OIKOS 104: 362-376, 2004 
Phenology of neotropical pepper plants (Piperaceae) and their 
association with their main dispersers, two short-tailed fruit bats, 
Cavollia pevspidllata and C. castanea (Phyllostomidae) 
Wibke Thies and Elisabeth K. V. Kalko 
Thies, W. and Kalko, E. K. V. 2004. Phenology of neotropical pepper plants 
(Piperaceae) and their association with their main dispersers, two short-tailed fruit 
bats, CaroUia perspicillata and C. castanea (Phyllostomidae). - Oikos 104: 362-376. 
To relate differences in phenological strategies of a group of closely related plants to 
biotic (pollinators, dispersers) and abiotic (water, light) factors, we studied leafing, 
flowering, and fruiting phenology of 12 species of Piper (Piperaceae) in a neotropical 
lowland forest in Panama for 28 months. We asked how Piper may partition time 
and vertebrate frugivores to minimize possible competition for dispersal agents. 
Based on habitat preferences and physiological characteristics we discriminate be- 
tween forest Piper species (eight species) and gap Piper species (four species). Forest 
Piper species flowered synchronously mostly at the end of the dry season. Gap Piper 
species had broader or multiple flowering peaks distributed throughout the year with 
a trend towards the wet season. Both groups of Piper species showed continuous fruit 
production. Fruiting peaks of forest Piper species were short and staggered. Gap 
Piper species had extended fruiting seasons with multiple or broad peaks. Both 
groups of Piper species also differed in their time of ripening and disperser spectrum. 
Forest Piper species ripened in late afternoon and had a narrow spectrum consisting 
mainly of two species of frugivorous bats: CaroUia perspicillata and C. castanea 
(Phyllostomidae). Fruits of gap Piper species, in contrast, ripened early in the 
morning and were eaten by a broader range of diurnal and nocturnal visitors, 
including bats, birds, and ants. We conclude that the differences in flowering 
phenology of forest and gap Piper species are primarily caused by abiotic factors, 
particularly the availability of water and light, whereas differences in fruiting patterns 
are mostly influenced by biotic factors. The staggered fruiting pattern of forest Piper 
species may reflect competition for a limited spectrum of dispersers. The long and 
overlapping fruiting periods of gap Piper species are associated with a larger 
spectrum of dispersers and may be a strategy to overcome the difficulty of seed 
dispersal into spatially unpredictable germination sites with suitable light conditions. 
W. Thies, Univ. of T?bingen, Animal Physiology, Auf der Morgenstelle 28, DE-72076 
T?bingen, Germany. - E. K. V. Kalko, Univ. of Ulm, Experimental Ecology, Albert- 
Einstein Allee 11, DE-89069 Ulm, Germany (elisabeth.kalko@biologie.uni-ulm.de) and 
Smithsonian Tropical Res. Inst., P. O. Box 2072, Balboa, Panam?. 
The many species of pantropical pepper plants (Piper, logical strategy. Fruits are harvested by a limited group 
Piperaceae) are important food resources for frugivo- of dispersers, mainly bats, in some cases also birds and 
rous insects and vertebrates (Gilbert 1980, Fleming ants (Fleming 1981, 1985, O'Donnell 1989). Patterns of 
1985), especially during seasonal scarcity of other fruits. fruit production and disperser spectrum of Piper con- 
Crops of individual Piper plants are small with a few trast strongly with those of "big-bang" plants such as 
fruits ripening in predictable time intervals over a pe- Neotropical figs (Moraceae: Ficus). Figs produce huge 
riod of several weeks or months, a steady-state pheno- fruit crops and attract a wide range of diurnal and 
Accepted 19 June 2003 
Copyright ? OIKOS 2004 
ISSN 0030-1299 
362 OIKOS 104:2 (2004) 
nocturnal dispersers (Janzen 1979, Milton et al. 1982, 
Terborgh 1983, 1986, Charles-Dominique 1993, Kalko 
et al. 1996b, Korine et al. 2000). 
Neotropical Piper species occur in a variety of habi- 
tats (Croat 1969, Fleming 1985, Fleming and Maguire 
1988, Greig 1991, Laska 1997). Members of this genus 
are well-known as pioneer plants that colonize gaps, 
clearings, and forest edges and are thus involved in 
regeneration processes and maintenance of plant diver- 
sity. Some Piper species also grow in closed forest. 
Denslow et al. (1990) and Greig (1991) separated shade 
tolerant or late successional "forest" species from shade 
intolerant or early successional "gap" species. In spite 
of their importance for frugivores and their crucial role 
in regeneration processes, little is known about the 
flowering, fruiting, and leafing pattern of most Piper 
species in local plant communities (Fleming 1981, 1985, 
Marquis 1988, Marinho-Filho 1991) and even less 
about the relationship between their fruiting phenology 
and their main dispersers (Fleming 1981, 1985). The 
temporal pattern of fruiting of animal-dispersed plants 
is not only vital for the successful dispersal of fruits and 
seeds and the establishment of seedlings, but also 
strongly influences the reproductive activity (Bonac- 
corso 1979, Wheelwright 1983, Dinerstein 1986, Levey 
1988, Worthington 1990, Sosa and Soriano 1996) and 
seasonal movements (Janzen 1967, Snow 1971, Hei- 
thaus et al. 1975, Karr 1976, Thompson and Willson 
1979, Stiles 1980, Pyke 1983, Wheelwright 1983) of the 
dispersers that depend on them. 
We studied the leafing, flowering, and fruiting phe- 
nology of 12 common Piper species in a tropical low- 
land forest on Barro Colorado Island (BCI) in Panama 
over a 28-month period. On BCI, a total of 22 species 
of Piper co-occurs. They form a distinctive group of 
understory plants that are insect-pollinated. Two 
frugivorous leaf-nosed bats (Phyllostomidae: Carollia 
perspicillata and C. castanea) that are common on BCI 
and whose diet contains large portions of Piper, are the 
main dispersers of their seeds (Bonaccorso 1979, Kalko 
et al. 1996a). The main objective of our study was to 
determine whether and how flowering, fruiting, and 
leafing differ among Piper species on BCI and to which 
extent abiotic (i.e. climatic) and biotic factors (i.e. 
pollinators, dispersers) may contribute to this pattern. 
Furthermore, we examine the role of Piper in the diet 
of its main dispersers, short-tailed fruit-bats (Carollia 
sp.), and assess the degree of specialization in the 
interaction between Piper and the two species of bats. 
79?5rW), the field station of the Smithsonian Tropical 
Research Institute in Panama. The island (15.6 km^) is 
covered with tropical moist, semi-deciduous forest of 
several successional stages (Foster and Brokaw 1990). 
To the northeast, about half of the island consists of 
secondary forest of about 100 to 200 years. Old forest 
of about 300 to 400 years covers most of the southwest- 
ern part of BCI and a small strip of old forest near the 
laboratory area. The laboratory area is cleared and is 
composed of early successional vegetation with pioneer 
plants, mainly Cecropia. Ochroma, Solarium, Piper, and 
Trema. 
Rainfall is seasonal, with a dry period between the 
end of December and the end of April and a wet period 
between May and most of December. Total annual 
rainfaU is 2650 mm (Windsor et al. 1990, Pat?n 1996). 
In the wet season, an average of 300 mm of rain falls 
per month, with a short, slightly drier period between 
August and September. 
Fruit availability and removal rate 
Fruit availability was determined for six Piper species 
(P. aequale, P. grande, P. culebranum., P. reticulaWm, P. 
dilatatum., and P. marginatum) by counting nightly 
production of ripe fruits. Removal rates of ripe fruits 
were determined for three Piper species (P. aequale, P. 
grande, and P. dilatatum.) by marking ripe fruits and 
counting the fruits remaining the next morning shortly 
after sunrise. 
Diet composition of C. perspicillata and C. 
castanea on BCI 
To determine the importance of Piper in the diet of 
their two main dispersers, we collected feces from 229 
bats (79 fecal samples of C. perspicillata and 150 fecal 
samples of C. castanea) concurrently with the pheno- 
logical study. Bats were caught in mist nets set over 
trails and streams at 15 localities on BCI. As some fecal 
samples of the bats contained more than one type of 
seed, we extracted a total of 271 seed samples from 229 
fecal samples (84 seed samples from C. perspicillata and 
187 seed samples from C. castanea). Seeds were iden- 
tified by comparing them with the seed collection of the 
BCI herbarium and with our identification key. Further 
details of capture data of the bats and of seasonal 
variation in diet composition will be given elsewhere. 
Material and Methods 
Study site and climate 
This study was conducted between February 1994 and 
December   1997  on  Barro  Colorado  Island  (9?09'N, 
Phenological patterns of Piper 
We tagged adult individuals (characterized by several 
stems or one main stem of more than 1 cm diameter at 
20 cm above ground) of 12 Piper species along small 
OIKOS 104:2 (2004) 363 
trails in a strip of old forest near the laboratory clearing 
and in the adjacent secondary forest, covering a total 
area of about 24 ha. Except for P. reticulatum., which is 
a small tree 3 to 7 m in height, all Piper species are 
shrubs with heights of 1 to 3 m. For each of the seven 
abundant species (P. aequale, P. carrilloanum, P. cordu- 
latum., P. dilatatum., P. grande, P. marginatum, and P. 
reticulatum), 15 individuals were tagged; for the five less 
common species (P. arboreum, P. culebranum., P. 
darienense, P. perlasense, and P. hispidum), we tagged 
as many adults as could be found along the trails (6 to 
11 individuals). This resulted in a total of 143 marked 
individuals and an average of 12 individuals per species 
(Table 1). Individuals that were lost due to falling 
branches or trees were replaced by nearby individuals 
of the same species. Tebbs (1990) regards P. perlasense 
as synonymous with P. aequale. Due to clear differ- 
ences in morphology and phenology on BCI (Croat 
1978), we treat both as separate species. 
Table 1. Spearman rank correlation coefficients (rj for the 
consistency of flowering and fruiting periods of gap and forest 
Piper species on BCI between the years 1994/95 (n = 7 
months), 1994/96 (n = 5) and 1995/96 (n = 8). p<0.05 (*); 
p<0.02 (**); p<0.01 (***). Numbers in parentheses indicate 
the numbers of individual plants monitored per species. 
