on the role of microenvironmental heterogeneity in the ecology and diversification of neotropical...

THE B O T A N I C A L R E V I E W Vor. 67 JANUARY-MARCH 2001 No. 1

On the Role of Microenvironmental Heterogeneity in the Ecology and Diversification of Neotropical

Rain-Forest Palms (Arecaceae)

J E N S - C H R I S T I A N SVENNING

Herbarium AA U Department of Systematic Botany

University of Aarhus Bygn. 137, Universitetsparken DK-8000 Aarhus C., Denmark

I.

II.

IIl. IV.

Abstract /Resumen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Large-Scale Pa lm-Environment Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Microenvironmenta l Heterogenei ty and Palm Autecology . . . . . . . . . . . . . . . . . . . . . . 5

A. Canopy Heterogenei ty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1. Light Response as a Dynamic Trait . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2. Individual Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

a. Survival and Damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

b. Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

c. Fecundi ty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

d. Postdispersal Seed Mortal i ty and Germinat ion . . . . . . . . . . . . . . . . . . . . . . . 10

3. Seed Dispersal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 4. Populat ion Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 5. Local Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

a. Canopy Gaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 b. Internal Gap Heterogeneity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

B. Conspecif ics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

1. Fecundi ty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2. Seed and Seedl ing Predation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3. Intraspecif ic Competi t ion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

C. Other Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 l . Compet i t ion from Trees and Shrubs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 2. Competi t ion among Palm Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3. Liana Interference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

4. Interference from Ant Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

D. Litter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Cop ies o f this i s sue [67(1)] may be purchased f rom the N Y B G Press, The N e w York Botanica l Garden, Bronx, NY 10458-5125, U.S.A. P lease inqui re as to pr ices .

The Botanical Review 67( 1): 1-53, January-March 200 I �9 2001 The New York Botanical Garden 1

2 THE BOTANICAL REVIEW

E. Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 1. Individual Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2. Local Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

F. Topography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 1, Individual Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

a. Uplands versus Wet Topographic Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 b. Heterogeneity within Uplands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2. Local Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 a. Patterns within Uplands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 b, Patterns within Floodplains and Swamps . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 c. Nonedaphic Topographic Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

G. Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 1. Individual Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

a. Fecundity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 b. Seed Predation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 c. Later Herbivory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

2. Seed Dispersal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 V. Microenvironmental Heterogeneity and Coexistence . . . . . . . . . . . . . . . . . . . . . . . . . . 29

A. Niche Differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 1. Light Niche Differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 2. Topographic and Edaphic Niche Differences . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3. Seed-Dispersal Niche Differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

B. Mass Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 1. Among-Habitat Mass Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 2. Among-Microhabitat Mass Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

C. Negative Density Dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 VI, Microenvironmental Heterogeneity and Speciation . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

A. Genetic Adaptation to Microenvironmental Heterogeneity . . . . . . . . . . . . . . . . . . 36 B. Sympatric Mechanisms for Reproductive Isolation . . . . . . . . . . . . . . . . . . . . . . . . 36 C. Parapatric Speciation and Edaphic Heterogeneity . . . . . . . . . . . . . . . . . . . . . . . . . 38 D, Parapatric Speciation and Canopy Heterogeneity . . . . . . . . . . . . . . . . . . . . . . . . . . 38 E. Alternative Evolutionary Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

VII. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 VIII. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

IX. Literature Cited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

I. Abstract

Microenv i ronmen ta l he terogenei ty is important in the eco logy and divers i f icat ion o f the

rich pa lm flora that inhabits neotropical rain forests. At s m a l l - - O . l - l O 2 m - - s c a l e s , neo-

t ropical rain fores ts exhibi t h igh he te rogene i ty in numerous envi ronmenta l factors: canopy

condi t ions , conspec i f ics , o ther plants , litter, soil factors, topography, and animal mutual is ts

and pests . These aspects o f mic roenv i ronmenta l he te rogenei ty affect the per formance and

the smal l - sca le dis t r ibut ion o f pa lms in numerous ways, often affect ing di f ferent spec ies dif-

ferently. Notably , even subtle envi ronmenta l variat ion can be o f crucial ecological impor-

tance. Microenvi ronmenta l heterogenei ty promotes the local coexis tence o f palm species by

niche d i f fe rences among the species and probably also by mass effects and negat ive dens i ty

dependence . Sympatr ic species o f the same growth form often differ in terms o f light require-

ments , edaph ic - topograph ic p re fe rences , and poss ib ly also in seed-dispersa l pat terns ,

whereas mass effects are likely to account for the local occurrence o f a share o f the rare spe-

MICROENVIRONMENTAL HETEROGENEITY IN NEOTROPICAL PALMS 3

cies. Density dependence seems to be frequent among large-seeded palms, but its importance needs to be assessed.

Microenvironmental heterogeneity is proposed to be an important diversity-generating factor in the neotropical palm flora through the process of parapatric speciation. This hypothesis is based on the observation that, in species-rich palm genera and species complexes, sympatric species or morphs often differ in edaphic-topographic preferences or in characteristics that confer differing light requirements and in traits that favor reproductive isolation.

Resumen

La heterogeneidad microambiental es importante en la ecologia y la diversificaci6n de la rica flora de palmas que habita los bosques hfimedos neotropicales. A pequefia escala (0.1-102 m), los bosques hfimedos neotropicales exhiben una alta heterogeneidad en numero- sos factores ambientales: las condiciones del dosel, los coespecificos, las otras plantas, la broza, los factores del suelo, la topografia, y los mutualistas y parisitos animales. Estos as- pectos de la heterogeneidad microambiental afectan, de muchas maneras, el funcionamiento y la distribuci6n a pequefia escala de las palmas, afectando a menudo distintas especies de modo diferente. Notablemente, afin una variaci6n ambiental sutil puede ser de importancia ecol6gica crucial.

La heterogeneidad microambiental promueve la coexistencia local de las especies de palmas por diferencias de nichos entre las especies y probablemente tambi6n por "mass effects" y dependencia negativa de la densidad. Las especies simpgtricas con la misma forma de creci- miento se diferencian a menudo en t6rminos de sus requerimientos de luz, preferencias ed~fico-topogrfificas y posiblemente tambi6n en los patrones de dispersi6n de semillas, mien- tras que "mass effects" probablemente explican la ocurrencia local de una parte de las especies raras. La dependencia de la densidad parece ser frecuente entre las palmas con semillas grandes, pero su importancia necesita ser evaluada.

Se propone que la heterogeneidad microambiental por el proceso de especiaci6n parapfitrica es un factor importante en la diversificaci6n de la flora neotropical de palmas. Esta hip6tesis se basa en la observaei6n de queen g6neros de palmas con muchas especies y en complejos de especies, las especies o formas simpfitricas se diferencian a menudo en preferencias edfifico-topogrfificas o en caracteristicas que les atribuyen diferentes requerimientos luminicos y en los rasgos que favorecen el aislamiento reproductivo.

II. Introduction

Palms (Arecaceae) form a systematically isolated group within Monocotyledones (Uhl & Dransfield, 1987; Duvall et al., 1993) and have fossil record going back to the early Upper Cretaceous (Daghlian, 1981 ; Muller, 1981; Uhl & Dransfield, 1987). Today, most palms are restricted to the Tropics and approximately 75% to tropical rain forests (Dransfield, 1978; Uhl & Dransfield, 1987). Palms are an important component of neotropical rain-forest ecosystems, often being abundant in the canopy as well as in the lower strata (e.g., Balslev et al., 1987; Faber-Langendoen & Gentry, 1991 ; Kahn & de Granville, 1992; Peres, 1994; Valencia et al., 1994; Terborgh et al., 1996; Borchsenius, 1997b; Cer6n & Motalvo A., 1997; Romo- leroux et al., 1997; Borchsenius et al., 1998). Palms are notorious for dominating poorly drained or flooded habitats (Beard, 1944; Kahn & de Granville, 1992), but the most abundant canopy tree in wet neotropical upland forests is also often a palm (e.g., Lieberman et al., 1985; Balslev et al., 1987; Faber-Langendoen & Gentry, 1991; Valencia et al., 1994). The


abundance of canopy palms in upland rain forests is a phenomenon largely confined to the Neotropics, Madagascar, New Caledonia, and a few other islands (Gentry, 1988). Palms often dominate the understory of the wetter neotropical rain forests (Kahn et al., 1988; Martinez-Ramos et al., 1988a; Hodel, 1992), where small understory palms can represent 60-70% of the understory plant cover (Kahn & de Granville, 1992) and where a single species, Chamaedorea elegans, can achieve densities of 5933 plants/ha (Hodel, 1992). The high density ofunderstory palms acts as a filter in tree recruitment, limiting tree seedling and sapling abundance (Pifiero et al., 1986; Denslow et al., 1991) and potentially influencing the composition of tree species (of. George & Bazzaz, 1999a, 1999b). Not only do palms have structural significance, their seeds and fruits are important food resources for many invertebrates and vertebrates, and some species are considered keystone species for frugivores (e.g., Uhl & Dransfield, 1987; Kahn & de Granville, 1992; Peres, 1994; Johnson, 1996; Hoch & Adler, 1997). Palms also provide resources to the human inhabitants of the neotropical rain forests, notably food, thatch, materials for construction and handicrafts, and medicine (e.g., Balslev & Barfod, 1987; Balick, 1988; Johnson, 1996; Borchsenius et al., 1998; Svenning & Balslev, 1998), and they have been an important food source for at least l 1,000 years (Roose- velt et al., 1996).

The ecological importance of palms has a long history in the Americas, where palm- dominated vegetation has existed for at least 71 million years (Wing et al., 1993). In South America, pollen records indicate the presence of several types of palms since the Upper Creta- ceous (Muller, 1981), and two extant groups ofneotropical rain-forest palms, Iriartea and le- pidocaryoid palms, have been known in northern South America since the early Tertiary (Daghlian, 1981; Muller, 1981; Rull, 1998). Two other groups of neotropical rain-forest palms, Chamaedorea and geonomoid palms, have questionable occurrences in the early Ter- tiary of southern North America (Daghlian, 1981), and numerous extant genera of rain-forest palms have more or less tentative late Tertiary pollen records from the Neotropics: Astro- caryum, A ttalea, A ttalea (as Maximiliana), Bactris, Chamaedorea, Cryosophila, Desmoncus, Euterpe, Iriartea, Manicaria, Mauritia, and Synechanthus (Hoorn, 1994; Graham & Dilcher, 1998). The long history of palms in neotropical rain forests has allowed them to adapt to the particular conditions there and has allowed other components of the rain-forest ecosystems to adapt to the palms (as Smythe [1989] suggested for the agouti [Dasyprocta spp.]-Astro- caryum spp. interaction).

Any single plant population or plant community is subject to a high degree of small-scale environmental heterogeneity (Fowler, 1988), and much of this variation affects the performance of individual plants (Hutchings, 1997). Such heterogeneity, created by edaphic conditions and disturbance, is known as a determinant of the small-scale distribution of palms as far back as the Upper Cretaceous (Wing et al., 1993). Small-scale environmental heterogeneity in tropical rain forests is greater than in other vegetation types (Ricklefs, 1977; Terborgh, 1985; Svenning, in prep.), and some of the most influential models proposed to explain the plant- species richness in tropical rain forests are based on this heterogeneity: niche differentiation (Ashton, 1969, 1989; Ricklefs, 1977; Rogstad, 1990; Clark & Clark, 1992; D. B. Clark et al., 1998; Kobe, 1999; Svenning, 1999a) and negative density dependence (Janzen, 1970; Con- nell, 1971; Wills et al., 1997). In addition to within-habitat heterogeneity, others have emphasized the importance of habitat diversity (Gentry, 1988; Tuomisto et al., 1995). Although many studies on the ecology of neotropical palms have been published during the last dec- ades, and the importance of small-scale environmental heterogeneity for palm community structure has been emphasized (e.g., Chazdon, 1996), an up-to-date review of the ecology of palms is lacking (Borchsenius et al., 1998).


Here, I review the literature on the influence ofmicroenvironmental heterogeneity on the individual performance, seed dispersal, and local distribution (autecology) of neotropical rain-forest palms. My aims in this review are to demonstrate the importance ofmicroenvironmental heterogeneity in the autecology of these plants and to discuss the implications of this insight for the understanding of the maintenance (synecology) and origin of the high palm- species richness in neotropical rain forests. To set the stage for the discussion of palm-microenvironment relationships, I offer a brief overview of larger-scale palm-environment patterns. Although my focus is on the Neotropics, I include works on the ecology of palms in other parts of the world. For this review, I define rain forest as humid, predominantly ever- green forests and microenvironmental heterogeneity as abiotic and biotic environmental variation at scales from 0.1 to 102 m. For New World palms, nomenclature follows Henderson et al. (1995); for Old World palms names, it follows the original source. 1 use "understory palm" to refer to species with an average adult height of <5 m, "midstory palm" to species with an average adult height of 5-14 m, and "canopy palms" to species of >15 m.

III. Large-Scale Palm-Environment Patterns

Approximately a third of the species in the family--550 species, according to Henderson et al. (1995); 1147, according to Moore (1973)--occur naturally in the Western Hemisphere. Large-scale patterns of palm diversity and abundance in the Neotropics are related primarily to the amount and seasonality of precipitation and to temperature, and secondarily to edaphic conditions. In the Neotropics there is a general increase in plant-species richness with increasing precipitation, reaching an asymptote at 4000-5000 mm/yr (Gentry, 1988; Givnish, 1999), and palm diversity follows this trend. Local palm-species richness and density are higher in the wetter western and central Amazonia than in the drier eastern Amazonia (Kahn et al., 1988; Kahn & de Granville, 1992). In western Amazonia more than 30 species of palms may coexist within just 50 ha of upland forest (Svenning, 199%). In Ecuador, palm-species richness increases with increasing precipitation surplus and absence of a dry season (Borchsenius & Skov, 1997; Skov & Borchsenius, 1997). The low drought tolerance of many palms is dra- matically illustrated by the decline of most palm populations in a Panamanian forest that has been increasingly subject to drought (Condit et al., 1996). Although the Andean palm flora is quite diverse, most neotropical palms occur in the lowlands (Moraes et al., 1995), and in Ec- uador palm-species richness thus increases with mean annual temperature (Borchsenius & Skov, 1997; Skov & Borchsenius, 1997). Apart from rainfall and temperature, neotropical palm diversity is also related to edaphic conditions, palms being particularly species rich on well-drained, unflooded, fertile soils (Gentry, 1988; Kahn et al., 1988; Kahn & de Granville, 1992). Nevertheless, the ecosystems that are most heavily dominated by palms are freshwater swamps (Beard, 1944; de Granville, 1984: Kahn & de Granville, 1992).

IV. Mieroenvironmental Heterogeneity and Palm Autecology

A. CANOPY HETEROGENEITY

Variation in canopy conditions is a major source of microenvironmental heterogeneity in tropical rain forests. This heterogeneity is created by treefalls (e.g., Denslow, 1987; Canham et al., 1990; Clark, 1990; Chazdon et al., 1996; Trichon et al., 1998), branch and palm leaf- falls (Clark, 1990; Trichon et al., 1998), and the size, location, and crown and leaf characteristics of the trees and other plants of which the canopy consists (e.g., Lieberman et al., 1989;


Kabakoff & Chazdon, 1996; Trichon et al., 1998). Canopy heterogeneity can be viewed as comprising heterogeneity in the canopy itself and heterogeneity generated by disturbance of the canopy. Canopy heterogeneity affects light intensity and quality (Kiltie, 1993; Chazdon et al., 1996; Kabakoff& Chazdon, 1996), soil conditions (Ricklefs, 1977; Denslow et al., 1998; Ostertag, 1998), microtopography (Clark, 1990), risk of physical damage (Aide, 1987; Sven- ning, in prep.), pest pressure (Schupp et al., 1989; Braker & Chazdon, 1993; Wenny & Levey, 1998), and seed dispersal (Schupp et al., 1989; Wenny & Levey, 1998). Not only do treefalls create microenvironmental heterogeneity relative to the conditions found in mature closed- canopy microsites, there is also much microenvironmental heterogeneity generated by variation among gaps in size, shape, orientation, depth, mode of tree death, height of surrounding vegetation, season of generation, and age (Grubb, 1977; Collins et al., 1985; Canham et al., 1990; Clark, 1990; Smith et al., 1992; Chazdon et al., 1996; Denslow et al., 1998). Treefall gaps also create a third kind of microenvironmental heterogeneity, namely internal gap heterogeneity, due to the microclimatic gradient from gap edge to gap center (Chazdon, 1986b; Becker et al., 1988; Chazdon et al., 1996), north-south asymmetry (especially away from the equator) in light levels (Canham, 1988), east-west asymmetry in the daily timing of maximum insolation (Smith et al., 1992), the numerous microenvironmental factors that differ among the root, bole, and crown zones oftreefall gaps (Brandani et al., 1988; Nfifiez-Farfan & Dirzo, 1988; Clark, 1990), and the location of remnant standing dead trees (Wenny & Levey, 1998). Finally, there is also strong canopy heterogeneity even in closed-canopy microsites, both as fine-scale (2.5-10 m) horizontal variation in light conditions (Chazdon, 1986b; Smith et al., 1992; D. B. Clark et al., 1996)---due mainly to variation in sunfleck activity (Chazdon & Pearcy, 1991), but also to variation in diffuse radiation (Chazdon, 1986b)--and as strong vertical gradients in light intensity and quality in tropical rain forests (e.g., Chazdon et al., 1996; D. B. Clark et al., 1996).

1. Light Response as a Dynamic Trait

Neotropical palms range from highly shade tolerant and shade requiring to needing high light levels, as will be discussed in the following sections. Still, the light response of rain- forest palms is not a static trait, and individual palms adapt physiologically and anatomically to the ambient light conditions (Hogan, 1988; Broschat et al., 1989). Moreover, shade tolerance will generally decline during ontogeny, as structural costs increase disproportionally with leaf and plant size (Givnish, 1979, 1982, 1988). This happens in the understory palms Geonoma cuneata, G. congesta, andAsterogyne martiana, in which the ratio of total leaf area to plant size declines with plant size (Chazdon, 1986a) and in which light-interception capacity increases at a slower rate than does crown size (Chazdon, 1985). The proportion ofbio- mass in live leaves similarly declines with size in two other understory palms (Pifiero et al., 1982; Ataroff & Schwarzkopf, 1994). Canopy palms also appear to become decreasingly shade tolerant with increasing size (de Granville, 1992), although to differing degrees (Sven- ning, 1999b). This is not always due only to increasing size; it can also be due to changes in crown architecture. Thus Iriartea deltoidea produces flat, monolayered leaves until it reaches a height of about 10 m, at which point it begins to produce bushy leaves (Svenning, pers. obs.), reflecting that in low-light environments a monolayered canopy is optimal because it minimizes self-shading, whereas in high-light environments a multilayered canopy is optimal because it increases photosynthetic capacity and drought resistance (Givnish, 1988; Crawley, 1997a). In sum, it is probably common for within-species shade tolerance to decline with size in palms, although physiological changes (cf. Grime, 1965; Woodward, 1990) can modify or


even reverse this pattern in some species. Clearly, the relationship of a given palm species to canopy heterogeneity needs to be evaluated over its whole ontogeny.

2. Individual Performance

Canopy heterogeneity affects all aspects of individual performance (survival, growth, fecundity, and establishment) of palms, not only through the direct effects of canopy heterogeneity on light conditions but also through its effects on other factors, such as animals, risk of physical damage, and interspecific interference.

a. Survival and Damage

Given the often very dark conditions in the rain-forest understory, one would expect shade to be an important mortality agent for palms (cf. Enright & Watson, 1992), but direct evidence thereof is lacking for neotropical palms. Thus survival of the understory palm Geonoma macrostachys var. macrostachys is not related to crown illumination (Svenning, in prep.), and juvenile survival of the canopy palm Socratea exorrhiza does not differ in mortality among low-and high-canopy microsites (Welden et al., 1991). In the last case the negative result may reflect the weak and non-monotonic relationship between canopy height and understory light availability (D. B. Clark et al., 1996), more than biology.

Contrary to the lack of evidence for direct shade-inflicted mortality, there is much evidence for increased damage and probably subsequent mortality in treefall gaps. Such damage has been found to result from four factors: solarization (cf. Salisbury & Ross, 1992), herbivory, falling canopy debris, and crown invasion.

Solarization, light-dependent inhibition of photosynthesis and chloroplast destruction (Salisbury & Ross, 1992), can be caused by the high light intensities in gaps. As water stress and high light act synergistically to produce solarization (Araus & Hogan, 1994), damage by high light intensities may be more important where there is a distinct dry season and may vary depending on local topographic or edaphic conditions. Shade-adapted understory palms are particularly susceptible to solarization, due partly to their canopy architecture, which maxi- mizes light interception and heating (cf. Chazdon, 1985; de Granville, 1992; Svenning, in prep.) and partly to their shade-adapted leaf anatomy (Lee, 1986). Consequently, seedlings of Geonoma congesta, G. cuneata, and Asterogyne martiana are strongly solarized even at 25.5% full sun (Chazdon, 1986b), and seedlings of Geonoma macrostachys var. macrostachys experience severe solarization when transplanted into the center of 200 m 2 gaps (Svenning, in prep.). Many Malayan understory palms also appear to be highly susceptible to solarization (Dransfield, 1969; Kiew, 1972). Although canopy palms are more tolerant, their susceptibility to solarization varies (Araus & Hogan, 1994).

Palms in the understory of tropical rain forests may experience complete or partial loss of leaves due to herbivory (Mendoza et al., 1987; Oyama & Mendoza, 1990; Oyama & Dirzo, 1991; Cintra, 1997a). The intensity and type ofherbivory may differ according to canopy heterogeneity, although the importance of such herbivory heterogeneity is largely unknown. To my knowledge, the only clear evidence is that provided by Braker and Chazdon's study (1993), which found that a gap-living grasshopper causes greater leaf damage in gap centers than in gap edges and damages some species of understory palms more than others. In the midstory Astrocaryum murumuru var. murumuru (as A. murumuru) seedling, leaf damage by crickets and seedling predation by agoutis, deer, and tapirs is common (Cintra, 1997a) but does not vary according to canopy heterogeneity (Cintra & Horna, 1997). Thus the effect of


canopy heterogeneity on the intensity and type ofherbivory varies among species and probably also among geographical sites due to variation in the herbivore assemblage.

Falling trees, branches, and large palm leaves are an important source of damage to palms in the rain-forest understory (Bullock, 1980; Mendoza et al., 1987; Martinez-Ramos et al., 1988a; Oyama & Mendoza, 1990; Chazdon, 1991a, 1992; Oyama & Dirzo, 1991). Neverthe- less, neotropical understory palms and canopy-palm juveniles are generally well adapted to resist and recover from such damage. Many species resist physical damage from canopy debris by having flexible stems, multiple stems, or meristems at or below ground level (e.g., Vandermeer, 1994) or are able to recover from even severe leaf area loss by increasing leaf production rates (Oyama & Mendoza, 1990; Chazdon, 1991a; Mendoza & Franco, 1992; cf. Ratsirarson et al., 1996) or reducing leaf abscission rates (Mendoza et al., 1987). Thus although 71% of the understory palms in a Nicaraguan rain forest were severely damaged by falling canopy debris during a hurricane, only 4% died (Vandermeer, 1994). Likewise, adult Astrocaryum mexicanum hit by falling trees or large branches are only rarely killed immediately, and subsequent survival is also high (Martinez-Ramos et al., 1988a). Still, even though damage from falling canopy debris may not be immediately lethal, it can cause severe leaf loss (cf. Aide, 1987) or bend stems to the ground (Martinez-Ramos et al., 1988a; Chazdon, 1991a), reducing photosynthesis and thereby reducing growth and survival directly (Men- doza et al., 1987), or probably more commonly reducing the ability to recover from further damage by causing the plant to mobilize its backup resources (cf. Ratsirarson et al., 1996). Defoliation also often halts or reduces reproduction (Mendoza et al., 1987; Ratsirarson et al., 1996; Cunningham, 1997), though not always (Oyama & Mendoza, 1990; Chazdon, 199 la). Thus, although neotropical palms are well adapted to cope with physical damage, such damage may still be demographically important. The risk of physical damage is influenced by canopy heterogeneity, as trees and branches are more likely to fall around preexisting gaps, and areas with high densities of large palms will have a high frequency of falling large palm leaves (Aide, 1987). Juveniles and adults ofGeonoma macrostachys var. macrostachys growing in gap- or building-phase forest experience a higher risk of heavy damage from falling branches and trees than are those growing in mature-phase forest, whereas the risk of major damage by other agents (herbivory, palm leaves, etc.) is unrelated to forest phase (Svenning, in prep.). Canopy heterogeneity thereby indirectly affects mortality, as mortality in this species is related to previous damage (Svenning, in prep.).

