988, Gentry 1989). Although the prevalence of allopatric spe- ciation among tropical woody plants needs far more comprehensive and careful study, for the mo- ment it appears reasonable to conclude with Eh- rendorfer (1982: 505) that for woody plants, "Ini- tial steps of speciation mostly are on the basis of Why Are There So Many Kinds of Tropical Trees? 463 allopatric (often peripatric) geographical or para- patric ecological differentiation." Second, how local is speciation? In woody plants, it is usually possible to hybridize members of populations many million years after they w^ere separated (Ehrendorfer 1982, Carr & Kyhos 1986, Baldwin et al. 1991). This rule needs more careful, and more comprehensive, testing. If true, then w^hen a barrier splits a tree population, different, mutually incompatible al?eles of the sort mentioned by Dobzhansky (1937) and Orr and Turelli (2001) do not accumulate fast enough to make hybrid breakdown a driving force in tree speciation. Most tree species do not arise because an impenetrable barrier split big populations of trees in half. Spe- ciation in woody plants must usually be a more "local" process, perhaps a form of "budding" (Steb- bins 1982) in which a subpopulation invades a new habitat, and selection, driven by the lower fitness of hybrids for either parent's niche, favors its re- productive isolation from the parent population. The final question is, do new species form in response to niche opportunity, or is speciation as little related to niche opportunity as mutation is to opportunities for organismic adaptation? One of the few relevant studies of speciation among trop- ical plants concerns speciation among large herbs, Costus (Costaceae) in Central America (Schemske 1981, Kay & Schemske 2003). As a rule, a Costus species is pollinated by either bees or humming- birds, rarely both. Sympatric Costus with different pollinators exchange few genes, even though arti- ficial hybrids do well in the laboratory (Kay & Schemske 2003). Is there selection against hybrid- izing? If so, it cannot be driven by the kind of genetic incompatibility invoked by Dobzhansky (1937); hybridizing must therefore be disadvanta- geous because hybrids are unfit for either parent's way of life (Schl?ter 1998). There is further evidence for selection against hybridizing in Costus. At La Selva, Costa Rica, a pair of bee-pollinated Costus avoid hybridizing by flowering at different times (Kay & Schemske 2003). In a pair of hummingbird-pollinated Costus species that flower simultaneously at both Barro Colorado and La Selva, pollen?stigma incompati- bility prevents hybridization between plants at the same site, but a plant of either species can be fer- tilized successfully by pollen from a plant of the other species at the other site (Kay 2002). The pol- len?stigma incompatibility occurs only when need- ed to prevent hybridizing. If cross-site hybrids grow well in the laboratory, hybridizing must be disad- vantageous because hybrids are fit for neither pa- rent's way of life. Now that access to the canopy has become more practicable (Ozanne et al. 2003), the studies of Schemske (1981) and Kay and Schemske (2003) on Costus can be repeated on tropical trees. As this has yet to happen, we confine ourselves to extrap- olating the conclusions of Kay and Schemske (2003) to those many woody genera with fully dis- tinct species, which are easily hybridized artificially but rarely hybridize in nature. When they are sym- patric, species in these genera flower at different times or use different pollinators (Ehrendorfer 1982, Gentry 1989). We therefore infer that, in these genera, speciation is a response to a niche opportunity and that selection reduces the proba- bility that sympatric species hybridize because hy- brids are unfit for either parent's way of life. What might drive speciation? What role might pest pressure play in speciation? "It is a common- place of African botany that genera very often have one or more species in the rain forest and other species in the savannah" (Gillett 1962). In South America, 14 of the 18 most common \voody spe- cies in a hectare of cerrado savanna near the Tropic of Capricorn censused by Silberbauer-Gottsberger and Eiten (1987: 77) are congeneric with some tree 10 cm DBH or greater in one of the three rain forest hectares north of Manaus censused by de Oliveira and Mori (1999). The woody genus Ru- prechtia (Polygonaceae), which originated in "cha- co" savannas over ten million years ago, has since radiated into deciduous dry forest. During the last million years, this genus has invaded and diversified in forests of Central America (Pennington et al. 2004). The vine genera Chaetocalyx and Nissolia have diversified back and forth from wet to dry forests during the last ten million years. How could a genus that evolved in rain forest compete suc- cessfully in dry forests or savannas with genera that evolved in those settings, or vice versa? A shift of habitats may be advantageous if it entails escape from pests (Gillett 1962; Janzen 1970: 523; Mitch- ell & Power 2003). Based on these limited data, it appears that (1) speciation among tropical w^oody plants is usually allopatric or parapatric; (2) new species arise locally, with, initially, small populations; and (3) speciation usually occurs in response to divergent selection be- tween a parental way of life and a new "niche op- portunity." (B) WHAT FACTORS INFLUENCE A REGION'S TREE DI- VERSITY?.?If each tree species initially spreads and 464 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit avoids subsequent extinction by exploiting a unique niche opportunity, then a region's tree di- versity will be enhanced by a greater number of local niche opportunities, which may arise either from a greater variety of habitats or from more opportunities per habitat. Large extent also enhances a region's tree di- versity. Other things being equal, larger regions should support higher regional diversity, and there- fore higher local diversity, than smaller ones (Ter- borgh 1973; Rosenzweig 1995: 287flf; Ricklefs 2004). Regional diversity represents a balance be- tween speciation (and some immigration) and ex- tinction (MacArthur & Wilson 1967, Terborgh 1973). In larger regions, extinction rate is lower because their species can spread more widely, re- ducing the probability of extinction by environ- mental variation (Rosenzweig 1995). In larger re- gions, speciation occurs more often. Larger regions offer more opportunities for isolation and diver- gence of peripheral tree populations (Ricklefs 2004) and more opportunities for speciation and diversification among herbivores and pathogens, thereby multiplying niche opportunities for trees (Ehrlich & Raven 1964, Dussourd & Eisner 1987, Becerra & Venable 1999, Becerra et al. 2001). Temperate zone forests in east Asia contain more species than forests of equal area and com- parable climate in North America (Latham ?C Ricklefs 1993b). This was true even before glacia- tions started in the Pleistocene, presumably because Eurasia has a far larger area of temperate zone hab- itat than North America. A hectare of everwet for- est in Amazonia contains more tree species than one in New Guinea (Table 