Species Year Flowering Fruiting 
Forest Piper species 
P. aequale (15) 94/95 0.64 0.79 * 
94/96 0.88 1.00 * 
95/96 0.70 0.88 ** 
P. arboreum (10) 0.69 0.71 
0.90 0.38 
0.60 0.42 
P. carrilloanum (15) 0.54 1.00 *** 
-0.03 1.00 * 
0.61 0.81 * 
P. cordulatum (15) 0.16 0.91 ** 
0.13 0.68 
0.93 *** 0.51 
P. darienense (11) 0.88 * 0.81 * 
0.85 0.53 
1.00 *** 0.78 * 
P. grande (15) 0.91 ** 0.78 
0.78 0.78 
0.69 1.00 *** 
P.perlasense (6) 0.35 1.00 *** 
0.6 1.00 * 
0.83 * 1.00 *** 
P. culebranum (8) 95/96 0.83 * 0.62 
Gap Piper species 
P. dilatatum (15) 94/95 0.61 0.71 
94/96 0.70 0.40 
95'96 0.98 *** 0.78 * 
P. hispidum (8) -0.62 0.38 
-0.03 0.25 
0.44 0.91 *** 
P. marginatum (15) 0.91 ** 0.79 * 
0.98 0.78 
0.73 0.96 *** 
P. reticulatum (15) 0.04 0.14 
0.13 0.08 
0.33 0.59 
We recorded the phenological data at bi-weekly in- 
tervals. During the 28-month study, February to April 
and June were monitored three times, the rest of the 
months were monitored twice. Piper culebranum was 
added in 1995. Numbers of immature and flowering 
inflorescences as well as immature and mature in- 
fructescences (hereafter called fruits) were classified into 
predetermined categories. To obtain an estimate of 
medium crop size of Piper species, the fruits of 150 
randomly chosen adults of P. aequale were counted 
shortly before ripening. Piper aequale is one of the most 
common Piper species on BCI and its average fruit 
number (11.7 + 14.6 fruits/plant) is within the range of 
the other monitored species. Based on this result, we 
classified the number of immature and flowering infl- 
orescences as well as immature and mature fruits into 
the categories "none" (no inflorescences, no fruits), 
"few" (1-10), and "many" ( > 10). For new leaves, we 
distinguished only between "present" and "absent". We 
choose these criteria because Piper in general has only 
few leaves and flushes them at very low rates. For data 
analysis the categories "none", "few", and "many" for 
flowers and fruits were transformed into the numbers 0, 
1 and 2, respectively. The categories "present" and 
"absent" for the new leaves were transformed into 0 
and 1. These values (phenological scores) were used for 
statistical analyses to test for consistency in flowering 
and fruiting patterns and differences in leafing, flower- 
ing and fruiting activities between the different Piper 
species. To obtain a monthly phenological pattern of 
individuals, the bi-weekly scores of an individual were 
averaged. To assess the monthly phenological pattern 
of a species, the monthly scores of all individuals of a 
species were averaged. To obtain the overall monthly 
phenological pattern for all Piper species, the mean 
monthly scores of all species were averaged. 
Statistical analysis 
Consistency in flowering and fruiting patterns (mea- 
sured as described above) between years was tested by 
the Spearman rank correlation test (Sokal and Rohlf 
1995). We applied Wilcoxon's signed-ranked test to 
determine if there were significant differences in flower- 
ing/fruiting activity of the 12 monitored species be- 
tween the first half (January to June; dry to early wet 
season) and the second half of the year (July to Decem- 
ber; wet season). To assess significant differences in the 
amount of yearly flower, fruit, and leaf production 
between forest and gap Piper species, we used the 
two-tailed Mann-Whitney U-test. We chose p < 0.05 as 
level of significance for all tests. 
To determine whether the level of flowering or fruit- 
ing overlap observed among Piper species differed sig- 
nificantly from expectations obtained by chance, we 
used a null model described in Pleasants (1980, 1990) 
364 OIKOS 104:2 (2004) 
and Wright and Calder?n (1995). Here, an overlap 
index (Pianka 1974) based on monthly values is calcu- 
lated for each pairwise combination of Piper species 
(Pleasants 1980, 1990). The mean overlap of these 
combinations was compared with 99 values of a ran- 
domization test as described in Wright and Calder?n 
(1995). The simulations randomized mean flowering 
and fruiting times while preserving the form of the 
observed distribution of flowering and fruiting around 
the year. Here, the year is not represented on a linear 
time scale, but on a circle with a rotation of 360? and 
an arbitrary origin. Mean flowering and fruiting times 
were determined as described in Wright and Calder?n 
(1995). The length of the flowering and fruiting season 
in our analyses encompasses all 12 months of the year 
because flowering and fruiting could be observed year- 
round in Piper. 
Flowering or fruiting times were regarded as "stag- 
gered" (temporally uniform) whenever the observed 
flowering or fruiting overlap value was smaller than 
that of at least 95 of the 99 simulations and as "aggre- 
gated" whenever the observed overlap value was 
greater than that of at least 95 simulations. 
J  FMAMJ  JASOND 
Fig. 1. Leafing, flowering, and fruiting activity of all Piper 
species studied on BCI (n = 12). Flowering and fruiting activi- 
ties were categorized as "none", "few", or "many" flowers or 
fruits present. Categories were transformed into phenological 
scores with numbers between 0 and 2. Leafing activity was 
categorized as "presence" or "absence" of new leaves and 
transformed into the numbers 0 and 1. The phenological 
scores from the same months of the 28-month study period 
(1994-1996) were averaged taking all Piper species together. 
Averaged monthly phenological scores of flowering and fruit- 
ing are depicted on the left y-axis, and the averaged monthly 
phenological scores of leafing on the right y-axis. 
between the two halves of the year (n : 
p = 0.9375). 
12, Ts = 37.5, 
Results 
Overall leaf production, flowering and fruiting 
activity of Piper 
On our scale from 0 to 1, leaf production of all Piper 
species was intermediate (mean value of 0.63) and 
almost constant throughout the year with a slight drop 
in April and an increase at the onset of the wet season 
in May (Fig. 1). There were two smaller peaks in the 
wet season, the first one in June and the second in 
September. However, there was no significant difference 
in leafing activity between the first and second half of 
the year (Wilcoxon's signed ranks test: n = 12, Ts = 22, 
p = 0.2094). 
Like leaf production, flowering was continuous 
throughout the year. On a scale from 0 to 2, overall 
activity was low with a mean value of 0.43 (Fig. 1). 
Annual flowering activity increased during the dry sea- 
son and peaked in April, the last month of the dry 
season, when leaf production was lowest. However, 
differences in flowering activity between the first, 
mostly dry and the second, wet half of the year were 
not significant (Wilcoxon's signed-ranks test n=ll, 
Ts= 17, p = 0.1823). 
The fruiting phenology of all Piper species together 
did not show a seasonal pattern on BCI (Fig. 1). On a 
scale from 0 to 2, ripe fruits were present at a very low 
level (mean value of 0.15) during the entire year with a 
slight increase at the end of the wet season in Novem- 
ber and December. Wilcoxon's signed-ranks test did 
not show a significant difference in fruit production 
Phenology of forest and gap Piper species 
Based on the major habitat types where the 12 moni- 
tored species of Piper occurred, we discriminated be- 
tween forest and gap Piper species. Forest Piper species 
are only found within the forest with low light condi- 
tions typical for the understory. Daws et al. (2002) 
measured in the understory on BCI 0.89% photosyn- 
thetically active radiation relative to full sunlight. On 
BCI, the group of forest Piper species includes P. 
aequale, P. arboreum, P. carrilloanum, P. cordulatum, P. 
culebranum, P. darienense, P. grande, and P. perlasense. 
With the exception of P. cordulatum, all other forest 
Piper species prefer steep, moist areas and thus are 
found mainly in ravines. Piper cordulatum occurs under 
more closed canopy in drier and flatter habitats. 
The group of gap Piper includes P. dilatatum, P. 
hispidum, P. marginatum, and P. reticulatum. With the 
exception of P. reticulatum, all gap Piper species are 
found in cleared, open areas (clearing sizes over 50 m^) 
with light levels typical for gap conditions. Daws et al. 
(2002) measured between 5.7% and 34.0% photosyn- 
thetically active radiation relative to full sunlight for 
gap sizes between 67 m^ and 320 m^. Piper reticulatum, 
which is a small tree 3 to 7 m in height, is intermediate 
in its habitat requirements. It grows within the forest, 
in forest gaps, and at forest edges. However, during the 
study period P. reticulatum produced inflorescences and 
fruits only in forest gaps and edges where the crowns of 
the trees were exposed to direct sunlight. We never 
observed flowering or fruiting trees within the forest. 
OIKOS 104:2 (2004) 365 
Leaf production, flowering, and fruiting activity of 
forest Piper species 
Leaf production in forest Piper species decreased dur- 
ing the dry season (Fig. 2a). It increased again to a high 
and even level during the wet season. Two smaller 
peaks of leaf production occurred in June and Septem- 
ber. There was no significant difference in leafing activ- 
ity between the two halves of the year (Wilcoxon's 
signed ranks test: n = 8, Ts = 10.5, p = 0.3270). 
Flowering of forest Piper species peaked strongly in 
the dry season between February and April (Fig. 2a). 
With the beginning of the wet season, the activity 
dropped sharply to almost none in June and July but 
increased slightly at the end of the wet season. Flower- 
ing activity was significantly higher in the first half of 
the year than in the second half (Wilcoxon's signed- 
ranks test; n = 8, Ts = 0, p = 0.0117). 