Crown invasion of short palms by lianas and pioneer trees may increase mortality in gaps and gap edges (Denslow, 1987; Bemal & Balslev, 1996), whereas the risk of liana infestation of tall palms may be at least as high in mature-phase forest as in gap edges (cf. Enright, 1992).

b. Growth

Canopy heterogeneity is crucial for the growth of many neotropical palms through its effect on light availability, but the effects are highly variable. Kahn (1986) suggested that canopy palms depend on large gaps for recruitment to the adult stage because of high light requirements for stem development, but such a simple relationship is not supported by the available growth studies. In Costa Rica, seedlings of Iriartea deltoidea (as I. gigantea) and Socratea exorrhiza (as S. durissima) decline under a closed canopy but grow when exposed to a minor gap, whereas small juveniles of these species appear to show positive growth even under a closed canopy (Vandermeer et al., 1974). In Welfia regia (as W. georgii), not only seedlings but also juveniles decline under a closed canopy but grow when exposed to a minor or major gap (Vandermeer et al., 1974). In Panama, nine-year-old juveniles of Socratea exor-

MICROENVIRONMENTAL HETEROGENEITY 1N NEOTROPICAL PALMS 9

rhiza had grown to a height of approximately 75 cm at 1% full sunlight, whereas similar-aged juveniles ofAttalea butyracea (as Scheelea zonensis) had only achieved about 25 cm (Araus & Hogan, 1994). Both species had grown to >4 m height at 70% full sunlight (Araus & Hogan, 1994). Thus, the importance of canopy heterogeneity for recruitment of canopy palms is clearly much less simple than Kahn originally proposed (1986). Some species probably do need large gaps for successful recruitment, but others need only minor gaps or are able to grow under even a closed canopy. Annual leaf production ofAttalea butyracea (as Scheelea zonensis) and the midstory palm Oenocarpus mapora (as O. panamanus) increases with increasing light availability in some years, but it does not in another midstory palm, Cryoso- phila warscewiczii (De Steven et al., 1987). Thus not only growth but also leaf production is affected differently by light conditions in different species of large palms. Canopy heterogeneity also affects growth of large palms through its effect on other plants. Thus seedlings of Welfia regia (as W. georgii) decline in major gaps, where they would otherwise grow, when they are overtopped by lianas or pioneer trees (Vandermeer et al., 1974).

Canopy heterogeneity also affects growth of understory palms. Even very shade-tolerant understory palms ( Geonoma congesta, G. cuneata, Asterogyne martiana) are light limited in closed understory microsites, their carbon gain being lineally related to (diffuse) light availability over 0.1-1.2 mol/mE/day, only being positive at daily total photosynthetic flux densities of>0.20 mol/m 2 ~1.0% full sunlight (Chazdon, 1986b). Thus growth responses to canopy heterogeneity are expected, and, accordingly, seedlings of these understory palms have increased growth under gap edges relative to closed understory conditions (Chazdon, 1986b). Seedling growth in another Geonoma, G. macrostachys vat. macrostachys, is affected by even more subtle canopy heterogeneity, growth being higher in medium-illumination (exposed to >1 small or >__1 large, lateral gaps) than in low-illumination (exposed to <l small, lateral gap) microsites (Svenning, in prep.). Light limitation and canopy-generated heterogeneity in growth occur not only in seedlings but also in larger individuals. In G. macrostachys var. macrostachys, growth of individuals of all sizes increases with increasing crown illumination from low to medium to high (direct vertical exposure), and at low crown illumination only seedlings and small juveniles on average maintain zero or positive growth (Svenning, in prep.). Leaf production also increases with light availability in larger individuals of three understory palms, including Geonoma cuneata, but not in two others (De Steven et al., 1987), although the negative results may reflect small sample sizes.

Ramet production ofclonal palms can also be subject to light limitation. In the understory palm Geonoma cf. aspidiifolia, the number of ramets per clump increases with increasing crown illumination in large clumps but not in small clumps, whereas it does not in two midstory palms (Svenning, 2000a), and ramet production in seedlings of the rattan Calamus caesius increases along a gradient of 6-63% canopy openness (Bonal, 1997).

c. Fecundity

Canopy-generated heterogeneity in light availability often affects not only growth but also fecundity of rain-forest palms, and again responses vary. Light limitation is especially clear for understory palms (Chazdon, 1986c; Cunningham, 1997; Svenning, in prep.). In Geonoma macrostachys var. macrostachys the probability of being fertile, the number ofinflorescences produced by fertile individuals, and inflorescence size all increase with increasing crown illumination (Svenning, in prep.). Still, light limitation of fecundity is apparently not ubiquitous (De Steven et al., 1987; Svenning, 2000a), although, again, the negative results may refect the limitations of the studies. The effect of canopy-generated light heterogeneity on growth


may also exhibit temporal variability. Thus individuals ofAstrocaryum mexicanum located in a gap showed an increased reproductive effort compared with individuals in a mature forest in 1977, but not in 1978 (Pifiero et al., 1982). Canopy-generated heterogeneity in light availability may also affect the fecundity ofmidstory palms. Light limitation has been documented for Oenocarpus mapora and Astrocaryum murumuru var. urostachys (as A. urostachys), but not for Phytelephas tenuicaulis or Astrocaryum standleyanum (De Steven et al., 1987; Svenning, 2000a). Although there are no detailed data on light limitation of fecundity in canopy palms, at least the onset of reproduction is influenced by canopy-generated heterogeneity in light availability. Both Oenocarpus bataua var. bataua (as Jessenia bataua) and Euterpe oleracea begin to reproduce at a much lesser height when growing under well-lit conditions (Strudwick & Sobel, 1988; Pedersen & Balslev, 1992). Light limitation on the fecundity of mature individuals is as yet largely unstudied. Seed production by adults of the subtropical canopy palm Rhopalostylis sapida increases with increasing canopy openness (Enright, 1992). Canopy- generated heterogeneity in light availability may also affect sex expression in palms that are capable of producing exclusively male or female inflorescences. In the midstory palmAttalea funifera female sex expression is favored by smaller and fewer neighboring trees, a pattern that probably reflects light limitation, because the cost of producing a female infructescence relative to a male inflorescence is 13:1 (Voeks, 1988).

Although canopy heterogeneity affects palm fecundity mainly through its effect on light levels, it may also act through other factors, such as predispersal seed predation (Martinez-Ramos et al., 1988b) and liana interference (Enright, 1992). Canopy heterogeneity may also influence pollination by affecting the behavior of the animal pollinators (cf. Kiltie, 1993) and thereby fecundity, which in many rain-forest palms is probably pollination limited (Cunningham, 1996).

d. Postdispersal Seed Mortality and Germination

Canopy heterogeneity can affect seed mortality and possibly also germination in neotropical rain-forest palms. Schupp and Frost (1989) suggested that higher seed and seedling mortality of large-seeded species in gaps due to rodent predation may be typical ofneotropical wet forests (but see Forget [1997] and below). Increased rodent predation in gaps can occur because rodents hide in the canopy debris and climber tangles there (Schupp et al., 1989). Seed predation of Welfia regia (as W. georgii) fits this hypothesis, for it is much higher in gaps than in closed understory (Schupp & Frost, 1989). In contrast, postdispersal seed predation, mainly by small mice, in Astrocaryum mexicanum is higher in mature forests than in building-phase forests and is lowest in gaps (S~inchez-Codero & Martinez-Gallardo, 1998), and seeds of As- trocaryum murumuru var. murumuru (as A. murumuru) survive longer under adults close to gaps than far from gaps (Cintra & Homa, 1997). Thus the hypothesis of Schupp et al. (1989) clearly only holds for some large-seeded palms and probably only in some forests. Schupp et al. (1989) also suggested that small seeds, which are not predated by rodents, may have higher seed survival in gaps, because increased light and lower humidity there inhibit plant patho- gens, but this hypothesis remains untested for palms. Internal gap heterogeneity may also affect seed mortality but does not do so in Astrocaryum murumuru var. murumuru (as A. murumuru), where seed survival does not differ between gap crown zone, gap bole zone, and understory microsites (Cintra & Homa, 1997). Among-gap heterogeneity in gap size does not affect seed survival in this species either (Cintra & Horna, 1997). The importance of canopy heterogeneity for seed germination in neotropical rain-forest palms has hardly been studied. In Geonoma macrostachys var. macrostachys germination and initial seedling survival do not differ between low and medium crown-illumination microsites (Svenning, in prep.). Given

MICROENV1RONMENTAL HETEROGENEITY IN NEOTROPICAL PALMS 11

the importance of canopy heterogeneity for seed mortality and germination in tropical rain- forest plants in general (Raich & Khoon, 1990; Ellison et al., 1993; V~tsquez-Y~nes & Orozco-Segovia, 1993; Schupp, 1995; Fredeen & Field, 1996), it is clear that there is a dearth of studies on these aspects of the ecology of neotropical rain-forest palms.

3. Seed Dispersal

Seed dispersal is an important component of plant-population ecology, because it determines the conditions under which the developing plant must grow (Schupp & Fuentes, 1995) and because the probability that a plant of a certain species will mature in a given spot depends on the joint probabilities of arrival and subsequent survival (Schupp et al., 1989). The vast majority of palms have animal-mediated seed dispersal, most commonly by birds and secondarily by mammals (Zona & Henderson, 1989). The particular dispersal syndrome will affect the local spatial occurrence of seed arrival of a given palm species. Here I shall show that animal-mediated seed dispersal probably is a powerful factor linking palm recruitment to microsites characterized by particular canopy conditions, depending on which palm species and animal dispersers are involved.

Canopy heterogeneity has been suggested to have a strong influence on seed dispersal (Schupp et al., 1989). According to Schupp et al. (1989), large seeds, dispersed mainly by tou- cans, guans, and monkeys, will arrive preferentially in closed-canopy microsites, as these animals prefer canopy trees as perches (and for movement) (Forget & Sabatier, 1997). Contrarily, small-animal-dispersed seeds, mainly dispersed by small birds and bats, should arrive preferentially to gap edges of maturing gaps and to a smaller degree inside maturing gaps, as these animals are most active in gaps but hide from predators in adjacent, less- exposed sites (Schupp et al., 1989). Only one study has directly tested these suggestions in the Neotropics. Seeds of the small-seeded tree Ocotea endresiana (Lauraceae) are dispersed preferentially to gap edges below standing dead trees (perches) by bellbirds (Procnias tricaruncu- lata), whereas four other bird species disperse the seeds at random with respect to canopy heterogeneity (Wenny & Levey, 1998). Thus seed dispersal in this species partly conforms to the suggestions by Schupp et al. (1989). Notably, seedlings from bellbird-dispersed seeds have lower mortality and better growth than do other seedlings (Wenny & Levey, 1998).

Although canopy heterogeneity probably affects the spatial occurrence of recruitment through its effect on seed dispersal in many palms, little information exists regarding this topic. Seedlings of the large-seeded canopy palm Oenocarpus bataua var. oligocarpa (as Jessenia bataua ssp. oligocarpa) appear to avoid gaps, occurring preferentially below large canopy trees used as perching sites by parrots feeding on nearby fruiting Oenocarpus bataua (Sist & Puig, 1987); that is, again, in accordance with Schupp et al. (1989). Given the high light requirements of this palm for stem growth (Sist & Puig, 1987; Svenning, 1999b), the favorability of this directed seed dispersal is questionable, although it is possible that Oenocar- pus bataua has best establishment under a relatively closed canopy (cf. Kahn & de Granville, 1992; Svenning, 1999b, 2000b). If seeds of small-seeded, bird-dispersed understory palms, such as species of Geonoma, Chamaedorea, Bactris, Asterogyne, Calyptrogyne, and Pres- toea (cf. Zona & Henderson, 1989; Vandermeer, 1993; Svenning, in prep.), are dispersed preferentially to gap edges and maturing gaps, as suggested by Schupp et al. (1989), this would probably indeed be favorable, beeause such mierosites are optimal for the growth of many of these palms (as discussed above). A likely example of gap-directed seed dispersal is known from the Paleotropics, where the palm civet (Paradoxurus hermaphroditus), which is an important disperser of many Asian rain-forest palms, usually defecates in gaps and thereby


probably provides gap-directed seed dispersal (Zona & Henderson, 1989). At least for many rattans, such dispersal will be advantageous (Dransfield, 1978; Bogh, 1996a).

Oilbirds (Steatornis caripensis) are probably efficient long-distance dispersers of palms (Snow & Snow, 1978; Herzog & Kessler, 1997), and it has been suggested that they affect the seedling distribution of Oenocarpus bataua var. bataua (as Jessenia bataua), dispersing seeds to their preferred feeding sites, in this case below Dacryodes sp. trees (Burseraceae) (Snow & Snow, 1978), Thus dispersal of large-seed palms may not only be preferentially toward closed canopy microsites but also be toward closed-canopy microsites below certain canopy-tree species (see section IV.G.2); in the case of oilbirds, to palms or trees of the fami- lies Burseraceae and Lauraceae, their preferred feeding trees (Herzog & Kessler, 1997). Given the diversity of tropical canopy trees, such behavior may generate tremendous microenvironmental heterogeneity for palm recruitment.

4. Population Dynamics

The previous sections show that canopy heterogeneity often affects the demographic rates of rain-forest palms, and one would therefore expect canopy heterogeneity to affect their population dynamics, too. The fact that canopy heterogeneity often affects palm distribution and abundance, as will be discussed in the next section, provides indirect evidence thereof, but several studies have also directly evaluated the importance of canopy heterogeneity for palm population dynamics (Martinez-Ramos et al., 1988b; Alvarez-Buylla & Slatkin, 1994; Frangi & Lugo, 1998; Svenning, in prep.). Although more studies are needed, it is evident from even these few studies that canopy heterogeneity can indeed play a central role in palm population dynamics.

The canopy palm Prestoea acuminata (as P. montana) increased strongly in basal area as well as in density in the six years following a hurricane, and its regeneration correlated strongly and positively with hurricane-induced canopy damage at the scale of 25 m 2 subplots, whereas topography and distance from the river were less important (Frangi & Lugo, 1998). Thus the population dynamics of this species seem to be controlled by the availability of open canopy conditions, although they are also influenced by other factors.

The importance of canopy-generated heterogeneity for population persistence has been evaluated for two neotropical rain-forest understory palms: Astrocaryum mexicanum (Mar- tinez-Ramos et al., 1988b; Alvarez-Buylla & Slatkin, 1994), and Geonoma macrostachys var. macrostachys (Svenning, in prep.). Martinez-Ramos et al. (1988b) investigated the importance of treefall gaps in the demography ofA. mexicanum. Transition matrix models were produced for mature forest, for a young gap, and for a somewhat older gap, and Martinez- Ramos et al. (1988b) concluded that the asymptotic population growth rate, t , differed among the three forest phases. However, a reanalysis of the data found that none of the population growth rates could be distinguished from population equilibrium, t = 1 (Alvarez-Buylla & Slatkin, 1994). Although this may reflect population-level indifference to canopy heterogeneity (but see Pifiero et al., 1986), it more probably reflects sample-size limitations or inade- quacy of the gap-understory design (cf. Lieberman et al., 1989; and the next example).

Also using matrix analysis, Svenning (in prep.) investigated the importance of crown illumination (cf. Clark & Clark, 1992; D. B. Clark et al., 1993; Davies et al., 1998) and forest phase in the demography of Geonoma macrostachys var. macrostachys. Population matrix models were analyzed for all individuals and for different subsets of individuals: with low crown illumination, with medium/high crown illumination, in gap/building-phase plots, or in mature-phase plots (Svenning, in prep.). The asymptotic population growth rate for the whole


population was 0.999 (95% confidence interval 0.983-1.014); that is, the population was at or very close to equilibrium. Under mature-phase conditions ~, was significantly higher and under gap/building-phase conditions ~, was significantly lower than for the whole population, although neither could be distinguished from 1. Thus, like the previous study, the gap- understory design yielded ambiguous results. But the crown illumination-based analyses had much clearer results, reflecting that the distinction between very shaded microsites and other microsites, rather than between very well lit microsites and other microsites as is usual, assessed for each individual crown, is a more relevant design for investigating the importance of canopy heterogeneity for shade-adapted understory plants. Geonoma macrostachys var. macrostachys could not persist under permanently low illumination (~, = 0.956, 95% confidence interval = 0.941-0.972), but it would increase in density under permanent medium/high illumination (~, = 1.080, 95% confidence interval = 1.051-1.104) (Svenning, in prep.). Thus canopy heterogeneity plays a central role in the population dynamics of this shade-tolerant understory palm, acting through severe light limitation of growth and fecundity in heavily shaded microsites (Svenning, in prep.). As Geonoma spp. are probably among the most shade-tolerant palms (cf. Chazdon, 1986b; Svenning, in prep.), canopy-generated heterogeneity in light availability is probably important for population performance of most rain- forest palms. Thus, local population instability of the canopy palm lriartea deltoidea at a site in Amazonian Ecuador (Svenning & Balslev, 1997) may be a result of temporal heterogeneity in canopy conditions.

5. Local Distribution

The previous sections show how canopy heterogeneity can, and probably often does, affect the demography and dispersal ofneotropical rain-forest palms and that it can be of crucial importance in their population dynamics. Here I shall show that the diverse and often strong effects of canopy heterogeneity, in particular through its effect on light availability, have the ability to affect the local distribution of neotropical palms. Because the demographic effects are highly variable among species, both in strength and quality, a similar diverse range of distribution patterns results. An important point is that temporal and spatial mass effects can mask strong demographic effects of canopy heterogeneity. Regarding the major palm growth forms, midstory palms and possibly climbing palms may tend to be dependent on open- canopied conditions, at least for recruitment, whereas understory palms and canopy palms have very diverse responses to canopy openness.

a. Canopy Gaps

1. Canopy palms. As already discussed, Kahn's hypothesis (1986) that canopy palms depend on large gaps for recruitment to the adult stage is not generally supported by the available growth studies. What, then, do the local distribution patterns of palms tell us regarding the gap dependency of large palms? Kahn and de Granville (1992) asserted that the canopy and midstory palms Attalea maripa (as Maximiliana maripa), A. speciosa (as Orbignya phalerata), Oenocarpus bacaba, O. distichus, and Socratea exorrhiza all occur only under canopy gaps, and Sist and Puig (1987) suggested that large juveniles of another canopy palm, Oenocarpus bataua var. oligocarpa (as Jessenia bataua ssp. oligocarpa), are associated with gaps. Further indications of the gap dependency of canopy palms are provided from outside the Neotropics (Weiner & Corlett, 1987; Enright & Watson, 1992). Rather contradictory results came from quantitative investigations of the canopy palms in an Ecuadorian upland for-


est (Svenning, 1999b, 2000b). There, small juveniles of Iriartea deltoidea, Oenocarpus bataua var. bataua, Attalea cf. maripa, and all canopy palms combined preferentially occur in mature-phase forest, whereas for larger sizes there is no general relationship of canopy palms to forest phase (Svenning, 1999b). At the scale of single understory plants, small juveniles of Oenocarpus bataua var. bataua occur irrespective of understory light availability, whereas small juveniles of/. deltoidea occur preferentially in microsites with somewhat elevated light levels (Svenning, 2000b), although still in mature-phase forest (Svenning, 1999b). As the juveniles grow in size, Oenocarpus bataua var, bataua becomes increasingly restricted to gap microsites, resulting in preadults and small adults being largely restricted to growing below major canopy gaps (Svenning, 1999b, 2000b). Although/. deltoidea juveniles also become increasingly associated with gaps, the majority of even 10-20 m tall individuals are found below a closed canopy (Svenning, 1999b). In sum, this more quantitative investigation shows that canopy palms range from only weakly to strongly gap dependent and that shade tolerance sometimes changes over ontogeny. It has also been suggested that clonal canopy palms are generally restricted to open-canopied forests and other habitats with permanent high light availability because the additional stems will be unable to develop due to shade in more closed forests (Kahn & de Castro, 1985), but this hypothesis still needs evaluation.

2. Midstory palms. Kahn and de Granville (1992) suggested that midstory palms generally tend to be most common in floodplain forest. Because floodplain forest is open canopied, this would occur if midstory palms were relatively light demanding, and a number of studies do show that canopy openness has a positive effect on the occurrence of midstory palms. Al- though small juveniles of Cryosophila guagara are not clumped in gaps, large individuals clump at a scale corresponding to the mean treefall gap size, and Cryosophila guagara therefore appears to depend on such gaps for trunk building, though not for seedling establishment (Richards & Williamson, 1975). In agreement with this interpretation, the similar Cryoso- phila warscewiczii in Costa Rica has seedlings and juveniles preferentially in larger gaps (relative to smaller gaps) in old-growth forest (as C. albida) (Brandani et al., 1988) and has abundant juveniles in secondary, but not in old-growth, forests (Guariguata et al., 1997), and in Panama it also appears to be most abundant in secondary forest (De Steven et al., 1987). Small juveniles ofAstrocaryum murumuru var. urostachys occur irrespective of small-scale variability in microsite light exposure, whereas larger juveniles appear to occur mainly in the more exposed microsites (Svenning, 2000b). This pattern confirms the result of another study at the same site that showed recruitment of large juveniles of the same species and of another clonal midstory palm, Phytelephas tenuicaulis, to occur mainly in low-canopy microsites (Svenning, 2000a). A high light demand has been even more convincingly shown for another midstory palm, Oenocarpus mapora, the density of seedlings and juvenile and adult clumps of this species being low in mature forest, intermediate in secondary forest, and very high in exposed, wind-thrown forest (De Steven, 1989; cf. Leigh et al., 1993). In this species the association with canopy heterogeneity arises due to a positive relation between higher light availability and fecundity recruitment, and probably ramet production (De Steven, 1989). Thus all evidence points toward relatively high light requirements for recruitment in midstory palms, although they may be able to persist as adults, at least for some time, in more heavily shaded microsites. At least for the more robust species, this may be due partly to relatively high light requirements for stem building (Svenning, 2000a). If midstory palms are more restricted to open-canopied habitats than are canopy palms, it could reflect the need for the midstory palms to reproduce below the canopy.

