Seed Dispersal, Seed Predation, and Seedling Recruitment of a Neotropical
Montane Tree
Daniel G. Wenny
Ecological Monographs, Vol. 70, No. 2. (May, 2000), pp. 331-351.
Stable URL:
http://links.jstor.org/sici?sici=0012-9615%28200005%2970%3A2%3C331%3ASDSPAS%3E2.0.CO%3B2-Y
Ecological Monographs is currently published by Ecological Society of America.
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained
prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in
the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/journals/esa.html.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic
journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers,
and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take
advantage of advances in technology. For more information regarding JSTOR, please contact support@jstor.org.
http://www.jstor.org
Mon Dec 3 10:39:04 2007
Ecologicai Monographs, 70(2), 2000, pp. 33 1-35 1
0 2000 by the Ecological Society of America
SEED DISPERSAL, SEED PREDATION, AND SEEDLING RECRUITMENT OF A NEOTROPICAL MONTANE TREE Department of Zoology, University of Florida, Gainesville, Florida 3261 1 USA
Abstract. Postdispersal fate of seeds from Ocotea endresiana (Lauraceae), a bird-dispersed
Neotropical montane tree, was studied in Costa Rica to determine the influence of seed dispersers, seed predators, and microhabitat characteristics on seedling recruitment. Particular
emphasis was placed on finding naturally dispersed seeds in order to study the link between
dispersal and postdispersal fate of seeds. Four species of birds (Emerald Toucanet, Aulacorhynchus prasinus; Resplendent Quetzal, Pharomachrus mocinno; Three-wattled Bellbird, Procnias tricarunculata; and Mountain Robin, Turdus plebejus) dispersed the seeds by regurgitation,
and one species (Black Guan, Chamaepetes unicolor), by defecation. Most seeds (80%) were
dispersed within 25 m of parent trees and under high (>92%) canopy cover. Bellbirds deposited
52% of the seeds they dispersed under habitual song perches in standing dead trees on the
edges of treefall gaps >25 m from parent trees. In contrast, the other four species dispersed
only 6% of the seeds they dispersed >25 m from parent trees, and <3% of the seeds were
dispersed to gaps. Thus, male bellbirds provided predictable, nonrandom dispersal to a different
microhabitat than the other four species.
Seed predation, germination, and one-year seedling survival were assessed for naturally
dispersed seeds and for seeds placed at randomly located sites. Seed predation during the
first 12 mo after dispersal was extremely high (99.7%). At least 50% of seed removal was
attributable to small rodents, such as Peromyscus mexicanus, with no evidence of scatterhoarding or secondary dispersal. Seeds dispersed >20 m from parent trees were removed
more slowly than seeds directly under the crown, but over the course of the fruiting season
seed predation approached loo%, throughout the study site. Seed predation did not differ
between gaps and forest understory.
Germination of seeds protected from mammals was high (70-90%), regardless of microhabitat or seed size. Seeds regurgitated by birds and seeds removed from fruits by hand
germinated equally well (98%), but seeds in intact fruits did not germinate. Seeds defecated
by guans and regurgitated by other species in the field had similar germination success.
Thus, dispersal by birds is important for removal of pulp but does not otherwise affect
germination. Seedlings were taller in gaps (<90% canopy cover) than in closed-canopy
forest (>90% cover). Fungal pathogens killed fewer seedlings in gaps than in understory,
while other causes of seedling mortality were similar among microhabitats. Consequently,
bellbirds, the most likely species to disperse seeds to gaps, had a disproportionate influence
on seedling recruitment. Nevertheless, the absolute number of seedlings was higher in
closed canopy forest, because the great majority of seeds (85%) landed there. The overall
pattern of recruitment, averaged among trees, had a large peak corresponding with bellbird
song perches in gaps, as well as a smaller peak near the parent tree corresponding with the
seeds dispersed by other species. The relative height of these peaks is opposite that of the
original pattern of dispersal. Frugivorous birds that use habitual perches may have disproportionate effects on plant recruitment. To test this idea further, it is necessary to study
each stage of the recruitment process, including seed dispersal, postdispersal seed survival,
and seedling survival.
Key words: Aulacorhynchus prasinus; birds, seed dispersal; Chamaepetes unicolor; Neotropical
cloud forest in Costa Rica; Ocotea endresiana (Lauraceae): Pharomachrus mocinno; Procnias tricarunculata; recruitment; rodents, seedpredation; seed dispersal andpredation; seedling survival; Turdus
plebejus.
INTRODUCTION
Seed dispersal determines the spatial arrangement
Manuscript received 7 August 1998; revised 10 March 1999;
accepted 2 April 1999; final version received 26 April 1999.
I Address all correspondence to: Illinois Natural History
Survey, P.O. Box 241, Savanna, Illinois 61074 USA.
E-mail: danwenny @internetni.com
and physical environment of seeds from which the next
cohort-of seedlings is selected. The links between seed
dispersal and seedling recruitment, however, are poorly
understood. Dispersal is the first of a series of events
that may be important in limiting recruitment, For vertebrate-dispersed plants, these stages include fruit removal, seed dissemination, postdispersal seed predation, potentially secondary dispersal, germination, and
332
Ecological Monographs
Val. 70, No. 2
DANIEL G. WENNY
seedling establishment. Previous studies have generally
focused on only one or a few of these stages (Herrera
et al. 1994, Schupp and Fuentes 1995).
Despite many studies on fruit consumption and seed
handling by birds and mammals, seed shadows (spatial
patterns of dispersed seeds) generated by animals are
poorly known (Willson 1993). While concentrations of
seeds under or near fruiting trees have been noted and,
indeed, are expected (Janzen 1970, Howe 1989, Willson 1993), the pattern of seed distribution farther from
fruiting trees is likely to be the most important part of
the seed shadow for plant fitness and population ecology (Portnoy and Willson 1993, Schupp and Fuentes
1995). In particular, the tail of the distribution is often
heterogeneous because of disperser behavior (McDonne11 and Stiles 1983, Hoppes 1988, Chavez-Ramirez and Slack 1994, Julliot 1996). Fruit-eating vertebrates differ in foraging behavior (Cruz 1981, Santana and Milligan 1984, Trainer and Will 1984, Moermond and Denslow 1985), fruit removal rates (Howe
and Vande Kerckhove 1980, Bronstein and Hoffman
1987, Englund 1993), seed-handling techniques (JanZen 1983, Levey 1987, Corlett and Lucas 1990, Stiles
and Rosselli 1993), and effects on seed germination
(Krefting and Roe 1949, Compton et al. 1996, Traveset
and Willson 1997). Few studies, however, have compared seed shadows generated by different dispersers
(but see Thomas et al. [I9881 and Chavez-Ramirez and
Slack [1994]), or the consequences of dispersal pattern
on recruitment (but see Howe [1989, 1 9 9 0 ~ 1Herrera
,
et al. [1994], Schupp [1995], and Schupp and Fuentes
[1995]).
After dispersal, seed predation and seedling mortality are often extensive for tropical trees (DeSteven and
Putz 1984, Howe et al. 1985, Chapman 1989, Hammond 1995, Cintra and Horna 1997, Peres et al. 1997)
and may be important in influencing the spatial pattern
of recruitment (Connell 1970, Janzen 1970, Harper
1977, Hubbell 1980, Clark and Clark 1984, Becker et
al. 1985, McCanny 1985, Howe 1989). In addition,
studies on forest dynamics have suggested that canopy
gaps are often crucial recruitment sites for tree seedlings (Hartshorn 1978, Denslow 1987, Swaine and
Whitmore 1988, Swaine 1996). Few studies, however,
have examined seed predation as a link between seed
dispersal and seedling recruitment. In fact, aside from
research on Virola in Panama (Howe and Vande Kerckhove 1980, Howe et al. 1985, Howe 1990b, 1993),
Phillyrea in Spain (Herrera et al. 1994, Jordano and
Herrera 1995), and Ficus in Borneo (Laman 1995,
1996a, b), I know of no studies that consider the various
stages leading to recruitment for any species of fleshyfruited, vertebrate-dispersed plant. Furthermore, some
plant species, typically considered to be bird or mammal-dispersed, may have a second (or even third) stage
of dispersal by an entirely different dispersal vector
(Roberts and Heithaus 1986, Clifford and Monteith
1989, Forget and Milleron 1991, Estrada et al. 1993,
Levey and Byrne 1993, Nogales et al. 1996, Wenny
1999). To understand the relative importance of seed
dispersers and seed predators in forest dynamics, it is
necessary to study each stage leading to recruitment
(and, ideally, to reproductive age).
The main factor that limits our understanding of the
link between seed dispersers and seedling recruitment
is the difficulty of finding dispersed seeds (Harper
1977, Janzen 1983, Schupp and Fuentes 1995). Thus,
most studies on vertebrate-dispersed plants have assessed the importance for plant fitness of different dispersers by the amount of fruit in the diet and gut treatment of seeds, and have not considered postdispersal
fate of seeds (reviewed by Howe [1986], Stiles [1989],
Jordano [1992], Willson [1992], and Levey et al.
[1994]). Similarly, studies on seed predation and seedling recruitment have relied on experimentally dispersed seeds, usually without data on the actual patterns
of seed rain (Price and Jenkins 1986, Crawley 1992,
Hulme 1993, Schupp 1995).
The objective of this study was to link patterns of
seed dispersal and seed predation with those of seedling
recruitment, by determining the locations of naturally
dispersed seeds, and then following the fate of those
seeds after dispersal. I studied a common tree, Ocotea
endresiana Mez (Lauraceae), in an old-growth montane
forest in Costa Rica, to answer the following questions:
(1) Do the five main fruit consumers of 0 . endresiana
generate different patterns of seed rain and, thus, subsequently influence germination and the pattern of
seedling recruitment? (2) What proportion of the dispersed seeds are subsequently killed by seed predators,
and what animals are the most important seed predators? (3) Do scatter-hoarding rodents provide a second
stage of dispersal? (4) What proportion of the dispersed
seeds germinate and establish as seedlings under different microhabitat conditions? To answer these questions, I used a combination of natural and manipulative
experiments to assess the types of sites to which seeds
are dispersed and how the habitat characteristics and
spatial arrangement of those sites relative to the parent
trees influence postdispersal survival. In a previous paper (Wenny and Levey 1998), based a subset of the
data from this study, I described some elements of the
first question and found that one of the five disperser
species deposited seed disproportionately in sites suitable for seedling survival. In the current paper, I examine this idea in more detail, by incorporating data
on seed predation and seedling establishment, as well
as additional data on dispersal not included in the previous paper.
Study site
This study was conducted during May 1993-July
1996 in the Monteverde Cloud Forest Preserve (10'12'
N, 84'42' W) in the Cordillera de Tilaran, northern
May 2000
CLOUD FOREST SEEDLING RECRUITMENT
Costa Rica. This 10 000-ha preserve is adjacent to other
protected lands encompassing one of the largest areas
of relatively unbroken forest in Costa Rica. The study
area contains the full complement of birds and mammals, including top predators that have historically occurred in the area (Young and McDonald 1999), although certainly the abundance of some species has
changed (Fogden 1993). Other characteristics of the
fauna are described by Nadkarni and Wheelwright
(1999).
The study area was in relatively undisturbed lower
montane rain forest (Hartshorn 1983) along the continental divide at 1600-m elevation. A 5-ha area, 500
m from the beginning of the Valley Trail (Sendero El
Valle), was mapped (using sighting compasses and
measuring tapes) and marked into 10 X 10 m quadrats
with PVC tubing at every grid point. The study area
was relatively level with a change in elevation of only
10 m from north to south (excluding stream banks).
The forest canopy is 25-30 m tall and dominated by
several species of Lauraceae, Sapium oligoneuron (Euphobiaceae), Ficus crassiuscula (Moraceae), Inga sp.
(Leguminosae), and Pouteria viridis (Sapotaceae). The
vegetation of the area is described in more detail by
Lawton and Dryer (1980) and Nadkarni et al. (1995).
Canopy gaps caused by falling trees and branches
are common and may be an important habitat for recruitment of some plants in the study area (Lawton and
Putz 1988, Lawton 1990). Although gaps may be
formed at any time of year, most are formed during the
dry season, when strong winds are more frequent (K.
G. Murray, personal communication). Gaps are also
formed under standing dead trees, especially Sapium,
which shed large limbs for several years before the
remainder of the trunk falls (personal observation; C.
Guindon, personal communication). Approximate
boundaries of gaps were sketched on to the map of the
study site, the area of each gap was estimated from the
map, and the proportion of the study site in gaps was
calculated from the sum of those estimates. Approximately 5.3% of the mapped study area was in gaps 2 1 0
m2, with vegetation 5 2 m tall (sensu Brokaw 1982),
including two Sapium gaps. Mean canopy cover (measured with a spherical densiometer 1.2 m above the
ground) in gaps and gap edges was 90.7% ( i 9 . 6 , range
60-92%, N = 78), while closed-canopy forest had
mean cover of 96.2% ( i 4 . 3 , range 91-100%, N = 234).
These values are similar to those at other Neotropical
forest sites (Howe et al. 1985, Levey 1988, Clark
1994).