Flowering of individual species of forest Piper species 
peaked mostly in April and May (Fig. 2b). Piper carril- 
  p. aequale 
 P. arboreum 
?^;i  P. carilloanum 
 P. cordulatum 
 P. culebranum 
 P. grande 
?i  P. daiienense 
 P.periasense 
Fig. 2. (a) Leafing, flowering, and fruiting activity of eight 
forest Piper species on BCI averaged for the same months over 
the 28 month study period (1994-1996); (b) flowering activity 
and (c) fruiting activity of individual forest Piper species 
(n = 8) averaged for each forest Piper separately. 
loanum and P. cordulatum started ?owering in the 
middle of the wet season and reached a peak at the 
beginning of the year, leading to the slight increase in 
overall ?owering activity in forest Piper species at the 
end of the year (Fig. 2a). In the randomization test, the 
observed overlap value ( = 0.1331) was greater than the 
simulated values of 84 runs, indicating a random distri- 
bution of flowering times within forest Piper species 
with only a slight tendency towards aggregation. 
Generally, fruiting activity of forest Piper species was 
relatively constant at a low level throughout the year 
(Fig. 2a). Fruiting activity in the first half of the year 
was not significantly different from fruiting activity in 
the second half (Wilcoxon's signed-ranks test; n = 8, 
Ts = 13.5, p = 0.4838). There was a small increase in 
fruit production at the end of the dry and at the 
beginning of the wet season. Highest fruit production 
occurred in November and December, while it dropped 
slightly in February and March. 
In contrast to leafing and flowering, fruiting activity 
of individual species showed a staggered pattern (Fig. 
2c). Almost every month a different species peaked in 
fruit production, resulting in the observed continuous 
fruiting activity of this group throughout the year. 
Piper aequale and P. grande, two morphologically 
rather similar and very common species, had distinct 
and narrow peaks in November and December, when 
almost all individuals of a species fruited syn- 
chronously, leading to the increase in fruit production 
in the overall activity of this group at the end of the 
year. 
In the randomization test, the observed overlap in 
fruiting times ( = 0.6563) was smaller than all 99 simu- 
lations, indicating a temporally uniform (staggered) 
distribution of the fruiting times of the eight species of 
forest Piper species. 
Ripening pattern, fruit availability, and removal 
rates of forest Piper species 
Fruits of forest Piper species started to mature in the 
late afternoon after 3:30 pm. A fruit that would be ripe 
in a given night could not be distinguished from imma- 
ture fruits the morning before maturation. Mature 
fruits were soft and produced a fruity smell (Thies et al. 
1998). Ripe fruits remained light green as unripe fruits 
and persisted for only one night. If a ripe fruit was not 
removed during this night, it fell off the next morning 
or was eaten by ants. At night, mostly bats and very 
rarely katydids (Tettigoniidae) or crickets (Gryllidae) 
were observed feeding on ripe fruits. The mean nightly 
availability of fruits was 0.2 + 0.1 ripe fruits per plant 
for P. aequale (n=1314 plants), 0.4 + 0.2 fruits per 
plant for P. grande (n = 322), and 0.4 + 0.2 fruits per 
plant for P. culebranum (n = 180). To conclude, the 
number of ripe fruits per night and plant was very low. 
366 OIKOS 104:2 (2004) 
but remained relatively constant from night to night 
throughout the fruiting period. 
Removal rates of marked ripe fruits were observed 
on 50 plants of P. aequale once a week over seven 
weeks in 1996, and over 9 weeks in 1997 as well as on 
26 plants of P. grande over 8 weeks in 1997. Bats 
removed entire ripe fruits in flight (Thies et al. 1998; 
pers. field obs.). The proportion of ripe P. aequale fruits 
taken at night was 88.6 + 9.0% in 1996, and 89.0 + 
18.2% in 1997. Nightly removal rate was equally high 
for P. grande (89.1 + 16.1%). 
Leaf production, flowering, and fruiting activity of 
gap Piper species 
Similarly to forest Piper species, leaf production in gap 
Piper species did not show a marked annual pattern 
(Fig. 3a). A slight drop in leaf production occurred in 
J2 2,0 
2 1,8 
Y? 1,6 
% 1,4 
1,2 
n 1,0 
B 0,8 
8 0,6 
O 0,4 
(11 0,2 
CL 0,0 
(a) 
2,0 
1,8 
m 1,6 
^ 1.4 
w 1,2 
en 1,0 
o 0,8 
?? 0,6 
a. 0,4 
0,2 
(b1 0,0 
2,0 
1,8 
m 1,6 
^ 1,4 
i? 1,2 
.S 
O) 1,0 
o 0,8 
ta 0,6 
CL 0,4 
0,2 
0,0 
T 1 1 T" 
 P. dilatatum 
 P. hispidum 
  P. marginatum 
??? P. reticulatum 
(c) 
Fig. 3. (a) Leafing, flowering, and fruiting activity of four gap 
Piper species on BCI averaged for the same months over the 
28 month study period (1994-1996); (b) flowering activity and 
(c) fruiting activity of individual gap Piper species (n = 4) 
averaged for each forest Piper separately. 
the second half of the dry season but not as pro- 
nounced as in forest Piper species. During the wet 
season, leafing activity was continuously at a medium 
level with a small peak in June and September. There 
was no difference in relative leaf production between 
forest and gap Piper (Mann-Whitney U-test: U = 59.5, 
p = 0.4689). However, because both groups of Piper 
have only few leaves and flush them at very low rates, 
these results are also applicable to absolute leaf produc- 
tion of forest and gap Piper. 
The flowering activity of gap Piper species was high 
during the year, suggesting high synchrony within spe- 
cies. However, in contrast to forest Piper species, flow- 
ering activity in gap Piper species was lowest during the 
dry season and highest during the wet season (Fig. 3a). 
Forest and gap Piper species differed significantly in 
flowering activity over the year (Mann-Whitney U-test; 
U = 5, p = 0.0001) with gap Piper species producing 
more inflorescences than forest Piper species. 
With the exception of P. reticulatum, each species of 
gap Piper produced flowers throughout the year with 
several subsequent months of high flowering activity 
(Fig. 3b). Flowering peaks were distributed broadly 
throughout the wet season and with a broad overlap in 
flowering activity between species (Fig. 3b). 
In the randomization test, the observed overlap in 
flowering times ( = 0.7094) was smaller than the simu- 
lated overlap in 94 simulations indicating that the over- 
lap in flowering times was not significantly different 
from a random distribution. However, a certain trend 
towards uniformity was visible. This result has to be 
viewed with caution because gap Piper consisted of 
only four species to run the randomization test and 
because, by averaging, multiple peaks converge into 
one. 
Fruit production of gap Piper species was signifi- 
cantly higher than in forest Piper species (Mann- 
Whitney U-test: U = 2, p = 0.0001); it was fairly 
homogeneous during the year, reaching the lowest 
value in September (Fig. 3a). In contrast to forest Piper 
species, fruit production was highest at the beginning of 
the dry season and then decreased slowly. 
Individual gap Piper species did not show the alter- 
nating fruiting activity between species with single, 
rather narrow peaks as seen in forest Piper species, but 
had multiple or broad fruiting peaks that overlapped 
widely between species (Fig. 3c). Within species, only 
one to five months per year were without fruit produc- 
tion in contrast to forest Piper, where individual species 
did not fruit for 7-11 months in a year. Piper reticula- 
tum, the intermediate species between forest and gap 
Piper species, showed three fruiting peaks, two higher 
ones in June and October and a smaller one in March. 
Piper hispidium. had fruiting peaks in March, May and 
September. Piper dilatatum and P. marginatum each 
had one broad peak per year. Piper dilatum fruited 
most heavily between July and October with highest 
OIKOS 104:2 (2004) 367 
fruiting activity in August. Piper marginatum. had maxi- 
mum fruiting activity between November and March. 
In the randomization test, the observed overlap in 
fruiting times ( = 0.4840) was lower than 99 simula- 
tions, indicating a uniform distribution for the four gap 
Piper species throughout the year. Although a uniform 
distribution of fruiting times was also found in forest 
Pipers, the fruiting pattern that leads to the uniform 
distribution in both groups of Piper is strikingly differ- 
ent. In forest Piper, individual species had only one 
rather narrow main fruiting peak per year with little 
overlap between species whereas in gap Piper, individ- 
ual species had multiple or broad fruiting peaks with 
large overlap between species. Both patterns result in 
continuous fruit production throughout the year (Fig. 
2a and 3a). 
Ripening pattern, fruit availability, and removal 
rates of gap Piper species 
With the exception of P. reticulaWm, 95% of gap Piper 
fruits started to ripen in the morning at about 06:00 am 
while 5% of the fruits began to ripen after 2:00 pm. 
Fruits of P. reticulatum ripened between 14:00 and 4:00 
pm as in forest Piper species. Ripe fruits smelled spicy 
and persisted for up to two days on the branch before 
they fell off. After the first night, however, they became 
mushy with a rotten smell. Fruits that were not eaten 
during the first day or night of maturation were not 
taken on the second day. 
Availability of gap Piper fruits was about 4 to 10 
times higher than that of forest Piper species, with 
0.9 + 0.5 ripe fruits per plant and day/night for P. 
dilatatum (n = 133 plants), 0.4 +1.4 fruits for P. mar- 
ginatum (n = 346), and 8.5 + 11.4 fruits for P. reticula- 
tum (n = 64). 
Fruits were eaten by ants, birds and bats. During the 
day, we frequently observed ants stripping off fruit 
tissue and seeds from the axis of the infructescence 
(rachis) that then remained empty on the branch. Occa- 
sionally, flycatchers (Tyrannidae) and tanagers 
(Thraupinae) took entire fruits during the day. The 
majority of the fruits, however, was removed by bats at 
night. 
Removal rates of marked fruits were determined 
quantitatively in 1997 on 44 plants of P. dilatatum once 
a week over 5 weeks. During the day, ants took 26.3 + 
16.9% of the available ripe fruits and birds 3.8 + 7.4%. 
At night, bats took 48.5 ? 20.3%. The fruits that re- 
mained untouched until the next morning (21.4 + 6.1%) 
fell off during the day. In contrast to forest Piper 
species, the available number of ripe fruits de- 
clined steadily during each species' fruiting season. 
Parallel to this decline, the removal rate by bats also 
declined. 
Differences in phenological patterns between years 
Each species of Piper that we monitored on BCI had at 
least one flowering and one fruiting period per year. 