3. Understory palms. How does canopy heterogeneity affect the distribution of neotropical understory palms? As already discussed, understory palms are vulnerable to solarization


when exposed to high light, but they are also unable to maintain self-sustainable populations in the abundant heavily shaded microsites (at least Geonoma macrostachys var. macrostachys [Svenning, in prep.]). Thus it may be expected that understory palms avoid large gaps as well as heavily shaded microsites. An investigation into palm-microenvironment relationships in Amazonian Ecuador showed rather conflictingly that forest phase at the 400 m 2 scale only affects the distribution of 2 of the 12 most common understory palms in an Amazonian upland forest (Svenning, 1999a). These two species are favored by a low canopy relative to a high canopy. Although this result could be due to the coarse spatial scale and the poor relationship between canopy height and understory light availability (cf. D. B. Clark et al., 1996), another investigation at the same site at the scale of single understory plants and using a fine-scale index of crown illumination gave a similar result (Svenning, 2000b; cf. Richards & Williamson, 1975; Welden et al., 1991). Rather than indicating that most understory palms can maintain themselves under any canopy condition, these results probably reflect the ability of these palms to persist, though gradually declining, in unfavorable microsites, as found for Geo- noma macrostachys var. macrostachys (Svenning, in prep.). There is no strong spatial corre- spondence with canopy heterogeneity in this species (Svenning, 1999a, 2000b), even though canopy heterogeneity is essential for its population performance (Svenning, in prep.). Still, more circumstantial evidence and anecdotal information suggest that the small-scale distribution of understory palms sometimes can be affected by canopy heterogeneity but that the strength and type of the relationship vary among the species. In an Andean rain forest, two of five understory palms, Chamaedorea linearis and C. pinnatifrons, occur more abundantly, as seedlings and juveniles as well as adults, in selectively logged old-growth forest than in un- touched old-growth forest (Svenning, 1998). Especially, the seedling density is increased by human disturbance, and the pattern probably reflects a positive response to a moderate opening of the canopy (Svenning, 1998). A positive effect of some canopy openness has also been reported for several other neotropical species (Pifiero et al., 1986; Kahn & Mejia, 1987; Oyama & Dirzo, 1991 ; Kahn & de Granville, 1992; Scariot, 1999), and Eugeissona triste has been observed to tolerate full sun and form dense colonies in selectively logged forests in Southeast Asia (Dransfield, 1978). Contrary to these patterns, a negative effect of canopy opening has been indicated for other neotropical species (Chazdon, 1986c), and a negative effect of canopy opening has also been noted for most Malayan understory palms (Dransfield, 1969; Kiew, 1972). Thus understory palms cover the whole range, from very shade-tolerant species to light-demanding species. Still, most understory palms are probably negatively affected by very open canopy conditions, such as those in large treefall gaps. Moderate canopy openness, small gaps, or gap edges will generally favor performance and recruitment but probably only affect the distribution of a minority of the most light-responsive species, except where canopy conditions are more permanently open, such as riverbanks or wind-exposed ridges. Thus, Kahn and de Granville (1992) suggested, high densities ofHyospathe elegans and Bactris gastoniana on a wind-exposed crest in French Guiana are due to a high frequency of treefall gaps.

4. Climbing palms. Kahn and de Granville (1992) also suggested thatDesmoncus spp., the only important climbing palms in the Neotropics, are associated with open canopy conditions. In disturbed forests Desmoncus polyacanthos has more abundant small individuals in early- successional upland forests than in mid-successional upland forests or floodplain forests, whereas larger individuals occur mainly where the canopy has more overlapping crowns (Troy et al., 1997). These patterns indicate that establishment is promoted by an open canopy and that later growth depends on trellis availability. This interpretation follows what is generally known for the paleotropical rattans (Putz, 1990; Tomlinson, 1990). Most likely, the dif-


ferent species of Desmoncus also differ in their light dependency and trellis requirements (cf. Bagh, 1996b), but this has not been studied. The slender Desmoncus mitis var. mitis may be more shade tolerant because of its smaller size (of. Chazdon, 1986c; Givnish, 1988) and may be able to use more weak trellises than can larger species, such as Desmoncuspolyacanthos.

b. Internal Gap Heterogeneity

Another aspect of canopy heterogeneity, internal gap heterogeneity, can also affect the spatial distribution of palms. Some species recruit preferentially in the root zone, others in the crown zone, and some have no preferences (Richards & Williamson, 1975; Brandani et al., 1988; Ntifiez-Farfan & Dirzo, 1988). Notably, seedlings and juveniles ofEuterpe precatoria var. longevaginata (as E. macrospadix) are overrepresented only in the crown zone of gaps created by Pentaclethra macroloba, Caesalpiniaceae (Brandani et al., 1988), probably reflecting advance regeneration under the open crown of this species (cf. Kabakoff & Chazdon, 1996). Thus microenvironmental heterogeneity associated with the different gap zones often, but not always, affects the occurrence ofneotropical palms, and the effect varies among species and can even depend on the species of the fallen tree.

B. CONSPECIFICS

Small-scale heterogeneity in the density and spatial location of conspecifics, from seeds to adults, causes spatial variability in the recruitment of some neotropical rain-forest palms by affecting fecundity, seed predation, germination, and seedling or juvenile performance. Still, the population-level consequences may often be slight, although this needs further study.

1. Fecundity

Although the topic is potentially important, only one study has documented the direct effect of conspecific neighbors on fecundity in neotropical rain-forest palms. In the protandrous, bat-pollinated understory palm Calyptrogyne ghiesbreghtiana, the number of fruits initiated per inflorescence is weakly, but positively, correlated with the local sex ratio (mean proportion of male-phase inflorescences in the site per night of female flowering in the focal plant) but is unrelated to the number of male inflorescences (Cunningham, 1995). These results indicate that competition among female-phase inflorescences for pollen is important in this species (Cunningham, 1995). How often intraspecific competition for pollination causes pollen limitation in neotropical palms remains to be seen. Many species are protandrous or protogynous and some are dioecious (notably Chamaedorea spp.), suggesting that pollen limitation could be widespread. Negative density effects on fecundity, such as through increased flower or predispersal seed predation, may also occur, but they have not been documented.

2. Seed and Seedling Predation

Seed and seedling predation may cause negative density-dependent recruitment in some neotropical palms, but the relationship can vary in time and possibly also among sites. Density-dependent recruitment appears to be associated mainly with large-seeded palms, although this conclusion is weakened by the dearth of studies evaluating the phenomenon in small-seeded species.


Large-seeded neotropical palms, notably of the genera Astrocaryum and Attalea, suffer high rates of seed predation by bruchid beetles (palm bruchids, Pachymerinae [Johnson et al., 1995]) and mammals (cf. Henderson, 1995), and density- and distance-dependent seed predation has been documented by detailed investigations of four species in this group: Attalea butyracea (as Scheelea zonensis and S rostrata) (Wilson & Janzen, 1972; Wright, 1983), A. maripa (Fragoso, 1997), Astrocaryum murumuru var. murumuru (as A. murumuru and A. macrocalyx)(Terborgh et al., 1993; Cintra, 1997b), and A. mexicanum (Sfinchez-Codero & Martinez-Gallardo, 1998). Based on these studies, recruitment of Attalea-Astrocaryum palms can be generalized to be strongly reduced where seed densities are high, especially up to 25 m or even 100 m away from adults. Thus both local seed density and distance to nearest adults have strong effects on the intensity of seed predation in these palms. Density- and distance- dependent predation on palm seeds is not limited to the Attalea and Astrocaryum genera. Thus seed mortality of Welfia regia (as W. georgii) is much higher in the understory beneath fruiting adults than 10 m away from the adults, probably due to rodent seed predation (Schupp & Frost, 1989). Indications of density- or distance-dependent seed predation have also been mentioned for a large-seeded, paleotropical rain-forest palm (Ratsirarson et al., 1996). Con- versely, predator satiation may sometimes cause seed survival to be positively density dependent, as found for seed predation ofAstrocaryum murumuru var. murumuru by bruchids (but not by mammals) (Cintra, 1997b) and as also suggested for Astrocaryum mexicanum. Pi- fiero and Sarukhfin (1982) stated that in this species squirrel seed predation is so high in parts of the forest where A. mexicanum has low densities that the species is declining there, whereas squirrels become satiated where fruit production is high, allowing a stable or increasing population in those areas. Nevertheless, this suggestion is not supported by the analyses of Coch- ran and Ellner (1992).

Density or distance dependence in seedling survival has received less attention than seed predation but has been documented in at least one case (Cintra & Homa, 1997; cf. De Steven, 1989). In one year, seedling survival ofAstrocaryum murumuru var. murumuru is higher far from adults than beneath them, presumably to due mammal herbivory, but no spatial pattern could be detected in the following year (Cintra & Horna, 1997). Although pest-mediated density or distance dependence has been documented for large-seeded palms, the population- level consequences remain to be evaluated. The high abundance of Attalea butyracea (as Scheelea zonensis) on small islands that have lost their mammalian seed predators, relative to areas with an intact seed-predator fauna, may indicate that density-dependent pest attacks can indeed limit the density of large-seeded palms, although differences in the disturbance regime may be more important in this case (Leigh et al., 1993).

In contrast, no negative distance or density dependence could be documented for seedling performance or recruitment of the small-seeded Geonoma macrostachys var. macrostachys, and seedling recruitment even correlated positively with initial seedling density (Svenning, in prep.). Although this pattern does not support density-dependent population regulation, a negative density effect may be overridden by clumping due to site favorability or seed dispersal (cf. Fowler, 1988).

3. Intraspecific Competition

Apart from pest pressure, intraspecific resource or interference competition may also be important for regulating palm densities (cf. Crawley, 1997b). Overall, intraspecific competition does seem to occur and sometimes even to be density limiting in neotropical rain-foresl palms (Yeaton, 1979; Sterner et al., 1986), although it seems likely to be important only in


abundant species, because vegetative competition in plants is generally limited to near neighbors (Armbruster, 1995; Crawley, 1997a). Still, density limitation is not strong even in some abundant palms (Cochran & Ellner, 1992; Svenning, in prep.).

In the abundant (up to 1200 adults per hectare) Astrocaryum mexicanum, juvenile survival is negatively density dependent (Martinez-Ramos et ah, 1988b; Cochran & Ellner, 1992), something that may reflect competition from adults, but the population level consequences are negligible (Cochran & Ellner, 1992). No studies have directly proved intraspecific resource competition in neotropical palms, although its occurrence has been implied by several studies of spatial pattems (Yeaton, 1979; Sterner et ah, 1986) or performance (Bovi et al., 1987; Svenning, in prep.; but not Vandermeer, 1977). Density-dependent thinning and spac- ing has been demonstrated for Socratea exorrhiza (as S. durissima) in Panama and Costa Rica and for Iriartea deltoidea (as I. gigantea) in Costa Rica (Yeaton, 1979; Sterner et ah, 1986). Yeaton (1979) suggested that the mechanism at the seasonal Panamanian site was root-level competition for water during the dry season, but this mechanism does not seem likely at the ever-wet Costa Rican site. In Geonoma macrostachys var. macrostachys, conspecific adult neighbors have a negative effect on growth, but this effect is too weak to cause negatively density-dependent recruitment despite the high abundance of this palm, 218 adults per hectare (Svenning, in prep.). A negative effect of neighboring adult understory palms may reflect competition for light, as overarching understory palms strongly reduce light availability, but adult effects may also reflect herbivore-mediated interference (Denslow et ah, 1991). The occurrence of intraspecific interference--seedlings being killed by falling adult leaves--has been shown more unambiguously than has resource competition. In Welfia regia (as W. georgii), seedlings experience a doubled mortality rate within 3 m of an adult compared with farther away, partly due to falling Welfia leaves (Vandermeer, 1977). This kind of interference is likely to be commonplace in palms (cf. De Steven, 1989; Ratsirarson et al., 1996), especially in large-leaved species (cf. Aide, 1987).

C. OTHER PLANTS

The multitude of plant species and growth forms in tropical rain forests is in itself a source of microenvironmental heterogeneity. Though likely to be important, this topic suffers from a dearth of studies. The more generalized effects of heterogeneity in canopy conditions, litter layer, and animal-mediated interference are covered in other sections and will not be discussed here.

1. Competition from Trees and Shrubs

Competition with trees and shrubs is potentially important but has only been little studied apart from diffuse competition for light, which is covered in section IV.A. Still, there is some evidence of belowground resource and aboveground interference competition. In Geonoma macrostachys var. macrostachys, growth (in leaf length) is negatively related to the presence of neighboring trees, whereas in one of two periods trees have a positive effect on spike length (Svenning, in prep.). The effect on growth perhaps reflects belowground resource competition between trees and this understory palm, for the analyses included a separate measure of light availability. Interference competition with trees and shrubs has also only been shown by one study: flower predation by katydids increases with increasing local obstruction of the inflorescence by surrounding vegetation in Calyptrogyne ghiesbreghtiana and reduces fruit initiation (Cunningham, 1995). Thus animal-mediated interference from neighboring plants can reduce palm fecundity.


2. Competition among Palm Species

Not only trees and shrubs but also palms of other species may cause interference or resource competition. Where canopy palms are common, the continual shedding of their large leaves is a major source of damage and mortality of palms in the understory (cf. Aide, 1987); and section IV.B.3). Consequently, smaller palms and other plants are often rare or lacking below the crown of large-leaved palms (Svenning, pers. obs.).

Though not studied, it is highly likely that understory palms compete with each other and may interfere with recruitment of larger palms. Overarching small understory palms and cyclanths cause a strong reduction in the already low understory light, whereas their effect on belowground resources appears negligible (Denslow et al., 1991). They may also increase litter cover, as ferns do (George & Bazzaz, 1999a). Although the reduction in light availability under the palms would be expected to cause poor tree seedling performance, as was indeed found, the reduced survival and growth of the tree seedlings mainly result from increased pest attack (Denslow et al., 1991). The importance of such competition is shown by the fact that tree regeneration is negatively related to the abundance of understory palms and cyclanths (Pifiero et al., 1986; Denslow et al., 1991). Although these studies have focused on the effect of understory palms on tree recruitment, recruitment of midstory and canopy palms and recruitment, as well as adult performance, of other species of understory palms would probably be similarly affected. The most likely competitive mechanisms appear to be competition for light and pest- and litter-mediated interference.

In a Costa Rican rain forest interspecific competition between the closely related stilt-root palms Iriartea deltoidea and Socratea exorrhiza was inferred from their local distributions (D. A. Clark et al., 1995). Probably due to past stem harvesting, I. deltoidea is missing from a small part of the forest on a soil type in which it otherwise occurs at high densities (D. A. Clark et al., 1995). Intriguingly, S. exorrhiza has higher densities on this soil type where I. deltoidea has been removed than where I. deltoidea has been left uncut. D. A. Clark et al. (1995) suggested that this may indicate competitive release of S. exorrhiza. This is possible, but an alternative explanation is more likely. As discussed in section IV.F, recruitment of S. exorrhiza is favored by treefall gaps, and this species would therefore be favored by selec- tive felling of trees of any species. Supporting this interpretation, S. exorrhiza exhibits strong recruitment in secondary forests but low recruitment in old-growth forest at the same site (Guariguata et al., 1997). Thus the increased abundance ofS. exorrhiza probably reflects general canopy opening rather than release from competition from I. deltoidea.

3. Liana Interference

As discussed above, lianas may reduce the survival, growth, and fecundity of palms. Such interference is probably most important for small palms and juveniles of large palms, for large palms are well equipped to avoid liana infestation, mainly due to the continual shedding of their long leaves, and therefore only experience low levels of crown infestation (Putz, 1984; Rich et al., 1987).

4. Interference f rom Ant Plants

A peculiar type of microenvironmental heterogeneity is that created by ant plants. Ants create bare clearings around host individuals of the treelet Duroia hirsuta, Rubiaceae, in Amazonia (Olesen et al., in press). These Duroia clearings average 58 m 2 and occupy 673


mS/ha at an Ecuadorian site (Olesen et al., in press) and thereby represent a significant type of microenvironmental heterogeneity. Palms are generally absent from the center of these clearings because of ant defoliation and the resulting high mortality, as found for transplanted Iri- artea deltoidea seedlings (Olesen et al., in press). Still, at least some understory palms appear to be better able to resist or avoid ant defoliation than most plants and are commonly found in the edges of Duroia clearings (Olesen et al., in press). The importance of Duroia clearings for local plant distributions has also been documented for two fern species (Tuomisto et al., 1998).

D. LITTER

Litter is a source of much microenvironmental heterogeneity in tropical rain forests, in terms of both litter cover and litter depth (Molofsky & Augsburger, 1992; Cintra, 1997a; Mar- tius & Bandeira, 1998). The litter layer varies according to the spatial location and species of trees and understory vegetation, topography, fallen trunks and branches, microtopography, and local edaphic conditions (Facelli & Pickett, 1991; Kohyama & Grubb, 1994; Cintra, 1997a), and litter depths can be largely spatially unpredictable at 1-20 m scales (Molofsky & Augsbur- ger, 1992). Litter cover and depth may also vary seasonally (Cintra, 1997a). Litter on the forest floor can be an important factor affecting recruitment by hiding seeds from seed and seedling predators, influencing microclimate and nutrient availability, reducing the light intensity and the red/far-red ratio, forming a physical barrier to germination, burying already established seedlings, or being allelopathic (Facelli & Pickett, 1991; Molofsky & Augsburger, 1992; Vfisquez-Y~ines & Orozco-Segovia, 1993; Kohyama & Grubb, 1994; Silver et al., 1994; Grubb, 1996; Cintra, 1997a; Crawley, 1997c). Still, its effects on palms have only been little studied. Bannister (1970) found that many seedlings ofPrestoea acuminata (as Euterpe globosa) are buried under litter after one year, most likely reducing their survival and growth. Seedlings of small-seeded palms, such as many understory species (e.g., Geonoma and Cha- maedorea spp.), may have difficulty penetrating thick litter layers and be particularly prone to litter burial (cf. Molofsky & Augsburger, 1992; Kohyama & Grubb, 1994; Grubb, 1996). Re- cruitment of these species may be limited in microsites where dense litter accumulates. Con- versely, litter cover has positive effects on both seed and seedling survival in the large-seeded Astrocaryum murumuru var. murumuru (Cintra, 1997a). The positive effect of leaf litter on seed and seedling survival mainly reflects protection from mammal seed predation; and, because litter does not form a barrier to seedling emergence in this large-seeded species, its recruitment is favored by a thick litter layer (Cintra, 1997a). Even the recruitment of large-seeded species may be obstructed by large, thick leaves, such as those ofMatisia spp. (Bombacaceae) and Sloanea spp. (Elaeocarpaceae) (Cintra, 1997a; Svenning, pers. obs.). Overall, litter heterogeneity may affect the ecology of rain-forest palms by generally favoring large-seeded species where the litter cover is relatively thick and favoring small-seeded species in microsites with a thin, open litter layer. In reality, patterns are probably more complicated, though. Hodel (1992) noted that two small-seeded understory palms, Chamaedorea pygmaea (as C stenocarpa) and C. stricta, with short, creeping stems, seem to prefer deep litter accumulations at the base of large, buttressed trees.

E. SOIL

Tropical rain forests exhibit strong fine-scale (10-105 m) variation in edaphic conditions, such as soil structure, vertical/lateral drainage, soil nutrients, and pH (Newbery & Proctor,


1984; Silver et al., 1994; Sabatier et al., 1997). Some, but not all, of this heterogeneity is related to topography (Newbery & Proctor, 1984; Sabatier et al., 1997). In temperate forests, soil parameters even exhibit very fine scale (0.1-10 m) heterogeneity, as they probably also do in tropical forests (cf. Richter & Babbar, 1991), and there is often little predictability in these parameters beyond a few meters (Lechowicz & Bell, 1991). Some of this edaphic variation arises from stem flow at the bases of tree trunks, litter from different plant species, gaps, tip-up mounds, animals, and microtopographic variation (Lechowiez & Bell, 1991; Richter & Babbar, 1991; cf. sections IV.A, IV.D). Edaphic heterogeneity affects the local distribution and abundance of tree species in tropical rain forests (Newbery & Proctor, 1984; Gartlan et al., 1986; Newbery et al., 1986; Rogstad, 1990; ter Steege et al., 1993; Korning et al., 1994; Dietrich et al., 1996; D. B. Clark et al., 1998). In a review of soil-plant relationships in tropical rain forests, Sollins (1998) concluded that although only few studies have clearly documented relationships between plant distributions and nutrient availability, this probably mainly reflects study limitations. Here I shall show that soil heterogeneity is likely to be an important factor in the ecology of neotropical rain-forest palms.


Growth of neotropical rain-forest palms can be nutrient limited, as shown by the increase in leaf production and number of leaves in the crown of the understory palm Reinhardtia gracilis after fertilization (Mendoza & Franco, 1992; cf. Pinard & Putz, 1992; McPherson & Wil- liams, 1996). If rain-forest palms frequently are nutrient limited, then heterogeneity in soil nutrients would be expected to affect their performance and possibly also their local distribution. Soil water conditions may also influence the performance of rain-forest palms, as will be discussed in section IV.F.


Throughout the world it is often stated that certain palm species in uplands are restricted to localized spots of high soil moisture or poor drainage (e.g., Kiew, 1972; Boydak, 1985; Kahn & de Granville, 1992; Dowe et al., 1996). Apart from anecdotal observations, quantitative studies from neotropical rain forests and elsewhere have documented that small-scale palm- soil relationships are indeed present, and also that the relationships are quite diverse. A study of the local tree-soil relationships in an upland rain forest in Amazonian Ecuador found that the local distribution of tree species, including two abundant canopy palms, corresponded with topsoil variation in clay content, aluminum, and soil nutrients (Koming et al., 1994). Iri- artea deltoidea is most common--in fact, is the most common tree--in a plot with moderate clay content, low aluminum content, and relatively high nutrient concentrations, whereas Oe- nocarpus bataua var. bataua (as Jessenia bataua) is abundant on a plot with very high clay content and low phosphorous concentration and in a plot with high aluminum content and no calcium (Korning et al., 1994). Thus I. deltoidea appears to specialize in the more favorable soil conditions, whereas O. bataua becomes abundant when soils are poor in nutrients, high in aluminum, or poorly drained (here, due to the high clay content). This interpretation is supported by the fact that O. bataua (as Jessenia bataua) is abundant in poorly drained sites throughout central and western Amazonia (var. bataua) and the Guianas (var. oligocarpa) (Kahn & de Granville, 1992; Sabatier et al., 1997; but cf. Svenning, 1999a). It remains possible, though, that the differing but overlapping edaphic preferences of these two species (e.g.,


Svenning, 1999a) merely reflects the more open canopy conditions under stressed soil conditions favoring the more light-demanding O. bataua (Svenning, 1999b).

Other types of palm-soil relationships have been documented for other neotropical palm species (Kahn & de Granville, 1992; D. A. Clark et al., 1995). Notably, the distribution and abundance of all five common large palms in a Costa Rican rain forest are related to soil types (D. A. Clark et al., 1995): two species are biased toward infertile soils; three species prefer fertile soils. Intriguingly, the distribution of/. deltoidea is skewed toward the more infertile soil at this Costa Rican site (D. A. Clark et al., 1995), whereas it prefers the more fertile soils in Amazonian Ecuador (Korning et al., 1994). These patterns may reflect the generally low soil fertility at the Amazonian site compared with the wide range, from infertile to fertile soils, at the Costa Rican site (Korning et al., 1994; D. A. Clark et al., 1995); namely, that/. deltoidea prefers soils of intermediate fertility. Studies outside neotropical rain forests have also found palm-soil relationships (Ayora & Orellana, 1993; Barrow et al., 1993; Olmsted & Alvarez-Buylla, 1995). Clearly, small-scale variability in soil conditions is probably an important factor in the ecology of many palms in neotropical rain forests and elsewhere, although detailed studies are needed to give real credit to this topic (cf. Sollins, 1998). If such studies use a phenomenological approach, it will be important to include other microenvironmental factors, particularly light, in order to be able to convincingly test for direct edaphic effects.