Most woody plants at Monteverde are animal dispersed, as is typical of Neotropical forests in general
and tropical montane forests in particular (Gentry 1982,
Howe and Smallwood 1982, Tanner 1982, Stiles 1985,
Levey and Stiles 1994). Bird dispersal predominates:
77% of the understory trees and shrubs and 63% of the
canopy trees have fruit morphology suggesting dispersal by birds (Stiles 1985). At least 70 resident and
333
migrant bird species feed on the fruits of > I 5 0 species
of trees and shrubs (Wheelwright et al. 1984). Most of
these birds regurgitate or defecate seeds intact (Wheelwright et al. 1984, Murray 1988).
The mean annual rainfall at 1520 m on the Pacific
slope 3 km from the study site is -2500 mm, with most
of the precipitation occurring during May-November.
Actual rainfall in the study site was probably >2500
mm, but the seasonal pattern was similar (Nadkarni
and Wheelwright 1999). Rain gauges underestimate the
amount of precipitation from mist and cloud interception, which contribute 5 5 0 % of the precipitation in
some Neotropical montane forests (Cavelier 1996).
Temperatures recorded in the forest understory at the
study site during this project ranged 15-22°C.
Study species
The Lauraceae is an important family in Neotropical
forests in terms of species richness, a food resource for
birds, and economic value (Wheelwright 1983, 1985a,
1991, Burger and van der Werff 1990, Gentry 1990,
Martinez-Ramos and Soto-Castro 1993, Guindon
1996). Members of the Lauraceae are also important
dietary components for frugivorous birds in Africa,
Southeast Asia, and Australia (Crome 1975, Snow
1981, Sun et al. 1997). Ocotea endresiana Mez (listed
as 0 . austinii in Wheelwright et al. [1984], Wheelwright [1985a, 19861) is a common canopy tree in montane forests in central and northwestern Costa Rica at
1100-2300 m elevation (Burger and van der Werff
1990). In the Monteverde area it occurs between 1550
and 1700 m along the continental divide. In the same
area, 20-30 other species of Lauraceae occur (Wheelwright 1985a, 1986, Haber 1991). Ocotea endresiana
is monecious and begins flowering in August, the midrainy season, and is pollinated by small flies and other
insects. Fruits ripen the following May and June in the
early rainy season. Ripe fruits (18 X 9 mm, 1.2 g) have
blue-black skin, lipid-rich pulp, and are held in a shallow reddish receptacle, typical of the Lauraceae genera
Ocotea and Nectandra (wheelwright et al. 1984, Burger and van der Werff 1990). Most of the volume of
the fruit is a single seed (15 X 8 mm, 0.75 g, fresh
mass) composed of a small embryo and two large cotyledons surrounded by a thin (0.3-mm) seed coat. Seeds
begin germination -6 wk after dispersal. Although
small compared to some other Lauraceae species, 0 .
endresiana fruits and seeds are among the largest at
Monteverde (Wheelwright et al. 1984). Twenty-one 0 .
endresiana trees >20 cm dbh were in the 5-ha study
site, and all but one produced fruit in both years. Fruits
and seeds for some experiments were collected from
17 additional trees outside the main study area. Large
fruit crops (21 10 i 1495 fruitsltree) were produced in
1993 and 1995, but not in 1994. Each 0 . endresiana
tree was a mean distance of 36.5 m (-C 12.1) from the
three closest conspecific adults.
The most common avian visitors to fruiting Laura-
DANIEL G. WENNY
334
ceae include the largest and most frugivorous species
at Monteverde (Wheelwright et al. 1984). Ocotea endresiana fruits are eaten primarily by five species of
birds: Emerald Toucanet (Ramphastidae: Aulacorhynchus prasinus), Resplendent Quetzal (Trogonidae:
Pharomachrus mocinno), Three-wattled Bellbird (Cotingidae: Procnias tricarunculata), Mountain Robin
(Turdidae: Turdus plebejus), and Black Guan (Cracidae: Chamaepetes unicolor), all of which breed in the
study site during the fruiting season. The first four species swallow fruits intact and regurgitate seeds in viable
condition. Seeds in this size range are generally regurgitated 15-30 min after ingestion (Wheelwright
1991). Guans defecate seeds in viable condition
(Wheelwright 1991). These bird species (especially
quetzals) typically remain in a fruiting tree after eating
several fruits, and often regurgitate seeds under the
same tree or nearby (Wheelwright 1983, 1991). Seed
processing times for guans are unknown, but they generally leave a fruiting tree before defecating the seeds
from that foraging bout (personal observation). Several
other bird species (Wheelwright et al. 1984), as well
as spider monkeys (Ateles geoffroyi), occasionally eat
0 . endresiana fruits and probably disperse viable seeds
(e.g., Chapman 1989).
Ecological Monographs
Vol. 70, No. 2
once every two weeks, so that, over the course of the
two-month season of fruit ripening, each 10 X 10 m
plot was checked at least four times. Bird observations
began whenever one of the five major consumers of 0 .
endresiana fruits was located, and continued as long
as the bird was visible or until the bird regurgitated or
defecated seeds. During 193 hr spent following birds,
184 seeds were found. I assumed that the distribution
of seed locations found by the combination of methods
would be representative of the entire population of
seeds.
In addition to these dispersed and nondispersed
seeds, other seeds were experimentally placed at randomly selected sites to determine if birds deposited
seeds in sites with specific or random characteristics.
Random numbers generated on a hand calculator were
used as site coordinates within the 5-ha grid. The postdispersal fate of seeds at these random locations was
compared to the fate of seeds at the dispersed and nondispersed locations.
Microhabitat characteristics
A combination of observational and experimental
data was collected to determine the probability of seedling establishment and one year survival for seeds dispersed by birds. In particular, the focus of this project
was to find the locations of seeds dispersed naturally,
and to compare the postdispersal fate of those seeds
with experimentally placed seeds, with the goal of determining the influence of dispersers on seedling recruitment. The locations of seeds naturally dispersed
by birds were classified in two categories, according
to their position relative to the parental tree crown.
Seeds were classified as nondispersed if directly under
the crown of a fruiting 0 . endresiana tree, or as dispersed if not under such a tree. Even though all dispersed and nondispersed seeds in this study had been
regurgitated or defecated by birds (and therefore "dispersed" sensu Janzen [1983]), it is generally believed
that seeds deposited under the parent trees have very
little chance of survival (e.g., Janzen 1970, Howe et
al. 1985, Hulme 1997). For convenience, I will refer
to sites of dispersed seeds as dispersed sites and of
nondispersed seeds as nondispersed sites.
For all seed locations, including random sites, I measured seed characteristics and microhabitat variables
that I thought might influence seed predation, germination, or seedling survival. Seed length and width
were measured with dial calipers, and seed mass was
measured 2 0 . 0 5 g with a spring balance. Canopy cover
was estimated with a spherical densiometer (Lemmon
1957). Leaf litter was the number of leaves pierced by
a metal stake thrust into the soil once at the site of
each seed. Vegetation density was the number of stems
within a 50-cm radius of the site. The distances to the
nearest woody stem > I m in height, tree >10 cm dbh,
trunk of fruiting Ocotea tree, and fallen log were measured with a fiberglass measuring tape. These variables
were selected based on their demonstrated importance
in previous studies. Seed size may influence the probability of seed predation (Price and Jenkins 1986,
Hulme 1993) or seedling size (Howe and Richter 1982).
Canopy cover (light availability) is known to be an
important factor for germination and tropical seedling
growth (Howe et al. 1985, Mulkey et al. 1996, Swaine
1996). Vegetation density and distance to objects may
influence rodent activity and seed predation (Smythe
1978, Kiltie 1981, Kitchings and Levey 1981). Finally,
leaf litter may influence seed predation or germination
(Schupp 1988, Molofsky and Augspurger 1992, Myster
and Pickett 1993).
Seed dispersal
Postdispersal seed predation
Seeds were located by following birds until they
dropped, regurgitated, or defecated seeds, and by systematic ground searches. These ground searches were
started at the base of a fruiting tree and proceeded along
10 m wide transects to 50-60 m from the trunk. It was
impossible to search the entire site with equal intensity,
but an effort was made to cover the entire site at least
At each dispersed, nondispersed, and random site, a
marked seed was used to assess rate of seed predation,
to identify seed predators, and to determine if secondary dispersal occurred. For this treatment, I used seeds
that were regurgitated by birds and collected under
fruiting trees adjacent to the study site. Seeds were
marked by gluing 50-75 cm of unwaxed dental floss
Experimental design
May 2000
CLOUD FOREST SEEDLING RECRUITMENT
to the seed, and tying -50 cm of flagging tape to the
distal end of the floss. Because the glue held best on
seeds with a dry seed coat, seeds were taken inside a
lab room and allowed to dry for 1-3 h before gluing.
Each marked seed was placed at a site the next morning.
To determine if presence of dental floss and flagging
tape influenced seed removal, I conducted a pilot study
in May 1993. Fifty marked seeds and 50 unmarked
seeds were placed singly at random locations and censused one week later. Removal of marked and unmarked seeds was not significantly different (49 and
47 seeds removed, respectively; x2 = 0.047, df = 1, P
> 0.10). Thus, the marking of seeds was assumed to
have no effect on seed removal.
Note that these marked seeds were placed next to
the cages protecting the original seeds deposited by
birds (see cage description in Methods: Germination
and seedling survival). The cages may have served as
cues for visually searching seed predators, but I believe
that is unlikely for two reasons. First, at the beginning
of the study in 1993, cages were novel items and would
not have been associated with seeds by the seed predators. Second, because most marked seeds were removed, but the cages were left at the sites for several
months, most cages were not good indicators of seed
availability. Olfactory cues left by marking and handling the seeds are another possible confounding factor,
but the frequent rains probably diminished these (Whelan et al. 1994).
In 1993, all marked seeds were censused once each
week for three weeks, and then at weeks 5 and 10.
Because most seeds were removed during the first week
after dispersal in 1993, marked seeds in 1995 were
checked more often: on days 1 , 3 , and 7, and once each
week afterwards until week 10. If a marked seed was
removed, the surrounding area was searched to find the
flagging tape-dental floss assembly. The end of the floss
where the seed had been attached was examined to
determine the fate of the seed. If a seed was entirely
removed, or if a piece of the seed coat remained attached to the floss, the seed was classified as killed by
seed predators, because captive Peromyscus treated 0.
endresiana seeds with dental floss in that manner when
they consumed them (Wenny, unpublished data). The
distance from the dispersal site to the predation site
was measured, and each site of a removed seed was
classified in the following categories: in burrow; in,
under, or near a fallen log ( > l o cm diameter); at the
base of a large tree; in dense vegetation (defined as
2 5 0 % cover of plants <50 cm tall within a 50-cm
radius of site, assessed visually); under or beside a
clump of fallen branches, or on the leaf litter. If a
marked seed was removed but not eaten, it was left in
the new location and included in subsequent censuses.
Distance effect.-Two
other experiments (one in
1995 and one in 1996) were conducted with marked
seeds to determine the effect of distance on seed removal, as well as whether most seed removal occurs
335
during the day or at night. In 1995, one seed was placed
at the base of each of 12 fruiting trees and at 20, 40,
60, and 80 m from the base of the tree. The orientation
of each transect was carefully selected to avoid approaching within 80 m of other fruiting 0. endresiana
trees. Seeds at these sites were censused on the same
schedule as the other marked sites in 1995. In 1996, a
similar experiment was conducted, except the seeds
were 5 and 40 m from each of 22 trees. Shorter and
fewer distances were used so that more trees could be
included, and because data from the previous year, as
well as other studies (e.g., Howe et al. 1985), indicated
that the distance effect can be detected beyond 20 m
from a conspecific adult. In this case, the compass directions were randomly selected, although some directions were discarded if the 40-m treatment was not 2 4 0
m from all fruiting conspecifics. This experiment was
run twice, once in the early fruiting season (late May)
and once two weeks later, during the peak fruiting season (mid-June). During the second of these trials, all
sites were checked at dusk and dawn for two consecutive days to determine whether most seed removal
occurred during the day or night. Removal during the
day likely could be attributed to the diurnal agouti,
while removal at night likely would be from nocturnal
species, such as Peromyscus or other small rodents.
Exclosures.-To determine the amount of seed predation attributable to large or small mammals, 38 sets
of exclosures were established in early 1995. Each set
had three treatments, each 1 m2 in area: (1) "no rodents" exclosures made of 1-cm2 galvanized wire
mesh, 0.9 m tall; (2) "small rodents only" exclosures
made of chicken wire mesh with 6-cm holes, 0.9 or
1.35 m tall; and (3) "all rodents" control plots with
only wooden stakes marking the corners. The bottom
edges of the wire exclosures were buried 5 cm below
ground and held with two or three 25-cm metal stakes
on each side. The corners were supported by 1.25-m
lengths of 10 m m thick metal stakes. The tops were
open to allow normal accumulation of fallen leaves.
Although small rodents can probably climb such cages,
the results show that the "no-rodent" treatment was
successful. Each set was located where convenient
(avoiding trees and fallen logs), within randomly located 10 x 10 m quadrats in the study site. One regurgitated 0. endresiana seed was placed in each exclosure and control plot in late June 1995. Each seed
was checked after 2, 4, and 8 d, 4 mo, and 1 yr.