3), presumably because Amazonia's total area of rain forest is far larger than New Guinea's. On the other hand, despite the small area and profound isolation of Madagascar's rain forest, a hectare of rain forest at 18?S musters 146 tree species 10 cm DBH or greater (Rakoto- malaza Sc Messmer 1999), far more than can be found in any hectare of temperate zone forest. Ex- cept for very small regions, limited regional area does not override the impact on local tree diversity of the variety of local niche opportunities. A region's tree diversity is also enhanced by a longer period of freedom from major environmen- tal change. It takes millions of years of freedom from such change for speciation and immigration to build up the tree diversity of a newly opened region to the level it can support. Tree diversity is lower in northern Europe than in the eastern Unit- ed States because glaciation devastated Europe's tree flora, and there has been too little time for immi- gration and speciation to replenish its diversity (Latham & Ricklefs 1993a). On the Hawaiian is- lands, the "Big Island" has 14 times Maui's area of alpine/subalpine habitat, but only 94 plant species in that habitat compared to Maui's 102?appar- ently because Maui's alpine/subalpine habitat is nearly tw^o million years old compared to less than 600,000 years for the "Big Island's" (Price 2004). For 1 ha plots in a given climate, tree diversity is lowest in areas such as Australia, south India, and west Africa (Table 3), where the forest retreated farthest into the most scattered refuges during Pleistocene glaciations (Motley 2000). The everwet upland forest at San Carlos de Rio Negro, Vene- zuela, has uncommonly low tree diversity for its climate (Table 3), apparently because the region was much drier during the Pleistocene: during the Last Glacial Maximum, sand dunes were active near San Carlos in what is now everwet forest (Car- neiro Filho et al. 2002). Tree diversity is greatest in Western Amazonia (Gentry 1988b), most of which was covered by rain forest all through the Pleisto- cene (Piperno 1997). Developing a western Ama- zonian level of tree diversity, however, does not re- quire the whole Cenozoic. Twenty million years ago, Malesia was covered by seasonal monsoon for- est; the diversity of Sarawak's everwet forest has de- veloped since then (Motley 2000). (C) ARE THE TROPICS A CRADLE OF TREE SPECIATION OR A MUSEUM OF TREE DIVERSITY?.?^We have seen how tree diversity is highest in those tropical re- gions where climates have varied least during the last 10 or 15 million years, and lowest in those areas where forests were most disrupted by the cy- clic revolutions of climate during the Pleistocene (Motley 2000). Indeed, Amazonian tree diversity was higher before the Pleistocene changes began than it is now (Hooghiemstra & van der Hammen 1998). The cyclical drying and wetting of tropical hab- itats during Pleistocene glaciations, whereby rain forests were fragmented during glacial periods (Guillaumet 1967, Haffer 1969, Prance 1982) and dry forests fragmented during interglacials (Pen- nington et al. 2004), enhanced speciation rates in some groups (Pennington et al. 2004). In Central America, especially, the expansion of depauperate rain forest during interglacials provided niche op- portunities for dry forest lineages to invade this rain forest (Pennington et al. 2004). Such specia- tion, however, did not compensate for the diversity lost through the extinctions caused by climate change. Why Are There So Many Kinds of Tropical Trees? 465 The latitudinal gradient in tree diversity is old. A hectare of forest on Barro Colorado includes 73 genera and 36 families among its 91 species 10 cm DBH or greater (Leigh et al. 2004), and a hectare of depauperate tropical forest in India's Western Ghats has 47 genera and 32 families among its 57 species 10 cm DBH or more (Parthasarathy & Karthikeyan 1997), more genera and families than a temperate zone hectare has tree species. Many of a tropical plot's tree genera and most of its families evolved long enough ago to have spread to more than one continent; 30 percent of the tree genera recorded from Makokou, Gabon, also occur in the Neotropics (Gentry 1982: 564). The number of lowland species that occur on both sides of the Ecuadorian Andes suggests that at least 30 percent of the plant species in Ecuador's lowland rain for- ests evolved before the Andes sundered these rain forests in the late Miocene (Jorgenson and L?on- Y?nez 1999, Raven 1999). Tropical forests are clearly museums of diversity. Tropical environments, however, especially those in the wet tropics, also offer plants more w^ays to achieve reproductive isolation, making specia- tion easier there (Fischer 1960). For example, a greater variety of pollinators and a longer growing season provide more scope for achieving reproduc- tive isolation by pollinator shift or change of flow- ering season (Gentry 1989, Kay &L Schemske 2003). Flow might these factors influence gradients in tree diversity? Some tropical lineages diversify much more rapidly than others. Inga (Leguminosae), with 300 species, began diversifying 10 or fewer million years ago (Richardson et al. 2001); Ficus (Moraceae), with 750 species, originated 90 million years ago (Machado et al. 2001). Do lineages diversify more rapidly if they have cheaper or more effective anti- herbivore defenses that allow them to invade al- ready occupied habitats (Ehrlich & Raven 1964)? In Central America, trees and canopy lianas usually belong to slowly diversifying lineages, centered on lowland Amazonia; treelets and shrubs usually be- long to rapidly diversifying clades centered on the foothills and lower slopes of the lower Andes (Gen- try 1982). Do the shrubs and treelets diversify more rapidly simply because their seeds disperse less efficiently (Givnish 1999)? What factors drive speciation in these "Andean " lineages? CONCLUSIONS AND OPEN QUESTIONS A test of Fiubbell's (2001) neutral theory suggests that tree diversity is no accident. Rather, each spe- cies has some setting where it increases when rare. It appears that neither disturbance nor microhabi- tat specialization can explain the diversity of trop- ical trees, even though larger-scale habitat and cli- mate differences are the primary causes of species turnover. Increased activity of specialized pests and pathogens in less seasonal climates appears to be a primary cause of the latitudinal gradient in tree diversity. Further work is needed to assess the role of pest pressure and other factors in generating within-tropical diversity gradients. To establish and develop these conclusions, we need to learn (1) the various ways habitat complexity can enhance tree diversity. Flow many tree species can coexist locally by this means?; (2) whether mutual repulsion among conspecifics is usually driven by specialist pests or pathogens; (3) whether or not the observed degree of mutual repulsion among conspecifics suf- fices to maintain observed tropical tree diversity; (4) whether leaves of wet tropical forests have suf- ficiently effective anti-herbivore defenses to reduce herbivory rates on young leaves by 75 percent com- pared to a forest with a four-month dry season; (5) how