The overall flowering activity of all Piper species 
showed a peak in the late dry season. Compared to 
1995 and 1996, the flowering peak of 1994 occurred one 
month earlier, in March. Because of this early peak in 
1994, the flowering patterns of 1994/95 and 1994/96 
were less consistent (Spearman rank correlation: flower- 
ing in 1994/95, r3 = 0.39 p > 0.05 n = 7, n.s.; flowering 
in 1994/96, r^ = 0.6 p > 0.05, n = 5, n.s.) than the flow- 
ering periods of 1995/96 (Spearman rank correlation: 
r^ = 0.83; n = 8; p = 0.02). Individual species of Piper 
varied in their flowering peaks from year to year, 
usually by a month, thus the consistency in flowering 
activity between years was low for many species (Table 
Overall fruiting activity of Piper species did not show 
a significant consistency between years because of the 
lack of pronounced peaks (Spearman rank correlation: 
fruiting in 1994/95, r, = 0.10, p > 0.05, n = 7, n.s.; fruit- 
ing in 1994/96, r, = 0.58, p > 0.05, n = 5, n.s.; fruiting in 
1995/96, r, = 0.64, p > 0.05, n = 8, n.s.). However, indi- 
vidual species had fruiting peaks. For most species 
these peaks were highly predictable and significantly 
consistent between years, even when the flowering 
peaks differed by a month from year to year (Table 1). 
Consistency between years was stronger for forest Piper 
species than for gap Piper species. Piper reticulatum 
showed the least consistency, since some individuals 
fruited more often per year than others. 
Duration of flowering and fruiting periods 
To categorize the duration of flowering and fruiting 
periods, we defined four categories. "Short" flowering 
or fruiting extends from 2 weeks to 4 months, "multi- 
ple, short" flowering or fruiting shows several periods 
per year that last 2 to 4 weeks each, "extended" flower- 
ing or fruiting ranges from 5 to 10 months, and "con- 
tinuous" refers to flowering or fruiting encompassing 
11 months to a whole year. 
Based on flowering periods, we found two groups in 
forest Piper species: one group with short flowering 
periods {P. aequale, P. arboreum, P. culebranum., P. 
grande, P. darienense and P. perlasense) and a second 
group with extended flowering periods (P. carrilloanum 
and P. cordulatum. Table 2). Gap Piper species showed 
more types of flowering duration with a trend to longer 
flowering periods than forest Piper species. Piper 
hispidum and P. marginatum had continuous flowering 
periods, while the flowering period of P. dilatatum. fell 
into the category "extended" with 10 months of flower- 
ing. Piper reticulatum. differed from all other Piper 
species in having multiple, short flowering periods that 
lasted only 2 to 4 weeks (Table 2). 
368 OIKOS 104:2 (2004) 
Table 2. Duration of flowering and fruiting within species in 
forest and gap Piper species on BCI. Numbers in parentheses 
indicate the number of flowering/fruiting months. 
Species Flowering Fruiting 
Forest Piper species 
P. aequale short (3^) short (4) 
P. arboreum short (2-4) short (2) 
P. carrilloanum extended (6-8) short (3) 
P. cordulatum extended (9) short (4-5) 
P. culebranum short (1-2) short (1-2) 
P. darienense short (1-2) short (3^) 
P. grande short (4-5) short (4) 
P. perlasense short (2-4) short (1) 
Gap Piper species 
P. dilatatum continuous (11-12) continuous (11) 
P. hispidum continuous (12) continuous (11) 
P. marginatum continuous (12) extended (8) 
P. reticulatum multiple, short (0.5- multiple, short 
1) (0.5) 
All forest Piper species had short fruiting periods 
that lasted either 1 to 4 months or were very short (i.e., 
2 to 4 weeks. Table 2). In gap Piper species, fruiting 
periods generally extended from 8 to 11 months (Table 
2). As in its flowering period, P. reticulatum differed 
from all other Piper species by having multiple, short 
fruiting periods of 2 to 4 weeks. 
3). Typically, the fruits of these species reached their 
final size and shape including seed set within two 
months after flowering. The fruits remained in this 
stage until ripening occurred very quickly at the end of 
the maturation time. The other half of the forest Piper 
species were divided into species with short maturation 
times preceded by an extended flowering period {P. 
carrilloanum. and P. cordulatum) and species with short 
maturation times following directly after a short flower- 
ing period (P. arboreum and P. culebranum). To sum- 
marize, all forest Piper species, except P. darienense, 
flowered together at the beginning of the rainy season 
and bore ripe fruits in the same year but at different 
times. Piper darienense was the only species that pro- 
duced ripe fruits at the beginning of the following year. 
In general, gap Piper species had short maturation 
times. In P. dilatatum, P. hispidum and P. marginatum, 
flowers and ripe fruits occurred simultaneously on the 
same branches (Table 3). These species also produced a 
few immature fruits shortly after the fruiting period. 
These fruits stayed immature for 2-4 months and 
ripened in the following fruiting period. Piper reticula- 
tum also had short maturation periods that lasted be- 
tween 2 weeks and 2 months and occurred shortly after 
or simultaneously with a very short flowering period 
(Table 3). 
Duration of fruit maturation 
The period of fruit maturation encloses the time be- 
tween the end of flowering and the onset of the produc- 
tion of ripe fruits. We distinguish two fruit maturation 
periods: a "short" period of 1-4 months, and an 
"extended" period lasting 5 months and more. Half of 
the forest Piper species had extended maturation times 
preceded by relatively short flowering periods {P. ae- 
quale, P. perlasense, P. grande and P. darienense. Table 
Diet of C. pevspicillata and C. castanea 
Ripe fruits of Piper were the core diet of both Carollia 
species. Diet analyses based on fecal samples revealed 
that large numbers of Piper seeds were present in both 
Carollias' year-round, including seeds of all Piper spe- 
cies that are common on BCI. In C. perspicillata, more 
than half (58.3%) of the total seed samples (n = 84) 
contained Piper. In C. castanea. Piper seeds occurred in 
92% of the total seed samples (n=187). However, 
although Piper formed the principal diet of the two 
Table 3. Maturation times (time between end of flowering and occurrence of first ripe fruits) of forest and gap Piper species 
fruits on BCI. For definitions see text. 
Species Flowering peak    Fruiting peak Maturation time 
Forest Piper species 
P. aequale April November 
P. arboreum April May 
P. carrilloanum January, April August 
P. cordulatum January April 
P. culebranum May July 
P. darienense May January next year 
P. grande April December 
P. perlasense April October 
Gap Piper species 
P. dilatatum 
P. hispidum 
P. marginatum 
P. reticulatum 
extended 
short, after a short flowering period 
short, after an extended flowering period 
short, after an extended flowering period 
short, after a short flowering period 
extended 
extended 
extended 
short, simultaneously with flowering 
short, simultaneously with flowering 
short, simultaneously with flowering 
short, after a short flowering period or simultaneously 
OIKOS 104:2 (2004) 369 
Carollia species, they differed in the consumption of 
forest versus gap Piper species. Carollia perspicillata 
consumed more gap Piper species (71.4%; total number 
of fecal samples containing Piper = 49), while C. cas- 
tanea ate equal parts of forest (43.6%) and gap Piper 
species (56.4%; total number of fecal samples contain- 
ing Piper = 172). 
Discussion 
General patterns of Piper phenology on BCI in 
comparison to other neotropical forests 
Our 28-month study of the phenology of 12 Piper 
species on BCI showed strong seasonality in flowering 
with a peak at the transition time between dry and wet 
season in April and May. Leafing and fruiting was less 
seasonal. Leaf production was almost constant with 
two small peaks in June and September. Fruiting oc- 
curred at a low but constant level throughout the year 
with a slight increase at the end of the year. This 
pattern coincides largely with the phenological patterns 
of other understory plants on BCI (Croat 1975, 1978, 
Foster 1990, Leigh and Windsor 1990, Leigh 1999). 
Flowering 
Compared with the flowering patterns of understory 
plants of a wet forest in La Selva and a dry forest in 
Guanacaste, both in Costa Rica (Opler et al. 1980), the 
flowering activity of Piper on BCI resembles more that 
of the highly seasonal Guanacaste dry forest that is 
influenced by the Pacific Ocean climate. While the 
flowering activity of understory plants in forests with 
relatively constant rainfall throughout the year such as 
the wet forest in La Selva shows a tendency towards 
aseasonality (see also Hilty 1980, Newstrom and 
Frankie 1994), flowering activity in forests with 
marked seasonality such as the dry forest in Gua- 
nacaste mostly peaks at the end of the dry season. It 
has been suggested that the onset of flowering is trig- 
gered by the first rainfalls at the end of the dry season 
(Janzen 1967, Opler et al. 1976, 1980, Borchert 1983). 
However, higher light levels during the dry season are 
also likely to affect time of flowering (van Schaik et al. 
1993, Wright and van Schaik 1994, Wright 1996). 
Fruiting 
In contrast to flowering, fruiting of understory plants 
does not show marked seasonal patterns, whether in 
wet or in dry forests (Opler et al. 1980, Foster 1990). 
As differences in rainfall and light conditions in the 
forest do not appear to strongly affect fruiting pat- 
terns, we conclude that the observed continuity of 
fruiting events throughout a year in Piper and other 
understory plants can be seen as a response of the 
plants  to   competition  for  the  same   seed  dispersers 
(Snow 1965, Janzen 1967, Hilty 1980, Rathcke and 
Lacey 1985, Foster 1990). 
Fruiting activity in Piper is constant throughout the 
year and increases slightly in November and Decem- 
ber. This increase is characteristic for understory plants 
on BCI and coincides with the period of fruit scarcity 
of canopy trees (Croat 1978, Foster 1990). Because of 
this continuity, Piper represents a reliable food source 
for frugivores year-round. This is well-documented in 
its continuous presences in the fecal samples of its 
main dispersers, the short-tailed fruit bats C. perspicil- 
lata and C. castanea. However, Carollia, in particular 
C. perspicillata, supplement their Piper diet with fruits 
from other understory plants as well as fruits from 
canopy trees (Bonaccorso 1979, unpubl.). Since fruit 
production of the latter decreases at the end of the 
year on BCI (Foster 1990), Piper becomes even more 
important for Carollia during this period. 