Germination and seedling survival
The original regurgitated or defecated seed at each
dispersed and nondispersed site, and each seed placed
at a randomly located site, was protected by a 4 X 4
X 2 cm cage made of 3-mm galvanized wire mesh held
in place by two metal stakes. Caged seeds were used
to determine germination rates and insect predation
rates, in the absence of mammalian seed predators.
Each site was checked weekly for 2 1 2 wk and monthly
DANIEL G. WENNY
thereafter, until June of the following year. Germination
was defined as the splitting of the seed coat and spreading of the cotyledons. Typically, the radicle had
emerged by the next census after germination, and a
week later the stem was visible. As each seed germinated and the shoot began to grow, the cage was removed to allow normal seedling growth. The seedling
location was marked with one of the stakes from the
cage. Causes of seedling mortality were classified as
mammal, insect, fungal pathogen, physical, or unknown. Mammals either ate the seed and left the damaged shoot behind (seed predators) or removed the entire shoot (herbivores). Some seeds that appeared to
have germinated were killed by beetle larvae (Heilipus
sp., Curculionidae) developing in the seed. Insectkilled seeds frequently developed a root, but never had
a shoot > 2 cm tall. Seedlings killed by fungal pathogens were characterized by a wilted and discolored
shoot above -4 cm (Augspurger 1990). Physical damage consisted of trampling by peccaries and other large
mammals, or damage from falling trees, branches, or
large leaves (Clark and Clark 1991). Unknown causes
of mortality included cases that fit more than one category where the sequence of events could not be determined, and cases that did not clearly fall into any
category. A seed was considered alive if the seed remained firm, even if the shoot had been eaten or otherwise damaged. Such seeds resprouted repeatedly
(personal observation). Seeds from 1995 that had not
germinated or resprouted after one year were cut open
and classified as insect-killed if filled with frass, or
viable if the embryo and cotyledons were not discolored or mealy in appearance.
Germination trials.-To determine the effect on 0.
endresiana germination of ingestion by birds and seed
burial (which I initially thought would apply to seeds
taken by mammals), trials were conducted in a greenhouse constructed of nylon window screen over a 12
X 6 m wooden frame on a concrete foundation set 25
cm into the ground. The roof was corrugated plastic.
Seeds were planted individually in cardboard milk cartons and cut plastic bottles of various sizes. All containers were washed with hot soapy water and dried in
the sun before use. Soil was collected from a nearby
secondary forest and mixed 3: 1 with sand.
Three seed treatments were compared: (1) naturally
regurgitated seeds, collected under fruiting trees or
along the trail to the study site; (2) seeds removed by
hand from ripe fruits, collected under fruiting trees;
and (3) entire ripe fruits, either fallen or dropped. Seeds
or fruits that were misshapen, diseased, or had signs
of insect infestation were not used. Each treatment had
35 seeds or fruits, placed on top of the soil. Additionally, 10 seeds were planted 5 cm deep to determine if
seeds could germinate after burial by scatter hoarding
rodents. Containers were watered as necessary to keep
the soil moist. Seeds were planted in June 1995, and
were checked weekly until early November 1995, when
Ecolog~calM onographs
Vol 70, No. 2
seeds that had not germinated were cut open and assessed for viability.
Seedling and sapling plots.-To determine if 0. endresiana seedlings and saplings are more likely to recruit under or away from conspecifics, seedlings and
saplings ( 5 3 m tall) were measured and mapped in
paired 10 X 10 m plots. For each of 10 trees, one plot
was located near the tree, with approximately half of
the plot directly under the crown. The second plot was
located 10-20 m from the edge of the first plot and at
least 15 m from the crown edge. No plots were located
in recent (<5 yr) canopy gaps or along streams.
Statistical analyses
The influence of the microhabitat variables on removal of marked seeds (at day 1 [I995 only] and week
2) and on germination, and 1-yr survival of caged seeds
were examined with multiple logistic regression using
SAS JMP (SAS Institute 1989). For each of the binomial (live or dead) response variables, data for the
two years were analyzed separately. Each model was
run, first with all predictor variables, and then with and
without each predictor variable to determine if the deletion of a given variable had a significant effect on
the amount of variation explained. This deletion procedure was repeated until only significant predictors
were retained (Trexler and Travis 1993). Models were
selected manually to avoid the problems of automatic
stepwise procedures (James and McCulloch 1990). The
predictor variables included three measures of seed size
(length, width, and mass), eight microhabitat characteristics (leaf litter, canopy cover, number of stems, and
distances to nearest caged seed, herbaceous stem,
woody stem, 10-cm tree, parent tree trunk, and fallen
log), date of dispersal (Julian date), and tree number.
Only single terms were included in the models as lackof-fit tests indicated that the single-term models were
adequate; thus, interaction terms were not required
(SAS Institute 1989).
The relationships of the microhabitat variables within and among the dispersed, nondispersed, and random
sites were examined with principal components analysis (PROC FACTOR) from the SAS statistical package (SAS Institute 1988). The mean loadings for the
three types of sites were compared with one-way analysis of variance from Super Anova (Abacus Concepts
1989). Type I11 sums of squares were used to compensate for the unequal sample sizes (Shaw and Mitchell-Olds 1993). Removal rates were compared among
treatments with survival analysis and ~ e h a n - ~ i l c o x o n
tests from SAS JMP (SAS Institute 1989) The GehanWilcoxon test places greater weight on early events,
than later events, and was deemed appropriate for this
study because most seeds were removed during the first
week (Pyke and Thompson 1986).
Parametric tests were used unless the data violated
the assumptions of normality and equal variance, in
which case nonparametric procedures were used. Data
May 2000
80
f
90
80
70
60
50
40
30
20
10
0
CLOUD FOREST SEEDLING RECRUITMENT
Dispersed ( N = 155)
Dispersed (N = 234)
Nondispersed ( N = 171)
IB Rand0111(N = 100)
Distance fro111parent tree (m)
FIG. 1. The distribution of seeds naturally regurgitated or
defecated by birds directly under the crown of a fruiting tree
(nondispersed), and away from fruiting trees (dispersed); and
the distribution of random sites at different distance categories from trunks of fruiting trees in (A) 1993 and (B) 1995.
Values on the x-axis represent the lnaxilnurn distance for each
category (5 = 0-5 m). The seed shadow is truncated at 70
m, because the abundance of Ocorea erldresiriizri trees in the
study site make sites >70 m very rare.
in the form of proportions were arcsine square root
transformed before analysis. Where multiple comparisons on a data set were involved, the alpha value was
adjusted according to the number of comparisons
planned (Bonferroni technique, Holm [1979]). Data
from the two years were very similar and were combined for analyses in which sample sizes would have
been low otherwise. Throughout this paper mean values
are followed by t 1 so.
337
m of the crown edge (1993,55%; 1995,45%), but some
seeds were as far as 70 m away. In rare cases (2.8%),
dispersed seeds were within 5 m of an 0. endresiana
trunk (Fig. 1). Despite their close proximity to the parent, they were classified as dispersed, because they
were not directly under the parent crown, which was
a result of asymmetrical crown geometry or a tilted
trunk (or both). Random sites were more evenly distributed than dispersed or nondispersed sites, but
showed a peak at 20-25 m in both years (Fig. 1). Most
seeds landed within 20 m of the fruiting trees, but the
tail of the seed distribution curve beyond 20 m accounted for 18% of the sites in 1993 and 21% in 1995.
Mean seed rain was highly variable when evaluated
among trees and years (Fig. 2). In particular, note that
seed distributions showed secondary peaks at 45-50
and 60-65 m in both years (Fig. 1). These peaks corresponded to habitual song perches used by bellbirds
(see Results; Microhabitat characteristics) that were
located 40-70 m from fruiting trees in the years this
study was conducted. A map and more information on
the locations of seeds are presented elsewhere (Wenny
1998).
Microhabitat characteristics
In a principal component analysis of the 1995 microhabitat characteristics, the first two components explained 63.8% of total variance; whereas, for 1993, the
first two components accounted for 45.5% of total variance. The relatively low amount of variance explained
is not entirely due to the inclusion of random sites
(Jackson 1993), because removal of random sites increased the total variance explained by the first two
components by only 2-3%. The pattern was similar in
both years: the first multivariate axis (PC 1) was char-
Seed dispersal
In 1993,284 seeds regurgitated or defecated by birds
were found (155 dispersed, 129 nondispersed), while
in 1995, 234 dispersed and 171 nondispersed seeds
were found. In 1993 and 1995, 95 and 100 random
sites, respectively, were established. Twenty-four percent of the 1993 seeds and 28% of the 1995 seeds were
found by following or observing birds (N = 184), and
the rest ( N = 505), by searching the ground. The total
number of seeds found (689) was 1% of the total seed
crop in the 5-ha study area in both years combined.
The dispersed seeds were most common within 10
-
Distance from crown edge (m)
FIG.2. The mean number of naturally regurgitated or defecated seeds per tree (+ 1 SII: N = 21) for both years combined
at different distances from the closest fruiting tree. Distance
class 0 includes all nondispersed seeds directly below parental
crowns. The other distance categories (5 = 0-5 rn, etc.) are
measured from the edge of the crown of the closest fruiting
Ocoren eizdresicii~atree. The values for each distance category
represent the number of seeds at a particular distance (means
calculated among all trees) for the entire study site.
338
Ecological Monographs
Vol. 70. No. 2
DANIEL G. WENNY
cept for some dispersed sites that have higher values
for the first component. These correspond to sites with
low canopy cover, relatively far from the parent trees
(Fig. 3). Seeds in the sites with values >2.0 in 1993,
and >1.0 in 1995, for PC 1 were all dispersed by bellbirds into large gaps surrounding song perches. Additionally, two of the 1993 seeds were dispersed by
Black Guans onto logs in another smaller gap. Bellbirddispersed seeds, seeds dispersed by other species, and
random sites had significantly different PC 1 loadings
= 7.9, P = 0.006) and 1995 (F,,
=
in 1993 (F,,
9.1, P < 0.001). Bellbird sites had higher factor loadings than other species' sites and random sites ( P <
0.002 in all cases), but sites of the other species and
random sites did not differ ( P = 0.2; see also Wenny
and Levey 1998).
The locations of dispersed, nondispersed, and random sites differed significantly in both years with regard to canopy cover, distance to parent, and number
of stems (Table 1). Dispersed seed sites had lower mean
canopy cover than nondispersed or random sites in both
years, while the nondispersed and random sites were
similar. In 1993, dispersed and random sites had similarly high numbers of stems compared to nondispersed
sites. In 1995, dispersed sites had more stems than both
the other sites, while random and nondispersed sites
had similar numbers of stems. The only other variable
that differed among the treatments was distance to closest stem in 1995. Dispersed sites were closer to stems
than nondispersed sites, while random sites did not
differ from the other two treatments (Table 1).
Combining the seeds from both years, and using only
seeds for which the disperser was known, the mean
canopy cover at sites of seeds dispersed by bellbirds
was significantly lower than of sites of seeds dispersed
by all the other species (89 and 96%, respectively;
Kruskal-Wallis test: x2 = 42.7, df = 1, P < 0.001).
Similarly, the mean distance from the closest parent
tree of bellbird sites was greater than any of the other
four species, (Kruskal-Wallis x2 = 66.06, df = 1, P <
0.001), but the difficulty in following the birds (especially robins, which tended to fly above the canopy)
biased the results in favor of bellbirds. Nevertheless,
the data show that bellbirds tend not to drop many seeds
near parent trees (Fig. 4). Considering that -5.3% of
the study area was in gaps, overall seed arrival in gaps
(12%) occurred more often than expected by chance
(x2 = 36.45, df = 1, P < 0.001), whereas the expected
number of random sites (4.1%) were in gaps (x2 =
0.557, df = 1, P > 0.05). Furthermore, bellbirds dispersed significantly more seeds to gaps than did the
other species (x2 = 39.3, df = 1, P < 0.001). However,
bellbirds dispersed seeds to only 2 of 29 gaps in the
study area.
,,
FIG. 3. Plot of sites with dispersed (open squares), nondispersed (solid squares), and rando~nlylocated seeds (crosses) from (A) 1993 and (B) 1995, on the first two axes determined by principal components analysis of habitat variables.
Axis 1 is primarily a function of canopy-cover leaf litter and
closest 10-cm (dbh) tree, and Axis 2 is primarily a function
of vegetation density and closest stem.
acterized by high negative loading of canopy cover, as
well as high positive loadings of leaf litter amount and
closest 10 cm dbh tree. In 1995 PC 1 also had a high
positive loading of distance to closest log. The second
axis (PC 2) was characterized by positive loading of
distance to closest stem in both years and negative
loading of vegetation density in 1995, but positive
loading in 1993. The mean loadings for PC I differed
significantly among the dispersed, nondispersed, and
random sites in 1993 (F,,,,, = 45.99, P < 0.0001), and
1995 (F,,,,, = 31.61, P < 0.0001). Each type of site
differed from the others (Fisher's LSD tests, P < 0.002
in all cases). Loadings for PC 2 differed only between
dispersed and nondispersed sites in 1995 (F,,,,, = 6.07,
P = 0.0025). This analysis shows that site characteristics of dispersed and nondispersed seeds were nonrandom.