much species turnover contributes to the lat- itudinal gradient in tree diversity; and (6) why spe- cies turnover is higher for herbs, shrubs, and tree- lets than for large trees. Local tree diversity is influenced by regional tree diversity. Regional tree diversity is higher in regions with more local niche opportunities (for which variety is greater at lower latitudes and in less seasonal environments), in larger regions, and in regions that have been free of major environ- mental shifts for a longer period. To attain a clearer understanding of the factors that maintain tree diversity, we must learn more about the mechanisms, and driving forces, of spe- ciation in woody plants. We must learn (1) if spe- ciation is usually allopatric; (2) how long a time after a tree population is split by an impenetrable barrier before hybrids between the two halves are no longer viable. If this time is very long, then reproductive isolation must usually be the outcome of natural selection driven by the unfitness of hy- brids for either parent's way of life; (3) what factors drive tree speciation? Is speciation usually associ- ated with invading new habitats?; and (4) why some lineages diversify rapidly, while other equally diverse lineages have diversified more slowly. Why the difference? Do cheaper, or more effective, anti- herbivore defenses allow a clade to diversify more rapidly? To understand why some tropical forests have so many kinds of trees, we will have to learn more 466 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit about the natural history of these trees, the history 2002 symposium, "Why are there so many kinds of trop- r .1 ? 1 ? _ j ^u J ? ? r c ^ ical trees?" That symposium was the original basis of this or their biomes, and the driving torces or tree evo- T^ T i. j ? , . paper. R. J. Marquis and two anonymous reviewers ^ ? ? helped greatly in bringing forth order into this paper C. Pizano kindly translated the abstract into Spanish. M. ACKNOWLEDGMENTS Brady E. Losos, and M. Massa guided the senior author through the mysteries of electronic editing. He is also J. Chave, P. Davidar, C. W. Dick, S. P. Hubbell, E. G. grateful to the plants and animals of Barro Colorado for Leigh, H. C. Muller-Landau, N. A. Pitman, H. ter Stee- continual reminders of what we tropical biologists must ge, J. Terborgh, and S. J. Wright contributed to the ATB explain. LITERATURE CITED ADLER, R R., AND J. MOSQUERA. 2000. Is space necessary? Interference competition and limits to biodiversity. Ecology 81: 3226-3232. AHUMADA, J. A., S. P. HUBBELL, R. CONDIT AND R. B. FOSTER. 2004. Long-term tree survival in a Neotropical forest: the influence of local biotic neighborhood. In E. Losos and E. G. Leigh Jr. (Eds.). Tropical forest diversity and dynamism, pp. 408?432. University of Chicago Press, Chicago, Illinois. AsHTON, P. S. 1964. Ecological Studies in the Mixed Dipterocarp Forests of Brunei State. Clarendon Press, Oxford, England. . 1988. Dipterocarp biology as a window to the understanding of tropical forest structure. Annu. Rev. Ecol. Syst. 19: 347-370. . 1989. Species richness in tropical forests. In L. B. Holm-Nielsen, I. C. Nielsen, and H. Balslev (Eds.). Tropical forests: Botanical dynamics, speciation and diversity, pp. 239?251. Academic Press, London, England. , AND P. HALL. 1992. Comparisons of structure among mixed dipterocarp forests of north-western Borneo. J. Ecol. 80: 459-481. AuGSPURGER, C. K. 1983. Seed dispersal of the tropical tree, Platypodiuni elegans, and the escape of its seedlings from fungal pathogens. J. Ecol. 71: 759?771. BALDWIN, B. C, D. W. KYHOS, J. DVORAK AND G. D. CARR. 1991. Chloroplast DNA evidence for a North American origin of the Hawaiian silversword alliance (Asteraceae). Proc. Nati. Acad. Sei. 88: 1840?1843. BALSLEV, H., J. LUTEYN, B. 0LLGAARD, AND L. B. HOLIM-NIELSEN. 1987. Composition and structure of adjacent unflooded and floodplain forest in Amazonian Ecuador Opera Bot. 92: 37?57. BASSET, Y. 2000. Insect herbivores foraging on seedlings in an unlogged rain forest in Guyana: Spatial and temporal considerations. Stud. Neotrop. Fauna Environ. 35: 115?129. BECERRA, J. X., AND D. L. VENABLE. 1999. Macroevolution of insect?plant associations: The relevance of host bioge- ography to host association. Proc. Nati. Acad. Sei. 96: 12626?12631. , , P. H. EVANS, AND W S. BO^)7ERS. 2001. Interactions between chemical and mechanical defenses in the plant genus Bursera and their implications for herbivores. Am. Zool. 41: 865?876. BECKER, P., P. E. RABENOLD, J. R. IDOL, AND A. P. SMITH. 1988. Water potential gradients for gaps and slopes in a Panamanian tropical moist forest's dry season. J. Trop. Ecol. 4: 173?184. BERMINGHAM, E., AND C. DICK. 2001. The Inga?Newcomer or museum antiquity? Science 293: 2214?2216. BERNAYS, E. A., AND D. H. JANZEN. 1988. Saturniid and sphingid caterpillars: Two ways to eat leaves. Ecology 69: 1153-1160. BoRMANN, F. H., AND M. F BuELL. 1964. Old-age stand of hemlock?northern hardwood forest in central Vermont. Bull. Torrey Bot. Club 91: 451-465. BROKAW, N. V. L. 1987. Gap-phase regeneration of three pioneer tree species in a tropical forest. J. Ecol. 75: 9?19. BROWN, W. H. 1919. Vegetation of the Philippine mountains. Bureau of Science, Manila, Philippines. BRUENIG, E. F 1996. Conservation and Management of Tropical Rainforests. CABI, Wallingford, England. BRUNIG, E. F. 1983. Vegetation structure and growth. In F. B. GoUey (Ed.). Tropical rain forest ecosystems: Structure and function, pp. 49?75. Elsevier, Amsterdam, The Netherlands. CARNEIRO FILHO, A., D. SCH\PARTZ, S. H. TATUMI, AND T ROSIQUE. 2002. Amazonian paleodunes provide evidence for drier climate phases during the late Pleistocene?Holocene. Quatern. Res. 58: 205?209. CARR, G. D., AND D. W. KYHOS. 1986. Adaptive radiation in the Hawaiian silversword alliance (Compositae-Madii- nae). II. Cytogenetics of artificial and natural hybrids. Evolution 40: 959?976. CHAVE, J., AND E. G. LEIGH JR. 2002. A spatially explicit neutral model of ?-diversity in tropical forests. Theor Popul. BioL 62: 153-168. CHESSON, P. L. 2000. Mechanisms of maintenance of species diversity. Ann. Rev. Ecol. Syst. 31: 343?366. , AND R. R. WARNER. 1981. Environmental variability promotes coexistence in lottery competitive systems. Am. Nat. 117: 923-943. CHRISTIE, D. A., AND J. J. ARMESTO. 2003. Regeneration microsites and tree species coexistence in temperate rain forests of Chilo? island, Chile. J. Ecol. 91: 776-784. CHUYONG, G. B., R. CONDIT, D. KENFACK, E. C. Losos, S. N. MOSES, N. C. SONG^)7E, AND D. W THOMAS. 2004. Korup forest dynamics plot, Cameroon. In E. Losos and E. G. Leigh Jr. (Eds.). Tropical forest diversity and dynamism, pp. 506?516. University of Chicago Press, Chicago, Illinois. Why Are There So Many Kinds of Tropical Trees? 467 CLARK, D. A., S. BROWN, D. W. KICKLIGHTER, J. Q. CHAMBERS, J. R. THOIMLINSON, J. NI, AND E. A. HOLLAND. 2001. Net primary production in tropical forests: An evaluation and synthesis of existing field data. Ecol. Appl. 11: 371-384. CLARK, D. B., AND D. A. CLARK. 1985. Seedling dynamics of a tropical tree: Impact of herbivory and meristem damage. Ecology GG: 1884-1892. CODY, M. L. 1966. A general theory of clutch size. Evolution 20: 174-184. CoLEY, P. D. 1983. Herbivory and defensive characteristics of tree species in a lowland tropical forest. Ecol. Monogr. 