Productivity and the onset of fruiting periods in 
Piper are very consistent from year to year in contrast 
to the overall fruit-production on BCI that shows large 
seasonal fluctuations, in particular in association with 
the El Ni?o Southern Oscillations (Wright et al. 1999). 
They do not seem to be strongly influenced by climatic 
changes such as prolonged rainy seasons or displaced 
dry seasons. This results in a stable resource base not 
only throughout one year but also over several years 
and could be one of the underlying reasons for the 
observed stability in the long-term population trends of 
both Carollia on BCI (Kalko et al. 1996a). 
Leafing 
Leaf production of pioneer plants in light gaps on BCI 
is strongly seasonal with most leaves flushing in the 
wet season (Coley 1990). Leafing in Piper resembles 
that of canopy trees on BCI with a major peak in May 
or June and a minor one in September (Leigh and 
Windsor 1990). The pronounced peak in leaf produc- 
tion of canopy trees on BCI in May/June is interpreted 
as a response to high herbivory pressure (Lieberman 
and Lieberman 1984, Aide 1988, 1993, Reich 1995, 
Coley and Barone 1996, Coley and Kursar 1996) where 
synchronized leafing may satiate herbivores. 
In contrast to canopy trees, leaf production of Piper 
is more even throughout the year and the peak in June 
is less pronounced. Species with less seasonal leaf flush 
appear to suffer less damage by herbivores which is 
corroborated by observations on young and old leaves 
oi Piper (Aide 1993). This is presumably due to chemi- 
cal defenses (Aide 1993) as extracts of plant material of 
some Piper species revealed antifertility effects on in- 
sect herbivores and insecticidal properties (Schultes and 
Raffauf 1990). The simultaneous decline in leaf pro- 
duction at the peak of flowering may facilitate resource 
and energy allocation necessary for the strong and 
short burst of flowers. 
370 OIKOS 104:2 (2004) 
Phenology of forest and gap Piper species 
As our study shows, forest and gap Piper species differ 
also in phenological patterns of flowering and fruiting. 
We ask what roles abiotic factors such as rainfall or 
light environment and biotic factors such as the com- 
munity of pollinators and dispersers might play for the 
timing of these reproductive activities. 
Flowering 
There was a strong tendency towards synchronous 
peaks of flowering at the end of the dry season in forest 
Piper species, while gap Piper species had broader or 
multiple peaks and flowered mostly in the wet season. 
In forests with a distinct dry and wet season, the highly 
synchronized and pronounced flowering peaks of many 
plants in the drier months of the year may reflect the 
higher abundance of insect pollinators (Wolda 1978, 
Foster 1990, Foster and Brokaw 1990, Smythe 1990, 
but see Borchert 1983). Synchronized flowering then 
may also enhance pollination success. As Piper is polli- 
nated by small insects such as Tr?gona bees, small 
beetles, and syrphid flies (Semple 1974, Fleming 1985, 
pers. obs.), the distinct flowering peak of Piper species 
on BCI can be interpreted as a response to pollinator 
availability. However, whereas pollinator availability 
during the dry season can explain the narrow dry 
season peak of flowering in forest Piper species, it is not 
sufficient to explain the flowering activity of gap Piper 
species that flowered mostly in the wet season. 
It has been predicted for plants of dry forests (Reich 
and Borchert 1984) and for plants of seasonal forest 
with limited access to dry season water (i.e. gap plants) 
(van Schaik et al. 1993, Wright and van Schaik 1994, 
Wright 1996), that water stress during the dry season 
should promote flower production during the wet sea- 
son. Plants, however, that have access to sufficient 
water during dry season should flower during peak 
irradiance when it is energetically more efficient to 
transfer assimilates produced during the peak of irradi- 
ance (drier months of the year) directly to growing 
organs. Therefore, many plants produce flowers in the 
dry season instead of storing assimilates for later 
translocation. However, this is only true for plants that 
do not suffer severe water stress in the dry season. 
Opler et al. (1978) showed that plants in dry forests 
that grow close to rivers flower during the dry season 
and use the peak of irradiance for higher photosynthe- 
sis rates whereas plants that grow farther away from 
water courses flower only at the onset of the rains. They 
avoid water stress but miss the peak of irradiance. 
These ideas are in accordance with our results for 
forest and gap Piper species. While forest Piper species 
on BCI occur mainly along streams and ravines and 
therefore have continuous access to water, they flower 
at highest light levels during the dry season. In contrast, 
gap Piper species in open habitats are presumably more 
water-limited (Greig 1991) and thus flower mostly dur- 
ing the wet season. 
We conclude that the observed difference in timing of 
flowering between forest and gap Piper species is re- 
lated primarily to abiotic constraints (water and light). 
The observed difference in the shape of the flowering 
peaks of forest and gap Piper species, however, could 
also be attributed to biotic factors. Whereas the highly 
synchronized and narrow flowering peaks of forest 
Piper species may reflect the higher availability of insect 
pollinators in the dry season, the broad and/or multiple 
flowering peaks of gap Piper species may be associated 
with a low but continuous abundance of insect pollina- 
tors throughout the wet season. 
Fruiting 
A different picture emerges from the fruiting patterns 
of forest and gap Piper species. Although all forest 
Piper species flowered in a narrow time window of two 
months, fruits matured asynchronously among species. 
Further, fruiting peaks were relatively short with a 
relatively uniform distribution throughout the year. 
This leads to a continuous and reliable fruit supply. On 
the other hand, fruiting periods of gap Piper species 
were extended and overlapped strongly. Snow (1965) 
proposed that displaced (staggered) fruiting seasons of 
plant species that share the same dispersers have 
evolved as a response to competition for these dis- 
persers. In spite of the presumed link between fruiting 
seasons and dispersers, little is known about the actual 
distribution of fruiting seasons, particularly for phylo- 
genetically related species of plants that share the same 
seed dispersers (Smythe 1970, McKey 1975, Hilty 1980, 
Cruz 1981, Wheelwright 1985, Smith-Ramirez et al. 
1998, Poulin et al. 1999), including bat-dispersed plants 
(Heithaus et al. 1975, Fleming 1985, Marinho-Filho 
1991). Poulin et al. (1999) studied the fruiting phenol- 
ogy of two common understory plants, Miconia and 
Psychotria, and found that these bird-dispersed shrubs 
benefit more from an aggregated fruiting pattern (Psy- 
chotria) when they have small crops produced mainly 
for nomadic frugivores, while segregated fruiting (Mi- 
conia) will be beneficial to plants with large crops taken 
mainly by sedentary frugivores. The few studies on 
bat-dispersed plants suggest that there is relatively low 
overlap between fruiting periods. Up to now, however, 
the hypothesis that fruiting seasons are staggered be- 
cause of competition for the same dispersers has not 
been verified as studies have either focused on only few 
and closely related plant species with shared dispersers 
or on many, mostly unrelated species of plants with 
multiple dispersers (Gleeson 1981). 
Our study tried to overcome these problems. Eight 
common to abundant forest Piper species on BCI are 
mainly dispersed by two closely related short-tailed 
fruit bats, C. castanea and C. perspicillata that are 
abundant on the island (Handley et al. 1991, Kalko et 
OIKOS 104:2 (2004) 371 
al. 1996a). Further, Piper fruits are the most important 
dietary item for both CarolUa. By displacing their fruit- 
ing peaks to different periods of the year, all forest 
Piper species can make use of the small spectrum of 
dispersers, thus avoiding overlap in fruit ripening with 
other congeners which could increase the risk of losing 
opportunities for dispersal because of disperser limita- 
tion. Thus, potential competition between Piper species 
for those dispersers can be reduced (Fleming 1981). It 
could be argued that the staggered fruiting periods are 
the result of species-specific differences in the time the 
fruits need to grow until they reach maturity. However, 
our results show that the fruits of half of the forest 
Piper with extended maturation times already reach 
their final size and shape shortly after flowering and 
remain in this stage for several more months until 
ripening occurs quickly within a few days. To our 
opinion, this further supports the hypothesis that stag- 
gered fruiting in forest Piper could be a response to 
disperser limitation. 
Another explanation could be that staggered fruiting 
in forest Piper species reduces competition for limited 
germination sites. This would imply temporally stag- 
gered germination in the different forest Piper species 
and scattered dispersal patterns by the bats. The latter 
holds true for both CarolUa as shown with radio-track- 
ing on BCI (unpubl.). Unfortunately, little is known 
about the germination patterns of Piper species. Gar- 
wood (1990) describes one main germination peak per 
year on BCI at the beginning of the rainy season. Here, 
moisture and gap formation due to increased numbers 
of tree falls are high (Brokaw 1990), thus providing 
good germination conditions. Therefore, if all forest 
Piper species germinate only during this peak, we 
would expect dormancy capabilities of seeds in those 
Piper species that fruit late in the wet season. 
Gap Piper species, on the other hand, have extended, 
overlapping fruiting seasons and are visited by bats, 
birds and ants (O'Donnell 1989, this study). Whereas 
bats and birds are known seed dispersers of Piper, ants 
are regarded mainly as seed predators (Perry and Flem- 
ing 1980) but might also act to a limited degree as seed 
dispersers of Piper (Dalling, pers. com.). It has been 
suggested (Howe 1977, Howe and Estabrook 1977, 
Howe and Smallwood 1982) that plants with specific 
requirements for seedling establishment, i.e. gaps whose 
formation is not predictable in space and time (Brokaw 
1986, Martinez-Ramos and Alvarez-Buylla 1986, Mur- 
ray 1986), should have longer fruiting seasons and use 
a broader spectrum of dispersers than plants whose 
habitat requirements are more predictable (e.g. forest 
understory). 
There are several advantages associated with a broad 
range of dispersers. Ants carry seeds only small dis- 
tances ranging from 1-12 meters (Heithaus 1986, 
Ohkawara and Higashi 1994, Ohkawara et al. 1996, 
Passos  and  Ferreira   1996,   Leal  and  Oliveira   1998). 
Thus, most seeds remain close to their parent plants. 