When the three treatments (dispersed, nondispersed,
and random) are identified on a plot of the first two
principal components, they showed broad overlap, ex-
,,,
Postdispersal seed predation
More than 50% of marked seeds were removed within one week in both years (Fig. 5). In 1993, seeds at
CLOUD FOREST SEEDLING RECRUITMENT
May 2000
339
TABLE1 . Summary of one-way analysis of variance tests for each habitat variable, compared among sites of dispersed (D)
and nondispersed (N) seeds, and randomly located sites (R) in 1993 and 1995.
Variable
Leaf litter
.
96)
Cano,
~ vcover (
~,
Vegetation density
Closest stem (cm)
Woody stem (cm)
Closest tree (m)
Closest log (m)
D
N
R
2.1
(1.3)
94.3,'
(3.6)
38.9"
(17.3)
10.8
(5.5)
61.9
(45.7)
1.50
(0.87)
1.47
(1.31)
1.9
(1.4)
95.6h
(1.5)
32.1h
(16.2)
12.2
(6.2)
58.9
(39.7)
1.68
(0.97)
1.30
(1.37)
2.4
(1.4)
95.8h
(1.7)
40.7,'
(20.4)
10.9
(6.2)
66.7
(42.3)
1.66)
(1.OO)
1.28
(1.25)
F2.
175
2.70
13.40'
7.98'
2.22
0.90
1.53
0.80
D
N
2.3
(1.3)
94.5"
(6.5)
26.8,'
(15.1)
11.6J
(6.4)
88.6
(62.5)
1.88
(1.08)
1.97
(1.74)
2.3
(1.2)
96.7h
(1.6)
19.9h
(10.2)
14.3h
(7.4)
75.5
(60.4)
1.92
(1.04)
1.78
(1.45)
R
2.1
(1.2)
96.5h
(3.1)
21.6h
(10.9)
13.3.1~
(12.0)
75.9
(55.8)
1.84
(1.08)
1.83
(1.66)
F2
502
0.91
12.40-115.50'r
5.46t
2.85
0.18
0.77
Notes: ANOVA tests had a Bonferroni-adjusted alpha value of 0.006. Mean values are shown for each treatment for each
variable. Standard deviations are in parentheses below each mean. Leaf litter was the number of leaves pierced by a metal
stake thrust into the soil at the site. Canopy cover was estimated with a spherical densiometer. Vegetation density was
estimated as the number of stems within a 50-cm radius of the site. The last four variables are distances to closest stem,
closest woody stem ( > I m height), closest tree ( > l o cm dbh), and closest log ( > l o cm). Within each year and variable for
which a significant difference among treatments was detected (indicated with a dagger ( t ) after the F value), results of post
hoc tests (Student-Newman-Kuels tests) are indicated with a letter after the mean. Means followed by different letters are
significantly different ( P < 0.05).
'r P < 0.006.
dispersed, nondispersed, and random sites were removed at similar rates, with dispersed seeds showing
a nonsignificant trend for slower removal than seeds at
nondispersed or random sites (Wilcoxon x2 = 5.36, df
= 2, P = 0.068). In 1995, dispersed seeds were removed more slowly than those at nondispersed and
random sites (Wilcoxon x2 = 9.09, df = 2, P = 0.011).
Ultimately, only 2 of 923 marked seeds (one dispersed
and one random site, both from 1995, and both '30
m from conspecific trees) in the entire study survived
to June 1996 (and germinated), for an overall predation
rate of 99.7%.
I found 75% of marked seeds after removal in 1993,
and 94% in 1995. In all cases, the seed was killed, and
1
"
Toucanet
FJ Robin
E! Quetzal
Z1
Tj
0
X
40
H Guan
C L
2
30
Belibird
$-
2 20
I0
n
0
10
20
30
40
50
60
Distance from crown edge (m)
70
FIG.4. The number of seeds dispersed at different distances from parental trunks by bellbirds, guans, quetzals, robins, and toucanets, for both years combined (N = 184).
in most cases (96%) entirely consumed. Few seeds
were eaten (13%) but not moved. The general pattern
was that seeds were taken a short distance (range O21 m, N = 761) to a site with cover, presumably where
the seed could be eaten in safety. The mean transport
distances of removed seeds (not counting seeds eaten
but not moved) were 2.5 m ( 2 3 . 8 , N = 286) in 1993
and 2.1 m ( 2 2 . 9 , N = 475) in 1995. Approximately
50% of those removed were found on the leaf litter,
but some were found in small burrows (<6 cm diameter), in or under fallen logs, or in crevices at the bases
of large trees (Fig. 6). Most seeds taken to burrows,
trees, and logs were clearly taken by small animals,
probably rodents, because the small diameter of the
opening would exclude larger species such as agoutis
(Dasyprocta p~cnctata).I found no evidence of scatterhoarding or secondary dispersal.
Significant predictors of survival of marked seeds to
2 wk in 1993 included more open canopy cover, deeper
leaf litter, and later dispersal date (Table 2 [seeds]). In
1995, single-day survival was significantly predicted
by greater seed mass, earlier dispersal date, and the
particular parent tree. At 2 wk, seeds farther from the
parent tree, dispersed later in the season, and at certain
parent trees were more likely to survive. Logistic regression models of survival past 2 wk were not significant, because so few seeds survived that long.
Distance Effect.-Seed
removal from transects radiating from fruiting 0. endresiana trees was slower
for seeds farther from the parent trees in both 1995 and
1996. The 1995 seed removal experiment, which ran
for 10 wk, showed that the effect of distance on seed
Ecological Monographs
Vol. 70. No. 2
DANIEL G. WENNY
100
Base of tree
1993
+ Dispersed (N = 148)
80
t
60
on
.c
.-
8
Burrow
Nondispersed (N = 120)
Log
* Random (N = 95)
Fallen branches
40
Dense vegetation
20
2
Leaf litter
0
%
O
'
i
l
20
40
60
80
on
0
1995
+-
2
80
+ Dispersed (N = 234)
60
* Random (N = 100)
g
t
Nondispersed (N = 17 1)
40
20
0
0
20
40
Time (d)
60
0
100
80
FIG. 5. Removal of marked seeds from sites of dispersed
(open squares), nondispersed (solid squares). and randomly
located (crosses) seeds in (A) 1993 and (B) 1995. Sample
sizes are shown in parentheses.
removal declined over time, and that seed predation
eventually approached 100% at all distances (Fig. 7a).
Only one seed (at 80 m) survived 10 wk. Removal of
seeds was significantly faster for seeds 5 4 0 m from
parents, than for seeds placed 60 or 80 m away (Wil-
20
40
Percentage of seeds
60
FIG. 6. Percentage of marked seeds found in different
locations after removal by animals. See text (Metlzodst Posrdispersal seer1 prerlntiorz) for definitions of categories.
coxon x2 = 4.5, df = 1, P = 0.03). The 1996 experiment
(Fig. 7b) showed that seed removal was significantly
faster for seeds experimentally dispersed during the
peak of fruiting than before the fruiting peak, but only
for seeds 5 m (under the crown) from parents (x' =
6.87, df = 1, P = 0.009), and not for seeds 40 m away
(x2 = 0.21, df = 1, P = 0.64). Removal rates did not
differ between seeds placed 5 and 40 m from parents
at either time in the fruiting season (early, X' = 0.65,
df = 1, P = 0.42; middle, x2 = 3.17, df = 1, P =
0.20).
The 1996 sample of 44 seeds was censused at dusk
(1800-1830) and dawn (0530-0600) for two consecutive days. Five seeds had been removed by the first
dusk census, two of which clearly were taken by small
rodents into small holes in fallen logs. By the next
dawn, 27 more seeds had been removed. One seed was
TABLE2. Results of logistic regressions of postdispersal survival of marked seeds and of seedlings against habitat variables.
2
Response period
1993
2-wk survival of seeds
r?
0.3 11
log likelihood
Predictors-1
X? 54.19":'::"
-
leaf litter"""
canopy cover*
dispersal date*""
canopy cover"
dispersal date:':
seed ~ n a s s * ~ :
+
-
I-yr survival of seedlingst
1995
I -d survival of seeds
2-wk survival of seeds
0.10
43.25"""
-
0.23
133.3***
-
0.3 1
144,5*:'::%
+
+ dispersal date'"':"
-
1-yr survival of seedlingst
0.13
67.52:"'":'.
-
tree number*
distance to parent***
dispersal date*":"
tree number*
canopy cover"
dispersal date''.'':
tree number*
Note: Only significant effects are listed.
"P < 0.05, #:"p< 0.01, *:w:P p 0,001,
7 The sign in front of each predictor variable indicates positive or negative correlation with the response variable.
i. These seeds were protected from seed predators with cages until germination. No variables significantly predicted germination success (not shown).
May 2000
CLOUD FOREST SEEDLING RECRUITMENT
34 1
mo, 82% of the seeds from the "small rodents only"
exclosure had been removed. Thus, although large
mammals may take some seeds, small mammals (presumably rodents) will eventually find and eat most
seeds.
Germination and seedling survival
-+
40 m. Peak
Time (d)
FIG.7 . (A) Removal of marked seeds experimentally dispersed 0. 20. 40, 60, and 80 m from trunks of the fruiting
Ocotea erzrlresicinci trees. starting in June 1995 (N = 12 at
each distance). (B) Removal of marked seeds at 5 and 4 0 m
from parent trees in the early fruiting season (late May) and
at the peak (mid-June) of the fruiting season. 1996 (N = 24
for each category). All seeds at 0 and 5 In were under the
parental crowns.
taken during that day and four more by the next dawn.
Thus, most seed removal (73%) took place at night (x'
= 25.9, df = 1, P < 0.001), probably by small rodents.
Exclosures.-Removal
rates differed significantly
among the three treatments; seeds accessible to all rodents were taken faster than the other two treatments
(Wilcoxon x2 = 16.7, df = 2, P = 0.007). After 8 d,
three seeds had been removed from the small mesh
exclosures, which were designed to exclude all rodents
(Fig. 8). Two of these exclosures had burrows inside
that were probably dug during the three months between construction of exclosures and the beginning of
the experiment. The third exclosure was hit by a fallen
tree, thus allowing access. Except for those three, seeds
were not removed from "no rodent" exclosures. Several other exclosures were damaged by falling branches, or breached by peccaries foraging for Inga pods,
but these incidents either occurred after the seed had
been removed or did not lead to seed removal. After
8 d all the seeds from the control plots had been removed, while 52% of the seeds from the "small rodents
only" exclosures had been removed (Fig. 8). After 4
Ocotea endresiana seeds began germinating -6 wk
after dispersal. The germination rate of caged seeds
ranged 70-95% in both years, and did not differ among
the seeds at dispersed, nondispersed, and random sites.
Of the seeds that germinated in 1993 and 1995, 35 and
27%, respectively, survived one year as seedlings.
Overall, 28% and 22% of seeds in 1993 and 1995,
respectively, germinated and survived one year. For the
1993 caged seeds, annual mean survival for the first
three years was 3 1% (Table 3). Seeds at four dispersed
and five random sites from 1993 survived until June
1996, for an overall 3-yr survival rate of 2%. Within
each stage (germination, I-, 2-, and 3-yr survival), the
three location treatments had similar survival rates except for I-yr survival in 1993 (F2,>,= 11.84, P =
0.0002). Seeds at dispersed sites had significantly higher 1-yr survival than those at nondispersed and random
sites ( P < 0.001). Random and nondispersed sites had
similar 1-yr survival ( P = 0.35). Germination success
of seeds defecated by guans did not differ from that of
seeds regurgitated by the other four species (x? = 0.98,
df = 1, P > 0.05).
None of the variables in the logistic regression models predicted germination success in either year. Significant predictors of 1-yr survival included lower canopy cover and later dispersal date in both 1993 and
1995 (Table 2 [seedlings]). Additionally, in 1995, 1-yr
survival was heterogeneous with respect to which par-
0
2
4
Time (d)
6
8
FIG. 8. Percentage of seeds remaining in three types of
exclosures after 8 d. For each treatment, N = 38. The "no
rodent" exclosure was designed to exclude all terrestrial animals. "small rodent" exclosures excluded large animals but
allowed access to small animals through 6-cm mesh, and "all
rodent" plots were controls with no exclosure.
Ecological Monographs
Vol. 70. No. 2
DANIEL G. WENNY
342
TABLE3. R e s ~ ~ lof
t s logistic regressions of postdispersal survival of seedlings predicted by habitat variables.
2 log
likelihood
Response
Predictorst
r?
X'
1993
1-yr survival
0.10
43.25***
-
canopy cover*
1995
1-yr survival
0.13
67.52***
-
canopy cover*
+ dispersal date*
+ dispersal date**
tree number*
Notes: These seeds were protected from seed predators with
cages until germination. Only significant effects are listed.
No variables significantly predicted germination success (not
shown).
* P < 0.05, "*P < 0.01. ***P < 0.001.
t The sign in front of each predictor variable indicates positive or negative correlation with the response variable.
ticular fruiting tree was closest, although no clear-cut
pattern was discernible in terms of crop size, tree size,
or proximity to other fruiting trees.
The causes of mortality of caged seeds and seedlings
during the first year differed among dispersed, nondispersed, and random sites in both 1993 (x2 = 29.14, df
= 6, P < 0.001) and 1995 (x' = 33.59, df = 6, P <
0.001; Fig. 9). Many seedlings were killed when seed
predators (presumably rodents) removed the seed and
severed the connection between the shoot and root.