53: 209-233. . 1987. Interspecific variation in plant anti-herbivore properties: The role of habitat quality and rate of dis- turbance. New Phytol. 106(suppl.): 251-263. . 1988. Effects of plant growth rate and leaf lifetime on the amount and type of anti-herbivore defense. Oecologia74: 531-536. , AND J. A. BARONE. 1996. Herbivory and plant defenses in tropical forests. Annu. Rev. Ecol. Syst. 27: 305? 335. , J. P. BRYANT, AND F. S. CHAP?N III. 1985. Resource availability and plant antiherbivore defense. Science 230: 895-899. , M. V. HELLER, R. AIZPRUA, B. ARA?Z, N. FLORES, M. CORREA, M. GUPTA, P. N. SOLIS, E. ORTEGA-BARR?A, L. I. RoiMERo, B. G?MEZ, M. RAMOS, L. CUBILLA-RIOS, T. L. CAPSON, AND T. A. KURSAR. 2003. Using ecological criteria to design plant collection strategies for drug discovery. Front. Ecol. Environ. 1: 421?428. CoNDiT, R., S. AGUILAR, A. HERNANDEZ, R. PEREZ, S. LAO, G. ANGEHR, S. P. HUBBELL, AND R. B. FOSTER. 2004. Tropical forest dynamics across a rainfall gradient and the impact of an El Ni?o dry season. J. Trop. Ecol. 20: 51-72. , P. S. AsHTON, P. BAKER, S. BUNYAVEJCHEWIN, S. GUNATILLEKE, N. GUNATILLEKE, S. P. HUBBELL, R. B. FOSTER, A. ITOH, J. V. LAFRANKIE, H. S. LEE, E. LOSOS, N. MANOKARAN, R. SUKUMAR, AND T YAMAKURA. 2000. Spatial pattern in the distribution of tropical tree species. Science 288: I4l4?1418. , , N. MANOKARAN, J. V. LAFRANKIE, S. P. HUBBELL, AND R. B. FOSTER. 1999. Dynamics of the forest communities at Pasoh and Barro Colorado: Comparing two 50-ha plots. Philosophic. Trans. R. Soc. Lond. B 354: 1739-1748. , S. P. HUBBELL, AND R. B. FOSTER. 1996. Changes in tree species abundance in a Neotropical forest: Impact of climate change. J. Trop. Ecol. 12: 231?256. , N. PITMAN, E. G. LEIGH JR., J. CHAVE, J. TERBORGH, R. B. FOSTER, P. NUNEZ V., S. AGUILAR, R. VALENCIA, G. VILLA, H. C. MULLER-LANDAU, E. LOSOS, AND S. P. HUBBELL. 2002. Beta-diversity in tropical forest trees. Science 295: 666-669. CoNNELL, J. H. 1971. On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. In P. J. den Boer and G. R. Gradwell (Eds.). Dynamics of populations, pp. 298?312. Centre for Agricultural Publication and Documentation,Wageningen, The Netherlands. . 1978. Diversity in tropical rain forests and coral reefs. Science 199: 1302?1310. , AND P. T GREEN. 2000. Seedling dynamics over thirty-two years in a tropical rain forest tree. Ecology 81: 568-584. CORNER, E. J. H. 1967. Ficus in the Solomon Islands and its bearing on the post-Jurassic history of Melanesia. Philosophic. Trans. R. Soc. Lond. B 253: 23-159. DALLING, J. W., H. C. MULLER-LANDAU, S. J. WRIGHT, AND S. P. HUBBELL. 2002. Role of dispersal in the recruitment limitation of Neotropical plant species. J. Ecol. 90: 714?727. DAVIDAR, P., AND J. P. PUYRAVAUD. 2002. Species turnover and ubiquitous tree species in the Western Ghats' rainforest of India. Association for Tropical Biology, annual meeting. Tropical forests: Past, present, future (program and abstracts), p. 25. DAVIES, S. J. 1998. Photosynthesis of nine pioneer Macaranga species from Borneo in relation to life history. Ecology 79: 2292-2308. , AND P. BECKER. 1996. Floristic composition and stand structure of mixed dipterocarp and heath forests in Brunei Darussalam. J. Trop. For. Sei. 8: 542?569. , P. A. PALMIOTTO, P. S. ASHTON, H. S. LEE, AND J. V. LAFRANKIE. 1998. Comparative ecology of 11 sympatric species o? Macaranga in Borneo: Tree distribution in relation to horizontal and vertical resource heterogeneity. J. Ecol. 86: 662-673. DE LIMA FILHO, D. DE A., F. D. DE A. MATOS, I. L. DO AMARAL, J. REVILLA, L. DE S. CO?LHO, J. F. RAMOS, AND J. L. DOS SANTOS. 2001. Inventario flor?stico de florests ombr?fila densa de terra firme, na regiao do Rio Uruca- Amazonas, Brasil. Acta Amaz. 31: 565?579. DE OLIVEIRA, A. A., AND S. A. MOR?. 1999. A Central Amazonian terra firme forest. I. High tree species richness on poor soils. Biodiv. Conserv. 8: 1219?1244. DICK, C. W, K. ABDUL-SALIM, AND E. BERMINGHAM. 2003. Molecular systematic analysis reveals cryptic Tertiary diversification of a widespread tropical rain forest tree. Am. Nat. 162: 691?703. DOBZHANSKY, T 1937. Genetics and the origin of species. Columbia University Press, New York, New York. . 1950. Evolution in the tropics. Am. Sei. 38: 209-221. DuMONT, J. F, S. LAMOTTE, AND F KAHN. 1990. Wetland and upland forest ecosystems in Peruvian Amazonia: Plant species diversity in the hght of some geological and botanical evidence. For. Ecol. Manage. 33/34: 125?139. DussouRD, D. E., AND T EISNER. 1987. Vein-cutting behavior: Insect counterploy to the latex defense of plants. Science 237: 898-901. 468 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit EGGELING, W. J. 1947. Observations on the ecology of the Budongo rain forest, Uganda. J. Ecol. 34: 20?87. EHRENDORFER, F. 1982. Speciation patterns in woody angiosperms of tropical origin. In C. Barrigozzi (Ed.). Mecha- nisms of speciation, pp. 479?509. Liss, New York, New York. EHRLICH, P. R., AND P. H. RAVEN. 1964. Butterflies and plants: A study in coevolution. Evolution 18: 586?608. EwENs, W. J. 1979. Mathematical population genetics. Springer, Berlin, Germany. FiLip, v., R. DiRzo, J. M. MAASS, AND J. SARUKH?N. 1995. Within- and among-year variation in the levels of herbivory on the foliage of trees from a Mexican tropical deciduous forest. Biotropica 27: 78?86. FISCHER, A. G. 1960. Latitudinal variations in organic diversity. Evolution 14: 64?81. FISHER, R. A. 1930. The genetical theory of natural selection. Clarendon Press, Oxford, England. FOSTER, R. B., J. ARCE B., AND T. W?CHTER. 1986. Dispersal and the sequential plant communities in Amazonian Peru floodplain. In A. Estrada and T. H. Fleming (Eds.). Frugivores and seed dispersal, pp. 357?370. Dr W. Junk, Dordrecht, The Netherlands. FRAGOSO, J. M. V., K. M. SILVIUS, AND J. A. CORREA. 2003. Long-distance seed dispersal by tapirs increases seed survival and aggregates tropical trees. Ecology 84: 1998?2006. GENTRY, A. FF. 1982. Neotropical floristic diversity: Phytogeographical connections between Central and South Amer- ica, Pleistocene climatic fluctuations, or an accident of the Andean orogeny? Ann. Mo. Bot. Gard. 69: 557? 593 . 1988a. Changes in plant community diversity and floristic composition on environmental and geographical gradients. Ann. Mo. Bot. Gard. 75: 1-34. . 1988b. Tree species richness of upper Amazonian forests. Proc. Nati. Acad. Sei. 85: 156?159. . 1989. Speciation in tropical forests. In L. B. FFolm-Nielsen, I. C. Nielsen, and FF. Balslev (Eds.). Tropical forests: Botanical dynamics, speciation and diversity, pp. 113?134. Academic Press, London, England. . 1990. (Ed.). Four Neotropical rainforests. Yale University Press, New FFaven, Connecticut. GiLLETT, J. B. 1962. Pest pressure, an underestimated factor in evolution. Syst. Assoc. Publ. 4: 37?46. GILBERT, G. S. 2002. Evolutionary ecology of plant diseases in natural ecosystems. Annu. Rev. Phytopathol. 40: 13? 43. , K. E. HARMS, D. N. HAMILL, AND S. P. HUBBELL. 