Frugivorous birds often also transport seeds only short 
distances since they tend to stay for a prolonged time in 
fruiting trees and frequently deposit seeds from a tree at 
the same site during fruit processing or defecation 
(Pratt and Stiles 1983, Stiles and White 1986, Murray 
1988, Thomas et al. 1988, Wheelwright 1991). However, 
it is unclear whether this behavior also applies for birds 
eating fruits from bushes (Murray 1988). In contrast to 
many frugivorous birds, frugivorous bats (Phyllostomi- 
dae) always carry fruits away from fruiting trees or 
bushes to a feeding roost (Fleming 1988, unpubl.). This 
behavior promotes dispersal of seeds away from the 
vicinity of the parent plant. The tiny seeds of Piper are 
either dropped during feeding at a temporary roost or 
they are swallowed and afterwards excreted undamaged 
with the feces (Fleming 1988, unpubl.). This way seeds 
can sometimes be transported over long distances. For- 
aging distances (and hence potential dispersal distances) 
of frugivorous neotropical bats vary between 50 m to 
several kilometers (Morrison 1980. Fleming 1988, 
Kalko et al. 1996b, unpubl.). 
Based on these observations, it appears that the 
dispersal of seeds of gap Piper species by birds and ants 
results mostly in short-range dispersal with a high 
probability that the seeds are deposited close to the 
parent plants. Although mortality of seeds and 
seedlings might be high closer to parent plants (Janzen 
1970, Augspurger 1984, Dalling et al. 1998, Harms et 
al. 2000, Wenny 2000), short-distance dispersal by ants 
and to some smaller extent also by birds could be 
advantageous for plants depending on gaps. The unpre- 
dictability of favorable germination sites for gap plants 
in space and time could favor seed dispersal systems 
where seeds that have a low probability of survival 
under closed canopy, are mostly dispersed locally 
within the patch (i.e. the gap) of the parent plant 
(Horvitz and Schemske 1986). 
Dispersal of gap Piper seeds by bats, on the other 
hand, may often lead to long-range dispersal and colo- 
nization of new sites. However, because nothing is 
known yet about the population dynamics of gap Piper 
species and the effects of seed pr?dation on distribu- 
tional patterns o? Piper seeds (Perry and Fleming 1980), 
our propositions about the effects of dispersal by differ- 
ent dispersers must remain tentative. 
Is the dispersal system Piper ? CarolUa specialized? 
As fecal analyses from several studies have revealed and 
our study confirms, both C. perspicillata and C. cas- 
tanea are the main dispersers of Piper in most parts of 
the neo tropics (Goodwin and Greenhall 1961, Fleming 
et al. 1972, Howell and Burch 1974, Bonaccorso 1979, 
de Foresta et al. 1984, Charles-Dominique 1986, 1991, 
Fleming 1988,  1991, Palmeirim et al.  1989, Marinho- 
372 OIKOS 104:2 (2004) 
Filho 1991, Ascorra and Wilson 1992, MuUer and dos 
Reis 1992, Gorchov et al. 1995). Fleming (1981) pro- 
posed that the dispersal system of Piper is specialized 
sensu model 1 proposed by Howe and Estabrook 
(1977). Specialized dispersal systems are characterized 
by plants that produce small numbers of nutrient rich 
fruits per day or night over a prolonged period and that 
are associated with a small number of reliable, special- 
ized dispersers (see also Howe 1993). Because our re- 
sults on BCI have shown clear differences in the 
phenologies and number of dispersal agents between 
gap and forest Piper species, this raises the question 
whether this proposition is valid for all Piper species. 
On BCI, forest Piper species meet all the conditions 
for a specialized system. Only about every third to fifth 
plant produces a ripe fruit per night. The time of 
ripening occurs in late afternoon which suggests that 
forest Piper species are highly specialized for bat disper- 
sal. Furthermore, our fecal samples demonstrate that 
only one species of bat, C. castanea, is the major seed 
disperser. Although the number of ripe fruits per night 
and plant is very low, the production of forest Piper 
fruits is predictable because the number of ripe fruits 
per night in a given area is almost constant throughout 
the fruiting period of a species. The fruits are probably 
a nutritionally rich food source with high protein con- 
tent (Dinerstein 1986, Fleming 1986, Herbst 1986, but 
see Bizerril and Raw 1997). Furthermore, each forest 
Piper species produces fruits at a predictable time of the 
year and fruit production is consistent from year to 
year. We conclude that these characteristics promote a 
specialization between forest Piper species and their 
dispersers, mainly C. castanea. 
By comparison with forest Piper species, gap Piper 
species on BCI do not fully meet the conditions for a 
specialized dispersal system. Gap Piper species produce 
two to ten times as many fruits per plant as forest Piper 
species. Fruit ripening starts in early morning and the 
seeds are dispersed by a wider range of dispersers: 
birds, bats, and probably also ants. The number of 
fruits available per plant from night to night is not as 
constant throughout the fruiting period as in forest 
Piper species, but shows a pronounced peak in the 
middle of the fruiting period with a subsequent decline 
thereafter. As indicated by removal rates, bats remove 
most of the fruits opportunistically at the peaks of fruit 
production. They remove proportionately more fruits 
when fruit numbers are high and less when fruit num- 
bers decline. Therefore, they are not very reliable visi- 
tors to gap Piper species. Carollia castanea probably 
eats gap Piper species only as an addition to other, 
more predictable forest Piper species. The diet of C. 
perspicillata includes, in addition to gap Pipers, large 
portions of other available fruits from shrubs and trees, 
most of them with large crops (unpubl.). Therefore, we 
propose that dispersal of gap Piper species resembles 
more a generalized system, corresponding to model 2 
proposed by Howe and Estabrook (1977). In general- 
ized systems, large crops of relatively nutritional poor 
fruits are produced over short periods, as in Miconia 
(Snow 1965) or Muntingia calabura (Fleming 1985) for 
a wide range of "opportunistic" frugivores (McKey 
1975, Howe and Estabrook 1977). However, more data 
are needed to substantiate the proposition for the two 
dispersal systems within Piper. For example, it is still 
unclear whether forest and gap Piper species differ in 
nutritional value, a prerequisite necessary for either a 
specialized or unspecialized system. 
Overall, it is necessary to test the general usefulness 
of this concept, i.e. the proposition of two dispersal 
systems (specialized versus generalized) for explaining 
the evolutionary and ecological relationships between 
plants and their dispersers. No studies are available so 
far that give strong support to the suggested patterns 
on a community level or offer convincing alternatives. 
Conclusions 
Our study on the phenology of Piper on BCI showed 
distinct differences in flowering and fruiting patterns 
between forest and gap Piper species. While the differ- 
ences in flowering appear to be influenced primarily by 
abiotic factors such as site-specific differences in water 
and light availability, the factors that influence fruiting 
patterns of Piper on BCI are more complex. We pro- 
pose that competition for a limited disperser spectrum 
favors the staggered fruiting pattern observed in forest 
Pipers. In contrast, the larger spectrum of possible seed 
dispersers as well as long and overlapping fruiting 
periods in gap Pipers appear to be associated with the 
low probability of seed dispersal into favorable germi- 
nation sites. Continuous availability of Piper fruits 
throughout the year permits dietary specialization of 
two bat species of the genus Carollia. We suggest that 
the relationship between fruit availability of Piper and 
the degree of dietary specialization of Carollia found in 
our study on BCI reflects a particularly tight associa- 
tion between plants and a small group of dispersers. 
Because of this tight association, we propose that the 
distribution and fruiting phenologies of sympatric Piper 
species influence the composition and abundance of 
coexisting Carollia. To substantiate these assumptions, 
studies comparing differences of Piper phenologies and 
dispersal modes at sites with different climatic condi- 
tions as well as different Piper and disperser assem- 
blages are crucial. Overall, it cannot be ruled out that 
phylogeny, in addition to ecological characteristics, 
might also influence the observed phonological differ- 
ences among Piper species as it has been demonstrated 
on the family level for the understory shrubs Piper and 
Miconia with regard to chemical plant defenses (i.e. 
phenolics,  Kochmer and Handel  1986,  Baldwin  and 
OIKOS 104:2 (2004) 373 
Schultz 1988). To test whether and how phylogenetic 
relationships may constrain the evolution of phenologi- 
cal patterns among Piper requires a sound phylogeny 
on the species level. As the systematics of the genus 
Piper are still in flux, the possible effects of phylogeny 
on phenology patterns cannot be fully addressed so far. 
Acknowledgements - We thank the Smithsonian Tropical Re- 
search Institute (STRI, Panama) for excellent logistics support 
and the many people who helped to accomplish this work, in 
particular Jim Dalling, Ted Fleming, Charles Handley, Egbert 
Leigh, Bob Stallard, Marco Tschapka and Njikoha Ebigbo for 
very valuable discussions and corrections on the manuscript, 
Joe S. Wright for the use of his program for null models in 
plant phenology and Bob Stallard for his help in applying it to 
our data, Peter Pilz for statistical advice, and Ingrid Kaipf for 
help with the drawings. This study was supported by a Ph.D. 
stipend from the Deutscher Akademischer Austauschdienst 
(DAAD) to Thies and by grants of the Deutsche Forschungs- 
gemeinschaft (DFG), Schwerpunktprogramm "Mechanismen 
der Aufrechterhaltung tropischer Diversit?t" to Schnitzler and 
Kalko. 
References 
Aide, T. M. 1988. Herbivory as a selective agent on the timing 
of leaf production in a tropical understory community. 
- Nature 336: 574-575. 
Aide, T. M. 1993. Patterns of leaf development and herbivory 
in a tropical understory community. - Ecology 74: 455- 
466. 
Ascorra, C. F. and Wilson, D. E. 1992. Bat frugivory and seed 
dispersal in the Amazon, Loreto, Peru. - Publ. Mus. Hist. 
Nat. UNMSM (A) 43: 1-6. 
Augspurger, C. 1984. Seedling survival of tropical tree species: 
interactions of dispersal distance, light-gaps, and patho- 
gens. - Ecology 65: 1705-1712. 