Mammalian herbivores also killed seedlings by entirely
removing the leaves and shoot. Mortality by the combination of mammalian seed predators and herbivores
was significantly different among the three location
treatments in 1995 (F2,17= 4.98, P = 0.012), but not
in 1993 (F2,24= 3.56, P = 0.044). In 1995, mammals
killed more seeds and seedlings at dispersed than at
random sites ( P < 0.05), but mortality caused by mammals at nondispersed sites did not differ from that at
either dispersed or random sites ( P > 0.05; Fig. 9).
Mortality caused by beetle larvae (Heilipus sp. and at
least one other unidentified species) differed among the
three treatments in both years (1993, F2,23= 4.85, P
= 0.012; 1995, F2,37 = 6.64, P = 0.003). In both years,
more nondispersed than dispersed seeds were killed by
beetles ( P < 0.05; Fig. 9). Mortality caused by fungal
pathogens, physical damage (falling branches, trampling), or unknown causes did not differ among treatments or years. For the three treatments, combined levels of mortality by each cause were similar between
the two years, except for insect predation, which was
significantly higher in 1993 vs. 1995 (t test; t = 1.97,
df = 52, P < 0.05), and mortality by mammals, which
was significantly higher in 1995 vs. 1993 ( t = 2.53, df
= 67, P < 0.01).
Mean canopy cover for seedlings that survived one
year at dispersed sites (94.3% -C 3.6, N = 86) was
significantly lower than for seedlings that died (95.2%
-1- 2.3, N = 303), whereas canopy cover did not differ
with 1-yr survival for nondispersed or random sites (2way ANOVA, F?,,, = 3.98, P = 0.019). Denser canopy
cover was also a significant predictor of higher mortality by fungal pathogens (logistic regression, P =
0.0013). Seedlings that survived one year were significantly taller at sites with less dense canopy cover (linear regression, height = 28.93-0.233x, r 2 = 0.284, P
< 0.001). Seedlings from seeds dispersed by bellbirds
were significantly taller than those at other species'
sites (t test = 2.37, N = 38, P = 0.02).
Germination trials.-Regurgitated
seeds and seeds
cleaned of pulp by hand had higher germination rates
than the seeds in intact fruits (x2 = 43.94, df = 2, P
< 0.001). Of 30 seeds in each of the first two treatments, all but one seed germinated, and that one failed
because it was infested with a beetle larva. In contrast,
seeds in intact fruits tended to mold; eight germinated,
and 22 appeared mealy and no longer viable after 12
wk. The eight seeds in fruits that germinated never
developed a shoot outside the fruit pulp and eventually
rotted. Thus, ingestion of seeds by frugivores was beneficial in terms of pulp removal, but probably did not
A) 1993
Dispersed (N = 116)
Nondispersed (N = 1 12)
Random (N = 73)
Mammals
C
-
Insects
Fungal
Physical
B) 1995
- ,
Dispersed (N = 186)
Nondispersed (N = 134)
Random (N = 82)
b
T
0.4
0.2
n
Mammals
Insects
Fungal
Physical
FIG. 9. Mean proportion per tree ( + 1 SD) of caged dispersed and nondispersed seeds, and randomly located seeds,
killed by different causes during the first year, for seeds from
(A) 1993 and (B) 1995. Cages were removed after germination and shoot development. Sample sizes of seeds or seedlings killed are shown in parentheses. Within each cause of
mortality, where a difference among sites was detected, bars
with different letters above are significantly different (Bonferroni-adjusted a = 0.0125). See text (Methods: Germirzntion
crnd seedling survival) for definitions of types of mortality.
CLOUD FOREST SEEDLING RECRUITMENT
May 2000
otherwise affect germination. Two of the 10 buried
seeds started to germinate (the seed coat appeared
cracked) and seemed viable, but the other eight seeds
rotted. After 4 mo, neither of the apparently germinating seeds had any sign of root development, nor had
the cotyledons begun to separate. Six of 10 buried seeds
had whitish mold on the seed coat. Thus, few buried
seeds appear to establish as seedlings.
Seedling and sapling plots.-Seedlings and saplings
were equally common in 1 0 X 1 0 m plots under (6.6
f 1.5 individuals) and away from (4.9 f 3.5) the parent
trees (paired t test; t = 1.15, P > 0.05). Individuals in
plots away from the parent trees, however, were taller
(79.9 t- 14.9 cm) than those near the parent trees (47.4
+- 7.2; Mann-Whitney U test, P = 0.009).
Recruitment
Based on the experiment with marked seeds, the
probability of recruitment is determined mostly by
postdispersal seed predators. Assuming the experiment
reflects the absolute level of seed predation, and all
seeds that escape predation germinate and survive, then
recruitment was 0% in 1993 and 0.4% in 1995. On the
other hand, if seed removal at 2 wk (75-93%) reflects
the ultimate pattern of survival (e.g., Li et al. 1996),
then it is possible to calculate the influence of stagespecific mortality patterns on the relative probability
of recruitment (Table 3, Fig. 10). The initial pattern of
seed rain shows a strong peak near the parent trees,
and secondary peaks at distances >40 m that correspond with bellbird song perches (Fig. 10a). The bellbird perches encompass a range of canopy cover values
from forest-gap edge (-92%) to gap centers (-70%).
Thus, seed rain is spread across a wider range of canopy
conditions with increasing distance from the parent
trees. Two-week seed survival is higher for seeds >10
m from fruiting trees, but relatively consistent across
the range of canopy cover (Fig. lob). Germination is
high for all seeds, regardless of distance from parent
or canopy cover (Fig. 10c). One-year seedling survival
is highest for sites >40 m from parents and <90%
canopy cover (Fig. 10d). The cumulative probability
of recruitment (the product of seed rain, 2-wk seed
survival, germination, and 1-yr survival) shows a peak
near the parent trees and shows secondary peaks corresponding to the bellbird perches (Fig. 10e). Note that
the probability of a seed dispersed by birds surviving
one year is <0.2% at any location.
The results illustrate three main points: (1) bellbirds
dispersed seeds in a different pattern than the other
species; (2) postdispersal seed predation is highest near
the parent trees, but is also high everywhere; and (3)
the pattern of recruitment is bimodal with a peak near
the parent tree in closed canopy forest, as well as another in gaps corresponding to bellbird perches.
343
Seed dispersal
Most 0. endresiana seeds handled by birds land under or just beyond the crowns of parent trees. This
distribution of seeds is typical for vertebrate-dispersed
plants (Dirzo and Dominguez 1986, Hoppes 1988,
Willson 1993, Laman 1 9 9 6 ~ ) The
.
abundance of 0.
endresiana adults in the study area makes dispersal
beyond 80 m from any conspecific tree virtually impossible. Considering the restricted elevational range
of this species, longer distance dispersal may lead to
arrival in habitats unsuitable for establishment (Wheelwright 1988). Thus, while regurgitating seeds just outside the crown of a fruiting tree may seem like poor
quality dispersal (McKey 1975), the sheer abundance
of seeds in that region
may lead to a small peak in
recruitment near the parent tree (Fig. 10e), as predicted
by Hubbell (1980) and Condit et al. (1992). The long
tail of the seed distribution, however, is apparently important for recruitment (see Clark 1998, Clark et al.
1998), because seeds dispersed farther from the parent
trees have a greater chance of arriving in a habitat
where the probability of seedling survival is higher
(i.e., a safe site).
Of the five species of birds that dispersed most of
the seeds in this study, all but one deposited most seeds
in closed canopy forest. Only male bellbirds frequently
dispersed seeds to tree fall gaps. In my study site, bellbird song perches were in dead Sapium oligoneuron
trees, bordering large tree fall gaps. Such sites are typical for bellbirds: Snow (1977) reported that bellbird
song perches at lower elevations in the Monteverde
area (where the forest is fragmented) are usually on
dead branches in tall trees on the forest edge. In other
Neotropical forests, three other species of bellbirds
(Procnias spp.) exhibit similar behavior in habitual use
of tall, exposed song perches, often on dead branches
(Snow 1961, 1970, 1 9 7 3 ~ )Snags
.
used by bellbirds in
my site had several branches, and the birds made use
of many different branches within each tree, such that
seeds under them were scattered over -25 m2, including a gradient of site conditions from gap to forest
understory. Male bellbirds also dispersed seeds in forest, as do the other species, but they typically spend
80-95% of the day in the vicinity of the song perch
(Snow 1977) and probably disperse a similar percentage of seeds they process under song perches. Female
bellbirds were rarely seen at the song perches, and then
only for a few minutes.
The influence of bellbirds on the pattern of seed fall
is shown by the slight increase in the number of seeds
40-65 m from the parent trees (Fig. 2). Bellbird perches
in my site happened to be far (>40 m) from any fruiting
0, endresiana trees. The other four bird species that
disperse 0.endresiana seeds occasionally perch on the
edges of gaps (personal observation), although
whether they do so more often than expected based on
perch availability is unknown. Because birds do not
DANIEL G. WENNY
Ecological Monographs
Vol. 70. No. 2
FIG. 10. (A) Mean seed rain, (B) 2-wk seed survival, (C)
germination, (D) I-yr seedling survival, and (E) cumulative
probability of recruitment, as functions of canopy cover and
distance from parent for both years combined, with means
evaluated among 21 trees. For A-D, each bar represents the
percentage of the seeds or seedlings that survived that stage
in each distancelcanopy cover category, averaged among
trees. Categories with fewer than three seeds were not included and are blank, while categories with a flat rectangle
are zero. Seed rain was determined from the original distribution of seeds naturally dispersed by birds. Two-week seed
survival was estimated from the removal of marked seeds.
Germination and seedling survival were estimated from the
seeds caged until after germination. Recruitment (the cumulative probability of surviving every stage) is calculated
as the product (A X B X C X D) for each distancelcover
category. For clarity, standard deviations are not shown.
May 2000
CLOUD FOREST SEEDLING RECRUITMENT
(cannot?) regurgitate in flight, and guans apparently do
not defecate in flight (personal observation), the most
likely way an 0. endresiana seed can arrive in a gap
is via male bellbirds. I never observed the other four
species on bellbird perches. Thus, 0. endresiana has
two types of dispersers that deposit seeds in two different, but, predictable patterns: male bellbirds depositing large numbers (52%) of the seeds they disperse
in and near gaps, and the other four species depositing
seeds in forest, most (75%) within 25 m of parent trees,
and fewer >25 m away (Wenny and Levey 1998). Additionally, on several occasions I followed quetzals and
guans from one fruiting 0. endresiana to another,
where they deposited seeds before eating fruits in the
second tree. Although I had no method of identifying
the source tree of the seeds, it is likely that dispersal
of seeds from one tree to another conspecific occurs
regularly (Wheelwright 1991, Gibson and Wheelwright
1995).
The patterns of seed distribution described here may
be similar for other trees in cloud forests, where these
disperser species occur during the breeding season, but
may be very different during the rest of the year. The
five species of birds that disperse 0. endresiana are
the most important fruit consumers of canopy trees in
the area (Wheelwright et al. 1984). Seeds from >20
other species of trees are dispersed by these birds, and
probably have similar patterns of seedfall as 0. endresiana, with a peak near the parent tree and a second
peak at bellbird perches. However, all five dispersers
(except perhaps toucanets) undertake elevational shifts
that correspond with fruit availability (Wheelwright
1983, Levey and Stiles 1992, Powell and Bjork 1995,
Guindon 1996) and leave the study area in late July or
early August. (In 1994 they left in June, when 0. endresiana failed to fruit). Male bellbirds may not provide such a pronounced peak in seed dispersal to song
perches after the breeding season, because they seem
to use many different perches in the nonbreeding season, and they tend to use subcanopy perches that are
not exposed (Stiles and Skutch 1989). Thus, seed fall
for the community as a whole is likely temporally, as
well as spatially, heterogeneous.
The implications of seed dispersal to habitual perches deserve further study. Do habitual perches represent
foci of plant recruitment, or of seed predation? Seed
dispersal to perches has been frequently studied in successional landscapes (Livingston 1972, McDonnell
1986, Holthuijzen et al. 1987, Guevara and Laborde
1993, McClanahan and Wolfe 1993, Robinson and
Handel 1993, Vieira et al. 1994, Nepstad et al. 1996,
Duncan and Chapma 1999). In these systems, directed
dispersal may play a key role in the restoration of abandoned pastures, logged forests, and other degraded
lands. It is possible that directed dispersal is more common in intact forests than previously expected. For example, manakins (Pipridae) and cocks-of-the-rock
(Rupicola: Cotingidae) choose lek perches that are
345
more sunlit than typical understory perches (Endler and
ThCry 1996), and such sites may provide growth advantages for seedlings. Snow (1961, 1970, 1973b)
notes that other species of bellbirds (Procnias) preferentially select perches on dead trees or branches, or
in sparsely vegetated trees. For these species, competition among males for females drives them to be as
conspicuous as possible. The fortuitous outcome of this
behavior may be a disproportionate effect on plant recruitment in the vicinity of their display areas (ThCry
and Larpin 1993). Whether habitual perches represent
foci of seedling recruitment (McDonnell and Stiles
1983, McDonnell 1986) or lead to density-dependent
seed and seedling mortality (Wheelwright 1988), needs
to be examined in more detail.