F. TOPOGRAPHY

As elsewhere, there is much local topographic variation in neotropical rain forests (slopes, plateaus, ridge tops, bottomlands, swamps, creeks [e.g., Lieberman et al., 1985; Becker et al., 1988; Basnet, 1992; Tuomisto et al., 1995; D. B. Clark et al., 1998; Svenning, 1999a]) and also microtopographic heterogeneity within these larger topographic units (small streams and poorly drained depressions [Lieberman et al., 1985; Svenning, 1999a; Svenning & Balslev, 1999], treefall pits and mounds [Putz, 1983], slope gradients [Poulsen & Balslev, 1991]). Although plants do not respond to topography per se, topography influences many factors that are important to plants: edaphic conditions, such as soil nutrients, pH, aluminum, texture, flooding, drainage, and drought susceptibility (e.g., Curi & Franz- meier, 1984; Rogstad, 1990; P. S. Ashton, 1992; Johnston, 1992; Kahn & de Granville, 1992; Silver et al., 1994; Botschek et al., 1996; Sabatier et al., 1997; Sollins, 1998); litter layer (Facelli & Pickett, 1991); disturbance regime and canopy heterogeneity (White, 1979; P. S. Ashton, 1992; Kahn & de Granville, 1992; D. B. Clark et al., 1996; Sollins, 1998); and biotic interactions (fungal attack, seed predation, or seed dispersal [Bodmer, 1991; Forget, 1994; Fragoso, 1997]). Thus topography influences most of the other types of microenvironmental heterogeneity discussed here. Given that topography combines multiple types of microenvironmental heterogeneity, it is no surprise that many studies have shown strong effects of topographic heterogeneity on the ecology of neotropical rain-forest palms. The many factors involved may also result in geographical inconsistencies in topographic preferences. For example, sensitivity to drought may cause species occurring in uplands in aseasonal wet areas to become restricted to poorly drained microsites in more seasonal areas. Such patterns are known for some neotropical rain-forest palms, including lriartea deltoidea (cf. Balslev et al., 1987; Henderson, 1990; Pinard, 1993; Peres, 1994; D. A. Clark et al., 1995; Chavez, 1996) and Oenocarpus bataua (cf. Kahn & de Castro, 1985; Peres, 1994; Sabatier et al., 1997; Svenning, 1999a), but they have not received direct study and may require more complex ex- planations (Svenning & Balslev, 1999).



a. Uplands versus Wet Topographic Units

Most information on the influence of topographic heterogeneity on the individual performance of neotropical rain-forest palms compares well-drained uplands with wet topographic units (floodplains or stream banks, for example). Several studies have found that flooding affects individual performance. Survivorship of transplanted lriartea deltoidea seedlings on an Amazonian floodplain is substantially negatively correlated with flooding level, and variability in flooding thereby causes both temporal and spatial heterogeneity in survivorship (Losos, 1995). In contrast, flooding does not affect seedling survivorship in As- trocaryum murumuru var. murumuru (as A. m. var.javarense) at the same site (Losos, 1995). Phytelephas aequatorialis has lower leaf-production rates in agroforestry plots that are frequently flooded than in drier plots, whereas inflorescence production is unaffected (Runk, 1998). Thus flooding may affect the performance of both seedling and adult palms, but species vary in their sensitivity to flooding.

Other abiotic factors may also influence performance along a topographic gradient toward a river or stream. Bernal (1998) suggested that river-channel migration is a major mortality factor for adult Phytelephas seemanii in floodplains. Changes in light availability over such gradients may also be important. Prestoea acuminata (as P. montana) has increased recruitment in a floodplain forest in locations of low microtopography, close to the river channel (Frangi & Lugo, ! 998). This probably mainly reflects a more open canopy close to river, as regeneration is more related to hurricane-induced canopy damage than to topography or distance from the river (Frangi & Lugo, 1998). Further illustrating how complex are the effects a topographic gradient may have, even biotic interactions may be affected. Thus seed mortality of the canopy palm Attalea butyracea (as Scheelea rostrata) is strongly affected by local topography in a Costa Rican forest (Janzen, 1971): only 1-6% of endocarps below adults in stream bottoms are attacked by bruchids, whereas >80% endocarps below adults just 15 m away and outside the stream bottom are infested (Janzen, 1971 ). Janzen (1971 ) suggested that the low infestation rate in stream bottoms results from fallen endocarps (including any bruchid larvae) being washed away regularly and from low rates of movement of adult bruchids between adult palms. From this discussion, it is clear that gradients in individual performance from wet to dry topographic units may reflect factors other than hydrology, notably canopy conditions and biotic interactions.

b. Heterogeneity within Uplands

The effect on individual performance of more subtle types of topographic heterogeneity-- within uplands, for example--have received little direct investigation, although, as can be seen from this section, this does not reflect lack of importance. In Geonoma macrostachys var. macrostaehys the density of juveniles and adults combined is not related to topographic position on well-drained hills, but this pattern hides counterbalancing topographic effects on recruitment and later survival (Svenning, in prep.). Seedling density is highest in the flatter and lower parts of well-drained hills, reflecting increased recruitment there, whereas survival of later stages is highest in the upper parts of the hills (Svenning, in prep.). The causal factors behind these patterns are not apparent. In Borneo, Calamus caesius experiences high mortality on upper slopes due to drought but performs well in damp parts of lower slopes (Drans- field, 1988). Thus, in this case, the demographic effects reflect differential drought


susceptibility among topographic positions (cf. Rogstad, 1990). In a seasonal rain forest in Panama, six slender understory palms experienced severe population declines in response to a 25-year drying trend, but five more massive palms were less negatively affected or were unaffected (Condit et al., 1996). Given this strong negative response, one would expect many understory palms, as well as other drought-sensitive palms, to become restricted to the microsites least susceptible to drying out in forests subject to periodic drought. The potential importance of this phenomenon is indicated by the fact that periodic droughts associated with E1 Nifio episodes not only affect Central American rain forests (Condit et al., 1996) but also have the ability to desiccate large areas of central and eastern Amazonian rain forest (Nepstad et al., 1999), and even western Amazonian rain forests are subject to short drought episodes every decade (Balslev et al., 1987).


a. Patterns within Uplands

Microenvironmental topographic heterogeneity affects the distribution of many neotropical rain-forest palms (e.g., Lieberman et al., 1985; Kahn, 1987; D. A. Clark et al., 1995; Svenning, 1999a; Svenning & Balslev, 1999) and has also been suggested as an important determinant of local palm distributions in the Paleotropics (e.g., Dransfield, 1969; Savage & Ashton, 1983). Illustrating the importance of topography in shaping local palm distributions, topography (mainly topographic position, but also inclination and the presence of small streams or periodically water-filled depressions) affects the distribution or abundance of 13 of the 25 more common taxa in 50 ha of upland forest in the Ecuadorian Amazon (Svenning, 1999a). Similar patterns have been reported from other neotropical rain forests (Kahn & de Castro, 1985; Kahn & de Granville, 1992; D. A. Clark et al., 1995; Sabatier et al., 1997). Not only topographic position and inclination but also altitude per se and slope aspect may affect palm distributions. In montane areas, small changes in altitude may also affect palm distributions (e.g., Kessler, 2000). According to Borchsenius (1997b), Socratea rostrata (nomenclature as in Borchsenius et al. [1998]) replaces Iriartea deltoidea as the dominant canopy palm on the western Andean slopes of Ecuador over an altitudinal interval of a few hundred meters (going up), apparently due to frequent cloud formation toward the upper end. Away from the equator or where winds come predominantly from a certain direction, slope aspect may also affect palm distributions. This has not been investigated for neotropical palms, but Rhapidophyllum hystrix has a strong preference for the west-northwest- facing slope, being nearly absent from the opposite slope, in a narrow ravine in Alabama (Clancy & Sullivan, 1990).

b. Pattems within Floodplains and Swamps

Microtopographic variation affects the distribution and abundance of palms within swamps and floodplains, many of these patterns probably reflecting differing degrees of tolerance of flooding, waterlogging, or drought (but see section IV.F.2.c). As far back as the lower Eocene, palms may have been associated with certain topographic microhabitats in floodplains (Daghlian, 1978). In a seasonal swamp forest in east-central Brazil, all five common species show microtopographic preferences: three for the low-lying areas around the central stream in all life stages (Euterpe oleracea only as adult), and two for the higher-lying areas away from the stream (Scariot et al., 1989). In a Costa Rican swamp forest, all four


abundant canopy palms also show microtopographic preferences: lriartea deltoidea (as. Igi- gantea), Welfia regia (as gy'. georgii), and Socratea exorrhiza (as S. durrissima) have a strong preference for poorly drained but firm sediments, being virtually absent from wetter microto- pograpbic units, whereasAstrocaryum alatum is abundant on both firm and soft sediments but nearly absent from water-covered areas (Lieberman et al., 1985). In well-drained floodplains at a site in the Colombian Amazon, 1. deltoidea prefers the better-drained, but often flooded, relatively fertile microsites, whereas S. exorrhiza is only weakly related to flooding and nutrients, although it prefers the poorest drained microsites (Duivenvoorden, 1995). On poorly drained floodplains at the same site, Mauritiaflexuosa prefers very poorly drained microsites, whereas Euterpe precatoria prefers frequently flooded microsites (Duivenvoorden, 1995). Similarly, in an Ecuadorian palm swamp, seedlings of the large canopy palm M. flexuosa are abundant on wet, soft sediments but absent from firm sediments (Svenning, pers. obs.). In upland swamps at the Colombian site, Mauritiella aculeata prefers the most poorly drained microsites, whereas Oenocarpus bataua var. bataua (as O. bataua) prefers the best-drained microsites (Duivenvoorden, 1995). In a seasonal swamp in Peru, Oenocarpus bataua var. bataua (asJessenia bataua var. bataua) likewise has a high density in the rarely flooded parts and a low density in small, frequently flooded depressions (Kahn & de Granville, 1992). On riverbanks in the Amazon estuary, the aroid Montrichardia sp. forms a dense outer zone, growing in mud that is subject to daily tidal flooding; behind it, on slightly higher, not daily flooded ground is a dense zone ofEuterpe oleracea; and on somewhat more protected ground Mauritiaflexuosa dominates (Strudwick & Sobel, 1988). From the above examples it is clear that palms divide swamps and floodplains into numerous microtopographic niches along the two hydrological niche axes, flooding and drainage, as do temperate meadow herbs (Silver- town et al., 1999). In some cases these microtopographic preferences may partly or com- pletely reflect less apparent factors, such as light availability, as I discuss in the following section

c. Nonedaphic Topographic Factors

I have already discussed how nonedaphic factors can create topographic gradients in individual performance, and here I shall discuss how two such factors, seed dispersal and canopy heterogeneity, can also cause topographic preferences. Animal-mediated seed dispersal may occur directionally to certain topographic microsites (cf. Mack, 1995). Although this phenomenon has not been directly studied for palms, it is probably important. Among large Ama- zonian ungulates only tapirs frequently defecate intact seeds and thus are the only large terrestrial mammal that acts as a longer-distance dispersal agent of large palm seeds (Bodmer, 1991 ; Fragoso, 1997). Because tapirs defecate mainly in areas that are flooded at least part of the year (Bodmer, 1991), the seeds of these palms are dispersed directionally to such wet sites. This directed dispersal by tapirs is probably general to the Neotropics, because defecation in wet areas is typical of neotropical tapirs, but they may also use additional microsites (Fragoso, 1997; cf. section IV.G). The following three examples show that canopy heterogeneity can also generate topographic preferences.

1. In a flat bottomland in an Ecuadorian rain forest, Iriartea deltoidea has higher adult densities on the banks of a small brook and also recruits preferentially here, even when adult abundance is factored out (Svenning & Balslev, 1999). These patterns may involve better drainage on the brook banks, relative to the flat areas away from the brook (Svenning & Balslev, 1999), for survival ofIriartea deltoidea seedlings is reduced by flooding (Losos, 1995). The permanently elevated light levels due to the less dense canopy above the brook are probably more im-


portant, though (Svenning & Balslev, 1999), because I. deltoidea recruits preferentially in microsites with slightly elevated light availability (Svenning, 1999b, 2000b).

2. In a Puerto Rican forest, the midstory palm Prestoea acuminata (as P. montana) prefers a valley bottom to the neighboring hill (Basnet, 1992; Johnston, 1992). Soil moisture, pH, and calcium content are highest in the bottomland, whereas soil aeration is poorest. Among these factors Johnston (1992) suggests that soil moisture determines the local distribution of P. acuminata. Because the valley bottoms at this site support lower stem densities, have fewer large trees, and have experienced more natural and human disturbance than have slopes and ridges (Basnet, 1992), increased light availability is an equally likely explanation for the high abundance ofP. acuminata there, especially because canopy opening is known to increase recruitment and growth in this species in Puerto Rico (Bannister, 1970; Lugo & Batlle, 1987; Frangi & Lugo, 1998). Supporting this interpretation, P. acuminata (as P. montana) is also abundant on steep slopes in Puerto Rico (e.g., Frangi & Lugo, 1998). Although canopy openness is probably the most important factor favoring this species, soil moisture may also have a direct effect. Thus in Ecuador Prestoea schultzeana occurs preferentially in low-lying, poorly drained microsites, as well as in low-canopy microsites (Svenning, 1999a, 2000b). Another factor that may contribute to the occurrence of P. acuminata (as Euterpe globosa) mainly along streams at low elevations in Puerto Rico is water-mediated seed dispersal from above (Bannister, 1970).

3. Socratea exorrhiza is a widespread canopy or midstory palm in neotropical rain forests (Henderson et al., 1995), and the topographic patterns recorded for this species are diverse and sometimes conflicting. It is abundant in seasonal swamps in Amazonia and French Gui- ana (Kahn & de Granville, 1992), even preferring the low-lying, streamside microsites in these swamps (Scariot et al., 1989), and is one of the few canopy palms that is frequent in permanent swamps (Kahn & de Granville, 1992). These patterns show that it must be tolerant of waterlogged soil conditions and flooding, probably because its stilt roots function as pnema- tophores (Kahn & de Granville, 1992). S. exorrhiza is not restricted to these wet areas, though. It is abundant on exposed ridges of low mountains in the eastern Andean foothills of Ecuador (Svenning, pers. obs.) and also occurs in upland forests throughout its range (Yeaton, 1979; Kahn & de Granville, 1992; Chavez, 1996; Guariguata et al., 1997; Svenning, 1999a). At the Tiputini River in lowland Amazonian Ecuador it exhibits an intriguing pattern: it is rather abundant on the floodplain and on a well-drained, unflooded crest bordering the river but scarce in the uplands away from the river (Svenning, pers. obs.). These diverse patterns probably reflect high light requirements for recruitment in this species. Though more shade tolerant and demanding than ,4ttalea butyracea (as Scheelea zonensis) (Hogan, 1988; Araus & Hogan, 1994), it has been considered to have relatively high light requirement for recruitment and to behave like a gap-dependent pioneer species in uplands (e.g., Yeaton, 1979; Kahn & de Granville, 1992). The available data support the conclusion that juveniles show a strong positive growth response to high light levels (Araus & Hogan, 1994), have much faster height growth rates in an abandoned cacao plantation than in old growth forest (Rich, 1986), and are abundant in 16-18-year-old secondary forests but scarce in neighboring old-growth forests on well-drained soil (Guariguata et al., 1997). Anatomical and architectural studies of its stem support the view that S. exorrhiza is a fast-growing pioneer species (Schatz et al., 1985; Rich, 1986). Juvenile S. exorrhiza (as S. durissima) also allocate proportionally less biomass to leaves than do juvenile Iriartea deltoidea (as I. gigantea) and juvenile, subterranean- stemmed Welfia regia (as W. georgii) (Schatz et al., 1985) and would therefore be expected to be less shade tolerant (cf. Chazdon, 1985, 1986a; Givnish, 1988). Thus the patterns of topographic preferences found in this species probably reflect a combination of high light require-


ments and tolerance ofwaterlogging and flooding, because swamp and floodplain forests are more open canopied than are upland forests (e.g., Kahn & de Granville, 1992) and because exposed crests have a higher incidence of treefall gaps.

G. ANIMALS

Spatial heterogeneity in palm dispersal and herbivory intensity mediated by animals re- sponding to other microenvironmental cues was discussed above. Here I shall elaborate on the role of animals. An important point is that animals may generate microenvironmental heterogeneity irrespective of preexisting heterogeneity simply by their intrinsic behavior. Some animals concentrate their movements in certain parts of a given habitat, at least for some time--for example, using certain tracks (cf. Forget and Sabatier, 1997: fig. 1) or sleeping sites (e.g., Julliot, 1997)--simply by habit. Although this type of heterogeneity may be quite important for the population and community ecology of palms (e.g., by limiting recruitment [Hubbell et al., 1999]), it remains largely unstudied.


a. Fecundity

Fruit initiation is pollen limited in the bat-pollinated understory palm Calyptrogyne ghiesbreghtiana (Cunningham, 1996). Because neotropical rain-forest palms do not produce fruits by selfing or apomixis and because geitonogamy is rare in many species (Eguiarte et al., 1992; Listabarth, 1993, 1999; Borchsenius, 1997a), pollen limitation may frequently be important, although it remains to be studied. Spatial heterogeneity in pollinator activity will be an important source of spatial heterogeneity in palm fecundity, if palms are often pollen limited.

b. Seed Predation

In the Neotropics palm seeds are predated by a variety of animals: palm bruchids (e.g., Janzen, 1971; Wright, 1983, 1990; Delobel et al., 1995; Johnson et al., 1995; Fragoso, 1997), scotylids (Janzen, 1972), curculionids (Oyama, 1991), various mice and rats (Oyama, 1991; Hoch & Adler, 1997; S~nchez-Codero & Martinez-Gallardo, 1998), squirrels (Kiltie, 1981; Galetti et al., 1992), agoutis and acouchis (Kiltie, 1981; Smythe, 1989), monkeys (Galetti et al., 1992), deer, peccaries, and tapirs (Kiltie, 1981; Smythe, 1989; Bodmer, 1991). As already discussed, seed-predation risk may depend on canopy conditions and be affected by density or distance effects, but it may also be affected by other microenvironmental factors. White- lipped peccaries search for scatter-hoarded palm fruits near objects on the forest floor (e.g., bases of large trunks, fallen logs, exposed roots) or beneath lianas and shrubs, and scatter hoarding in this case thus increases seed predation risk (Kiltie, 1981). Seed-predation intensity may also vary seasonally, and this may interact with microsite conditions to cause spatial heterogeneity in seed-predation risk (Wright, 1990).

c. Later Herbivory

Mammal herbivory can sometimes be an important source of mortality and damage in neotropical understory palms. Seedlings can, as discussed in section IV.B.2, be wholly predated by mammalian herbivores, such as agoutis, deer, and tapirs, but larger individuals can also be


damaged or killed by mammalian herbivores (De Steven & Putz, 1985; De Steven, 1989; cf. Bullock, 1980). In periods of fruit shortage on Barro Colorado Island, understory palms are damaged and sometimes killed by monkeys and peccaries feeding on the terminal bud (De Steven & Putz, 1985). Some species are damaged much more commonly than are others, some of which are seldom attacked (De Steven & Putz, 1985). Although it has not been studied, it is quite likely that the intensity of mammal herbivory varies not only temporally but also in space and that it therefore constitutes a source of microenvironmental heterogeneity. Thus patches of dense undergrowth may offer palms more protection from mammal herbivory by causing the palms to be more cryptic (cf. George & Bazzaz, 1999b) and obstructing the movement of the larger mammals.

As discussed in previous sections, invertebrate herbivory can be an important source of mortality and damage, and the risk of invertebrate herbivory is also likely to be influenced by the herbivores' behavior, such as their microhabitat preferences (Braker & Chazdon, 1993; cf. George & Bazzaz, 1999b).

2. Seed Dispersal

Seed dispersal by frugivorous (including granivorous) animals is the main mode of dispersal for palms and for tropical plant species in general (Zona & Henderson, 1989; Julliot, 1997). Spatially heterogeneous behavior of the frugivores can thus be an important determinant of recruitment pattems (cf. Mack, 1995; Fragoso, 1997; Julliot, 1997) and increases small-scale spatial heterogeneity (Julliot, 1997).

Scatter hoarding by rodents is an important source of microenvironmental heterogeneity that affects palm recruitment. In the Neotropics large palm fruits are often scatter hoarded by various types of rodents: agoutis (Kiltie, 1981; Smythe, 1989; Forget, 1991 ; Bemal, 1998), acouchis (Forget, 1991), spiny rats (Forget, 1991 ; Hoch & Adler, 1997), and squirrels (Galetti et al., 1992). The activity of these animals results in spatially heterogeneous seed shadows, for they cache near objects on the forest floor (Kiltie, 1981; Forget, 1991; Hoch & Adler, 1997). Contrary to agoutis and acouchis, spiny rats do not bury seeds in the soil; they hide them beneath litter (Forget, 1991), thereby dispersing them to a different microenvironment. Rodent seed dispersal may vary from site to site as a function of microhabitat and rodent activity (Forget, 1991). Thus acouchis may be less frequent visitors to an Astrocaryum paramaca palm growing in open understory, because they prefer foraging in dense sites. Availability of alternative food sources may also influence rodent activity (Forget, 1991).

Other aspects of the behavior of terrestrial mammals may also introduce spatial heterogeneity into the seed-dispersal pattern. For example, mammal-dispersed palms invaded an abandoned plantation mainly along a narrow footpath, reflecting the preferential use of this path by their mammal dispersers (Vandermeer, 1993). This result suggests that seeds of these palms under natural conditions may arrive along tracks created by tapirs, peccaries, and other large mammals. Bird-dispersed palms do not exhibit this invasion pattern (Vandermeer, 1993).

An even more striking example of the microenvironmental heterogeneity introduced by animal-mediated seed dispersal and its ecological importance is the tapir dispersal of Attalea maripa (as Maximiliana maripa) seeds in a Brazilian forest (Fragoso, 1997). There, tapirs defecate not only in wet areas (cf. section IV.F.2.c) but also in upland "latrines" next to an emergent tree, nearly always Couratari multiflora, Lecythidaceae (Fragoso, 1997). Seeds of A. maripa are dispersed short distances (mostly <__5 m) by various pulp-consuming mammals (rodents, monkeys, tapirs, deer, peccaries) and long distances by tapirs to latrines in wet areas


or upland Couratari latrines (Fragoso, 1997). As a consequence of the intense seed and seedling predation below adults (cf. section IV.B.2), densities of seedlings and juveniles are higher in latrine sites (Fragoso, 1997). Control sites (bases of emergent trees 60-150 m away from latrines and usually >500 m away from adult aggregations) have intermediate seedling densities, probably reflecting secondary dispersal by rodents from Couratari latrines and low seed and seedling predation (Fragoso, 1997). Thus long-distance dispersal by tapirs to the bases of large trees of a particular tree species interacts with short-distance dispersal by rodents and other mammals and with adult density-dependent seed predation to cause the clumped distribution of adult A. maripa at this site (Fragoso, 1997).

Monkeys are probably another source of spatial heterogeneity in palm recruitment. Spider monkeys (Ateles paniscus) generate spatial heterogeneity in seed dispersal and subsequent seedling establishment, as shown for Virola sp. (Myristicaceae), by providing directed dispersal toward microsites below tall trees, which they prefer for movement and foraging, and by habitually moving along particular arboreal pathways (Forget & Sabatier, 1997). Howler monkeys (Alouatta seniculus) provide directed seed dispersal to their sleeping sites, where 65% of their defecation occurs due to their low digestive rates (Julliot, 1997). This seed dispersal resulted in nearly four times as many seedlings of six non-palm tree species being dispersed by these monkeys below their sleeping sites as in areas away from their sleeping sites, and it also resulted in clumped seedling distributions within the sleeping sites (Julliot, 1997). The effect of monkey dispersal on recruitment has not been studied for palms, even though numerous neotropical rain-forest palms are dispersed by spider, howler, or other monkeys (Zona & Henderson, 1989).

V. Microenvironmental Heterogeneity and Coexistence

Neotropical rain forests, especially the wettest and warmest, harbor a high local species richness of palms and of plants in general. How is the local coexistence of these numerous palm species, as well as their coexistence with the rest of the plant community, possible? Here I discuss how microenvironmental heterogeneity may contribute to local coexistence of palms through niche differences, mass effects, and negative density dependence. Other factors, such as light limitation or dispersal limitation, may also be important because they slow exclusion processes (Wright, 1991; Hubbell et al., 1999), as suggested for understory palms (Svenning, in prep.), but will not be discussed further.