2001. Effects of seedling size, El Ni?o drought, seedling density, and distance to nearest conspecific adult on 6-year survival of Ocotea whitei seedlings in Panama. Oecologia 127: 509-516. , S. P. HUBBELL, AND R. B. FOSTER. 1994. Density and distance-to-adult effects of a canker disease of trees in a moist tropical forest. Oecologia 98: 100?108. GivNisH, T. J. 1999. On the causes of gradients in tropical tree diversity. J. Ecol. 87: 193?210. GLEASON, H. A. AND A. CRONQUIST. 1964. The natural geography of plants. Columbia University Press, New York, New York. GOOD, R. 1964. The geography of the flowering plants, 3rd edition. Longmans, Green, and Co., London, England. Gosz, J. R., G. E. LIKENS, AND F. H. BORMANN. 1972. Nutrient content of litter fall on the Hubbard Brook Exper- imental Forest, New Hampshire. Ecology 53: 769?784. GRIER, C. C, AND R. S. LOGAN. 1977. Old-growth Pseudotsuga rnenziesii corarnxxnvxies of a western Oregon watershed: Biomass distribution and production budgets. Ecol. Monogr. 47: 373?400. GRUBB, P. J. 1977. The maintenance of species-richness in plant communities: The importance of the regeneration niche. Biol. Rev. 52: 107-145. , J. R. LLOYD, T D. PENNINGTON, AND T C. WHITMORE. 1963. A comparison of montane and lowland rain forest in Ecuador I. The forest structure, physiognomy and floristics. J. Ecol. 51: 567?601. GuARiGUATA, M. R. 1998. Response of forest tree saplings to experimental mechanical damage in lowland Panama. For. Ecol. Manage. 102: 103-111. GuiLLAUMET, J.-L. 1967. Recherches sur la V?g?tation et la Flore de la R?gion du Bas-Cavally (C?te-D'Ivoire). ORS- TOM, Paris, France. HAFFER, J. 1969. Speciation in Amazonian forest birds. Science 165: 131?137. HAMMOND, D. S., AND V. K. BROWN. 1998. Disturbance, phenology and life-history characteristics: Factors influencing distance/density-dependent attack on tropical seeds and seedlings. In D. M. Newbery, H. H. T Prins, and N. D. Brown (Eds.). Dynamics of tropical communities, pp. 51?78. Blackwell Science, Oxford, England. HARMS, K. E., S. JOSEPH WRIGHT, O. CALDER?N, A. HERN?NDEZ, AND E. A. HERRE. 2000. Pervasive density-dependent recruitment enhances seedling diversity in a tropical forest. Nature 404: 493?495. HARRISON, R. D., A. ABDUL HAMID, T KENTA, J. LAFRANKIE, H.-S. LEE, H. NAGAMASU, T NAKASHIZUKA, AND P. PALMIOTTO. 2003. The diversity of hemi-epiphytic figs {Ficus: Moraceae) in a Bornean lowland rain forest. Biol. J. Linn. Soc. 78: 439-455. HAWKINS, B. A., R. FIELD, H. V. CORNELL, D. J. CURRIE, J.-F GU?GAN, D. M. KAUFMAN, J. T KERR, G. G. MITTELBACH, T. OBERDORFF, E. M. O'BRIEN, E. E. PORTER, AND J. R. G. TURNER. 2003. Energy, water, and broad-scale geographic patterns of species richness. Ecology 84: 3105?3117. HooGHiEMSTRA, H., AND T. VAN DER HAMMEN. 1998. Ncogcnc and Quaternary development of the Neotropical rain forest: The forest refugia hypothesis, and a literature overview. Earth Sei. Rev. 44: 147?183. HOWE, H. F. 1977. Bird activity and seed dispersal of a tropical wet forest tree. Ecology 58: 539?550. , E. W. SCHUPP, AND L. C. WESTLEY. 1985. Early consequences of seed dispersal for a Neotropical tree {Virola surinamensis). Ecology 66: 781?791. HUBBELL, S. P. 1979. Tree dispersion, abundance, and diversity in a tropical dry forest. Science 203: 1299?1309. Why Are There So Many Kinds of Tropical Trees? 469 . 1997. A unified theory of biogeography and relative species abundance and its application to tropical rain forests and coral reefs. Coral Reefs 16(suppl): S9?S21. . 1998. The maintenance of diversity in a Neotropical tree community: Conceptual issues, current evidence, and challenges ahead. In F. Dallmeier and J. A. Comiskey (Eds.). Forest biodiversity research, monitoring and modeling: Conceptual background and Old World case studies, pp. 17?44. UNESCO, Paris, France and Parthenon Publishing, Pearl River, New York. . 2001. The unified neutral theory of biodiversity and biogeography. Princeton University Press, Princeton New Jersey. , J. A. AHUMADA, R. CONDIT, AND R. B. FOSTER. 2001. Local neighborhood effects on long-term survival of individual trees in a Neotropical forest. Ecol. Res. 16: 859?875. , R. CONDIT, AND R. B. FOSTER. 1990. Presence and absence of density dependence in a Neotropical tree community. Philosophic. Trans. R. Soc. Lond. B 330: 269?281. , AND R. B. FOSTER. 1986. Commonness and rarity in a Neotropical forest: Implications for tropical tree conservation. In M. E. Soul? (Ed.). Conservation biology: The science of scarcity and diversity, pp. 205?231. Sinauer Associates, Sunderland, Massachusetts. HUSTON, M. A. 1994. Biological diversity: The coexistence of species on changing landscapes. Cambridge University Press, Cambridge, England. HYATT, L. A., M. S. ROSENBERG, T G. HOWARD, G. BOLE, W. FANG, J. ANASTASIA, K. BROWN, R. GRELLA, K. HINMAN, J. P. KuRDZiEL, AND J. GuREVicH. 2003. The distance dependence prediction of the Janzen?Connell hypothesis: A meta-analysis. Oikos 103: 590-602 JANZEN, D. H. 1967. Why mountain passes are higher in the tropics. Am. Nat. 101: 233?249. 1970. Herbivores and the number of tree species in tropical forests. Am. Nat. 104: 501?528. 1974. Tropical blackwater rivers, animals, and mast fruiting by the Dipterocarpaceae. Biotropica 6: 69?103. 1980. Specificity of seed-attacking beetles in a Costa Rican deciduous forest. J. Ecol. 68: 929?952. 1981. Patterns of herbivory in a tropical deciduous forest. Biotropica 13: 271?282. 1984. Two ways to be a tropical big moth: Santa Rosa saturniids and sphingids. Oxf Surv. Evol. Biol. 1: 85-140. . 1988. Ecological characterization of a Costa Rican dry forest caterpillar fauna. Biotropica 20: 120?135. JOHNSON, P. L., AND W. T SWANK. 1973. Studies of cation budgets in the southern Appalachians on four experimental watersheds with contrasting vegetation. Ecology 54: 70?80. JORGENSEN, P. M., AND S. LE?N-YANEZ. 1999. Catalogue of the vascular plants of Ecuador. Missouri Botanical Garden, St. Louis, Missouri. JUNK, W J. 1989. Flood tolerance and tree distribution in Central Amazonian floodplains. In L. B. Holm-Nielsen, I. C. Nielsen, and H. Balslev (Eds.). Tropical forests: botanical dynamics, speciation and diversity, pp. 47?64. Academic Press, London, England. KAY, K. 2002. Evidence for reinforcement of speciation in pollinator-sharing Neotropical Costus. Association for Tropical Biology, annual meeting. Tropical forests: Past, present, future (program and abstracts), p. 54. KAY, K. M., AND D. W. SCHEIMSKE. 2003. Pollinator assemblages and visitation rates for 11 species of Neotropical Costus (Costaceae). Biotropica 35: 198?207. KELLY, C. K., AND M. G. BOWLER. 2002. Coexistence and relative abundance in forest trees. Nature 417: 437?440. , H. B. SMITH, Y M. BUCKLEY, R. CARTER, M. FRANCO, W JOHNSON, T JONES, B. MAY, R. P. ISHIWARA, A. P?REZ-JIM?NEZ, A. S. MAGALLANES, H. STEERS, AND C. WATERMAN. 2001. Investigations in commonness and rarity: A comparative analysis of co-occurring, congeneric Mexican trees. Ecol. Lett. 4: 618?627. KiMURA, M., AND T OHTA. 1973. The age of a neutral mutant persisting in a finite population. Genetics 75: 199? 212. KING, D. A. 1993. A model analysis of the influence of root and foliage allocation on forest production and com- petition between trees. Tree Physiol. 12: 119?135. . 