Seed predation
Removal of marked seeds after dispersal in this study
always resulted in predation. Although seeds were taken into burrows and logs, I never found any treatment
of seeds indicative of scatter-hoarding, such as burial
in the soil (Smythe 1978, Hallwachs 1986, Forget 1990,
1993), or under piles of leaf litter (Forget 1991). In
addition, the failure of buried seeds to germinate and
survive in the greenhouse experiment suggests that
scatter-hoarding would not be advantageous for 0. endresiana recruitment. Only two species of mammals
known to scatter-hoard seeds occur in the study site:
spiny pocket mice (Heteromys desmarestianus) and
agoutis (Dasyprocta punctata). Both species are much
less common here than at lower elevations (personal
observation; K. G. Murray, personal communication).
Although some species of squirrels (Sciurus) and deer
mice (Peromyscus) are known to cache seeds in other
regions of the world (Vander Wall 1990), information
on tropical species (in this site, S. deppei and P. mexicanus) is lacking (Emmons 1990).
The main seed predators in this study were probably
rodents, particularly the deer mouse Peromyscus mexicanus, which is by far the most common terrestrial
rodent in the study area (Anderson 1982, Langtimm
1992). Based on the exclosure experiment, however, it
is possible that species larger than mice (i.e., agouti,
paca, peccary, wood-quail) are also important seed
predators. Because removal of seeds from the open
control plots was about twice that of the "small rodents
only" plots, agoutis and other large mammals may have
been responsible for as much as 50% of 0.endresiana
seed predation. Nevertheless, the exclosures also
showed that, in the absence of agoutis, small rodents
will find and eat most of the seeds, albeit over a longer
time period. Additionally, the findings that most
marked seeds from the removal experiment were taken
at night, and that removal of seeds always resulted in
predation rather than scatter-hoarding, suggest that
agoutis (which are mostly diurnal and known scatterhoarders [Smythe 1978, Hallwachs 19861, may not be
taking many seeds. However, few data exist on the
Ecological Monographs
Vol. 70, No. 2
DANIEL G. WENNY
extent to which agoutis eat but do not scatter-hoard,
certain species (e.g., Smythe 1978).
The high degree of postdispersal seed predation reported here is not unusual (Harper 1977, Cavers 1983).
Recent reviews have found 90-100% losses of seeds
to predators in 22 of 62 studies (Crawley 1992, Hulme
1993). Previous studies on seed predation of Lauraceae
found 96-100% predation (Wheelwright 1988, Chapman 1989, Holl and Lulow 1997). The high level of
seed predation in my study site could be explained by
dense vegetation and cloudy conditions, providing
small rodents protection from predators (Bowers and
Dooley 1993, Vhsquez 1996). High predation is probably not a consequence of high rodent populations due
to lack of predators, because potential predators of mice
(such as owls, forest-falcons, and cats) are relatively
common (personal observation). During the day, light
levels in the study area are often reduced because of
cloudy weather (Cavelier 1996, Chazdon et al. 1996),
as well as due to dense vegetation. Indeed, on a few
occasions, mice were seen during the day (especially
Peromyscus and Scotinomys) and some marked seeds
were removed by small rodents during the day (see
Results: Postdispersal seed predation: Distance effect).
In addition, local naturalists report seeing more mice
on misty or rainy days and nights than during clear
weather (T. Guindon, personal communication). Thus,
the high levels of predation for 0. endresiana could
be a result of longer activity patterns by small rodents,
and such levels may be typical for many large-seeded
tree species in cloud forests.
Seedling survival and recruitment
Most of the seeds protected from rodents germinated.
High germination rates for Lauraceae are apparently
typical (Wheelwright 1985b). Some seeds that germinated were eventually killed by beetle larvae (Heilipus
sp. Curculionidae) developing in the cotyledons. After
the cages were removed from the growing seedlings,
the seeds as well as the seedlings were susceptible to
predation by mammals. It is difficult to determine the
role of each species involved, but the suite of species
that kill seedlings is probably larger than the suite of
seed predators, and could include pacas (Agouti paca),
brocket deer (Mazama americana), tapir (Tapirus bairdii), and peccaries (Tayassu tajacu), in addition to
agoutis and smaller rodents.
Insects were most likely to kill nondispersed seeds
or seedlings, while mammals killed all types of seeds
(dispersed, nondispersed, and random). This finding is
consistent with Howe (1993) and Terborgh et al.
(1993), who found that insect-caused mortality was distance dependent, but mortality by mammals was not.
In New Guinea, Merg (1994) found the opposite pattern: insects tended to kill dispersed seeds and mammals killed nondispersed seeds. In contrast to insects
and mammals, fungal pathogens killed proportionately
more 0. endresiana seedlings in closed-canopy forest
than in gaps. Augspurger (1984) also showed fungal
pathogens of seedlings were less prevalent in gaps than
forest understory, for nine species of tropical trees.
The greater height of seedlings and saplings in plots
>20 m from the parent trees, compared to those in
plots beneath the crowns, suggests that seedling mortality was higher closer to the parent trees. Seedling
ages are not known, but this difference suggests that
longer term survival several years after dispersal is
higher for seedlings >20 m from the crown of parent
trees than for seedlings beneath crowns. Thus, seed
dispersal away from the crown of a conspecific appears
to increase the chance of recruitment over the long term
(Clark and Clark 1984, Li et al. 1996).
The overall pattern of recruitment with respect to
distance from parent trees and canopy cover was bimodal. This pattern was caused by the combination of
distance-dependent seed predation (Fig. lob) and the
influence of canopy cover on seedling survival (Fig.
10d), despite the fact that most seeds landed close to
the parent trees in closed-canopy forest (Fig. 10a). This
pattern is an example of spatial discordance caused by
the lack of congruity among the stages leading to recruitment (Henera et al. 1994, Jordano 1995). The occurrence of such discordance emphasizes the importance of stage-specific survival patterns on patterns of
recruitment and the need for data on the sequential
stages of plant reproduction rather than on only one or
a few stages. Bellbirds are clearly an important part of
the dispersal system of 0. endresiana, as over half
(52%) of the seeds they dispersed landed in a zone of
higher recruitment.
Conclusion
The overall conclusion of this study is that the pattern of recruitment of 0. endresiana depends on the
combined effects of seed dispersers, seed predators,
and seedling mortality. Selection on plant traits occurs
during each stage, and selection during sequential stages may be opposed (Wheelwright and Orians 1982,
Herrera 1985, 1986, Wheelwright 1988, Herrera et al.
1994). For example, high seed predation overall, and
the slight preference for larger seeds by seed predators,
may select for smaller seeds or larger seed crops, while
seed dispersers may prefer larger fruits, which have
larger seeds, but occur in smaller fruit crops (but see
Howe and Vande Kerckhove [1980, 19811, Wheelwright [I9911 and Mazer and Wheelwright [1993]). On
the other hand, seed size did not influence germination,
seedling height, or seedling survival. In addition, the
bimodal spatial pattern of recruitment may represent
disruptive selection on seedling traits. Some trees were
visited by all five species of dispersers, while the trees
far from bellbird perches tended not to be visited by
bellbirds. Thus, seeds dispersed to gaps beneath bellbird perches were mostly from a subset of the available
trees. The extent to which such differences in dispersal
and subsequent recruitment affect gene flow is poorly
May 2000
C L O U D FOREST SEEDLING R E C R U I T M E N T
understood (Gibson and Wheelwright 1995, Hamrick
and Nason 1996)' Further studies that 'Ompare
dispersal patterns and the subsequent stages leading to
recruitment at different sites, as well as over longer
time periods, are especially needed.
ACKNOWLEDGMENTS
Victor PerCz, Jason Bennett, and Wendy Gibbons provided
crucial field assistance. I thank the Organization for Tropical
Studies ( O T S ) ,Monteverde Institute, Tropical Science Center,
University o f Florida Department o f Zoology, and especially
the Monteverde community for logistic support. Doug Levey
helped with all aspects o f the project, including funding, fieldwork, analysis, and manuscript preparation. He deserves credit for the success o f the project, whereas the mistakes are
entirely m y own. Comments from Karen Bjorndal, Colin
Chapman, David Clark, David Gorchov, Susan Moegenburg,
and Jack Putz greatly improved the manuscript. I also thank
Greg Murray, Nat Wheelwright, Bill Haber, and Carlos, Benito, and TomBs Guindon for advice and encouragement.
Funding from the Organization for Tropical Studies (Pew
Charitable Trust and Jesse Smith Noyes Foundation), the
American Museum o f Natural History (Chapman Fund), Sigma Xi (Grants-in-Aid o f Research), and National Science
Foundation (Dissertation Improvement Grant DEB 93-2 1553)
is gratefully acknowledged.
Abacus Concepts. 1989. SuperANOVA: accessible general
linear modeling. Abacus Concepts, Berkeley, California.
USA.
Anderson, S . D. 1982. Comparative population ecology o f
Perornyscus mexicanus in a Costa Rican wet forest.. Dissertation. University o f Southern California, Los Angeles,
California, U S A .
Augspurger, C . K. 1984. Seedling survival o f tropical tree
species: interactions o f dispersal distance, light-gaps, and
pathogens. Ecology 65:1705-1 7 12.
Augspurger, C . K. 1990. T h e potential impact o f fungal pathogens on tropical plant reproductive biology. Pages 237245 in K. S . Bawa and M. Hadley, editors. Reproductive
ecology o f tropical forest plants. UNESCOIParthenon Publishing Group, Paris, France.
Becker, P., L. W . Lee, E. D. Rothman, and W . D. Hamilton.
1985. Seed predation and the coexistence o f tree species:
Hubbell's models revisited. Oikos 44:382-390.
Bowers, M. A , , and J. L . Dooley. 1993. Predation hazard and
seed removal b y small mammals: microhabitat versus patch
scale effects. Oecologia 94:247-254.
Brokaw, N. V . L . 1982. T h e definition o f tree fall gap and
its e f f e c t on measures o f forest dynamics. Biotropica 14:
158-160.
Bronstein, J. L., and K. H o f f m a n . 1987. Spatial and temporal
variation in frugivory at a Neotropical fig, Ficus pertusa.
Oikos 49:261-268.
Burger, W . , and H. van der W e r f f . 1990. Lauraceae. Flora
Costaricensis 80. Fieldiana Botany ( N e w Series) 23: 1-121.
Cavelier, J . 1996. Environmental factors and ecophysiological processes along altitudinal gradients in wet tropical
mountains. Pages 399-439 in S . S . Mulkey, R . L. Chazdon,
and A . P. Smith, editors. Tropical forest plant ecophysiology. Chapman and Hall, New Y o r k , New Y o r k , U S A .
Cavers, P. B . 1983. Seed demography. Canadian Journal o f
Botany 61:3578-3590.
Chapman, C . A . 1989. Primate seed dispersal: the fate o f
dispersed seeds. Biotropica 21: 148-154.
Chavez-Ramirez, F., and R . D. Slack. 1994. Effects o f avian
foraging and post-foraging behavior on seed dispersal patterns o f Ashe juniper. Oikos 71:40-46.
347
Chazdon, R. L., R. W . Pearcy, D. W . Lee, and N. etcher.
1996. Photosynthetic responses o f tropical forest plants to
contrasting light environments. Pages 5-55 in S . S . Mulkey,
R , L , Chazdon, and A , P, Smith, editors, Tropical forest
plant ecophysiology. Chapman and Hall, New Y o r k , New
York, U S A .
Cintra, R . , and V . Horna. 1997. Seed and seedling survival
o f the palm Astrocaryum murumuru and the legume
tree
Dipteryx micrarztha in gaps in Amazonian forest. Journal
o f Tropical Ecology 13:257-277.
Clark, D. A . 1994. Plant demography. Pages 90-105 in L .
A . McDade, K. S . Bawa, H. A . Hespenheide, and G . S .
Hartshorn, editors. La Selva: ecology and natural history
o f a neotropical rainforest. University o f Chicago Press,
Chicago, Illinois, U S A .
Clark, D. A , , and D. B. Clark. 1984. Spacing dynamics o f a
tropical tree: evaluation o f the Janzen-Connell model.
American Naturalist 124:769-788.
Clark, D. B . , and D. A . Clark. 1991. T h e impact o f physical
damage on canopy tree regeneration in tropical rain forest.
Journal o f Ecology 79:447-457.
Clark, J . S . 1998. W h y trees migrate so fast: confronting
theory with dispersal biology and the paleorecord. American Naturalist 152:204-224.
Clark, J . S . , E. Macklin, and L. W o o d . 1998. Stages and
spatial scales o f recruitment limitation in southern Appalachian forests. Ecological Monographs 68:213-235.
Clifford,H. T., and G . B . Monteith. 1989. A three-phase seed
dispersal mechanism in Australian quinine bush (Peltostigma pubescens Domin). Biotropica 21:284-286.
Compton, S . G., A . J . F. K. Craig, and I . W . R. Waters. 1996.
Seed dispersal in an African fig tree: birds as high-quantity,
low-quality dispersers? Journal o f Biogeography 23:553563.
Condit, R., S . P. Hubbell, and R. B. Foster. 1992. Recruitment
near conspecific adults and the maintenance o f tree and
shrub diversity in a neotropical forest. American Naturalist
140:261-286.