A. NICHE DIFFERENCES

Niche differentiation along microenvironmental gradients may be important for promot- ing coexistence of plant species in tropical rain forests (e.g., Dobzhansky, 1950; Ashton, 1969; Grubb, 1977, 1996; Ricklefs, 1977; Connell, 1978; Terborgh, 1985; P. M. S. Ashton, 1992; D. B. Clark et al., 1998; Svenning, 1999a). Although the role of microenvironmental heterogeneity in the maintenance of tropical plant-species richness is controversial (Denslow, 1987; Phillips et al., 1994; Grubb, 1996; Hubbell et al., 1999; Kobe, 1999), it is worth noting that canopy heterogeneity creates wider microenvironmental gradients in tropical rain forests than in other forest types. Closed understory microsites in tropical rain forests are more heavily shaded than are similar microsites in other forests, often receiving <1% full sunlight (Chazdon & Fetcher, 1984; Terborgh, 1985; Canham et al., 1990). Conversely, the centers of large treefall gaps in tropical forests receive more light than do similar microsites in forests away from the Tropics, and treefall gaps may also create wider gradients in soil conditions in


tropical forests than in temperate forests (Ricklefs, 1977). The previous sections show that microenvironmental heterogeneity in many abiotic and biotic factors affects the individual performance and small-scale distribution ofneotropical palms and that it does so in diverse ways. Thus niche differences have a large potential for being a major contributing factor to the local coexistence of numerous palm species, as well as their coexistence with the rest of the plant community. Niche differences between two species could arise by natural selection minimiz- ing interspecific competition (e.g., Giller, 1984) or by parapatric speciation (cf. section VI), but they may also be largely coincidental, simply reflecting different phylogenetic histories. Competitive interactions among neotropical palm species most likely do occur, but they are probably important mainly among abundant species (section IV.C). Nevertheless, competitive interactions among scarcer species could also occur due to competition for particular but sparse microsites, competition for pollinators and pollen-mediated interference, competition for seed dispersers, pest-mediated interference, or mycorrhizal interactions (Feinsinger, 1987; Armbruster, 1995; D. A. Clark et al., 1995; Svenning, 1999a). Thus adaptive niche differentiation could well be an important phenomenon even among species with low densities.

In species-rich neotropical palm communities, palms are represented by four broad growth-form categories: understory, midstory, canopy palms, and climbing palms (e.g., Svenning, 1999a). Part of their local coexistence may simply reflect the fact that these different growth forms coexist because the dominant growth form cannot occupy all of the space (Grubb, 1977) and because the first three growth forms represent specializations in different parts of the vertical light gradient (cf. Terborgh, 1973). As Terborgh (1985) argued, tropical forests provide more vertical light niche space than do forests at higher latitudes, so vertical light niche differentiation may allow more species to coexist in tropical forests. Because palms may coexist due to growth-form differences, I shall focus on niche differences among palms of similar growth forms along horizontal microenvironmental gradients in the rest of this section. Although no study has provided clear evidence of the importance of niche differences for the coexistence of palm species, many studies from neotropical rain forests indicate that niche differences are indeed important. The most direct evidence concerns niche differences in the utilization oftopographic-edaphic and canopy heterogeneity (mainly light); other microenvironmental heterogeneity (e.g., litter) remains largely unexplored.

1. Light Niche Differences

Light niche differences may result from species differing in their light requirements consis- tently throughout the life cycle, or just in a part of the life cycle. A special case is when species change rank in relative shade tolerance during the life cycle (Gmbb, 1996). The two most common canopy palms in an Amazonian forest exhibit such a crossover (Svenning, 1999b, 2000b). Small juveniles of Oenocarpus bataua occur irrespective of light availability, whereas small juveniles oflriartea deltoidea occur preferentially in microsites with somewhat elevated light levels (Svenning, 2000b). As the juveniles grow in size, O. bataua becomes increasingly restricted to gap microsites, resulting in preadults and small adults being largely restricted to growing below major canopy openings (Svenning, 1999b, 2000b). Iriartea deltoidea juveniles also become increasingly associated with gaps, but the majority of even 10-20 m tall individuals are found below a closed canopy (Svenning, 1999b). The understory species Prestoea schultzeana also decreases in shade tolerance with increasing size (Svenning, 2000b), whereas the two midstory species Astrocaryum murumuru var. urostachys (as A. urostachys) and Phyte- lephas tenuicaulis are more gap requiring as juveniles than as adults (Svenning, 2000a). Thus crossovers in shade tolerance may well be common (cf. section IV.A. 1).

MICROENV1RONMENTAL HETEROGENEITY IN NEOTROPICAL PALMS 31

Although ontogenetic changes in shade tolerance make it difficult to categorize a given species as shade tolerant or shade intolerant, some generalizations can be made if shade tolerance is judged by the minimum light level or gap size needed and the maximum tolerated for successful recruitment to the adult stage. Using this definition, neotropical canopy palms range from relatively shade tolerant (e.g., Iriartea deltoidea) to requiring high light levels (e.g., major treefall gaps [Attalea butyracea, Oenoearpus bataua, Soeratea exorrhiza]; cf. sections IV.A, IV.F.2.c). Notably, juveniles of Attalea butyracea avoid dehydration and chronic strong photoinhibition even under 70% full sunlight, whereas Socratea exorrhizaju- veniles are permanently stressed under these conditions (Araus & Hogan, 1994). Thus even gap-requiring species can differ in light requirements. Understory palms show similar ranges in shade tolerance, whereas midstory and possibly also climbing palms may be more ecologically restricted (see section IV.A). Among understory palms a species like Prestoea schultzeana needs significant canopy openings (Svenning, 1999a, 2000b), whereas a species like Geonoma macrostachys var. maerostaehys is damaged by light intensity in even a small treefall gap (Svenning, in prep.). Still, the last species cannot sustain itself under long-term low illumination (Svenning, in prep.), and these extremely shaded microhabitats are open to smaller, more shade-tolerant species (cf. Chazdon, 1985, 1986a, 1986c), such as Geonoma stricta var. strieta. Based on architectural and ecophysiological studies of three understory palms, Chazdon (1986c) concluded that the 1 m tall Geonoma cuneata can reproduce under heavier shade than can the 1.5-2 m tall Asterogyne martiana, which can reproduce under shadier conditions than can the 3-5 m tall Geonoma congesta. Thus palms show very diverse light responses, and their rankings in relative shade tolerance sometimes change during ontogeny. Niche differences along the light gradient are therefore likely to be important factor in the local coexistence of palms in tropical rain forests.

2. Topographic and Edaphic Niche Differences

Although light niche differences are important in the coexistence of palm species, there is also evidence for the importance of niche differences in the topographic and edaphie dimensions. Here I discuss these two factors together, for they are often closely related and difficult to separate. Thus, 13 of the 24 more abundant palm taxa in an upland forest in Amazonian Ec- uador are distributed according to topographic-edaphic factors, and the main structure in the palm-species composition at the 400 m 2 scale is that determined by topography (Svenning, 1999a). Even though no topographic-edaphic niche differentiation is apparent among canopy palms, antagonistic patterns, some of them very strong, are present in two species pairs of midstory palms (Astrocaryum murumuru versus Geonoma maxima, Phytelephas tenuicaulis versus Geonoma maxima) and six species pairs of understory palms (Geonoma cf. aspidiifolia versus G. macrostachys var. nov., Geonoma cf. aspidiifolia versus Prestoea schultzeana, Hyospathe elegans versus Geonoma macrostachys var. nov., Hyospathe elegans versus Prestoea schultzeana, Geonoma triglochin versus G. macrostachys var. nov., and Geonoma triglochin versus Prestoea schultzeana). This is a conservative record of niche differences because many more patterns are partly nonoverlapping, and the bimodal pattern found for At- talea sp. (as A. indet.) most likely resulted from a preference for upper hills by A ttalea maripa and for valley bottoms by Attalea insignis (Svenning, pers. obs). Thus topographic-edaphic niche differences clearly contribute to the local species richness of this palm community. Other neotropical studies have generally documented or indicated a similar amount of topographic-edaphic niche differences (but not always; cf. Scariot et al., 1989) and, in some cases, also among canopy palms, not only in well-drained uplands but also in swamp or flood-


plain forests (see sections IV.E, IV.F), and similar niche differences are probably also found among Asian rain-forest palms (Dransfield, 1969). Overall, microenvironmental niche differences with regard to topographic or edaphic conditions clearly appear to be important for species coexistence of neotropical rain-forest palms of similar growth form. Nonetheless, most of the studies also indicate that these niche differences alone can only explain a minor part of the coexisting pairs of species, although more detailed evaluation of hydrological conditions and soil nutrients may reveal a higher degree of niche separation in these dimensions (cf. Sil- vertown et al., 1999).

3. Seed-DispersalNicheDifferences

Dispersal of seeds of different palm species to different areas of the forest at any given point in time would promote coexistence. This phenomenon occurs when two palm species have different dispersal agents and these offer directed seed dispersal to different microsites. It also occurs when the dispersal agents do not use the same areas of the forests for other rea- sons- for example, if howler monkeys exclude other frugivores from their sleeping sites. Al- though such phenomena appear quite probable, given the seed-dispersal patterns discussed above, no data are available to evaluate this hypothesis at the moment.

B. MASS EFFECTS

The floristic composition of palm communities differs strongly among the major habitat types found in neotropical rain forests (e.g., Kahn & de Castro, 1985; Balslev et al., 1987; Kahn & de Granville, 1992; Peres, 1994; Duivenvoorden, 1995), a well-known phenomenon otherwise not considered in this review. In the Neotropics, notably in western Amazonia, such habitat types are diverse and occur as an intricate landscape-scale mosaic (Gentry, 1988; Duivenvoorden, 1995; Tuomisto et al., 1995). Considering this, together with the common microenvironmental edaphic-topographic specializations discussed above, it is clear that mass effects (Shmida & Ellner, 1984) may be frequent in neotropical rain-forest palms, as Gentry (1988) suggested for neotropical rain-forest plants in general. Mass effect is the establishment of a species, as a result of propagule influx from adjacent favorable sites, in sites where it cannot be self-maintaining (Shmida & Ellner, 1984). In neotropical rain forests mass effects are likely to occur at two scales: among habitats and among microhabitats (Svenning, 1999a), the former dependent on landscape-scale environmental heterogeneity and the latter dependent on microenvironmental heterogeneity. If such mass effects occur, their effect will be to increase the species richness at scales below the scale at which the mass effect occurs. Although no direct proof exists for the occurrence of mass effects in neotropical palm communities, much evidence indicates their likely importance at both scales.

1. Among-Habitat Mass Effects

Species that are common in a particular habitat often occur as rare individuals in other, nearby habitats. At a site in the Peruvian Amazon, two of the more abundant midstory or canopy palms, Astrocaryum murumuru and Mauritiaflexuosa, are restricted to a particular habitat type, whereas the three other abundant species have distributions that indicate mass effects (Gentry, 1988). Euterpeprecatoria is frequent in upland plots on rich soil but rare in swamps or floodplain forests and in upland forests on sandy soil. lriartea deltoidea is abundant on rich upland soils and rich floodplain soils but rare in upland sandy soil plots, whereas Socratea ex-


orrhiza is abundant on rich floodplain soils, less frequent on rich upland soils, and rare in swamps and upland sandy soil plots (Gentry, 1988).

Several studies have found that species typical of wet habitats occur as rare individuals in dry uplands. In an upland Amazonian forest, five of the rare species, including E. precatoria and S. exorrhiza, are much more abundant in nearby swamps or floodplains (Svenning, 1999a). E. precatoria shows a similar pattern at two other sites, too. It is abundant (3417 indi-

viduals/ha) in seasonal swamps a forest in central Brazil but is found only as much rarer seedlings and juveniles (46-129/ha) on the well-drained soils of adjacent hills (Kahn & de Castro, 1985). At a site in Colombia, E. precatoria and I. deltoidea are represented by rare large individuals in well-drained uplands but are abundant in poorly drained creek-side forests (Dui- venvoorden, 1995). Euterpe oleracea exhibits a similar pattern in French Guiana, where it is represented by a low density of seedlings and juveniles (17/ha) in an upland forest but is abundant in a nearby seasonal swamp (Kahn & de Granville, 1992). Several other examples exist for other species in other sites. The understory palm Astrocaryum alatum occurs at low densities in upland areas in a Costa Rican rain forest but is abundant in nearby swamps (D. A. Clark et al., 1995), and another understory palm, Leopoldinia piassaba, has high densities of seedlings, juveniles, and adults on gley and podzol soils in a floodplain forest, whereas only a few seedlings and small juveniles occur on a nearby hill (Lescure et al., 1992). In Peru, only seedlings and juveniles of Oenocarpus bataua var. bataua (as Jessenia bataua var. bataua) are found on clayey, well-drained soil. Adults are abundant on permanently waterlogged soil and on irregularly waterlogged, white sand soil; densities are intermediate on dry white sand (Kahn & de Granville, 1992). Another case of mass effect has been suggested for this species. Seedlings of Oenocarpus bataua var. bataua (as Jessenia bataua) on a dry ridge were suggested to represent long-distance dispersal by oilbirds to this "rather unsuitable site" (Snow & Snow, 1978). Given the preference for well-drained hilltops shown by this species in Ecuador (Svenning, 1999a), this example is less convincing, though. Finally, Vandermeer (1993) suggested that the rather sparse occurrence of Chamaedorea tepejilote >25 m away from rivers may depend on seed dispersal from the dense riverside population. Many of the above examples indicate mass effects created when wet source habitats maintain sink populations in drier habitats. The dearth of examples of mass effects in the opposite direction may reflect the fact that seeds and seedlings of most upland species are killed by waterlogging or flooding (cf. Iri- artea deltoidea in Losos [1995]).

A striking example of among-habitat mass effect of another type also exists. Kessler (2000) recently found indications of widespread upslope altitudinal mass effects among palms in a transect from 300 to 3950 m above sea level in the Bolivian Andes. As a strong indication of upslope mass effect, plots with only juvenile palms are significantly most common toward the upper altitudinal range limits, and in 8 of 25 palm species (42%) the uppermost individuals found are all juveniles, occurring 100-700 m above the adult range limit (Kessler, 2000). There is no evidence ofdownslope mass effect. Intriguingly, only weak indications of mass effects are evident in a palm community in the Ecuadorian Andes, 1248-1938 m above sea level (Svenning, 1998). Only in one of the five common species do seedlings or juveniles occur outside the adult height range (Svenning, unpubl.). Four palm species occur at both locations: Geonoma undata exhibits potential upslope mass effect at both sites; Chamaedorea linearis and Prestoea acuminata both exhibit strong upslope mass effect in Bolivia, but no mass effect in Ecuador; and no indication for mass effects are found for Chamaedoreapinnatifrons in either site (Kessler, unpubl.; Svenning, unpubl.). Thus upslope mass effects may be present in some Andean rain forests but absent even for the same species in other forests. Kessler (unpubl.) suggested long-distance dispersal by oilbirds (Her-


zog & Kessler, 1997) to be the cause of the mass effects at the Bolivian site. Because these birds roost in large colonies in caves (e.g., Snow & Snow, 1978; Herzog & Kessler, 1997), there is likely to be high spatial variability in the intensity and direction of oilbird seed dispersal, which could explain the mass effect differences.

2. Among-Microhabitat Mass Effects

Mass effects across microenvironmental topographic-edaphic mosaics may be frequent, too. At this smaller scale, palm species with a certain microhabitat affinity are often found as rare adults or only as seedlings and juveniles in other microhabitats. Thus, in an upland Amazo- nian forest all 13 species that exhibit topographic preferences also sometimes occur even in their most avoided topographic position (Svenning, 1999a). Likewise, Euterpe precatoria var. longevaginata (as E. macrospadix) occurs as rare individuals on one soil type in an upland Costa Rican rain forest but is abundant on three other soil types (D. A. Clark et al., 1995). There is no indication of edaphic-topographic mass effects for the four other abundant midstory and canopy palms at this site, though. In a seasonal swamp forest in east-central Brazil, adults of Euterpe oleracea are abundant on low-lying grounds but virtually absent on higher ground, whereas seedlings are abundant throughout (Scariot et al., 1989). Socratea exorrhiza and Geonoma baculifera are likewise abundant on low-lying ground but relatively rare on higher ground; no mass effects were evident for a fourth abundant species (Scariot et al., 1989). Thus, among-microhabitat mass effects appear to occur frequently in both dry and wet habitats.

The most direct evidence for mass effects concerns spatiotemporal mass effects (Shmida & Ellner, 1984) in relation to microenvironmental heterogeneity in canopy conditions. Geonoma macrostachys var. macrostachys occurs in dark microsites, where its growth and fecundity are so light limited that it would not be able to permanently maintain its population (Svenning, in prep.). Because rain-forest canopy conditions are highly dynamic, such spatiotemporal mass effects are probably frequent.

Although dynamic demographic studies are needed to unambiguously document the occurrence of mass effects (Shmida & Ellner, 1984), the discussed patterns are strong indications that this phenomenon, in combination with landscape-scale and microenvironmental topographic-edaphic heterogeneity, may be a significant factor in the small-scale coexistence of numerous palm species in neotropical rain forests. Spatiotemporal mass effects in relation to canopy heterogeneity are also likely to be commonplace and need further investigation.

C. NEGATIVE DENSITY DEPENDENCE

Microenvironmental heterogeneity created by the spatial occurrence of conspecifics may contribute to the local coexistence of neotropical palms, if negative density effects (including distance effects) keep populations at relatively low densities (cf. Janzen, 1970; Connell, 1971; Wills et al., 1997). Givnish (1999) suggested that increasing density-dependent mortality due to pest pressure is a major causal factor in the increase in the species richness of woody plants (and palms; cf. section III) with increasing precipitation and soil fertility in the Tropics. Thus density dependence may explain these large-scale diversity gradients. Strong negative density dependence mediated by seed and seedling predation by mammals and bruchid beetles is probably widespread, if not ubiquitous, in the large-seeded Astrocaryum and A ttalea genera (cf. section IV.B.2). Still, these demographic effects do not necessarily have strong population-level consequences (cf. Cochran & Ellner, 1992), and their general importance for other large-seeded palms also needs to be investigated. Density effects strong enough to limit


population densities have not been documented for small-seeded palm species. Although this may mainly reflect the dearth of studies, it is worth noting that a detailed demographic study of the abundant Geonoma macrostachys var. macrostachys did not find strong density effects (Svenning, in prep.). The preliminary conclusions are that strong density dependence may be common in large-seeded species but unimportant in smaller-seeded, mainly understory, species. As the large-seeded Attalea and Astrocaryum genera do not show a strong diversity gradient with respect to precipitation, whereas understory palms do (cf. Wessels Boer, 1968; Kahn & de Granville, 1992; Henderson et al., 1995), these preliminary conclusions do not support Givnish's hypothesis (1999). Also, the importance of negative density dependence in the mutual coexistence of even large-seeded palms has not yet been investigated. Are the pests that mediate intraspecific density dependence species specific, or do they attack all or several species of large-seeded palms at a given site? In the latter case, pests would mediate interspecific interference rather than promote mutual coexistence.

VI. Microenvironmental Heterogeneity and Speeiation

Microenvironmental heterogeneity may not only be important in maintaining neotropical palm-species richness, it may also be a strong diversity-generating factor. Here I shall advance the argument that habitat and microenvironmental gradients in edaphic and canopy conditions through the process ofparapatric speciation may be important for the generation of much of the high species richness in neotropical palms. Still, other evolutionary processes, notably allopatric speciation in association with climatic variability and dispersal barriers (e.g., Ashton, 1969; Sytsma & Schaal, 1985; Gentry, 1989; Bush, 1994; Haffer, 1997; Niklas, 1997; Colinvaux, 1998), have probably also been involved in the diversification of neotropical palms, although probably mainly at the level of among-region differentiation (cf. Hender- son, 1995; and, for Old World palms, cf. Dransfield, 1999).

Parapatric speciation is speciation without geographical isolation by disruptive selection across environmental gradients (Gentry, 1989); that is, a special type of sympatric speciation (cf. Niklas, 1997). Parapatric speciation has also been termed "ecological speciation" (Mo- rell, 1999). Recently, parapatric speciation has received theoretical support (Dieckmann & Doebeli, 1999; Kondrashov & Kondrashov, 1999; Tregenza & Butlin, 1999), and new empiri- cal studies also point to the importance of ecological specialization in speciation (Orr & Smith, 1998; Morell, 1999). Parapatric speciation has been suggested as a major speciation mode in rain-forest animals (Endler, 1982; Smith et al., 1997; but see Cracraft & Prum, 1988; Patton & Smith, 1992; Haffer, 1997), and there are many indications of the importance of parapatric speciation in the diversification of plants, including neotropical rain-forest palms, as I shall discuss here. Based on his observations of edaphic habitat specialization among closely related neotropical plant species, Gentry (1989) suggested that parapatric speciation has been a major factor in the generation of neotropical plant-species richness. Contrary to temperate forests, a large part of tree-species richness in tropical rain forests is due to series of closely related sympatric species (e.g., Fedorov, 1966; Rogstad, 1989; Davies et al., 1998). This pattern is repeated in tropical palms by the occurrence of a series of sympatric conge- neric species (especially Bactris, Chamaedorea, and Geonoma) and the occurrence of species-rich complexes with more or less sympatric morphs, notably Bactris spp. and Geonoma spp. but also including Hyospathe elegans (Skov & Balslev, 1989; Hodel, 1992; Henderson, 1995; Borchsenius, 1997a, 1999; Knudsen, 1999; Listabarth, 1999). This phenomenon is also found in the Paleotropics--for example, the genera Neophloga and Dypsis (Dransfield, 1989). These patterns are consistent with the importance of parapatric specia-


tion, or other sympatric speciation modes, in the generation of tropical species richness in palms as well as in plants in general.

Parapatric speciation involves two processes: genetic adaptation to environmental heterogeneity and the development of reproductive isolation among the resulting differentially adapted subpopulations (e.g., Dieckmann & Doebeli, 1999; Kondrashov & Kondrashov, 1999; Tregenza & Butlin, 1999). Here I shall review the evidence for the occurrence of these two processes in neotropical rain-forest palms and then consider whether the patterns of diversification in neotropical rain-forest palms are consistent with parapatric speciation.

A. GENETIC ADAPTATION TO MICROENVIRONMENTAL HETEROGENEITY

Plants frequently adapt genetically to local conditions, such as light intensity, soil conditions, interspecific and intraspecific competitors, soil microorganisms, pollinating vectors, and pests at scales as small as 10-100 m or even less than 1 m (e.g., Luescher & Jacquard, 1991; Linhart & Grant, 1996). Small-scale heterogeneity in forest-floor edaphic conditions has a strong effect on the performance of herbs and is sufficiently predictable to lead to local (<10 m) genetic specialization (Lechowicz & Bell, 1991). Small-scale adaptive genetic differentiation is not limited to herbs but also occurs in woody species (Linhart & Grant, 1996). Notably, genetic adaptation to microenvironmental heterogeneity in edaphic conditions, aspect, elevation, and parasitism has even been seen in conifers such as Pinus spp. (Pinaceae), well known to produce copious amount of widely dispersed pollen, at scales of<l 0 m to >100 m (Aitken & Libby, 1994; Linhart & Grant, 1996). Large, effective population sizes and high levels of genetic variation are typical of tropical tree species and allow natural selection to be very efficient (Eguiarte et al., 1992). Thus, most likely, genetic differentiation occurring within microenvironmental heterogeneity, at scales of <lm to >100 m, is common in forest plants, including those of tropical rain forests. Important determinants of the scale of genetic differentiation are plant stature and life history, herbaceous plants and trees often exhibiting genetic differentiation at scales of <-50 m and 100-300 m, respectively (Linhart & Grant, 1996). Thus microenvironmental genetic adaptation may be more prominent in understory palms than in taller palms. Although evidence for microscale genetic adaptation in palms is scarce, the data available indicate that it does occur. Based on seed-and pollen-dispersal distances, Eguiarte et al. (1993) estimated the genetic neighborhood area for the understory palm Astrocaryum mexicanum to be 2551 m 2 and the neighborhood effective population size to be 102-895 individuals. These estimates for genetic neighborhood area and effective population size are relatively large (cf. Levin, 1988) and suggest that microevolution by natural selection may be important in this species (Eguiarte et al., 1993). Supporting this argument there is indeed significant but low genetic differentiation among four sites separated by 200-560 m for adults and, to a lesser degree, also for seeds, and the increase in differentiation over ontogeny suggests local adaptation by natural selection (Eguiarte et al., 1992).