Baldwin, I. T. and Schultz, J. C. 1988. Phylogeny and the 
patterns of leaf phenolics in gap- and forest-adapted Piper 
and Miconia understory shrubs. - Oecologia 75: 105-109. 
Bizerril, M. X. and Raw, A. 1997. Feeding specialization of 
two species of bats and fruit quality of Piper arboreum in a 
Central Brazilian gallery forest. - Rev. Biol. Trop. 45: 
913-918. 
Bonaccorso, F. J. 1979. Foraging and reproductive ecology in 
a Panamanian bat community. - Bull. Florida State Mus. 
24: 359-408. 
Borchert, R. 1983. Phenology and control of flowering in 
tropical trees. - Biotropica 15: 81-89. 
Brokaw, N. V. L. 1986. Seed dispersal, gap colonization, and 
the case of Cecropia insignis. - In: Estrada, A. and Flem- 
ing, T. H. (eds). Frugivores and seed dispersal. Dr. W. 
Junk, Dordrecht, pp. 323-331. 
Brokaw, N. V. L. 1990. Ca?da de ?rboles: frequencia, cronolo- 
g?a y consequencias. - In: Leigh, E. G., J.r., Rand, A. S. 
and Windsor, D. M. (eds). Ecolog?a de un Bosque Tropi- 
cal. Ciclos estacionales y cambios a largo plazo. Smithso- 
nian Press, pp. 163-172. 
Charles-Dominique, P. 1986. Inter-relations between frug?vo- 
rous vertebrates and pioneer plants: Cecropia, birds and 
bats in French Guiana. - In: Estrada, A. and Fleming, T. 
H. (eds). Frugivores and seed dispersal. Dr. W. Junk, 
Dordrecht, pp. 119-135. 
Charles-Dominique, P. 1991. Feeding strategy and activity 
budget of the frugivorous bat Carollia perspicillata (Chi- 
roptera: Phyllostomidae) in French Guiana. - J. Trop. 
Ecol. 7: 243-256. 
Charles-Dominique, P. 1993. Speciation and coevolution: an 
interpretation of frugivory phenomena. - In: Fleming, T. 
H. and Estrada, A. (eds), Frugivory and seed dispersal. 
- Vegetatio 107/108. Kluwer, pp. 75-84. 
Coley, P. D. 1990. Tasas de herb?vorismo en diferentes ?rboles 
tropicales. - In: Leigh, E. G., J.r., Rand, A. S. and 
Windsor, D. M. (eds). Ecolog?a de un Bosque Tropical. 
Ciclos estacionales y cambios a largo plazo. Smithsonian 
Press, pp. 191-200. 
Coley, P. D. and Barone, J. A. 1996. Herbivory and plant 
defenses in tropical forests. - Ann. Rev. Ecol. Syst. 27: 
305-335. 
Coley, P. D. and Kursar, T. A. 1996. Anti-herbivore defenses 
of young tropical leaves: physiological constraints and 
ecological trade-offs. - In: Smith, A. P., Mulkey, S. S. and 
Chazdon, R. L. (eds). Tropical forest plant ecophysiology. 
Chapman & Hafl, pp. 305-336. 
Croat, T. B. 1969. Seasonal flowering behavior in central 
Panama. - Ann. Missouri Bot. Gard. 56: 295-307. 
Croat, T. B. 1975. Phenolog?cal behavior of habitat and 
habitat classes on Barro Colorado Island (Panama Canal 
Zone). - Biotropica 7: 270-277. 
Croat, T. B. 1978. Flora of Barro Colorado Island. - Stanford 
Univ. Press. 
Cruz, A. 1981. Bird activity and seed dispersal of a montane 
forest tree (Dunalia arborescens) in Jamaica. - Biotropica 
13 (Suppl.): 34-44. 
Dalling, J. W., Swaine, M. D. and Garwood, N. C. 1998. 
Dispersal patterns and seed bank dynamics of pioneer trees 
in moist tropical forest. - Ecology 79: 564-578. 
Daws, M. I., Burslem, D. F. R. P., Crabtree, L. M. et al. 2002. 
Differences in seed germination responses may promote 
coexistence of four sympatric Piper species. - Funct. Ecol. 
16: 258-267. 
Denslow, J. S., Schultz, J. C, Vitousek, P. M. et al. 1990. 
Growth responses of tropical shrubs to treefall gap envi- 
ronments. - Ecology 71: 165-179. 
de Foresta, H., Charles-Dominique, P., Erard, C. et al. 1984. 
Zoochorie et premiers stades de la regeneration naturelle 
apr?s coupe en for?t guyanaise. - Revue Ecologie (Terre e 
Vie) 39: 369-400. 
Dinerstein, E. 1986. Reproductive ecology offrait bats and the 
seasonality of fruit production in a Costa Rican cloud 
forest. - Biotropica 18: 307-318. 
Fleming, T. H. 1981. Fecundity, fruiting pattern, and seed 
dispersal in Piper amalago (Piperaceae), a bat-dispersed 
tropical shrub. - Oecologia 51: 42-46. 
Fleming, T. H. 1985. Coexistence of five sympatric Piper 
(Piperaceae) species in a tropical dry forest. - Ecology 66: 
688-700. 
Fleming, T. H. 1986. Opportunism versus specialization: the 
evolution of feeding strategies in frugivorous bats. - In: 
Estrada, A. and Fleming, T. H. (eds). Frugivores and seed 
dispersal. Dr. W. Junk, Dordrecht, pp. 105-118. 
Fleming, T. H. 1988. The short-tailed fruit bat. A study of 
plant-animal interactions. - Univ. of Chicago Press. 
Fleming, T. H. 1991. The relationship between body size, diet 
and habitat use in frugivorous bats, genus Carollia (Phyl- 
lostomidae). - J. Mammal. 72: 493-501. 
Fleming, T. H., Hooper, E. T. and Wilson, D. E. 1972. Three 
Central American bat communities, structure, reproductive 
cycles and movement patterns. - Ecology 53: 555-569. 
Fleming, T. H. and Maguire, J. 1988. Patterns of diversity, 
density, and distribution in a wet tropical forest Piper 
flora. - Bufl. Ecol. Soc. Am. 69: 133. 
Foster, R. B. 1990. Ciclo estacional de ca?da de frutos en la 
isla de Barro Colorado. - In: Leigh, E. G., J.r., Rand, A. 
S. and Windsor, D. M. (eds). Ecolog?a de un Bosque 
Tropical. Ciclos estacionales y cambios a largo plazo. 
Smithsonian Press, pp. 219-242. 
Foster, R. B. and Brokaw, N. V. L. 1990. Estructura e historia 
de la vegetaci?n de la isla de Barro Colorado. - In: Leigh, 
E. G., J.r., Rand, A. S. and Windsor, D. M. (eds). Ecolo- 
g?a de un Bosque Tropical. Ciclos estacionales y cambios a 
largo plazo. Smithsonian Press, pp. 113-128. 
374 OIKOS 104:2 (2004) 
Garwood, N.C. 1990. Ciclo estacional de germinaci?n de 
semillas en un bosque semicaducifolio tropical. - In: 
Leigh, E. G., J.r., Rand, A. S. and Windsor, D. M. (eds). 
Ecolog?a de un Bosque Tropical. Ciclos estacionales y 
cambios a largo plazo. Smithsonian Press, pp. 243-255. 
Gilbert, L. E. 1980. Food web organization and the conserva- 
tion of neotropical diversity. - In: So?le, M. E. and 
Wilcox, B. A. (eds). Conservation biology, an evolution- 
ary-ecological perspective. Sinauer, pp. 11-33. 
Gleeson, S. K. 1981. Character displacement in flowering 
phenologies. - Oecologia 51: 294-295. 
Goodwin, G. G. and Greenhall, A. M. 1961. A review of the 
bats of Trinidad and Tobago. - Bull. Am. Mus. Nat. Hist. 
122: 191-301. 
Gorchov, D. L., Cornejo, F., Ascorra, C. F. et al. 1995. 
Dietary overlap between frugivorous birds and bats in the 
Peruvian Amazon. - Oikos 74: 235-250. 
Greig, N. 1991. Ecology of co-occurring species of neotropical 
Piper (Piperaceae): distribution, reproductive ecology, and 
seed pr?dation. PhD thesis, Univ. of Texas, Austin. 
Handley, C. O., Wilson, D. E. and Gardner, A. L. 1991. 
Demography and natural history of the common fruit bat, 
Artibeus jamaicencis, on Barro Colorado Island, Panam?. 
- Smithsonian Contr. Zool. 511: 1-173. 
Harms, K. E., Wright, S. J., Calder?n, O. et al. 2000. Perva- 
sive density-dependent recruitment enhances seedling diver- 
sity in a tropical forest. - Nature 404: 493-495. 
Heithaus, E. R. 1986. Seed dispersal mutualism and the popu- 
lation density of Asarwn canadense, an ant-dispersed plant. 
- In: Estrada, A. and Fleming, T. H. (eds). Frugivores and 
seed dispersal. Dr. W. Junk, Dordrecht, pp. 199-210. 
Heithaus,   E.   R.,   Fleming,  T.   H.   and  Opler,  P.  A.   1975. 
Foraging patterns and resource utilization in seven species 
of bats in a seasonal tropical forest. - Ecology 56: 841- 
854. 
Herbst, L. H. 1986. The role of nitrogen from fruit pulp in the 
nutrition   of  the   frugivorous   bat   Carollia perspicillata. 
- Biotropica 18: 39-44. 
Hilty, S. L. 1980. Flowering and fruiting periodicity in a 
premontane rain forest in pacific Colombia. - Biotropica 
12: 292-306. 
Horvitz, C. C. and Schemske, D. W. 1986. Seed dispersal and 
environmental heterogeneity in a neotropical herb: a model 
of population and patch dynamics. - In: Estrada, A. and 
Fleming, T. H. (eds). Frugivores and seed dispersal. Dr. W. 
Junk, Dordrecht, pp. 169-186. 
Howe, H. F. 1977. Bird activity and seed dispersal of a 
tropical wet forest tree. - Ecology 58: 539-550. 