Connell, J . H. 1970. O n the role o f natural enemies in preventing competitive exclusion in some marine animals and
in rain forest trees. Pages 298-312 in P. J. den Boer and
G . R . Gradwell, editors. Dynamics o f populations. PUDOC,
Wageningen, T h e Netherlands.
Corlett, R. T., and P. W . Lucas. 1990. Alternative seed-handling strategies in primates: seed-spitting by long-tailed
macaques (Macaca fascicularis). Oecologia 82:166-17 1.
Crawley, M. J . 1992. Seed predators and plant population
dynamics. Pages 157-191 irz M . Fenner, editor. Seeds: the
ecology o f regeneration in plant communities. C A B International, Wallingford, U K .
Crome, F. H. J. 1975. T h e ecology o f fruit pigeons in tropical
northern Queensland. Australian Wildlife Research 2: 155185.
Cruz, A . 1981. Bird activity and seed dispersal o f a montane
forest tree (Dunalia arborescens) in Jamaica. Biotropica
12S:34-44.
Denslow, J . S . 1987. Tropical rainforest gaps and tree species
diversity. Annual Review o f Ecology and Systematics 18:
431-451.
DeSteven, D., and F. E. Putz. 1984. Impact o f mammals on
early recruitment o f a tropical tree, Dipteryx panamensis,
in Panama. Oikos 43:207-2 16.
Dirzo, R . , and C . A . Dominguez. 1986. Seed shadows, seed
predation and the advantages o f dispersal. Pages 237-249
in A . Estrada and T. H. Fleming, editors. Frugivores and
seed dispersal. Dr. W . Junk Publishers, Dordrecht, T h e
Netherlands.
Duncan, R . S., and C . A . Chapman. 1999. Seed dispersal and
potential forest succession in abandoned agriculture in tropical Africa. Ecological Applications 9:998-1008.
348
DANIEL G. WENNY
Emmons, L. H. 1990. Neotropical rainforest mammals. University of Chicago Press, Chicago, Illinois, USA.
Endler, J. A., and M. Thery. 1996. Interacting effects of lek
placement, display behavior, ambient light, and color patterns in three neotropical forest-dwelling birds. American
Naturalist 148:421-452.
Englund, R. 1993. Fruit removal in Viburrzurlz opulus: copious
seed predation and sporadic massive seed dispersal in a
temperate shrub. Oikos 67:503-5 10.
Estrada, A , , G . Halffter, R. Coates-Estrada, and D. A. Meritt,
Jr. 1993. Dung beetles attracted to mammalian herbivore
(Alouatta palliata) and omnivore (Nasua narica) dung in
the tropical rain forest of Los Tuxtlas, Mexico. Journal of
Tropical Ecology 9:45-54.
Fogden, M. 1993. An annotated checklist of the birds of
Monteverde and Pefias Blancas. Green Mountain Publications, Monteverde, Costa Rica.
Forget, P.-M. 1990. Seed-dispersal of V o u a c a p o ~ ~americana
a
(Caesalpiniaceae) by caviomorph rodents in French Guiana. Journal of Tropical Ecology 6:459-468.
Forget, P.-M. 1991. Scatter hoarding of Astrocaryum p a r a rnaca by Proechimys in French Guiana: comparison with
Myoprocta exilis. Tropical Ecology 32:155-167.
Forget, P.-M. 1993. Postdispersal predation and scatter hoarding of Dipteryx panamensis (Papilionaceae) seeds by rodents in Panama. Oecologia 94:255-26 1.
Forget, P.M., and T. Milleron. 1991. Evidence for secondary
dispersal by rodents in Panama. Oecologia 87:596-599.
Gentry, A. H. 1982. Patterns of neotropical plant species
diversity. Evolutionary Biology 15: 1-84.
Gentry, A. H. 1990. Floristic similarities and differences between southern Central America and upper and central
Amazonia. Pages 141-159 in A. H. Gentry, editor. Four
neotropical forests. Yale University Press, New Haven,
Connecticut, USA.
Gibson, J. P., and N. T. Wheelwright. 1995. Genetic structure
in a population of a tropical tree Ocotea tenera (Lauraceae):
influence of avian seed dispersal. Oecologia 103:49-54.
Guevara, S., and J. Laborde. 1993. Monitoring seed dispersal
at isolated standing trees in tropical pastures: consequences
for local species availability. Vegetatio 107/108:319-338.
Guindon, C. F. 1996. The importance of forest fragments to
the maintenance of regional biodiversity in Costa Rica.
Pages 168-186 in J. Schelhas and R. Greenberg, editors.
Forest patches in tropical landscapes. Island Press, Washington, D.C., USA.
Haber, W. A. 1991. Lista provisional de las plantas de Monteverde, Costa Rica. Brenesia 34:63-120.
Hallwachs, W. 1986. Agoutis (Dasyprocta punctata): inheritors of guapinol (Hymenaea courbaril: Leguminosae).
Pages 285-304 in A. Estrada and T. H. Fleming, editors.
Frugivores and seed dispersal. Dr. W. Junk Publishers, Dordrecht, The Netherlands.
Hammond, D. S. 1995. Postdispersal seed and seedling mortality of tropical dry forest trees after shifting agriculture,
Chiapas, Mexico. Journal of Tropical Ecology 11:295-3 13.
Hamrick, J. L., and J. D. Nason. 1996. Consequences of
dispersal in plants. Pages 203-236 in 0 . E . Rhodes, R. K.
Chesser, and M. H. Smith, editors. Population dynamics in
ecological space and time. University of Chicago Press,
Chicago, Illinois, USA.
Harper, J. L. 1977. Population biology of plants. Academic
Press, London, UK.
Hartshorn, G . S . 1978. Tree falls and tropical forest dynamics. Pages 617-638 in P. B. Tomlinson and M. H. Zimmerman, editors. Tropical trees as living systems. Cambridge University Press, Cambridge, UK.
Hartshorn, G. S. 1983. Plants. Pages 118-157 it? D. H. JanZen, editor. Costa Rican natural history. University of Chicago Press, Chicago, Illinois, USA.
Ecological Monographs
Vol. 70. No. 2
Herrera, C. M . 1985. Determinants of plant-animal coevolution: the case of mutualistic dispersal of seeds by vertebrates. Oikos 44: 132-141.
Herrera, C. M . 1986. Vertebrate-dispersed plants: why they
don't behave the way they should. Pages 5-18 in A. Estrada
and T. H. Fleming, editors. Frugivores and seed dispersal.
Dr. W. Junk Publishers, Dordrecht, The Netherlands.
Herrera, C. M., P. Jordano, L. Lopez-Soria, and J. A. Amat.
1994. Recruitment of a mast-fruiting, bird-dispersed tree:
bridging frugivore activity and seedling establishment.
Ecological Monographs 64:3 15-344.
Holl, K. D., and M. E. Lulow. 1997. Effects of species, habitat, and distance from edge on postdispersal seed predation
in a tropical forest. Biotropica 29:459-468.
Holm, S. 1979. A simple sequentially rejective multiple test
procedure. Scandinavian Journal of Statistics 6:65-70.
Holthuijzen, A. M . A , , T. L. Sharik, and J. D. Fraser. 1987.
Dispersal of eastern red cedar (Juniperus virginiarza) into
pastures: an overview. Canadian Journal of Botany 65:
1092-1095.
Hoppes, W. G . 1988. Seedfall pattern of several species of
bird-dispersed plants in an Illinois woodland. Ecology 69:
320-329.
Howe, H. F. 1986. Seed dispersal by fruit-eating birds and
mammals. Pages 123-189 in D. R. Murray, editor. Seed
dispersal. Academic Press, Sydney, Australia.
Howe, H. F. 1989. Scatter- and clump-dispersal and seedling
demography: hypothesis and implications. Oecologia 79:
4 17-426.
Howe, H. F. 1 9 9 0 ~ .Seed dispersal by birds and mammals:
implications for seedling demography. Pages 191-218 in
K. S. Bawa and M. Hadley, editors. Reproductive ecology
of tropical forest plants. UNESCOIParthenon Publishing
Group, Paris, France.
Howe, H. F. 1990b. Survival and growth of juvenile Virola
surinamensis in Panama: effects of herbivory and canopy
closure. Journal of Tropical Ecology 6:259-280.
Howe, H. E 1993. Aspects of variation in a neotropical seed
dispersal system. Vegetatio 107/108:149-162.
Howe, H. E , and W. M . Richter. 1982. Effects of seed size
on seedling size in Virola surinamerzsis; a within and between tree analysis. Oecologia 53:347-35 1.
Howe, H. F., E. W. Schupp, and L. C. Westley. 1985. Early
consequences of seed dispersal for a neotropical tree (Virola surinamensis). Ecology 66:78 1-79 1.
Howe, H. F., and J. Smallwood. 1982. Ecology of seed dispersal. Annual Review of Ecology and Systematics 13:201228.
Howe, H. F., and G . A. Vande Kerckhove. 1980. Nutmeg
dispersal by tropical birds. Science 210:925-926.
Howe, H. F., and G . A. Vande Kerckhove. 1981. Removal
of wild nutmeg (Virola surinun~ensis)crops by birds. Ecology 62:1093-1106.
Hubbell, S. P. 1980. Seed predation and the coexistence of
tree species in tropical forests. Oikos 35:214-229.
Hulme, P. E. 1993. Postdispersal seed predation by small
mammals. Symposium of the Zoological Society of London
65:269-287.
Hulme, P. E . 1997. Postdispersal seed predation and the establishment of vertebrate-dispersed plants in Mediterranean
scrublands. Oecologia 111:91-98.
James, F. C., and C . E. McCulloch. 1990. Multivariate analysis in ecology and systematics: panacea or pandora's box?
Annual Review of Ecology and Systematics 21:129-166.
Janzen, D. H. 1970. Herbivores and the number of tree species in tropical forests. American Naturalist 104:501-528.
Janzen, D. H. 1983. Dispersal of seeds by vertebrate guts.
Pages 232-262 it? D. J. Futuyma and M . Slatkin, editors.
Coevolution. Sinauer Associates, Sunderland, Massachusetts, USA.
May 2000 C L O U D FOREST SEEDLING RECRUITMENT
Jordano, P. 1992. Fruits and frugivory. Pages 105-156 irz M .
Fenner, editor. Seeds: the ecology of regeneration in plant
communities. CAB International, Wallingford, UK.
Jordano, P. 1995. Spatial and temporal variation in the avianfrugivore assemblage of Prurzus mahaleb: patterns and consequences. Oikos 71:479-491.
Jordano, P., and C. M . Herrera. 1995. Shuffling the offspring:
uncoupling and spatial discordance of multiple stages in
vertebrate seed dispersal. Ecoscience 2:230-237.
Julliot, C. 1996. Seed dispersal by red howling monkeys
(Alouatta setziculus) in the tropical rain forest of French
Guiana. International Journal of Primatology 17:239-258.
Kiltie, R. A. 1981. Distribution of palm fruits on a rainforest
floor: why white-lipped peccaries forage near objects. Biotropica 13:141-145.
Kitchings, J. T., and D. J. Levey. 1981. Habitat patterns in
a small mammal community. Journal of Mammalogy 62:
814-820.
Krefting, L. W., and E . I. Roe. 1949. The role of some birds
and mammals in seed germination. Ecological Monographs
19:270-286.
Laman, T. G . 1995. Ficus stupenda germination and seedling
establishment in a Bornean rain forest canopy. Ecology 76:
261 7-2626.
Laman, T. G . 1996a. Ficus seed shadows in a Bornean rain
forest. Oecologia 107:347-355.
Laman, T. G . 1996b. Specialization for canopy position by
hemiepiphytic Ficus species in a Bornean rain forest. Journal of Tropical Ecology 12:789-803.
Langtimm, C. A. 1992. Specialization for vertical habitats
within a cloud forest community of mice. Dissertation. University of Florida, Gainesville, Florida, USA.
Lawton, R. 0 . 1990. Canopy gaps and light penetration into
a wind-exposed tropical lower montane rain forest. Canadian Journal of Forest Research 20:659-667.
Lawton, R., and V. Dryer. 1980. The vegetation of the Monteverde Cloud Forest Preserve. Brenesia 18:101-1 16.
Lawton, R. O., and F. E. Putz. 1988. Natural disturbance and
gap-phase regeneration in a wind-exposed tropical cloud
forest. Ecology 69:764-777.
Lemmon, P. E. 1957. A new instrument for measuring forest
overstory density. Journal of Forestry 55:667-668.
Levey, D. J. 1987. Seed size and fruit-handling techniques
of avian frugivores. American Naturalist 129:471-485.
Levey, D. J. 1988. Tropical wet forest tree fall gaps and
distributions of understory birds and plants. Ecology 69:
1076-1089.
Levey, D. J., and M. M. Byrne. 1993. Complex ant-plant
interactions: rain forest ants as secondary dispersers and
postdispersal seed predators. Ecology 74: 1802-1 8 19.
Levey, D. J., T. C. Moermond, and J. S. Denslow. 1994.
Frugivory: an overview. Pages 282-294 irz L. A. McDade,
K. S . Bawa, H. A. Hespenheide, and G . S. Hartshorn, editors. La Selva: ecology and natural history of a neotropical
rainforest. University of Chicago Press, Chicago, Illinois,
USA.