B. SYMPATRIC MECHANISMS FOR REPRODUCTIVE ISOLATION

Local genetic differentiation on microenvironmental and habitat gradients does not necessarily lead to speciation, for it may be overwhelmed by the homogenizing effects of gene flow. Nevertheless, local selection frequently appears to be strong enough in plants to cause genetic differentiation even in the face of high levels of gene flow (Linhart & Grant, 1996); and if genetic adaptation to different microhabitats does occur, then natural selection will favor barriers to gene flow among the differentially adapted subpopulations, because hybrids


would be maladaptive (Tregenza & Butlin, 1999). The ever-changing climate and other forms of environmental instability may cause frequent genetic reshuffling and thus prevent local genetic adaptation from resulting in speciation, if the taxa have not become reproductively isolated (e.g., Roy et al., 1996). Simple mechanisms, involving few loci, for achieving fast reproductive isolation would thus be highly conducive to parapatric speciation. Modeling further predicts that parapatric speciation requires that reproductive isolation be achieved by divergence in a relatively small number of mating trait loci even when environmental instability is not considered (Dieckmann & Doebeli, 1999; Kondrashov & Kondrashov, 1999). Such simple mechanisms have been documented for neotropical palms.

It appears that differences in reproductive biology, such as scent or flowering phenology, sufficient to prevent gene flow may arise quickly in geonomoid palms. In this group the chemical composition of the floral scent varies among taxa, probably causing differences in pollinator faunas that may prevent gene flow (Knudsen, 1999). Because floral scent varies not only among genera and species but also strongly within species complexes--Geonoma cuneata and G. macrostachys, for example--and among geographical subpopulations (Knud- sen, 1999), it seems to provide a mechanism by which reproductive isolation may arise quickly. Reproductive isolation may also occur due to differences in seasonal (e.g., Geonoma macrostachys var. acaulis [as G. acaulis] and var. macrostachys [as G. macrostachys] [Lis- tabarth, 1993, 1999]) and diurnal timing of anthesis (e.g., G. cuneata var. sodiroi and the closely related G. irena [Borchsenius, 1997a]). The genetic differences behind these reproductive differences are probably small (Borchsenius, 1997a), so reproductive isolation and thus speciation may be rapid.

Palms are dispersed mainly by birds and mammals (Zona & Henderson, 1989), and these seed-dispersal mechanisms may promote parapatric speciation. Seed dispersal by birds or other animals may cause speciation when important dispersers provide directed seed dispersal to certain (micro)habitats (cf. sections IV.A.3, IV.F.2.c, IV.G.2) and also mainly forage therein. For example, if a species has two important seed dispersers, one that prefers wet microsites and another that prefers hilltops, gene flow would be limited between the two subpopulations, subject to different selection pressures, and parapatric speciation could result. In agreement therewith, Givnish (1999) argued that seed dispersal by forest-interior birds may accelerate speciation in rain-forest understory plants because those birds have a poor ability to cross habitat barriers. He even mentioned as possible examples the species-rich understory palm genera Chamaedorea and Geonoma from the Neotropics and Dypsis from Madagascar, as well as the species-rich understory dicot genera Piper (Piperaceae), Psychotria (Rubi- aceae), and Solanum (Solanaceae). In support of this hypothesis, Eriksson and Bremer (1991) found that among shrubby Rubiaceae, genera with fleshy, mainly bird-dispersed fruits are much more species rich than are genera with dry, abiotically dispersed fruits. Although the authors suggested increased dispersal ability and allopatric speciation to be the cause, these patterns are also consistent with habitat-restricted dispersal and parapatric speciation. More passively, seed dispersal may also promote local differentiation, and thus parapatric speciation, simply by being highly limited (Linhart & Grant, 1996), as Eguiarte et al. (1993) found for ,4strocaryum mexicanum.

Overall, it seems likely that local genetic adaptation to habitat and microenvironmental heterogeneity is frequent in neotropical rain-forest palms, that natural selection subsequently favors reproductive isolation among the morphs, and that this can be achieved quite rapidly by differentiation in various apparently labile reproductive traits, as well as being promoted by seed dispersal. Parapatric speciation thus appears to be a probable mode of speciation in neotropical rain-forest palms, but what is the actual evidence thereof?.


C. PARAPATRIC SPECIATION AND EDAPHIC HETEROGENEITY

The edaphic parapatric-speciation model proposed by Gentry (1989) has been supported by the finding that edaphic specialization is frequent among pteridophytes, melastomes, and trees of neotropical lowland rain forests (e.g., ter Steege et al., 1993; Tuomisto & Ruokolainen, 1993; Tuomisto et al., 1995; Tuomisto & Poulsen, 1996; D. B. Clark et al., 1998). Support is also provided by the finding that the evolution of temperate meadow herbs in many phyloge- netically independent cases has been shaped by a trade-offbetween drought and flooding tolerance at the scale of the microenvironmental edaphic heterogeneity found within meadows (Silvertown et al., 1999). As discussed above, similar patterns, both among and within habitats, are found in palms (sections IV.E, IV.F, V.A.2). Even more indicative of such parapatric speciation is the occurrence of clear edaphic specialization found even within a species complex (Polyalthia hypoleuca, Annonaceae) ofpaleotropical rain-forest trees (Rogstad, 1990). Similar edaphic specialization even within genera and species complexes is also found in neotropical palms, notably Attalea, Astrocaryum, Bactris, and Geonoma (Henderson, 1995). In an upland site in Amazonian Ecuador the nine sympatric Geonoma taxa exhibit very diverse patterns of edaphic specialization (Svenning, 1999a): Geonoma cf. aspidiifolia and G. maxima are most common on hills, avoiding valley bottoms; G. macrostachys var. nov. prefers valley bottoms and poorly drained microsites; G. triglochin avoids poorly drained areas; and G. stricta var. stricta avoids sloping areas. Geonoma brongniartii appears to be present only as rare "sink" individuals in valley bottoms, representing mass effects from nearby abundant floodplain populations; and Geonoma stricta var. piscicauda is too rare for its edaphic preferences to be assessed (Svenning, 1999a). Geonoma macrostachys var. macrostachys is ubiquitous, but most common in valley bottoms and on hilltops (Sverming, 1999a), and appears to be represented by two different morphs in these microsites, a large one that is most frequent on hilltops and a small one that is very abundant and nearly restricted to valley bottoms (Svenning, pers. obs.). Similarly, Borchsenius (1999) noted that sympatric morphs of Geonoma cuneata have different microenvironmental topographic preferences. Thus, as exemplified by Geonoma macrostachys and G. cuneata, even within-habitat edaphic specialization among morphs of a species complex occurs. Edaphic specialization within species complexes is also known from the Malayan Iguanura geonomaeformis-wallichiana species complex (Dransfield, 1969). The occurrence of edaphic specialization within genera and even species complexes is in accordance with parapatric speciation. Interestingly, one of the mechanisms for achieving reproductive isolation discussed above, floral scent, is involved in maintaining the genetic integrity of the small and large G. macrostachys var. macrostachys morphs: although the two morphs are extremely similar except for adult size, they differ in floral-scent chemistry (Knudsen, 1999). Thus both of the processes that are involved in parapatric speciation, genetic adaptation to environmental heterogeneity and protection of the resulting genotypes by reproductive barriers, are indicated in the case of the G. macrostachys var. macrostachys morphs.

D. PARAPATRIC SPECIATION AND CANOPY HETEROGENEITY

Although parapatric speciation by edaphic specialization may well be an important speciation mode in neotropical palms, parapatric speciation occurring in microenvironmental heterogeneity in light availability is probably even more important. Many species-rich genera of understory palms exhibit a huge diversity in plant size, leaf size, and morphology, notably Bactris, Chamaedorea, and Geonoma in the Neotropics, and Areca, Dypsis, Gronophyllum, lguanura, Licuala, Neophloga, Pinanga, Ptychosperma, and Salacca in the Paleotropics


(Dransfield, 1989; Chazdon, 1991 b; Hodel, 1992). Dransfield (1989) suggested that this diversification evolved through neoteny, but that has yet to be proved (Chazdon, 1991b). For Geonoma (23 species), Chazdon (1991b) showed that leaf morphology is constrained by leaf size and that leaf size is constrained by plant size, both within and among species. A principal component analysis (PCA) of the morphological characteristics of adults of the different species produced a first axis corresponding to overall plant size (Chazdon, 1991b); and Bor- chsenius (1999), again using PCA, similarly found that four sympatric morphs in the Geonoma cuneata complex differ primarily in size. Thus much of the diversification of understory palms simply represents evolutionary divergence in size. Chazdon (1985, 1986a, 1986c) showed that variation in plant size in two Geonoma spp. and the related Asterogyne martiana causes variation in the whole-plant biomass-specific efficiency of light interception. Because these palms have similar physiological photosynthetic capabilities (Chazdon, 1986b), the smaller the plant, the higher its ability to maintain a positive carbon balance in shaded conditions (Chazdon, 1985, 1986a; cf. Givnish, 1988). Species that achieve reproductive maturity at a small size are thus be able to grow and reproduce at lower light levels than are larger species, whereas larger species are better able to exploit higher light levels (Chazdon, 1985, 1986a, 1986c). Thus changes in adult plant size have created the evolutionary potential for adaptation to various parts of the heterogeneous understory light environment and have probably been important in the radiation of Geonoma (Chazdon, 1986c, 1991b) and other understory palms. Variation in size and thus probably in shade tolerance is not limited to small understory palms but is also found in species-rich neotropical rain-forest genera composed mainly of canopy, midstory, and massive understory palms, notably Astro- caryum, Attalea, Oenocarpus, Syagrus, and Wettinia (Henderson et al., 1995). Kahn (1986) suggested that short-stemmed species of Attalea and Astrocaryum have arisen as an adaptation to closed-canopy conditions where light is insufficient to allow building a massive aerial stem. Leaf-architecture variation also suggests that light niche specialization has been important in the diversification of larger neotropical palms. Canopy palms (e.g., Astrocaryum chambira, A. standleyanum, and A. vulgare) tend to have multilayered leaf architectures: leaflets overlapping and oriented in multiple directions, or vertically oriented leaflets. Species of lower strata (e.g., Astrocaryum mexicanum, A. murumuru, and A, paramaca) tend to have leaflets horizontally oriented and nonoverlapping, thereby forming a monolayer (de Gran- ville, 1992; cf. section IV.A. 1 ). Specialization in microenvironmental light heterogeneity has clearly been important in the evolution of neotropical rain-forest palms, and parapatric speciation seems likely to have been important in the diversification of Geonoma and other species-rich genera.

E. ALTERNATIVE EVOLUTIONARY SCENARIOS

From the above discussion it seems clear beyond a doubt that adaptation to edaphic and light heterogeneity has been an important evolutionary force in neotropical palms. Still, this has not necessarily involved parapatric speciation; it may merely reflect niche differentiation among competing species already in existence (cf. Whittaker, 1965; Rogstad, 1990; Sven- ning, 1999a); that is, species that have arisen through other mechanisms (e.g., Sytsma & Schaal, 1985; Bush, 1994; Haffer, 1997; Niklas, 1997). Nevertheless, the occurrence of numerous sympatric morphs of several Geonoma spp., often exhibiting clear niche differences and mechanisms that promote reproductive isolation, is suggestive of parapatric speciation. Additional support is provided by the fact that, at least in the Geonoma cuneata complex, these morphs appear to be a mainly local phenomenon (Borchsenius, 1999). Even so, these


morphs (species) may also have arisen through sympatric speciation by stochastic changes in the apparently labile reproductive systems and may have become niche differentiated only after speciation. Thus parapatric speciation as an important speciation mechanism in neotropical palms is a hypothesis that needs future evaluation.

VII. Conclusions

Palms are an ecologically important and species-rich component of neotropical rain forests and have probably been so since the Upper Cretaceous. In this review I have shown that small-scale environmental heterogeneity is very important in the ecology of neotropical rain- forest palms, contributes to the local coexistence of sometimes more than 30 palm species, and probably is an important factor in the evolutionary diversification of these palms.

The neotropical rain forests are extremely heterogeneous environments at scales of 0.1-102 m. Numerous factors capable of affecting plant performance and distribution contribute to this microenvironmental heterogeneity: canopy structure and dynamics, intraspecific density and distance effects, interspecific plant resource and interference competition, litter layer, soil conditions, topography, and animal mutualists and pests. All of these factors affect individual performance and sometimes also the local distribution of neotropical rain- forest palms, and they do so in ways that often differ among different palm species. Even subtle microenvironmental differences can be crucial in determining the performance or distribution of palms, such as small differences in light availability in closed-canopy conditions, internal gap heterogeneity, objects on the forest floor, or the presence of particular canopy- tree species.

Microenvironmental heterogeneity promotes the local coexistence of palm species in several ways: by niche differences among the species, mass effects, and intraspecific negative density dependence. Though still not sufficiently studied, niche differences in terms of light requirements and edaphic-topographic preferences have been shown for sympatric species of understory, midstory, and canopy palms. Niche differences in terms of seed-dispersal patterns appear quite likely, too, but are purely speculative at this point. Demographic, architectural, and ecophysiological studies indicate that understory palms partition even the darkest part of the forest light gradient: the smaller the species, the shadier the sites it can exploit. Although the importance of mass effects has yet to be rigorously proved, local palm distributions suggest that they may account for the occurrence of a significant portion of the rarer species in many habitats and microhabitats. Strong intraspecific density dependence seems to be the rule among large-seeded palms, at least the generaAstroearyum and,4ttalea. It may not be an important phenomenon among small-seeded palms, and its importance for the mutual coexistence of even large-seeded palms has not yet been investigated.

Microenvironmental heterogeneity seems to be important not only for maintaining the richness of neotropical palm species but also for generating diversity. In species-rich palm genera and species complexes, sympatric species or morphs often differ in edaphic- topographic preferences or in characteristics, notably size, conferring differing light requirements; and sympatric morphs of several Geonoma species complexes exhibit divergence in reproductive traits that promote reproductive isolation. These patterns suggest parapatric, rather than allopatric, speciation by divergent selection over microenvironmental gradients.

To proceed farther in understanding the importance of microenvironmental heterogeneity in the ecology and evolution of neotropical rain-forest palms, I suggest that three lines of study are particularly needed.


1. Studies of niche differentiation at the scale of a complete palm community, or at least for all species of a given major growth form. These studies should include detailed assess- ments of the effects of subtle heterogeneity in light levels (e.g., Clark & Clark, 1992; Sven- ning, 2000b), litter layer (e.g., Molofsky & Augsburger, 1992; Cintra, 1997a), edaphic conditions (e.g., Rogstad, 1990; D. A. Clark et al., 1995; D. B. Clark et al., 1998; Silvertown et al., 1999), and the spatiotemporal patterns of seed dispersal (e.g., Julliot, 1997; Wenny & Levey, 1998) and, preferably, evaluate the population-level consequences of any differences (el. Svenning, in prep.).

2. Studies of pest-mediated density dependence, not only for large-seeded palms but also for small-seeded ones. These studies need to analyze the effects of both intraspecific and interspecific density and should include scales of>100 m as well as 1-10 m (cf. Cintra, 1997b; Fragoso, 1997).

3. Finally, to test the parapatric-speciation model, local-scale studies of species complexes (e.g., Bactris simplicifrons, Geonoma cuneata, or Geonoma macrostachys), integrating microenvironmental niche differentiation (cf. Rogstad, 1990; Clark & Clark, 1992; D. A. Clark et al., 1995; Davies et al., 1998; Svenning, 1999a, 2000b), reproductive biology (cf. Lista- barth, 1993, 1999; Borchsenius, 1997a; Knudsen, 1999), and population genetics (cf. Patton & Smith, 1992; Aitken & Libby, 1994; Linhart & Grant, 1996) are needed.

VIII. Acknowledgments

I thank Henrik Balslev for his comments on earlier drafts of this manuscript, Michael Kessler for providing me with his unpublished manuscript, and Elvira Balslev for improving the Spanish abstract. I am also grateful to the Centre for Tropical Biodiversity (Danish Natu- ral Science Research Council, grant #11-0390), Svend G. Fiedler og hustrus legat til fremme af botanisk og ark~eologisk forskning, the European Science Foundation's Programme on Tropical Canopy Research, and the Faculty of Natural Sciences at the University of Aarhus for economic support, INEFAN for research permits, and Pontificia Universidad Cat61ica del Ecuador for providing research facilities in Ecuador.

IX. Literature Cited

Aide, T. M. 1987. Limbfalls: A major cause of sapling mortality for tropical forest plants. Biotropica 19: 284-285.

Aitken, S. N. & W. J. Libby. 1994. Evolution of the pygmy-forest edaphic subspecies of Pinus contorta across an ecological staircase. Evolution 48: 1009-1019.

Alvarez-Buylla, E. R. & M. Slatkin. 1994. Finding confidence limits on population growth rates: Three real examples revisited. Ecology 75: 255-260.

Araus, J. L. & K. P. Hogan. 1994. Leaf structure and patterns of photoinhibition in two neotropical palms in clearings and forest understory during the dry season. Amer. J. Bot. 81 : 726-738.

Armbruster, W. S. 1995. The origins and detection of plant community structure: Reproductive versus vegetative processes. Folio Geobot. Phytotax. 30: 483-497.

Ashton, P. M. S. 1992. Establishment and early growth of advance regeneration of canopy trees in moist mixed-species forest. Pp. 101-122 in M. J. Kelty, B. C. Larson & C. D. Oliver (eds.), The ecology and silviculture of mixed-species forests: A Festschrift for David M. Smith. Kluwer Academic, Dordrecht, Netherlands.

Ashton, P. S. 1969. Speciation among tropical forest trees: Some deductions in the light of recent evidence. Biol. J. Linn. Soc. 1: 155-196.

42 THE BOTANICAl_. REVIEW

- - . 1989. Species richness in tropical forests. Pp. 239-251 in L. B. Holm-Nielsen, I. C. Nielsen & H. Bal- slev (eds.), Tropical forests: Botanical dynamics, speciation and diversity. Academic Press, London.

- - - . 1992. The structure and dynamics of tropical rain forest in relation to tree species richness. Pp. 53-64 in M. J. Kelty, B. C. Larson & C. D. Oliver (eds.), The ecology and silviculture of mixed- species forests: A Festschrift for David M. Smith. Kluwer Academic, Dordrecht, Netherlands.

Ataroff, M. & T. Schwarzkopf. 1994. Vegetative growth in Chamaedorea bartlingiana. Principes 38: 24-32.

Ayora, N. N. & R. Orellana. 1993. Physicochemical soil factors influencing the distribution of two coastal palms in Yucatan, Mexico. Principes 37: 82-91.

Baliek, M. J. (ed.). 1988. The palnv--Tree of life: Biology, utilization, and conservation: Proceedings of a symposium at the 1986 Annual Meeting of the Society for Economic Botany held at the New York Botanical Garden, Bronx, New York, 13-14 June 1986. Advances in Economic Botany 6. New York Botanical Garden, Bronx.

Balslev, H. & A. Barfod. 1987. Ecuadorean palms: An overview. Opera Bot. 92: 17-35. - - , d. Luteyn, B. Oligaard & L. B. Holm-Nielsen. 1987. Composition and structure of adjacent

unflooded and floodplain forest in Amazonian Ecuador. Opera Bot. 92: 37-57. Bannister, B. A. 1970. Ecological life cycle ofEuterpeglobosa Gaertn. Pp. B299-B314 in H. T Odum

& R. F. Pigeon (eds.), A tropical rainforest: A study of irradiation and ecology at E1 Verde, Puerto Rico. U.S. Atomic Energy Commission, Office of Information Services, Oak Ridge, TN.

Barrow, P., G. Duff, D. Liddle & J. Russell-Smith. 1993. Threats to monsoon rainforest habitat in northern Australia: The case of Ptychosperma bleeseri Burrer (Arecaceae). Austral. J. Ecol. 18: 463-471.

Basnet, K. 1992. Effect of topography on the pattern of trees in tabonuco (Dacryodes excelsa) dominated rain forest in Puerto Rico. Biotropica 24: 31-42.

Beard, J. S. 1944. Climax vegetation in tropical America. Ecology 25: 127-158. Becker, P., P. E. Rabenold, J. R. Idol & 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. Bernal, R. 1998. Demography of the vegetable ivory palm Phytelephas seemannii in Colombia, and the

impact of seed harvesting. J. Appl. Ecol. 35: 64-74. - - & H. Balslev. 1996. Strangulation of the palm Phytelephas seemannii by the pioneer tree Ce-

cropia obtusifolia: The cost of efficient litter trapping. Ecotropica 2:177-184. Bodmer, R. E. 1991. Strategies of seed dispersal and seed predation in Amazonian ungulates. Biotropica

23: 255-261. Bngh, A. 1996a. Abundance and growth of rattans in Khao Chong National Park, Thailand. Forest Ecol.

Managem. 84: 71-80. - - - . 1996b. The reproductive phenology and pollination biology of four Calamus (Arecaceae) spe-

cies in Thailand. Principes 40: 5-15. Bonal, D. 1997. Influence of some in situ environmental factors on growth performances of Calamus

caesius. J. Trop. Forest Sci. 9: 369-378. Borehsenius, F. 1997a. Flowering biology of Geonoma irena and G. cuneata var. sodiroi (Arecaceae).

PI. Syst. & Evol. 208: 187-196. - - . 1997b. Palm communities in western Ecuador. Principes 41: 93-99. - - . 1999. Morphological variation in Geonoma cuneata in western Ecuador. Pp. 131-139 in

A. Henderson & F. Borchsenius (eds.), Evolution, variation, and classification of palms. New York Botanical Garden Press, Bronx.

- - & F. Skov. 1997. Ecological amplitudes of Ecuadorian palms. Principes 41: 179-183. - - , H. B. Pedersen & H. Balslev. 1998. Manual to the palms of Ecuador. AAU Reports. Dept. of

Systematic Botany, University of Aarhus, Aarhus, Denmark. Botschek, J., J. Ferraz, M. Jahnel & A. Skowronek. 1996. Soil chemical properties ofa toposequence

under primary rain forest in the Itacoatiara vicinity (Amazonas, Brazil). Geoderma 72:119-132. Bovi, M. L. A., L. A. Sies, M. Cardoso & J. Cione. 1987. Densidade de plantio de palmiteiro en regime

de sombreamento permanente. Bragantia 46: 329-341. Boydak, M. 1985. The distribution of Phoenix theophrasti in the Datqa Peninsula, Turkey. Biol. Conser-

vation 32: 129-135.


Braker, E. & R. L. Chazdon. 1993. Ecological, behavioural and nutritional factors influencing use of palms as host plants by a neotropical forest grasshopper. J. Trop. Ecol. 9:183-197.

Brandani, A., G. S. Har tshorn & G. H. Orians. 1988. Internal heterogeneity of gaps and species richness in Costa Rican tropical wet forest. J. Trop. Ecol. 4:99-119.

Broschat, T. K., H. Donselman & D. MeConneil. 1989. Light acclimatization in Ptychosperma elegans. HortScience 24: 267-268.

Bullock, S. H. 1980. Demography of an undergrowth palm in littoral Cameroun. Biotropica 12: 247-255.

Bush, M. B. 1994. Amazonian speciation: a necessarily complex model. J. Biogeogr. 21: 5-17. Canham, C. D. 1988. An index for understory light levels in and around canopy gaps. Ecology 69:

1634-1638. - - , J. S. Denslow, W. J. Platt, J. R. Runkle, T. A. Spies & P. S. White�9 1990. Light regimes be-

neath closed canopies and tree-fall gaps in temperate and tropical forests. Canad. J. Forest Res. 20: 620-631.

Cer6n, C. E. & C. Motalvo A. 1997. Composici6n y estructura de una hect~rea de bosque en la Ama- zonia e c u a t o r i a n a ~ o n informaci6n etnobot~inica de los Huaorani. Pp. 153-172 in R. Valencia & H. Balslev (eds.), Estudios sobre diversidad y ecologia de plantas: Memorias del iI Congreso Ecua- toriano de Botfinica realizado en la Pontificia Universidad Cat61ica del Ecuador, Quito, 16-20 octubre 1995. Pontificia Universidad Cat61ica del Ecuador, Quito.