1994. Influence of light level on the growth and morphology of saplings in a Panamanian forest. Am. J. Bot. 81: 948-957. KoHYAMA, T, E. SUZUKI, T PARTOMIHARDJO, T YAMADA, AND T KUBO. 2003. Tree species differentiation in growth, recruitment and allometry in relation to maximum height in a Bornean mixed dipterocarp forest. J. Ecol. 91: 797-806. LATHAM, R. E., AND R. E. RICKLEFS. 1993a. Continental comparisons of temperate-zone tree species diversity. In R. E. Ricklefs and D. Schl?ter (Eds.). Species diversity in ecological communities, pp. 294?314. University of Chicago Press, Chicago, Illinois. , AND . 1993b. Global patterns of tree species richness in moist forests: Energy-diversity theory does not account for variation in species richness. Oikos 67: 325?333. LEE, H. S., S. J. DAVIES, J. V. LAFRANKIE, S. TAN, T YAMAKURA, A. ITOH, T OHKUBO, AND P. S. ASHTON. 2002. Floristic and structural diversity of mixed dipterocarp forest in Lambir Hills National Park, Sarawak. J. Trop. For. Sei. 14: 379-400. , S. TAN, S. J. DAVIES, J. V. LAFRANKIE, P. S. ASHTON, T. YAMAKURA, A. ITOH, T. OHKUBO, AND R. HARRISON. 2004. Lambir Forest Dynamics Plot, Sarawak, Malaysia. In E. Losos and E. G. Leigh Jr. (Eds.). Tropical forest diversity and dynamism, pp. 527?539. University of Chicago Press, Chicago, Illinois. LEIGH, E. G. JR. 1999. Tropical forest ecology. Oxford University Press, New York, New York. LEIGH, E. G. JR., S. LAO, R. CONDIT, S. P. HUBBELL, R. B. FOSTER AND R. PEREZ. 2004. Barro Colorado Island Forest 470 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit Dynamics Plot, Panama. In E. Losos and E. G. Leigh Jr. (Eds.). Tropical forest diversity and dynamism, pp. 451?463. University of Chicago Press, Chicago, Illinois. , AND S. J. WRIGHT. 1990. Barro Colorado Island and tropical biology. In A. H. Gentry (Ed.). Four Neotropical rainforests, pp. 28?47. Yale University Press, New Haven, Connecticut. LiEBERMAN, D., G. S. HARTSHORN, M. LiEBERiMAN, AND R. PERALTA. 1990. Fotest dynamics at La Selva Biological Station, 1969-1985. In A. H. Gentry (Ed.). Four Neotropical rainforests, pp. 509-521. Yale University Press, New Haven, CT. LiEBERiMAN, M., D. LiEBERMAN, G. S. HARTSHORN, AND R. PERALTA. 1985. Small-scale altitudinal variation in lowland wet tropical forest vegetation. J. Ecol. 73: 505?516. LIKENS, G. E., F. H. BORMANN, R. S. PIERCE, J. S. EATON, AND N. M. JOHNSON. 1977. Biogeochemistry of a forested ecosystem. Springer, New York, New York. LIMA DA SILVA, A. S., P. L. B. LISBOA, AND U. N. MACIEL. 1992. Diversidade floristica e estrutura em floresta densa da hacia do Rio Juru?-AM. Bol. Mus. Para. Emilio Goeldi Ser. Bot. 8: 203-258. LoTT, E. J., S. H. BuLLocK, AND J. A. SOLIS-MAGALLANES. 1987. Floristic diversity and structure of upland and arroyo forests of coastal Jalisco. Biotropica 19: 228?235. MACARTHUR, R. H. 1961. Population effects of natural selection. Am. Nat. 95: 195-199. MACARTHUR, R. H., AND E. O. WILSON. 1967. The theory of island biogeography. Princeton University Press, Prince- ton, New Jersey. MACHADO, C. A., E. JOUSSELIN, F KJELLBERG, S. G. COMPTON, AND E. A. HERRE. 2001. Phylogenetic relationships, historical biogeography and character evolution of fig-pollinating wasps. Proc. R. Soc. Lond. B 268: 685? 694. MAKANA, J.-R., T B. HART, D. E. HIBBS, AND R. CONDIT. 2004. Stand structure and species diversity in the Ituri Forest Dynamics Plots: A comparison of monodominant and mixed forest stands. In E. Losos and E. G. Leigh Jr. (Eds.). Tropical forest diversity and dynamism, pp. 159?174. University of Chicago Press, Chicago, Illinois. MARQUIS, R. J., AND H. E. BRAKER. 1994. Plant?herbivore interactions: Diversity, specificity, and impact. In L. A. McDade, K. S. Bawa, H. A. Hespenheide and G. S. Hartshorn (Eds.). La Selva: Ecology and Natural History of a Neotropical Rain Forest, pp. 261?281. University of Chicago Press, Chicago, Illinois. MARQUIS, R. J., I. R. DINIZ, AND H. C. MOR?IS. 2001. Patterns and correlates of interspecific variation in foliar insect herbivory and pathogen attack in Brazilian Cerrado. J. Trop. Ecol. 17: 127?148. MARQUIS, R. J., H. C. MOR?IS, AND I. R. DINIZ. 2002. Interactions among Cerrado plants and their herbivores: Unique or typical? In P. S. Oliveira and R. J. Marquis (Eds.). The Cerrados of Brazil, pp. 306?328. Columbia University Press, New York, New York. MARTINEZ-YRIZAR, A., J. M. MAASS, L. A. PEREZ-JIMENEZ, AND J. SARUKHAN. 1996. Net primary productivity of a tropical deciduous forest ecosystem in western Mexico. J. Trop. Ecol. 12: 169?175. MEDINA, E., AND E. CUEVAS. 1989. Patterns of nutrient accumulation and release in Amazonian forests of the Upper Rio Negro basin. In J. Proctor (Ed.). Mineral nutrients in tropical forest and savanna ecosystems, pp. 217? 240. Blackwell Scientific, Oxford, England. MITCHELL, C. E., AND A. G. POWER. 2003. Release of invasive plants from fungal and viral pathogens. Nature 421: 625-627. MOLES, A. T, AND M. WESTOBY. 2003. Latitude, seed pr?dation and seed mass. J. Biogeogr. 30: 105?128. MoLOFSKY, J., AND C. K. AuGSPURGER. 1992. The effect of leaf litter on early seedling establishment in a tropical forest. Ecology 73: 68-77. , R. DuRRETT, J. DusHOFF, D. GRIFFEATH, AND S. LEVIN. 1999. Local frequency dependence and global coexistence. Theor. Pop. Biol. 55: 270?282. MONK, C. D., AND F P. DAY JR. 1985. Vegetation analysis, primary production and selected nutrient budgets for a southern Appalachian oak forest: A synthesis of IBP studies at Coweeta. For. Ecol. Manage. 10: 87?113. MoRLEY, R. J. 2000. Origin and evolution of tropical rain forests. John Wiley and Sons, Chichester, England M?LLER, M. J. 1982. Selected climatic data for a global set of standard stations for vegetation science. Dr W. Junk, The Hague, The Netherlands. MULLER-LANDAU, H. C, S. J. WRIGHT, O. CALDER?N, AND R. B. FOSTER. 2002. Assessing recruitment limitation: Concepts, methods and case studies from a tropical forest. In D. J. Levey, W R. Silva, and M. Galetti (Eds.). Seed dispersal and frugivory: Ecology, evolution and conservation, pp. 1?17. CABI, Wallingford, England. MuRALi, K. S., AND R. SuKUMAR. 1993. Leaf flushing phenology and herbivory in a tropical dry deciduous forest, southern India. Oecologia 94: 114?119. MURPHY, P. G., AND A. E. LUGO. 1986. Structure and biomass of a subtropical dry forest in Puerto Rico. Biotropica 18: 89-96. NAGYLAKI, T 1976. The decay of genetic variability in geographically structured populations. II. Theor. Pop. Biol. 10: 70-82. NAKASHIZUKA, T 2001. Species coexistence in temperate, mixed deciduous forests. Trends Ecol. Evol. 16: 205?210. NASON, J. D., E. A. HERRE, AND J. L. HAMRICK. 1998. The breeding structure of a tropical keystone plant resource. Nature 391: 685-687. NovoTNY, v., S. E. MILLER, Y BASSET, L. CIZEK, P. DROZD, K. DARROW, AND J. LEPS. 2002. Predictably simple: Assemblages of caterpillars (Lepidoptera) feeding on rainforest trees in Papua New Guinea. Proc. R. Soc. Lond. B 269: 2377-2344. OKSANEN, L., S. D. FRETWELL, J. ARRUDA, AND P. NIEMELA. 1981. Exploitation ecosystems in gradients of primary productivity. Am. Nat. 118: 240-261. Why Are There So Many Kinds of Tropical Trees? 471 ORR, H. A., AND M. TURELLI. 