Levey, D. J., and F. G . Stiles. 1992. Evolutionary precursors
of long-distance migration: resource availability and movement patterns in neotropical landbirds. American Naturalist
140:447-476.
Levey, D. J., and E G . Stiles. 1994. Birds: ecology, behavior,
and taxonomic affinities. Pages 217-228 in L. A. McDade,
K. S . Bawa, H. A. Hespenheide, and G. S. Hartshorn, editors. La Selva: ecology and natural history of a neotropical
rainforest. University of Chicago Press, Chicago, Illinois,
USA.
Li, M., M. Lieberman, and D. Lieberman. 1996. Seedling
demography in undisturbed tropical wet forest in Costa
Rica. Pages 285-314 in M. D. Swaine, editor. Ecology of
349
tropical forest tree seedlings. UNESCOIParthenon. Paris,
France.
Livingston, R. B. 1972. Influence of birds, stones, and soil
on the establishment of pasture juniper, Jurziperus cornmunis, and red cedar, J. virgitziarza in New England pastures. Ecology 53: 1141-1 147.
Martinez-Ramos, M., and A. Soto-Castro. 1993. Seed rain
and advanced regeneration in a tropical rain forest. Vegetatio 107/108:299-3 18.
Mazer, S . J., and N. T. Wheelwright. 1993. Fruit size and
shape: allometry at different taxonomic levels in bird-dispersed plants. Evolutionary Ecology 7:556-575.
McCanny, S. J. 1985. Alternatives in parent-offspring relationships in plants. Oikos 45:148-149.
McClanahan, T. R., and R. W. Wolfe. 1993. Accelerating
forest succession in a fragmented landscape: the role of
birds and perches. Conservation Biology 7:279-288.
McDonnell, M . J. 1986. Old field vegetation height and the
dispersal pattern of bird-disseminated woody plants. Bulletin of the Torrey Botanical Club 113:6-l l .
McDonnell, M. J., and E. W. Stiles. 1983. The structural
complexity of old field vegetation and the recruitment of
bird-dispersed plant species. Oecologia 56:109-116.
McKey, D. 1975. The ecology of coevolved seed dispersal
systems. Pages 159-191 irz L. E. Gilbert and P. H. Raven,
editors. Coevolution of plants and animals. University of
Texas Press, Austin, Texas, USA.
Merg, K. E 1994. Utility of seed dispersal in escape from
seed predators: an experiment in a lowland rainforest of
New Guinea. Thesis. University of Florida, Gainesville,
Florida, USA.
Moermond, T. C., and J. S . Denslow. 1985. Neotropical avian
frugivores: patterns of behavior, morphology, and nutrition,
with consequences for fruit selection. Ornithological
Monographs 36:865-897.
Molofsky, J., and C. K. Augspurger. 1992. The effect of leaf
litter on early seedling establishment in a tropical forest.
Ecology 73:68-77.
Mulkey, S . S., R. L. Chazdon, and A. P. Smith, editors. 1996.
Tropical forest plant ecophysiology. Chapman and Hall,
New York, New York, USA.
Murray, K. G. 1988. Avian seed dispersal of three neotropical
gap-dependent plants. Ecological Monographs 58: 17 1198.
Myster, R. W., and S. T. A. Pickett. 1993. Effects of litter,
distance, density and vegetation patch type on postdispersal
tree seed predation in old fields. Oikos 66:381-388.
Nadkarni, N. M., T. J. Matelson, and W. A. Haber. 1995.
Structural characteristics and floristic composition of a
Neotropical cloud forest, Monteverde, Costa Rica. Journal
of Tropical Ecology 11:48 1-495.
Nadkarni, N. M . , and N. T. Wheelwright, editors. 1999. Monteverde: the ecology and conservation of a tropical cloud
forest. Oxford University Press, Oxford, UK.
Nepstad, D. C., C. Uhl, C. A. Pereira, and J. M . C. da Silva.
1996. A comparative study of tree establishment in abandoned pasture and mature forest of eastern Amazonia. Oikos 76:25-39.
Nogales, M., F. M . Medina, and A. Valido. 1996. Indirect
seed dispersal by the feral cats Felis catus in island ecosystems (Canary Islands). Ecography 19:3-6.
Peres, C. A , , L. C. Schiesari, and C. L. Dias-Leme. 1997.
Vertebrate predation of Brazil nuts (Bertholletia excelsa,
Lecythidaceae), an agouti-dispersed Amazonian seed crop:
a test of the escape hypothesis. Journal of Tropical Ecology
13:69-79.
Portnoy, S., and M . F. Willson. 1993. Seed dispersal curves:
behavior of the tail of the distribution. Evolutionary Ecology 7:25-44.
Powell, G . V. N., and R. Bjork. 1995. Implications of intra-
350
DANIEL G. WENNY
tropical migration on reserve design: a case study using
Phnromachrus mocitztzo. Conservation Biology 9:354-362.
Price, M . V., and S . H. Jenkins. 1986. Rodents as seed consumers and dispersers. Pages 191-235 itz D. R. Murray,
editor. Seed dispersal. Academic Press, Sydney, Australia.
Pyke, D. A,, and J. N. Thompson. 1986. Statistical analysis
of survival and removal rate experiments. Ecology 67:240245.
Roberts, J. T., and E. R. Heithaus. 1986. Ants rearrange the
vertebrate-generated seed shadow of a neotropical fig tree.
Ecology 67: 1946-1 05 1.
Robinson, G. R., and S. N. Handel. 1993. Forest restoration
on a closed landfill: rapid addition of new species by bird
dispersal. Conservation Biology 7:271-278.
Santana, C. E., and B. G . Milligan. 1984. Behavior of toucanets, bellbirds, and quetzals feeding on lauraceous fruits.
Biotropica 16: 152-154.
SAS Institute. 1988. SASISTAT user's guide. SAS Institute,
Cary, North Carolina, USA.
SAS Institute. 1989. JMP user's guide, version two. SAS
Institute, Cary, North Carolina, USA.
Schupp, E. W. 1988. Factors affecting postdispersal seed survival in a tropical forest. Oecologia 76:525-530.
Schupp, E. W. 1995. Seed-seedling conflicts, habitat choice,
and patterns of plant recruitment. American Journal of Botany 82:399-409.
Schupp, E. W., and M. Fuentes. 1995. Spatial pattern of seed
dispersal and the unification of plant population ecology.
Ecoscience 2:267-275.
Shaw, R. G., and T. Mitchell-Olds. 1993. ANOVA for unbalanced data: an overview. Ecology 74:1638-1645.
Smythe, N. 1978. The natural history of the Central American
agouti (Dnsyprocta putzctata). Smithsonian Contributions
to Zoology 257: 1-52.
Snow, B. K. 1961. Notes on the behavior of three Cotingidae.
Auk 78:150-161.
Snow, B. K. 1970. A field study of the Bearded Bellbird in
Trinidad. Ibis 112:299-329.
Snow, B. K. 1973n. Notes on the behavior of the White
Bellbird. Auk 90:743-75 1 .
Snow, B. K. 1977. Territorial behavior and courtship of the
male Three-wattled Bellbird. Auk 94:623-645.
Snow, D. W. 1973b. Distribution, ecology and evolution of
the bellbirds (Procnias, Cotingidae). Bulletin of the British
Museum (Natural History) 25:369-391.
Snow, D. W. 198 1. Tropical frugivorous birds and their food
plants: a world survey. Biotropica 13:l-14.
Stiles, E. W. 1989. Fruits, seeds, and dispersal agents. Pages
87-122 it1 W. G. Abrahamson, editor. Plant-animal interactions. McGraw-Hill, New York, New York, USA.
Stiles, F. G . 1985. On the role of birds in the dynamics of
neotropical forests. Pages 49-59 in A. W. Diamond and T.
E. Lovejoy, editors. Conservation of tropical forest birds.
International Council for B?rd Preservation, Cambridge,
UK.
Stiles, F. G., and L. Rosselli. 1993. Consumption of fruits of
the Melastomataceae by birds: how diffuse is coevolution.
Vegetatio 107/108:57-73.
Stiles, F. G., and A. F. Skutch. 1989. A guide to the birds of
Costa Rica. Cornell University Press, Ithaca, New York,
USA.
Sun, C., T. C. Moermond, and T. J. Givnish. 1997. Nutritional
determinants of diet in three turacos in a tropical montane
forest. Auk 114:200-211.
Swaine, M. D., editor. 1996. Ecology of tropical forest seedlings. UNESCOIParthenon, Paris, France.
Swaine, M. D., and T. C. Whitmore. 1988. On the definition
of ecological species groups in tropical rain forests. Vegetatio 75:s 1-86.
Tanner, E . V. J. 1982. Species diversity and reproductive
Ecological Monographs
Vol. 70, No. 2
mechanisms in Jamaican trees. Biological Journal of the
Linnean Society 18:263-278.
Terborgh, J., E. Losos, M. P. Riley, and 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.
Thtry, M., and D. Larpin. 1993. Seed dispersal and vegetation dynamics at a cock-of-the-rock's lek in the tropical
forest of French Guiana. Journal of Tropical Ecology 9:
109-116.
Thomas, D. W., D. Cloutier, M. Provencher, and C. Houle.
1988. The shape of bird- and bat-generated seed shadows
around a tropical fruiting tree. Biotropica 20:347-348.
Trainer, J. M., and T. C. Will. 1984. Avian methods of feeding
on Bursern simnrubn (Burseraceae) fruits in Panama. Auk
101: 193-195.
Traveset, A,, and M. E Willson. 1997. Effect of birds and
bears on seed germination of fleshy-fruited plants in temperate rainforests of southeast Alaska. Oikos 80:89-95.
Trexler, J. C., and J. Travis. 1993. Nontraditional regression
analyses. Ecology 74: 1629-1637.
Vander Wall, S. B. 1990. Food hoarding in animals. University of Chicago Press, Chicago, Illinois, USA.
Vasquez, R. A. 1996. Patch utilization by three species of
Chilean rodents differing in body size and mode of locomotion. Ecology 77:2343-235 1.
Vieira, I. C. G., C. Uhl, and N. D. 1994. The role of the
shrub Cordin multisipcatn Cham. as a "succession facilitator" in an abandoned pasture, Paragominas, Amazonia.
Vegetatio 115:91-99.
Wenny, D. G . 1998. Seed dispersal and postdispersal seed
fate of four tree species in a neotropical cloud forest. Dissertation. University of Florida, Gainesville, Florida, USA.
Wenny, D. G . 1999. Two-stage dispersal of Cuarea glnbra
and C. kunthinnn (Meliaceae) in Monteverde, Costa Rica.
Journal of Tropical Ecology 15:481-496.
Wenny, D. G., and D. J. Levey. 1998. Directed seed dispersal
by bellbirds in a tropical cloud forest. Proceedings of the
National Academy of Sciences USA 95:6204-6207.
Wheelwright, N. T. 1983. Fruits and the ecology of Resplendent Quetzals. Auk 100:286-301.
Wheelwright, N. T. 1 9 8 5 ~ .Competition for dispersers, and
the timing of flowering and fruiting in a guild of tropical
trees. Oikos 44:465-477.
Wheelwright, N. T. 1985b. Fruit size, gape width, and the
diets of fruit-eating birds. Ecology 66:808-818.
Wheelwright, N. T. 1986. A seven-year study of individual
variation in fruit production in tropical bird-dispersed tree
species in the family Lauraceae. Pages 19-35 in A. Estrada
and T. H. Fleming, editors. Frugivores and seed dispersal.
Dr. W. Junk Publishers, Dordrecht, The Netherlands.
Wheelwright, N. T. 1988. Four constraints on coevolution
between fruit-eating birds and fruiting plants: a tropical
case study. Proceedings of the International Ornithological
Congress XIX:827-845.
Wheelwright, N. T. 1991. How long do fruit-eating birds stay
in the plants where they feed? Biotropica 23:29-40.
Wheelwright, N. T., W. A. Haber, K. G . Murray, and C. A.
Guindon. 1984. Tropical fruit-eating birds and their food
plants: a survey of a Costa Rican lower montane forest.
Biotropica 16: 173-192.
Wheelwright, N. T., and G . H. Orians. 1982. Seed dispersal
by animals: contrasts with pollen dispersal, problems of
terminology, and constraints on coevolution. American
Naturalist 119:402-413.
Whelan, C. J., M . L. Dilger, D. Robson, N. Hallyn, and S.
Dilger. 1994. Effects of olfactory cues on artificial-nest
experiments. Auk 111:945-952.
Willson, M . E
1992. The ecology of seed dispersal. Pages
61-85 in M. Fenner, editor. Seeds: the ecology of regen-
May 2000
CLOUD FOREST SEEDLING RECRUITMENT
eration in plant communities. CAB International, Wallingford, UK.
Willson, M. E 1993. Dispersal mode, seed shadows, and
colonization patterns. Vegetatio 107/108:261-280.
Young, B., and D. McDonald. 1999. Lessons from 25 years
35 1
of bird research at Monteverde. In N. M . Nadkarni and N.
T. Wheelwright, editors. Monteverde: the ecology and conservation of a tropical cloud forest. Oxford University
Press. Oxford, UK.