Chavez, F. 1996. Estudio preliminar de la familia Arecaceae (Palmae) en el Parque Nacional del Manu (Pakitza y Cocha Cashu). Pp. 141-168 in D. E. Wilson & A. Sandoval (eds.), Manu: The biodiversity of southeastern Peru. Smithsonian Institution, Washington, DC.

Chazdon, R. L. 1985. Leaf display, canopy structure, and light interception of two understory palm species. Amer. J. Bot. 72: 1493-1502.

- - - . 1986a. The costs of leaf support in understory palms: Economy versus safety. Amer. Naturalist 127: 9-30.

�9 1986b. Light variation and carbon gain in rain forest understorey palms. J. Ecol. 74:995-1012. �9 1986c. Physiological and morphological basis of shade tolerance in rain forest understory

palms. Principes 30: 92-99. �9 1991 a. Effects of leaf and ramet removal on growth and reproduction of Geonoma congesta, a

clonal understorey palm. J. Ecol. 79:1137-1146. �9 1991b. Plant size and form in the understory palm genus Geonoma: Are species variations on a

theme? Amer. J. Bot. 78: 680-694. �9 1992. Patterns of growth and reproduction of Geonoma congesta, a clustered understory palm.

Biotropica 24:43-51. �9 1996. Spatial heterogeneity in tropical forest structure: Canopy palms as landscape mosaics.

Trends Ecol. Evol. 11 : 8-9. - - & N. Feteher. 1984. Light environments of tropical forests. Pp. 27-36 in E. Medina, H. A. Moo-

ney& C. V~squez-Y~nes (eds.), Physiological ecology of plants of the wet Tropics: Proceedings of an international symposium held in Oxatepec and Los Tuxtlas, Mexico, June 29 to July 6, 1983. W. Junk, The Hague, Netherlands.

- - & R. W. Pearcy. 1991. The importance of sunflecks for forest understory plants. BioScience 41: 760-766.

, , D. W. Lee & N. Feteher. 1996. Photosynthetic responses of tropical forest plants to contrasting light environments. Pp. 5-55 in S. S. Mulkey, R. L. Chazdon & A. P. Smith (eds.), Tropical forest plant ecophysiology. Chapman & Hall, New York.

Cintra, R. 1997a. Leaf litter effects on seed and seedling predation of the palm Astrocaryum murumuru and the legume tree Dipteryx micrantha in Amazonian forest. J. Yrop. Ecol. 13: 709-725.

- - . 1997b. A test of the Janzen-Connell model with two common tree species in Amazonian forest. J. Trop. Ecol. 13: 641-658.

- - & V. Horna. 1997. Seed and seedling survival of the palm Astrocayum murumuru and the legume tree Dipteryx micrantha in gaps in Amazonian forest. J. Trop. Ecol. 13: 257-277.

Claney, K. E. & M. J. Sullivan�9 1990. Demography of the needle palm, Rhapidophyllum hystrix, in Mis- sissippi and Alabama. Principes 34: 64-78.


Clark, D. A. & D. B. Clark. 1992. Life history diversity of canopy and emergent trees in a neotropical rain forest. Ecol. Monogr. 62: 315-344.

- - , D. B. Clark, R. Sandoval M. & M. V. Castro C. 1995. Edaphic and human effects on landscape-scale distributions of tropical rain forest palms. Ecology 76: 2581-2594.

Clark, D. B. 1990. The role of disturbance in the regeneration ofneotropical moist forests. Pp. 291-315 in K. S. Bawa & M. Hadley (eds.), Reproductive ecology of tropical forest plants. UNESCO, Paris.

- - . , D. A. Clark & P. M. Rich. 1993. Comparative analysis of microhabitat utilization by saplings of nine tree species in neotropical rain forest. Biotropica 25: 397-407.

., , -, S. Weiss & S. F. Oberbauer. 1996. Landscape-scale evaluation of understory light and canopy structure: Methods and application in a neotropical lowland rain forest. Canad. J. Forest Res. 26: 747-757.

�9 , & J. M. Read. 1998. Edaphic variation and mesoscale distribution of tree species in a neotropical rain forest. J. Ecol. 86:101-112.

Cochran, M. E. & S. EIIner. 1992. Simple methods for calculating age-based life history parameters for stage-structured populations. Ecol. Monogr. 62: 345-364.

Colinvaux, P. A. 1998. A new vicariance model for Amazonian endemics. Global Ecol. & Biogeogr. Lett. 7: 95-96.

Collins, B. S., K. P. Dunne & S. T. A. Pickett. 1985. Responses of forest herbs to canopy gaps. Pp. 217-234 in S. T. A. Pickett & P. S. White (eds.), The ecology of natural disturbance and patch dynamics. Academic Press, Orlando, FL.

Condit, R., S. P. Hubbell & R. B. Foster. 1996. Changes in tree species abundance in a neotropical forest: Impact of climate change. J. Trop. Ecol. 12: 231-256.

Connell, J. H. 1971. On the role of natural enemies in preventing competitive exclusion in some marine animals and rain forest trees. Pp. 298-312 in P. J. Den Boer & G. R. Gradwell (eds.), Dynamics of populations: Proceedings of the Advanced Study Institute on Dynamics of Numbers in Popula- tions, Oosterbeek, the Netherlands, 7-18 September 1970. Centre for Agricultural Publishing and Documentation, Wageningen, Netherlands.

- - . 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302-1310. Cracraft, J. & R. O. Prum. 1988. Patterns and processes of diversification: Speciation and historical

congruence in some neotropical birds. Evolution 42: 603-620. Crawley, M. J. 1997a. Life history and environment. Pp. 73-131 in M J. Crawley (ed.), Plant ecology.

Ed. 2. Blackwell Science, Oxford. - - (ed.). 1997b. Plant ecology. Ed. 2. Blackwell Science, Oxford. - - . 1997c. The structure of plant communities. Pp. 475-531 in M. J. Crawley (ed.), Plant ecology.

Ed. 2. Blackwell Science, Oxford. Cunningham, S. A. 1995. Ecological constraints on fruit initiation by Calyptrogyne ghiesbreghtiana

(Arecaceae): Floral herbivory, pollen availability, and visitation by pollinating bats. Amer. J. Bot. 82: 1527-1536.

- - . 1996. Pollen supply limits fruit initiation by a rain forest understorey palm. J. Ecol. 84: 185-194.

- - . 1997. The effect of light environment, leaf area, and stored carbohydrates on inflorescence production by a rain forest understory palm. Oecologia 111 : 36-44.

Curi, N. & D. P. Franzmeier. 1984. Toposequence of oxisols from the central plateau of Brazil. Soil Sci. Soc. Amer. J. 48: 341-346.

Daghlian, C. P. 1978. Coryphoid palms from the lower and middle Eocene of southeastern North Amer- ica. Palaeontographica, Abt. B, Pal~tophytol. 166: 44-82.

- - . 1981. A review of the fossil record of monocotyledons. Bot. Rev. (Lancaster) 47: 517-555. Davies, S. J., P. A. Palmiotto, P. S. Ashton, H. S. Lee & J. V. LaFrankie. 1998. Comparative ecology

of I 1 sympatric species of Macaranga in Borneo: Tree distribution in relation to horizontal and vertical resource heterogeneity. J. Ecol. 86: 662-673.

De Granville, J.-J. 1984. Monocotyledons and pteridophytes as indicators of environmental constraints in the tropical vegetation. Candollea 39: 265-269.

- - . . 1992. Life forms and growth strategies of Guianan palms as related to their ecology. Bull. Inst. Fran. l~tudes Andines 21: 533-548.


De Steven, D. 1989. Genet and ramet demography of Oenocarpus mapora ssp. mapora, a clonal palm of Panamanian tropical moist forest. J. Ecol. 77: 579-596.

- - & F. E. Putz. 1985. Mortality rates of some rain forest palms in Panama. Principes 29:162-165. ~ , D. M. Windser, F. E. Putz & B. de Leon. 1987. Vegetative and reproductive phenologies of a

palm assemblage in Panama. Biotropica 19: 342-356. Delobel, A., G. Couturier, F. Kahn & J. A. NUsson. 1995. Trophic relationships between palms and

bruchids (Coleoptera: Bruchidae: Pachymerini) in Peruvian Amazonia. Amazoniana 8: 209-219. Denslow, J. S. 1987. Tropical rainforest gaps and tree species diversity. Ann. Rev. Ecol. Syst. 18:

431-451. - - , E. Newell & A. M. Ellison. 1991. The effect of understory palms and cyclanths on the growth

and survival of Inga seedlings. Biotropica 23: 225-234. - - . , A. M. Ellison & R. E. Sanford. 1998. Treefall gap size effects on above- and below-ground

processes in a tropical wet forest. J. Ecol. 86: 597-609. Dieckmann, U. & M. Doebeli. 1999. On the origin of species by sympatric speciation. Nature 400:

354-357. Dietrich, W. E., D. M, Windsor & T. Dunne. 1996. Geology, climate, and hydrology of Barro Colorado

Island. Pp. 21-46 in E. G. Leigh Jr., A. S. Rand & D. M. Windsor (eds.), The ecology of a tropical forest: Seasonal rhythms and long-term changes. Smithsonian Institution Press, Washington, DC.

Dobzhansky, T. 1950. Evolution in the Tropics. Amer. Sci. 32: 209-221. Dowe, J. L., J. Benzie & E. Ballment. 1996. Ecology and genetics of Carpoxylon macrospermum H.

Wendl. & Drude (Arecaceae), an endangered palm from Vanuatu. Biol. Conservation 79:205-216. Dransfield, J. 1969. Palms in the Malayan forest. Malayan Nat. J. 22: 144-151.

�9 1978. Growth forms of rain forest palms. Pp. 247-268 in P. B. Tomlinson & M. H. Zimmermann (eds.), Tropical trees as living systems: The proceedings of the fourth Cabot symposium held at Harvard Forest, Petersham, Massachusetts, on April 26-30, 1976. Cambridge University Press, Cambridge, England.

~ . 1988. Prospects for rattan cultivation. Pp. 190-200 in M. J. Balick (ed.), The palm--Tree of life: Biology, utilization, and conservation. Advances in Econonmic Botany 6. New York Botanical Garden, Bronx.

- - . 1989. Speciation patterns in the palms of Madagascar. Pp. 153-171 in L. B. Holm-Nielsen, I. C. Nielsen & H. Balslev (eds.), Tropical forests: Botanical dynamics, speciation and diversity. Aca- demic Press, London.

~ . 1999. Species and species concepts in Old World palms. Pp. 5-20 in A. Henderson & F. Bor- chsenius (eds.), Evolution, variation, and classification of palms. New York Botanical Garden Press, Bronx.

Duivenvoorden, J. F. 1995. Tree species composition and rain forest-environment relationships in the middle Caquetfi area, Colombia, NW Amazonia. Vegetatio 120:91-113.

Duvall, M. R., M. T. Clegg, M. W. Chase, W. D. Clark, W. J. Kress, H. G. Hills, L. E. Eguiarte, J. F. Smith, B. S. Gaut, E. A. Zimmer & G. H. Learn Jr. 1993. Phylogenetic hypotheses for the monocotyledons constructed from rbcL sequence data. Ann. Missouri Bot. Gard. 80: 607-619.

Eguiarte, L. E., N. Perez-Nasser & D, Piilero. 1992. Genetic structure, outcrossing rate and heterosis in Astrocaryum mexicanum (tropical palm): Implications for evolution and conservation. Heredity 69: 217-228.

- - , A. Btirqez, J. Rodrignez, M. Martinez-Ramos, J. Sarukh~in & D. Pifiero. 1993. Direct and indirect estimates of neighborhood and effective population size in a tropical palm, Astrocaryum mexicanum. Evolution 47: 75-87.

Ellison, A. M., J. S. Denslow, B. A. Loiselle & D. Bren6s M, 1993. Seed and seedling ecology ofneotropical Melastomataceae. Ecology 74:1733-1749.

Endler, J. A. 1982. Pleistocene forest refuges: Fact or fancy? Pp. 641-657 in G. T. Prance (ed.), Biologi- cal diversification in the Tropics: Proceedings of the fifth international symposium of the Associa- tion for Tropical Biology, held at Macuto Beach, Caracas, Venezuela, February 8-13, 1979. Columbia University Press, New York.

Enright, N. J. 1992. Factors affecting reproductive behaviour in the New Zealand nikau palm, Rhopa- lostylis sapida Wendl. et Drude. New Zealand J. Bot. 30: 69-80.


- - & A. D. Watson�9 1992. Population dynamics of the nikau palm, Rhopalostylis sapida (Wendl. et Drude), in a temperate forest remnant near Auckland, New Zealand. New Zealand J. Bot. 30: 29-43.

Eriksson, O. & B. Bremer. 1991. Fruit characteristics, life forms, and species richness in the plant family Rubiaceae. Amer. Naturalist 138: 751-761.

Faber-Langendoen, D. & A. H. Gentry. 1991. The structure and diversity of rain forests at Bajo Calima, Choc6 region, western Columbia. Biotropica 23: 2-I 1.

Facelli, J. M. & S. T. A. Pickett. 1991. Plant litter: Its dynamics and effects on plant community structure. Bot. Rev. (Lancaster) 57: 1-32.

Fedorov, A. A. 1966. The structure of the tropical rain forest and speciation in the humid Tropics. J. Ecol. 54:1-11.

Feinsinger, P. 1987. Effects of plant species on each other's pollination: is community structure influenced? Trends Ecol. Evol. 2: 123-126.

Forget, P.-M. 1991. Scatterhoarding ofAstrocaryum paramaca by Proechimys in French Guiana: Com- parison with Myoprocta exilis. Trop. Ecol. 32: 155-167.

�9 1994. Recruitment pattern of Vouacapoua americana (Caesalpiniaceae), a rodent-dispersed tree species in French Guiana. Biotropica 26: 408-419.

- - - . 1997. Effect of microhabitat on seed fate and seedling performance in two rodent-dispersed tree species in rain forest in French Guiana. J. Ecol. 85: 693-703.

- - & D. Sabatier. 1997. Dynamics of the seedling shadow of a frugivore-dispersed tree species in French Guiana. J. Trop. Ecol. 13: 767-773.

Fowler, N. 1988. The effects of environmental heterogeneity in space and time on the regulation of populations and communities. Pp. 249-269 in A. J. Davy, M. J. Hutchings & A. R. Watkinson (eds.), Plant population ecology: The 28th Symposium of the British Ecological Society, Sussex, 1987. Blackwell Scientific, Oxford.

Fragoso, J. M. V. 1997. Tapir-generated seed shadows: Scale-dependent patchiness in the Amazon rain forest. J. Ecol. 85: 519-529.

Frangi, J. L. & A. E. Lugo. 1998. A floodplain palm forest in the Luquillo Mountains of Puerto Rico five years after Hurricane Hugo. Biotropica 30: 339-348.

Fredeen, A. L. & C. B. Field. 1996. Ecophysiological constraints on the distribution of Piper species. Pp. 597-618 in S. S. Mulkey, R. L. Chazdon & A. P. Smith (eds.), Tropical forest plant ecophysiology. Chapman & Hall, New York.

Galettl, M., M. Paschoal & F. Pedroni. 1992. Predation on palm nuts (Syagrus romanzoffiana) by squirrels (Sciurus ingrami) in south-east Brazil. J. Trop. Ecol. 8: 121-123.

Gartlan, J. S., D. M. Newbery, D. W. Thomas & P. G. Waterman. 1986. The influence of topography and soil phosphorous on the vegetation of Korup Forest Reserve, Cameroun. Vegetatio 65: 131-148.

Gentry, A. H. 1988. Changes in plant community diversity and floristic composition on environmental and geographical gradients. Ann. Missouri Bot. Gard. 75: 1-34.

- - . . 1989. Speciation in tropical forests. Pp. 113-134 in L. B. Holm-Nielsen, 1. C. Nielsen & H. Bals- lev (eds.), Tropical forests: Botanical dynamics, speciation and diversity. Academic Press, Lon- don.

George, L. O. & F. A. Bazzaz. 1999a. The fern understory as an ecological filter: Emergence and establishment of canopy-tree seedlings. Ecology 80: 833-845.

- - & .1999b. The fern understory as an ecological filter: Growth and survival of canopy- tree seedlings. Ecology 80: 846-856.

Giller, P. S. 1984. Community structure and the niche. Chapman & Hall, London. Givnish, T. 1979. On the adaptive significance of leaf form. Pp. 375-407 in O. T. Solbrig, S. Jain, G. B.

Johnson & P. H. Raven (eds.), Topics in plant population biology. Columbia University Press, New York.

- - . . 1982. On the adaptive significance of leaf height in forest herbs. Amer. Naturalist 120:353-381. - - . . 1988. Adaptation to sun and shade: A whole-plant perspective. Austral. J. P1. Physiol. 15:

63-92. - - - . 1999. On the causes of gradients in tropical tree diversity. J. Ecol. 87: 193-210.


Graham, A. & D. L. I)ileher. 1998. Studies in neotropical paleobotany, XII. A palynoflora from the Pliocene Rio Banano Formation of Costa Rica and the Neogene vegetation of Mesoamerica. Amer. J. Bot. 85: 1426-1438.

Grime, J. P. 1965. Shade tolerance in flowering plants. Nature 208: 161-163. Gruhb, P. J. 1977. The maintenance of species-richness in plant communities: The importance of the re-

generation niche. Biol. Rev. Cambridge Philos. Soc. 52: 107-145. - - . 1996. Rainforest dynamics: The need for new paradigms. Pp. 215-233 in D. S. Edwards, W. E.

Booth & S. C. Choy (eds.), Tropical rainforest research--Current issues: Proceedings of the con- ference held in Bandar Seri Begawan, April 1993. Kluwer Academic, Dordrecht, Netherlands.

Guariguata, M. R., R. L. Chazdon, J. S. 1)enslow, J. M. Dupuy & L. Anderson. 1997. Structure and floristics of secondary and old-growth forest stands in lowland Costa Rica. P1. Ecol. 132:197-120.

Haffer, J. 1997, Alternative models of vertebrate speciation in Amazonia: An overview. Biodiv. & Con- serv. 6: 451-476.

Henderson, A. 1990. Arecaceae, I. Introduction and the Iriarteinae. Flora Neotropica Monograph 51. New York Botanical Garden, Bronx.

- - . . 1995. The palms of the Amazon. Oxford University Press, New York. , G. Galeano & R. Bernal. 1995. Field guide to the palms of the Americas. Princeton University

Press, Princeton, NJ. Herzog, S. K. & M. Kessler. 1997. Dieta de una colonia de gu~charos (Steatornis caripensis) en el Par-

que Nacional Carrasco, Cochabamba, Bolivia. Ecol. Bolivia 30: 69-73. Hoeh, G. A. & G. H. Adler. 1997. Removal of black palm (Astrocaryum standleyanum) seeds by spiny

rats (Proechimys semispinosus). J. Trop. Ecol. 13: 51-58. Hodel, D. R. 1992. Chamaedorea palms: The species and their cultivation. International Palm Society,

Lawrence, KS. Hogan, K. P. 1988. Photosynthesis in two neotropical palm species. Func. Ecol. 2: 371-377. Hoorn, C. 1994. An environmental reconstruction of the palaeo-Amazon River system (Middle-Late

Miocene, NW Amazonia). Palaeogeogr. Palaeoclimatol. Palaeoecol. 112:187-238. Hubbell, S. P., R. B. Foster, S. T. O'Brien, K. E. Harms, R. Condit, B. Weehsler, S. J. Wright &

S. Loo de Lao. 1999. Light-gap disturbances, recruitment limitation, and tree diversity in a neotropical forest. Science 283: 554-557.

Hutehings, M. J. 1997. The structure of plant populations. Pp. 325-358 in M. J. Crawley (ed.), Plant ecology. Blackwell Science, Oxford.

Janzen, D. H. 1970. Herbivores and the number of tree species in tropical forests. Amcr. Naturalist 104: 501-528.

- - . 1971. The fate of Scheelea rostrata fruits beneath the parent tree: Predispersal attack by bruchids. Principes 15: 89-101.

- - . 1972. Association of a rainforest palm and seed-eating beetles in Puerto Rico. Ecology 53: 258-261.

Johnson, C. D., S. Zona & J. A. Nilsson. 1995. Bruchid beetles and palm seeds: Recorded relationships. Principes 39: 25-35.

Johnson, D. (ed.). 1996. Palms: Their conservation and sustained utilization. Status survey and conservation action plan. IUCN, Gland, Switzerland.

Johnston, M. H. 1992. Soil-vegetation relationships in a tabonuco forest community in the Luquillo Mountains of Puerto Rico. J. Trop. Ecol. 8: 253-263.

Julliot, C. 1997. Impact of seed dispersal by red howler monkeys Alouatta seniculus on the seedling population in the understorey of tropical rain forest. J. Ecol. 85:431-440.

Kabakoff, R. P. & R. L. Chazdon. 1996. Effects of canopy species dominance on understory light availability in low-elevation secondary forest stands in Costa Rica. J. Trop. Ecol. 12: 779-788.

Kahn, F. 1986. Life forms of Amazonian palms in relation to forest structure and dynamics. Biotropica 18: 214-218.

�9 . 1987. The distribution of palms as a function of local topography in Amazonian terra-firme forests. Experientia (Basel) 43: 251-259.

- - & A. de Castro. 1985. The palm community in a forest of central Amazonia, Brazil. Biotropica 17: 210-216.


- - & J.-J. de Granville. 1992. Palms in forest ecosystems of Amazonia. Ecological Studies, 95. Springer-Verlag, Berlin.

- - & K. Mejia. 1987. Notes on the biology, ecology, and use of a small Amazonian palm: Lepido- caryurn tessmannii. Principes 31: 14-19.

, - - & A. de Castro. 1988. Species richness and density of palms in terra firme forests of Amazonia. Biotropica 20: 266-269.

Kessler, M. 2000. Upslope-directed mass effect in palms along an Andean elevational gradient: A cause for high diversity at mid-elevation? Biotropica 32: 756-758.

Kiew, R. 1972. The natural history of Iguanura geonomaeformis Martius: A Malayan undergrowth palmlet. Principes 16: 3-10.

Kiltie, R. A. 1981. Distribution of palm fruits on a rain forest floor: Why white-lipped peccaries forage near objects. Biotropica 13: 141-145.

- - . . 1993. New light on forest shade. Trends Ecol. Evol. 8: 39-40. Knudsen, J. T. 1999. Floral scent chemistry in geonomoid palms (Palmae: Geonomeae) and its impor-

tance in maintaining reproductive isolation. Pp. 14 l-168 in A. Henderson & F. Borchsenius (eds.), Evolution, variation, and classification of palms. New York Botanical Garden Press, Bronx.

Kobe, R. K. 1999. Light gradient partitioning among tropical tree species through differential seedling mortality and growth. Ecology 80:187-201.

Kohyama, T. & P. J. Grubb. 1994. Below- and above-ground allometries of shade-tolerant seedlings in a Japanese warm-temperate rain forest. Func. Ecol. 8: 229-236.

Kondrashov, A. S. & F. A. Kondrashov. 1999. Interactions among quantitative traits in the course of sympatric speciation. Nature 400:351-354.

Korning, J., K. Thomsen, K. Dalsgaard & P. Nernberg. 1994. Characters of three udults and their relevancc to the composition and structure of virgin rain forest of Amazonian Ecuador. Geoderma 63: 145-164.

Lechowicz, M. J. & G. Bell. 1991. The ecology and genetics of fitness in forest plants, II. Microspatial heterogeneity of the edaphic environment. J. Ecol. 79: 687-696.

Lee, D. W. 1986. Unusual strategies of light absorption in rain-forest herbs. Pp. 105-131 in T. J. Givnish (ed.), On the economy of plant form and function: Proceedings of the Sixth Maria Moors Cabot Symposium, Evolutionary Constraints on Primary Productivity, Adaptive Patterns of En- ergy Capture in Plants, Harvard Forest, August 1983. Cambridge University Press, Cambridge, England.