2001. The evolution of postzygotic isolation: Accumulating Dobzhansky?Mullet in- compatibilities. Evolution 55: 1085-1094. OTTO, S. P., AND J. WHITTON. 2000. Polyploid incidence and evolution. Ann. Rev. Genet. 34: 401?437. OzANNE, C. M. P., D. ANHUF, S. L. BOULTER, M. KELLER, R. L. KITCHING, C. K?RNER, F. C. MEINZER, A. W. MITCHELL, T. NAKASHIZUKA, P. L. SILVA DIAS, N. E. STORK, S. J. WRIGHT, AND M. YOSHIMURA. 2003. Biodiversity meets the atmosphere: A global view of forest canopies. Science 301: 183?186. P?GALA, S. W., C. D. CANHAIM, J. SAPONARA, J. A. SILANDER JR., R. K. KOBE, AND E. RIBBENS. 1996. Forest models defined by field measurements: Estimation, error analysis and dynamics. Ecol. Monogr. GG: 1?43. PAINE, R. T. 1966. Food web complexity and species diversity. Am. Nat. 100: 65?75. PARKER, G. G. 1994. Soil fertility, nutrient acquisition and nutrient cycling. In L. A. McDade, K. S. Bawa, H. A. Hespenheide and G. S. Hartshorn (Eds.). La Selva: Ecology and natural history of a Neotropical rain forest, pp. 54?63. University of Chicago Press, Chicago, Illinois. PARTHASARATHY, N., AND R. KARTHIKEYAN. 1997. Biodiversity and population density of woody species in a tropical evergreen forest in Courtallum reserve forest. Western Ghats, India. Trop. Ecol. 38: 297?306. PELISSIER, R., AND B. RIERA. 1993. Dix ans de dynamique d'une for?t dense humide de Guyane Fran?aise. Rev. Ecol. Terre Vie 48: 21-33. PENNINGTON, R. T., M. LEVIN, D. E. PRADO, C. A. PENDRY, S. K. PELL, AND C. A. BUTTERWORTH. 2004. Historical climate change and speciation: Neotropical seasonally dry forest plants show patterns of both Tertiary and Quaternary diversification. Philosophic. Trans. R. Soc. Lond. B 359: 515?538. PETERS, H. A. 2003. Neighbour-regulated mortality: The influence of positive and negative density dependence on tree populations in species-rich tropical forests. Ecol. Lett. 6: 757?765. PHILLIPS, O. L., R NUNEZ V., A. L. MONTEAGUDO, A. P. CRUZ, M.-E. C. ZANS, W G. SANGHEZ, M. YLI-HALLA, AND S. ROSE. 2003. Habitat association among Amazonian tree species: A landscape-scale approach. J. Ecol. 91: 757-775. PIANKA, E. R. 1994. Evolutionary Ecology, 5th edition. Harper Collins, San Francisco, California. PiPERNO, D. R. 1997. Phytoliths and microscopic charcoal from leg 155: A vegetational and fire history of the Amazon basin during the last 75 K. Y. Proc. Ocean Drilling Prog. Sei. Results 155: 411-418. PITMAN, N. C. A., J. TERBORGH, M. R. SILMAN, AND P. NUNEZ V. 1999. Tree species distributions in an upper Amazonian forest. Ecology 80: 2561?2661. , , , , D. A. NEILL, C. E. CER?N, W. A. PALACIOS, AND M. AULESTIA. 2001. Dominance and distribution of tree species in upper Amazonian terra firme forests. Ecology 82: 2101?2117 2002. A comparison of tree species diversity in two upper Amazonian forests. Ecology 83: 3210?3224. , , , A. THOMPSON, R. SHARITZ, B. ALLEN, C. CER?N, D. NEILL, W PALACIOS, AND M. AULESTIA. 2002. Habitat diversity, species turnover, and ubiquitous tree species among the trees of Western Amazonia. Association for Tropical Biology, annual meeting. Tropical forests: past, present, future (Program and Ab- stracts), p. 90. POTTS, M. D., P. S. ASHTON, L. S. KAUFMAN, AND J. B. PLOTKIN. 2002. Habitat patterns in tropical rain forests: A comparison of 105 plots in northwest Borneo. Ecology 83: 2782?2797. PRANCE, G. T (ED.) 1982. Biological diversification in the tropics. Columbia University Press, New York, New York. PRICE, J. P. 2004. Floristic biogeography of the Hawaiian islands: Influences of area, environment and paleogeography. J. Biogeogr. 31: 487-500. PROCTOR, J., J. M. ANDERSON, P. CHAI, AND H. W VALLAGK. 1983. Ecological studies in four contrasting lowland rain forests in Gunung Mulu National Park, Sarawak. I. Forest environment, structure and floristics. J. Ecol. 71: 237-260. PUIG, H., AND J. P. LESGURE. 1981. Etude de la variabilit? floristique dans la r?gion de la piste de Saint-Elie. Bulletin de liaison de groupe de travail sur l'?cosyst?me forestier Guyanais (ECEREX) 3: 26?29. PUTZ, F. E. 1983. Treefall pits and mounds, buried seeds, and the importance of soil disturbance to pioneer trees on Barro Colorado Island, Panama. Ecology 64: 1069?1074. , AND S. APPANAH. 1987. Buried seeds, newly dispersed seeds, and the dynamics of a lowland forest in Malaysia. Biotropica 19: 326?333. PYKE, C. R., R. CONDIT, S. AGUILAR, AND S. LAO. 2001. Floristic composition across a climatic gradient in a Neotropical lowland forest. J. Veg. Sei. 12: 553-566. RAKOTOMALAZA, P. J., AND N. MESSMER. 1999. Structure and floristic composition of the vegetation in the R?serve Naturelle Int?grale d'Andohahela, Madagascar. Fieldiana Zool. 94: 51?96. RAUSHER, M. D. 1981. The effect of native vegetation on the susceptibility o? Aristolochia reticulata (Aristolochiaceae) to herbivore attack. Ecology 62: 1187-1195. RAVEN, P. H. 1999. Foreword. In P. M. Jorgensen and S. Leon-Yanez. Catalogue of the vascular plants of Ecuador, pp. vi-viii. Missouri Botanical Garden, St. Louis, Missouri. REGAL, P. J. 1977. Ecology and evolution of flowering plant dominance. Science 196: 622?629. REICH, P. B., M. B. WALTERS, AND D. S. ELLSWORTH. 1992. Leaf life-span in relation to leaf, plant and stand char- acteristics among diverse ecosystems. Ecol. Monogr. 62: 365?392. RICHARDS, P. W. 1952. The tropical rain forest. Cambridge University Press, Cambridge, England. RICHARDSON, J. E., R. T PENNINGTON, T D. PENNINGTON, AND P. M. HOLLINGSWORTH. 2001. Rapid diversification of a species-rich genus of Neotropical rain forest trees. Science 293: 2242?2245. 472 Leigh, Davidar, Dick, Puyravaud, Terborgli, ter Steege, and Wriglit RicKLEFS, R. E. 1977. Environmental heterogeneity and plant species diversity: A hypothesis. Am. Nat. Ill: 376? 381. . 2004. A comprehensive framework for global patterns in biodiversity. Ecol. Lett. 7: 1?15. , AND D. SCHL?TER. 1993. Species diversity: regional and historical influences. In R. E. Ricklefs and D. Schl?ter (Eds.). Species diversity in ecological communities, pp. 350?363. University of Chicago Press, Chicago, Illinois. RjDLEY, H. N. 1930. The dispersal of plants throughout the world. Lovell Reeve, Ashford, Kent, England. RoHDE, K. 1992. Latitudinal gradients in species diversity: The search for the primary cause. Oikos 65: 514?527. RosENZWEiG, M. L. 1995. Species diversity in space and time. Cambridge University Press, Cambridge, England. RuoKOLAiNEN, K., H. TuoMisTo, J. VoRMisTo, AND N. PiTMAN. 2002. Two biases in estimating range size of Amazonian plant species. J. Trop. Ecol. 19: 935-942. SABATIER, D., AND M.-F. PR?VOST. 1989. Quelques donn?es sur la composition floristique et la diversit? des peuple- ments forestiers de Guyane Fran?aise. Bois For?ts Trop. 219: 31?56. SALOIM?O, R. DE P. 1991. Uso de parcelas permanentes para estudos da vegeta?ao da floresta tropical ?mida. L Municipio de Marab?, Para. Bol. Mus. Para. Emilio Goeldi, S?r. Bot. 7: 543-604. SARRAILH, J. M. 1989. LOp?ration ECEREX. Bois For?ts Trop. 219: 79-97. ScHEiNER, S. M., AND J. M. REY-BENAYAS. 1994. Global patterns of plant diversity. Evol. Ecol. 8: 331?347. ScHEMSKE, D. W. 1981. Floral convergence and pollinator sharing in two bee-pollinated tropical herbs. Ecology 62: 946-954. SCHL?TER, D. 1998. Ecological causes of speciation. In D. J. Howard and S. H. Berlocher (Eds.). Endless forms, pp. 114?129. Oxford University Press, New York, New York. SCHULZE, M. D., AND D. F. WHITACRE. 