Leigh, E. G., Jr., S. J. Wright, E. A. Herre & F. E. Putz. 1993. The decline of tree diversity on newly isolated tropical islands: A test of a null hypothesis and some implications. Evol. Ecol. 7:76-102.

Lescure, J.-P., L. Emperaire & C. Franciscon. 1992. Leopoldinia piassaba Wallace (Arecaceae): A few biological and economic data from the Rio Negro region (Brazil). Forest Ecol. Managem 55: 83-86.

Levin, D. A. 1988. Local differentiation and the breeding structure of plant populations. Pp. 305-329 in L. D. Gottlieb & S. K. Jain (eds.), Plant evolutionary biology. Chapman & Hall, London.

Lieberman, M., D. Lieberman, G. S. Hartshorn & R. Peralta. 1985. Small-scale altitudinal variation in lowland wet tropical forest vegetation. J. Ecol. 73: 505-516.

- - , D. Lieberman & R. Peralta. 1989. Forests are not just Swiss cheese: Canopy stereogeometry of non-gaps in tropical forests. Ecology 70: 550-552.

Linhart, Y. B. & M. C. Grant. 1996. Evolutionary significance of local genetic differentiation in plants. Ann. Rev. Ecol. Syst. 27: 237-277.

Listabarth, C. 1993. Pollination in Geonoma macrostachys and three congeners, G. acaulis, G. gracilis, and G. interrupta. Bot. Acta 106: 496-506.

- - - . 1999. Pollination studies of palm populations: a step toward the application of a biological species concept. Pp. 79-93 in A. Henderson & F. Borchsenius (eds.), Evolution, variation, and classification of palms. New York Botanical Garden Press, Bronx.

Losos, E. 1995. Habitat specificity of two palm species: Experimental transplantation in Amazonian successional forests. Ecology 76: 2595-2606.

Luescher, A. & P. Jacquard. 1991. Coevolution between interspecific plant competitors? Trends Ecol. Evol. 6: 355-358.


Lugo, A. E. & C. T. R. Batlle. 1987. Leaf production, growth rate and age of the palm Prestoea montana in the Luquillo Experimental Forest, Puerto Rico. J. Trop. Ecol. 3: 151-161.

Mack, A. L. 1995. Distance and non-randomness of seed dispersal by the dwarf cassowary Casuarinus bennetti. Ecography 18: 286-295.

Martinez-Ramos, M., E. Alvarez-Buylla, J. Sarukhitn & D. Pifiero. 1988a. Treefall age determina- tion and gap dynamics in a tropical forest. J. Ecol. 76: 700-716.

- - , J. Sarukhitn & D. Pifiero. 1988b. The demography of tropical trees in the context of forest gap dynamics: the case ofAstrocaryum mexicanum at Los Tuxtlas tropical rain forest. Pp. 293-313 in A. J. Davy, M. J. Hutchings & A. R. Watkinson (eds.), Plant population ecology: The 28th Sympo- sium of the British Ecological Society, Sussex, 1987. Blackwell Scientific, Oxford.

Martius, C. & A. G. Bandeira. 1998. Wood litter stocks in tropical moist forest in central Amazonia. Ecotropica 4:115-118.

McPherson, K. & K. Williams. 1996. Establishment growth of cabbage palm, Sabal palmetto (Are- caceae). Amer. J. Bot. 83: 1566-1570.

Mendoza, A. & M. Franco. 1992. Integraci6n clonal en una palma tropical. Bull. Inst. Fran. l~tudes An- dines 21: 623~535.

- - . , D. Pifiero & J. Sarukhiin. 1987. Effects of experimental defoliation on growth, reproduction and survival ofAstrocaryum mexicanum. J. Ecol. 75: 545-554.

Molofsky, J. & C. K. Augsburger. 1992. The effect of leaf litter on early seedling establishment in a tropical forest. Ecology 73: 68-77.

Moore, H. E. 1973. The major groups of palms and their distributions. Gentes Herb. 11: 27-140. Moraes, M., G. Galeano, R. Bernal, H. Balslev & A. Henderson. 1995. Tropical Andean palms (Are-

caceae). Pp. 473-487 in S. P. Churchill, H. Balslev, E. Forero & J. L. Luteyn (eds.), Biodiversity and conservation of neotropical montane forests: Proceedings of the Neotropical Montane Forest Biodiversity and Conservation Symposium, the New York Botanical Garden, 21-26 June 1993. New York Botanical Garden, Bronx.

Morell, V. 1999. Ecology returns to speciation studies. Science 284:2106-2108. Muller, J. 1981. Fossil pollen record of extant angiosperms. Bot. Rev. (Lancaster) 47: 1-142. Nepstad, D. C., A. Verlssimo, A. Alencar, C. Nobre, E. Lima, P. Lefebvre, P. Schlesinger, C. Potter,

P. Moutinho, E. Mendoza, M. Cochrane & V. Brooks. 1999. Large-scale impoverishment of Amazonian forests by logging and fire. Nature 398: 505-508.

Newbery, D. M. & J. Proctor. 1984. Ecological studies in four contrasting lowland rain forests in Gunung Mulu National Park, Sarawak, IV. Associations between tree distribution and soil factors. J. Ecol. 72: 475-493.

- - - , J. S. Gartlan, D. B. McKey & P. G. Waterman. 1986. The influence of drainage and soil phosphorous on the vegetation of Douala-Edea Forest Reserve, Cameroun. Vegetatio 65: 149-162.

Niklas, K. J. 1997. The evolutionary biology of plants. University of Chicago Press, Chicago. Nfiilez-Farfan, J. & R. Dirzo. 1988. Within-gap heterogeneity and seedling performance in a Mexican

tropical forest. Oikos 51: 274-284. Olesen, J. M., A. M. Christensen, L. I. EskUdsen, J.-C. Svenning & R. Lindberg. In press. Plants in

the garden of devil: Intruders of an ant-plant mutualism. Ecotropica. OImsted, L & E. R. AIvarez-Buylla. 1995. Sustainable harvesting of tropical trees: Demography and

matrix models of two palm species in Mexico. Ecol. Appl. 5: 484-500. Orr, M. R. & T. B. Smith. 1998. Ecology and speciation. Trends Ecol. Evol. 13: 502-506. Ostertag, R. 1998. Belowground effects of canopy gaps in tropical wet forest. Ecology 79:1294-1304. Oyama, K. 1991. Seed predation by a curculionid beetle on the dioecious palm Chamaedorea tepejilote.

Principes 35: 156-160. - - & R. Dirzo. 1991. Ecological aspects of the interaction between Chamaedorea tepejilote, a dioe-

cious palm, and Calyptocephala marginipennis, a herbivorous beetle, in a Mexican rain forest. Principes 35: 86-93.

- - & A. Mendoza. 1990. Effects of defoliation on growth, reproduction, and survival ofa neotropical dioecious palm, Chamaedorea tepefilote. Biotropica 22: 119-123.

Patton, J. L. & M. F. Smith. 1992. mtDNA phylogeny of Andean mice: A test of the diversification across ecological gradients. Evolution 46:174-183.


Pedersen, H. B. & H. Balslev. 1992. The economic botany of Ecuadorean palms. Pp. 173.-191 in M. J Plotkin & L. M. Famolare (eds.), Sustainable harvest and marketing of rain forest products. Island Press, Washington, DC.

Peres, C. A. 1994. Composition, density, and fruiting phenology of arborescent palms in an Amazonian terra firme forest. Biotropica 26: 285-294.

Phillips, O. L., P. Hall, A. H. Gentry, S. A. Sawyer & R. V~tsquez. 1994. Dynamics and species richness of tropical rain forests. Proc. Natl. Acad. U.S.A. 91: 2805-2809.

Pinard, M. 1993. Impacts of stem harvesting on populations oflriartea deltoidea (Palmac) in an extrac- tive reserve in Acre, Brazil. Biotropica 25 :2 -14

- - & F. E. Putz. 1992. Population matrix models and palm resource management Bull. Inst. Fran. l~tudes Andines 21: 637-649.

Pifiero, D. & J. Sarukh/m. 1982. Reproductive behaviour and its individual variation in a tropical palm, Astrocaryum mexicanum. J. Ecol. 70: 461-472.

& P. Alberdi. 1982. The costs of reproduction in a tropical palm, Astrocaryum mexicanum. J. Ecol. 70: 473-481.

- - - , M. Martlnez-Ramos, A. Mendoza, E. Alvarez-Buylla & J. Sarukh~in. 1986. Demographic studies in Astrocaryum mexicanum and their use in understanding community dynamics. Principes 30:108-116.

Poulsen, A. D. & H. Balslev. 1991. Abundance and cover of ground herbs in an Amazonian rain forest. J. Veg. Sci. 2: 315-322.

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.

�9 1984. How trees avoid and shed lianas. Biotropica 16: 19-23. �9 1990. Growth habits and trellis requirements of climbing palms (Calamus spp.) in North-eastern

Queensland. Austral. J. Bot. 38: 603-608. Raich, J. W. & G. W. Khoon. 1990. Effects of canopy openings on tree seed germination in a Malaysian

dipterocarp forest. J. Trop. Ecol. 6: 203-217. Ratsirarson, J., J. A. Silander Jr . & A. F. Richard. 1996. Conservation and management of a threat-

ened Madagascar palm species, Neodypsis decaryi, Jumelle. Conservation Biol. 10: 40-52. Rich, P. M. 1986. Mechanical architecture of arborescent rain forest palms. Principes 30:117-131. - - , S. Lure, L. Mufios E. & M. Quesada A. 1987. Shedding of vines by the palms Welfiageorgii

and lriartea gigantea. Principes 31 : 31-40. Richards, P. & G. B. Williamson. 1975. Treefalls and patterns of understory species in a wet lowland

tropical forest. Ecology 56:1226-1229. Richter, D. D. & L. I. Babbar . 1991. Soil diversity in the Tropics. Advances Ecol. Res. 21: 315-

389. Ricklefs, R. E. 1977. Environmental heterogeneity and plant species diversity: A hypothesis. Amer.

Naturalist 111:376-381. Rogstad, S. H. 1989. The biosystematics and evolution of the Polyalthia hypoleuca species complex

(Annonaceae) of Malesia, I. Systematic treatment J. Arnold Arbor. 70:153 246. - - . 1990. The biosystematics and evolution of the Polyalthia hypoleuca species complex (Annona-

ceae) of Malesia, II. Comparative distributional ecology. J. Trop. Ecol. 6 :387-408 Romoleroux, K., R. Foster, R. Valencia, R. Condit, H. Balslev & E. Losos. 1997. Arboles y arbustos

(dap>l cm) encontrados en dos hect~reas de un bosque de la Amazonia ecuatoriana. Pp. 189-215 in R. Valencia & H. Balslev (eds.), Estudios sobre diversidad y ecologla de plantas: Memorias del II Congreso Ecuatoriano de Bot~inica realizado en la Pontificia Universidad Cat61ica del Ecuador, Quito, 16-20 oetubre 1995. Pontificia Universidad Cat61ica del Ecuador, Quito.

Roosevelt, A. C., M. L. da Costa, C. L. Machado, M. Michab, N. Mercier, H. Valladas, J. Feathers, W. Barnett, M. I. da Silveira, A. Henderson, J. Silva, B. Chernoff, D. S. Reese, J. A. Holman, N. Toth & K. Schick. 1996. Paleoindian cave dwellers in the Amazon: The peopling of the Ameri- cas. Science 272: 373-384.

Roy, K., J. W. Valentine, D. Jablonski & S. M. Kidwell. 1996. Scales of climatic variability and time averaging in Pleistocene biotas: Implications for ecology and evolution. Trends Ecol. Evol. 11: 458-463.


Rull, V. 1998. Biogeographical and evolutionary considerations ofMauritia (Arecaceae), based on paly- nological evidence. Rev. Paleobot. Palynol. 100: 109-122.

Rank, J. V. 1998. Productivity and sustainability of a vegetable ivory palm (Phytelephas aequatorialis, Arecaceae) under three management regimes in northwestern Ecuador. Econ. Bot. 52:168-182.

S a b a t i e r , D., M. Grimaldi, M.-F. Pr6vost, J. Guillanme, M. Godron, M. Dosse & P. Curtal. 1997. The influence of soil cover organization on the floristic and structural heterogeneity of a Guianan rain forest. PI. Ecol. 131: 81-108.

Salisbury, F. B. & C. W. Ross. 1992. Plant physiology. Ed. 4. Wadsworth, Belmont, CA. S~inchez-Codero, V. & R. Martinez-Gallardo. 1998. Postdispersal fruit and seed removal by forest-

dwelling rodents in a lowland rainforest in Mexico. J. Trop. Ecol. 14: 139-151. Savage, A. J. P. & P. S. Ashton. 1983. The population structure of the double coconut and some other

Seychelles palms. Biotropica 15:15-25. S e a r i o t , A. 1999. Forest fragmentation effects on palm diversity in central Amazonia. J. Ecol. 87: 66-76. - - - , A. T. O. Filho & E. Lleras. 1989. Species richness, density and distribution of palms in an east-

ern Amazonian seasonally flooded forest. Principes 33:172-179. Schatz, G. E., G. B. Williamson, C. M. Cogswell & A. Stam. 1985. Stilt roots and growth of arboreal

palms. Biotropica 17: 206-209. Sehnpp, E. W. 1995. Seed-seedling conflicts, habitat choice, and patterns of plant recruitment. Amer.

J. Bot. 82: 399-409. - - & E. J. Frost. 1989. Differential predation of Welfia georgii seeds in treefall gaps and the forest

understory. Biotropica 21:200-203. - - & M. Fnentes. 1995. Spatial patterns of seed dispersal and the unification of plant population

ecology, l~coscience 2: 267-275. - - - , H. F. Howe, C. K. Aogsborger & D. J. Levey. 1989. Arrival and survival in tropical treefall

gaps. Ecology 70: 562-564. Shmlda, A. & S. EIIner. 1984. Coexistence of plant species with similar niches. Vegetatio 58: 29-55. Silver, W. L., F. N. Seatena, A. H. Johnson, T. G. Siccama & M. J. Sanehez. 1994. Nutrient availabil-

ity in a montane wet tropical forest: Spatial patterns and methodological considerations. P1. & Soil 164: 129-145.

Silvertown, J., M. E. Dodd, D. J. G. Gowing & J. O. Mountford. 1999. Hydrologically defined niches reveal a basis for species richness in plant communities. Nature 400: 61-63.

Sist, P. & H. Puig. 1987. Rtgtntration, dynamique des populations et disstmination d'un palmier de Guyane fran~aise: Jessenia bataua (Mart.) Burret subsp, oligocarpa (Griseb. & H. Wendl.) Balick. Bull. Mus. Natl. Hist. Nat., B, Adansonia 3: 317-336.

Skov, F. & H. Balslev. 1989. A revision ofHyospathe (Arecaceae). Nord. J. Bot. 9: 189-202. - - & F. Borchsenius. 1997. Predicting plant species distribution patterns using simple climatic pa-

rameters: A case study of Ecuadorian palms. Ecography 20: 347-355. Smith, A. P., K. P. Hogan & J. R. Idol. 1992. Spatial and temporal patterns of light and canopy struc-

ture in a lowland tropical moist forest. Biotropica 24:503-511. Smith, T. B., R. K. Wayne, D. J. Girman & M. W. Bruford. 1997. A role for ecotones in generating

rainforest biodiversity. Science 276:1855-1857. Smythe, N. 1989. Seed survival in the palmAstrocaryum standleyanum: Evidence for dependence upon

its seed dispersers. Biotropica 21:50-56. Snow, D. W. & B. K. Snow. 1978. Palm fruits in the diet of the oilbird, Steatornis caripensis. Principes

22: 107-109. Soilins, P. 1998. Factors influencing species composition in tropical lowland rain forest: Does soil mat-

ter? Ecology 79: 23-30. Sterner, R. W., C. A. Ribic & G. E. Sehatz. 1986. Testing for life historical changes in spatial patterns

of four tropical tree species. J. Ecol. 74: 621-633. Strudwick, J. & G. L. Sobel. 1988. Uses ofEuterpe oleracea Mart. in the Amazon estuary, Brazil. Pp.

225-253 in M. J. Balick (ed.), The palm--Tree of life: Biology, utilization, and conservation. Ad- vances in Economic Botany 6. New York Botanical Garden, Bronx.

S v e n n i n g , J.-C. 1998. The effect of land-use on the local distribution of palm species in an Andean rain forest fragment in northwestern Ecuador. Biodiv. & Conserv. 7: 1529-1537.


- - . . 1999a. Microhabitat specialization in a species-rich palm community in Amazonian Ecuador. J. Ecol. 87: 55-65.

- - . . 1999b. Recruitment of tall arborescent palms in the Yasuni National Park, Amazonian Ecuador: Are large treefall gaps important? J. Trop. Ecol. 15: 355-366.

- - . . 2000a. Growth strategies ofclonal palms (Arecaceae) in a neotropical rain forest, Yasuni, Ecua- dor. Austral. J. Bot. 48: 167-178.

- - . . 2000b. Small canopy gaps influence plant distributions in the rain forest understory. Biotropica 32: 252-261.

- - . . In prep. Crown illumination limits the population growth rate of a neotropical understory palm (Geonoma macrostachys, Arecaceae). PI. Ecol.: submitted.

- - & H. Balslev. 1997. Small-scale demographic disequilibrium oflriartea deltoidea (Arecaceae) in Amazonian Ecuador. Pp. 263-274 in R. Valencia & H. Balslev (eds.), Estudios sobre diversidad y ecologia de plantas: Memorias del II Congreso Ecuatoriano de Botgtnica realizado en la Pontiflcia Universidad Cat61ica del Ecuador, Quito, 16-20 octubre 1995. Pontiflcia Universidad Cat61ica del Ecuador, Quito.

- - & .1998. The palm flora of the Maquipucuna montane forest reserve, Ecuador. Principes 42: 218-226.

- - & .1999. Microhabitat-dependent recruitment oflriartea deltoidea (Arecaceae) in Ama- zonian Ecuador. Ecotropica: 5: 69-74.

Sytsma, K. J. & B. A. Sehaal. 1985. Genetic variation, differentiation, and evolution in a species complex of tropical shrubs based on isozymic data. Evolution 39: 582-593.

Ter Steege, H., V. G. Jetten, A. M. Polak & M. J. A. Werger. 1993. Tropical rain forest types and soil factors in a watershed area in Guyana. J. Veg. Sci. 4: 705-716.

Terborgh, J. 1973. On the notion of favorableness in plant ecology. Amer. Naturalist 107:481-501. �9 1985. The vertical component of plant species diversity in temperate and tropical forests. Amer.

Naturalist 126: 760-776. , E. Losos, M. P. Riley & M. B. Riley. 1993. Predation by vertebrates and invertebrates on the

seeds of five canopy tree species of an Amazonian forest. Vegetatio 107/108: 375-386. -, R. B. Foster & P. Nufiez V. 1996. Tropical tree communities: A test of the nonequilibrium hy-

pothesis. Ecology 77: 561-567. Tomlinson, P. B. 1990. The structural biology of palms. Clarendon Press, Oxford; Oxford University

Press, New York. Tregenza, T. & R. Butlin. 1999. Speciation without isolation. Nature 400:311-312. Trichon, V., J.-M. N. Walter & Y. Laumonier. 1998. Identifying spatial patterns in the tropical rain

forest structure using hemispherical photographs. P1. Ecol. 137: 227-244. Troy, A. R., P. M. S. Ashton & B. C. Larson. 1997. A protocol for measuring abundance and size of a

neotropical liana, Desmoncus polyacanthos (Palmae), in relation to forest structure. Econ. Bot. 51 : 339-346.

Tuomisto, H. & A. D. Poulsen. 1996. Influence of edaphic specialization on pteridophyte distribution in neotropical rain forests. J. Biogeogr. 23: 283-293.

- - & K. Ruokolainen. 1993. Distribution of Pteridophyta and Melastomataceae along an edaphic gradient in an Amazonian rain forest. J. Veg. Sci. 4: 25-34.

, , R. Kalliola, A. Linna, W. Danjoy & Z. Rodriguez. 1995. Dissecting Amazonian biodiversity. Science 269: 63-66.

- - , A. D. Poulsen & R. C. Moran. 1998. Edaphic distribution of some species of the fern genus Adiantum in western Amazonia. Biotropica 30: 392-399.

Uhl, N. W. & J. Dransfield. 1987. Genera palmarum: A classification of palms based on the work of Harold E. Moore, Jr. L. H. Bailey Hortorium; International Palm Society, Lawrence, KS.

Valencia, R., H. Balslev & G. Paz y Mifio. 1994. High tree alpha-diversity in Amazonian Ecuador. Bio- div. & Conserv. 3: 21-28.

Vandermeer, J. H. 1993. Successional patterns of understory palms in an old cacao plantation on the Caribbean coast of Costa Rica. Principes 37: 73-79.

- - . . 1994. Effects of Hurricane Joan on the palms of the Caribbean coast rainforest of Nicaragua. Principes 38: 182-189.


- - . 1977. Notes on the density dependence in Welfia georgii Wendl. ex Burret (Palmae): A lowland rainforest species in Costa Rica. Brenesia 10/11: 9-15.

- - , d. Stout & G. Miller. 1974. Growth rates of Welfia georgii, Soeratea durissima, and lriartea gigantea under various conditions in a natural rainforest in Costa Rica. Principes 18:148-154.

Visquez-Y~nes, C. & A. Orozco-Segovia. 1993. Patterns of seed longevity and germination in the tropical rainforest. Ann. Rev. Ecol. Syst. 24: 69-87.

Voeks, R. A. 1988. Changing sexual expression of a Brazilian rain forest palm (A ttaleafunifera Mart.). Biotropica 20:107-113.

Weiner, J. & R. T. Corlett. 1987. Size structure ofLivistona endauensis populations at four sites on Gunung Janing Barat, Johore, Malaysia. Malayan Nat. J. 41 : 297-302.

Welden, C. W., S. W. Hewett, S. P. Hubbell & R. B. Foster. 1991. Sapling survival, growth, and recruitment: Relationship to canopy height in a neotropical forest. Ecology 72: 35-50.

Wenny, D. G. & D. J. Levey. 1998. Directed seed dispersal by bellbirds in a tropical cloud forest. Proc. Natl. Acad. U.S.A. 95: 6204~207,

Wessels Boer, J. G. 1968. The geonomoid palms. Verh. Kon. Ned. Akad. Wetensch. Afd., Natuurk., Tweede Sect., 58 (1). Noord-Hollandsche U.M., Amsterdam.

White, P. S. 1979. Pattern, process, and natural disturbance in vegetation. Bot. Rev. (Lancaster) 45: 229-299.

Whittaker, R. H. 1965. Dominance and diversity in land plant communities. Science 147: 250-260. Wills, C., R. Condit, R. B. Foster & S. P. Hubbell. 1997. Strong density- and diversity-related effects

help to maintain tree species diversity in a neotropical forest. Proc. Natl. Acad. U.S.A. 94: 1252-1257.

Wilson, D. E. & D. H. Janzeu. 1972. Predation on Scheelea palm seeds by bruchid beetles: Seed density and distance from the parent palm. Ecology 53: 954-959,

Wing, S. L., L. J. Hiekey & C. C. Swisher. 1993. Implications of an exceptional fossil flora for Late Cretaceous vegetation. Nature 363: 342-344.

Woodward, F. L 1990. From ecosystems to genes: The importance of shade tolerance. Trends Ecol. Evol. 5: 111-115.

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.

- - . 1990. Cumulative satiation of a seed predator over the fruiting season of its host. Oikos 58: 272-276.

- - . 1991. Seasonal drought and the phenology of understory shrubs in a tropical moist forest. Ecol- ogy 72: 1643-1657.

Yeaton, R~ I. 1979. Intraspecific competition in a population of the stilt palm, Socratea durissima (Oesrt.) Wendl. on Barro Colorado Island, Panama. Biotropica 11: 155-158.

Zoua, S. & A. Henderson. 1989. A review of animal-mediated seed dispersal of palms. Selbyana 11 : 6-21.

on the role of microenvironmental heterogeneity in the ecology and diversification of neotropical...

Documents