1999. A classification and ordination of the tree community at Tikal National Park, Peten, Guatemala. Bull. Fla. Mus. Nat. Hist. 41: 169-297. ScHUUR, E. A. G. 2003. Productivity and global climate revisited: The sensitivity of tropical forest growth to precip- itation. Ecology 84: 1165-1170. ScRiBER, J. M. 1973. Latitudinal gradients in larval feeding specialization of the world Papilionidae (Lepidoptera). Psyche 80: 355-373. . 1995. Overview of swallowtail butterflies: Taxonomic and distributional latitude. In]. M. Scriber, Y Tsubaki, and R. C. Lederhouse (Eds.). Swallowtail Butterflies, pp. 3?8. Scientific Publishers, Gainesville, Florida. SHANAHAN, M., AND S. G. COMPTON. 2001. Vertical stratification of figs and fig-eaters in a Bornean lowland rain forest: How is the canopy different? Plant Ecol. 153: 121?132. SHEA, K., AND P. CHESSON. 2002. Community ecology theory as a framework for biological invasions. Trends Ecol. Evol. 17: 170-176. SHMIDA, A., AND M. V. WILSON. 1985. Biological determinants of species diversity. J. Biogeogr. 12: 1?20. SHUTTLEWORTH, W. J. 1988. Evaporation from Amazonian rainforest. Proc. Roy. Soc. Lond. B 233: 321?346. SiLBERBAUER-GoTTSBERGER, I., AND G. EiTEN. 1987. A hectare of Cerrado. I. General aspects of the trees and thick- stemmed shrubs. Phyton 27: 55?91. SKELLAM, J. G. 1951. Random dispersal in theoretical populations. Biometrika 38: 196?218. SoLLiNs, P., C. C. GRIER, F M. McCoRisoN, K. CROMACK, JR., R. FOGEL, AND R. L. FREDRIKSEN. 1980. The internal element cycles of an old-growth Douglas-fir ecosystem in western Oregon. Ecol. Monogr. 50: 261?285. STEBBINS, G. L. 1982. Plant speciation. In C. Barrigozzi, (Ed.). Mechanisms of speciation, pp. 21?39. Liss, New York, New York. SVENNING, J.-C. 1999. Microhabitat specialization in a species-rich palm community in Amazonian Ecuador. J. Ecol. 87: 55-65. . 2001. On the role of microenvironmental heterogeneity in the ecology and diversification of Neotropical rain-forest palms (Arecaceae). Bot. Rev. 67: 1?53. TERBORGH, J. 1973. On the notion of favorableness in plant ecology. Am. Nat. 107: 481?501. . 1985. The vertical component of plant species diversity in temperate and tropical forests. Am. Nat. 126: 760-776. , R. B. FOSTER, AND P. NUNEZ V. 1996. Tropical tree communities: A test of the nonequilibrium hypothesis. Ecology 77: 561-567. , E. Losos, M. P. RiLEY, AND M. BoLA?os RiLEY. 1993. Ptedation by vertebrates and invertebrates on the seeds of five canopy tree species of an Amazonian forest. Vegetatio 107/108: 375?386. TER STEEGE, H., AND D. S. HAMMOND. 2001. Character convergence, diversity, and disturbance in tropical rain forest in Guyana. Ecology 82: 3197-3212. , N. PITMAN, D. SABATIER, H. CASTELLANOS, P. VAN DER HOUT, D. C. DALY, M. SILVEIRA, O. PHILLIPS, R. VASQUEZ, T VAN ANDEL, J. DUIVENVOORDEN, A. A. DE OLIVEIRA, R. EK, R. LILWAH, R. THOMAS, J. VAN ESSEN, C. BAIDER, P. MAAS, S. MORI, J. TERBORGH, PERCY NUNEZ V., H. MOGOLL?N, AND W. MORAWETZ. 2003. A spatial model of tree a-diversity and tree density for the Amazon. Biodiversity and Conservation 12: 2255? 2277. THOMAS, S. C, AND F A. BAZZAZ. 1999. Asymptotic height as a predictor of photosynthetic characteristics in Malaysian rain forest trees. Ecology 80: 1607?1622. TILMAN, D. 1982. Resource competition and community structure. Princeton University Press, Princeton, New Jersey. . 1994. Competition and biodiversity in spatially structured habitats. Ecology 75: 2?16. , AND S. P?GALA. 1993. The maintenance of species richness in plant communities. In R. E. Ricklefs and D. Schl?ter (Eds.). Species diversity in ecological communities, pp. 13?25. University of Chicago Press, Chicago, Illinois. Why Are There So Many Kinds of Tropical Trees? 473 TuoiMisTo, H., K. RuoKOLAiNEN, M. AGUILAR, AND A. SARIMIENTO. 2003. Floristic patterns along a 43-km long transect in an Amazonian rain forest. J. Ecol. 91: 743?756. , , AND M. YLI-HALLA. 2003. Dispersal, environment, and floristic variation of Western Amazonian forests. Science 299: 241-244. UHL, C, AND P. G. MURPHY. 1981. Composition, structure, and regeneration of a tierra firme ?otest in the Amazon basin of Venezuela. Trop. Ecol. 22: 219?237. VAN BAEL, S. A., J. D. BRAWN, AND S. K. ROBINSON. 2003. Birds defend trees from herbivores in a Neotropical forest canopy. Proc. Nad. Acad. Sei. 100: 8304-8307. VAN SCHAIK, C. P., AND E. MIRMANTO. 1985. Spatial variation in the structure and litterfall of a Sumatran rain forest. Biotropica 17: 196-205. V?SQUEZ-MART?NEZ, R., AND O. L. PHILLIPS. 2000. Allpahuayo: Floristics, structure, and dynamics of a high-diversity forest in Amazonian Peru. Ann. Mo. Bot. Gard. 87: 499-527. WARING, R. H., N. C. COOPS, J. L. OHMANN, AND D. A. SARR. 2002. Interpreting woody plant richness from seasonal ratios of photosynthesis. Ecology 83: 2964?2970. WEBB, C. O., AND D. R. PEART. 1999. Seedling density dependence promotes coexistence in Bornean rain forest trees. Ecology 80: 2006-2017. WHITTAKER, R. H., F. H. BORMANN, G. E. LIKENS, AND T. G. SICCAMA. 1974. The Hubbard Brook ecosystem study: forest biomass and production. Ecol. Monogr. 44: 233?254. WILLIG, M. R., D. M. KAUFMAN, AND R. D. STEVENS. 2003. Latitudinal gradients in biodiversity: pattern, process, scale and synthesis. Ann. Rev. of Ecol. Evol. Syst. 34: 273?309. WILLIS, J. C. 1922. Age and area. Cambridge University Press, Cambridge, England. WILLS, C, AND R. CONDIT. 1999. Similar non-random processes maintain diversity in two tropical rainforests. Proc. R. Soc. Lond. B 266: 1445-1452. , , R. B. FOSTER, AND S. P. FFUBBELL. 1997. Strong density- and diversity-related effects help to maintain tree diversity in a Neotropical forest. Proc. Nati. Acad. Sei. 94: 1252?1257. , AND D. R. GREEN. 1995. A genetic herd-immunity model for the maintenance of MHC polymorphism. Immunol. Rev. 143: 263-292. WoLDA, H. 1983. Spatial and temporal variation in abundance in tropical animals. In S. L. Sutton, T C. Whitmore and A. C. Chadwick (Eds.). Tropical rain forest: Ecology and management, pp. 93?105. Blackwell Scientific, Oxford. WONG, M., S. J. WRIGHT, S. P. HUBBELL, AND R. B. FOSTER. 1990. The spatial pattern and reproductive consequences of outbreak defoliation in Quararihea asterolepis, a tropical tree. J. Ecol. 78: 579?588. WRIGHT, D. D., J. H. JESSEN, P. BURKE, AND H. G. DE SILVA GARZA. 1997. Tree and liana enumeration and diversity on a one-hectare plot in Papua New Guinea. Biotropica 29: 250?260. WRIGHT, D. H. 1983. Species-energy theory: An extension of species-area theory. Oikos 41: 496?506. WRIGHT. S. J. 1983. The dispersion of eggs by a bruchid beetle among Scheelea palm seeds and the effect of distance to the parent palm. Ecology 64: 1016?1021. . 2002. Plant diversity in tropical forests: A review of mechanisms of species coexistence. Oecologia 130: 1? 14. , H. C. MULLER-LANDAU, R. CONDIT, AND S. P. HUBBELL. 2003. Gap-dependent recruitment, realized vital rates, and size distributions of tropical trees. Ecology 84: 3174?3185. ZiMiMERMANN, M. H., AND C. L. BROWN. 1971. Trees: Structure and function. Springer, New York, New York.