Howe, H. F. 1993. Specialized and generalized dispersal sys- 
tems. Where does "the paradigm" stand? - In: Fleming, T. 
H. and Estrada, A. (eds), Frugivory and seed dispersal. 
Vegetatio 107/108. Kluwer, pp. 3-13. 
Howe, H. F. and Estabrook, G. F. 1977. On intraspecific 
competition for avian dispersers in tropical trees. - Am. 
Nat. Ill: 817-832. 
Howe, H. F. and Smallwood, J. 1982. Ecology of seed disper- 
sal. - Annu. Rev. Ecol. Syst. 13: 201-228. 
Howell, D. J. and Burch, D. 1974. Food habits of some Costa 
Rican bats. - Rev. Biol. Trop. 21: 281-294. 
Janzen, D. H. 1967. Synchronization of sexual reproduction of 
trees within the dry season in Central America. - Evolu- 
tion 21: 620-637. 
Janzen, D. J. 1970. Herbivores and the number of tree species 
in tropical forests. - Am. Nat. 104: 501-528. 
Janzen, D. H. 1979. How to be a fig. - Annu. Rev. Ecol. Syst. 
10: 13-51. 
Kalko, E. K. V., Handley, C. O. and Handley, D. 1996a. 
Organization, diversity, and longterm dynamics of a 
neotropical bat community. - In: Cody, M. and Small- 
wood, J. (eds), Longterm studies in vertebrate communi- 
ties. Academic Press, pp. 503-553. 
Kalko, E. K. V., Herre, E. A. and Handley, C. O., J.r. 1996b. 
Relation of fig fruit characteristics to fruit-eating bats in 
the New and Old World tropics. - J. Biogeogr. 23: 565- 
576. 
Karr, J. R. 1976. Seasonality, resource availability and com- 
munity diversity in tropical bird communities. - Am. Nat. 
110: 973-994. 
Kochmer, J. P. and Handel, S. N. 1986. Constraints and 
competetion   in   the   evolution   of  flowering   phenology. 
- Ecol. Monogr. 56: 303-325. 
Korine, C, Kalko, E. K. V. and Herre, E. A. 2000. Fruit 
characteristics and factors affecting fruit removal in a 
Panamanian community of strangler figs. - Oecologia 123: 
560-568. 
Laska, M. S. 1997. Structure of understory shrub assemblages 
in adjacent secondary and old growth tropical wet forests, 
Costa Rica. - Biotropica 29: 29-37. 
Leal, I. R. and Oliveira, P. S. 1998. Interactions between 
fungus-growing ants (Attini), fruits and seeds in cerrado 
vegetation in southeast Brazil. - Biotropica 30: 170-178. 
Leigh, E. G. 1999. Tropical forest ecology. A view from Barro 
Colorado Island. - Oxford Univ. Press, pp. 1-245. 
Leigh, E. G. and Windsor, D. M. 1990. Producci?n del bosque 
y regulaci?n de consumidores primarios de la isla de Barro 
Colorado. - In: Leigh, E. G., Jr., Rand, A. S. and Wind- 
sor, D. M. (eds). Ecolog?a de un Bosque Tropical. Ciclos 
estacionales y cambios a largo plazo. Smithsonian Press, 
pp. 179-190. 
Levey, D. J. 1988. Tropical wet forest treefall gaps and distri- 
butions of understory birds and plants. - Ecology 69: 
1076-1089. 
Lieberman, D. and Lieberman, M. 1984. The causes and 
consequences of synchronous flushing in a dry tropical 
forest. - Biotropica 16: 193-201. 
Marinho-Filho, J. S. 1991. The coexistence of two frugivorous 
bat species and the phenology of their food plants in 
Brazil. - J. Trop. Ecol. 7: 59-67. 
Marquis, R. J. 1988. Phenological variation in the neotropical 
understory shrub Piper arieianum: causes and conse- 
quences. - Ecology 69: 1552-1565. 
Martinez-Ramos, M. and Alvarez-Buylla, E. 1986. Seed dis- 
persal, gap dynamics and tree recruitment: the case of 
Cecropia obtusifolia at Los Tuxtlas, Mexico. - In: Estrada, 
A. and Fleming, T. H. (eds), Frugivores and seed dispersal. 
Dr. W. Junk, Dordrecht, pp. 333-346. 
McKey, D. 1975. The ecology of coevolved seed dispersal 
systems. - In: Gilbert, L. E. and Raven, P. H. (eds), 
Coevolution of animals and plants. Univ. of Texas Press, 
pp. 159-191. 
Milton, K., Windsor, D. M., Morrison, D. W. et al. 1982. 
Fruiting  phenologies   of two   neotropical   Ficus   species. 
- Ecology 63: 752-762. 
Morrison, D. W. 1980. Foraging and day-roost dynamics of 
canopy fruit bats in Panama. - J. Mammal. 61: 20-29. 
Murray, K. G. 1986. Consequences of seed dispersal for 
gap-dependent plants: relationships between seed shadows, 
germination requirements, and forest dynamic processes. 
- In: Estrada, A. and Fleming, T. H. (eds). Frugivores and 
seed dispersal. Dr. W. Junk, Dordrecht, pp. 187-198. 
Murray, K. G. 1988. Avian seed dispersal of three neotropical 
gap-dependent plants. - Ecol. Monogr. 58: 271-298. 
M?ller, M. F. and dos Reis, N. R. 1992. Partic?o de recursos 
alimentares entre quatro especies de morcegos frug?voros 
(Chiroptera, Phyllostomidae). - Rev. Bras. Zool. 9: 345- 
355. 
Newstrom, L. E. and Frankie, G. W. 1994. A new classifica- 
tion for plant phenology based on flowering patterns in 
lowland tropical rain forest trees at La Selva, Costa Rica. 
- Biotropica 26: 141-159. 
O'Donnell, S. 1989. A comparison of fruit removal by bats 
and birds from Piper hispidum SW. (Piperaceae), a tropical 
second growth shrub. - Brenesia 31: 25-32. 
Ohkawara, K. and Higashi, S. 1994. Relative importance of 
ballistic and ant dispersal in two diplochorous Viola species 
(Violaceae). - Oecologia 100: 135-140. 
OIKOS 104:2 (2004) 375 
Ohkawara, K., Higashi, S. and Ohara, M. 1996. Effects of 
ants, ground beetles and the seed-fall patterns of myrmeco- 
chory    of   Erythronium   japonicum    Decne.    (Liliaceae). 
- Oecologia 106: 500-506. 
Opler, P. A., Frankie, G. W. and Baker, H. G. 1976. Rainfall 
as a factor in the release, timing, and synchronization of 
anthesis by tropical trees and shrubs. - J. Biogeogr. 3: 
231-236. 
Opler, P. A., Frankie, G. W. and Baker, H. G. 1980. Compar- 
ative phenological studies of treelet and shrub species in 
tropical wet and dry forests in the lowlands of Costa Rica. 
- J. Ecol. 68: 167-188. 
Palmeirim, J., Gorchov, D. L. and Stoleson, S. 1989. Trophic 
structure of a neotropical frugivore community: is there 
competition between birds and bats? - Oecologia 79: 403- 
411. 
Passes, L. and Ferreira, S. O. 1996. Ant dispersal of Crot?n 
priscus (Euphorbiaceae) seeds in a tropical semideciduous 
forest in southeastern Brazil. - Biotropica 28: 697-700. 
Paton, S. 1996. Meteorological and hydrological summary for 
Barro Colorado Island. - Smithsonian Tropical Res. Inst. 
Perry, A. E. and Fleming, T. H. 1980. Ant and rodent 
pr?dation on small, animal-dispersed seeds in a dry tropical 
forest. - Brenesia 17: 11-22. 
Pianka, E. R.  1974. Niche overlap and diffuse competition. 
- Proc. Nati Acad. Sei. 71: 2141-2145. 
Pleasants, J. M. 1980. Competition for bumblebee pollinators 
in rocky mountain plant communities. - Ecology 61: 
1446-1459. 
Pleasants, J. M. 1990. Null-model tests for competitive dis- 
placement: the fallacy of not focusing on the whole com- 
munity. - Ecology 71: 1078-1084. 
Poulin, B., Wright, S. J., Lefebvre, G. et al. 1999. Interspecific 
synchrony and asynchrony in the fruiting phonologies of 
congenric bird-dispersed plants in Panama. - J. Trop. 
Ecol. 15: 213-227. 
Pratt, T. K. and Stiles, E. W. 1983. How long fruit-eating 
birds stay in the plants where they feed: implications for 
seed dispersal. - Am. Nat. 122: 797-805. 
Pyke, G. H. 1983. Seasonal pattern of abundance of hon- 
eyeaters and their resources in heathland areas near Syd- 
ney. - Aust. J. Ecol. 8: 217-233. 
Rathcke, B. and Lacey, E. P. 1985. Phenological patterns of 
terrestrial plants. - Annu. Rev. Ecol. Syst. 16: 179-214. 
Reich, P. B. 1995. Phenology of tropical forests: patterns, 
causes, and consequences. - Can. J. Bot. 73: 164-174. 
Reich, P. B. and Borchert, R. 1984. Water stress and tree 
phenology in a tropical dry forest in the lowlands of Costa 
Rica. - J. Ecol. 72: 61-74. 
Schultes, R. E. and Raffauf, R. F. 1990. The healing forest. 
Medicinal and toxic plants of the northwest Amazonia. 
- Dioscorides Press, Portland, Oregon. 
Semple, K. S. 1974. Pollination in Piperaceae. - Ann. Mis- 
souri Bot. Gard. 61: 868-871. 
Smith-Ramirez, C, Armesto, J. J. and Figueroa, J. 1998. 
Flowering, fruiting and seed germination in Chilean rain 
forest Myrtaceae: ecological and phylogenetic constraints. 
- Plant Ecol. 136: 119-131. 
Smythe, N. 1970. Relationships between fruiting seasons and 
seed dispersal methods in a neotropical forest. - Am. Nat. 
104: 25-35. 
Smythe, N. 1990. Abundancia estacional de insectos nocturnos 
en un bosque neotropical. - In: Leigh, E. G., J.r., Rand, 
A. S. and Windsor, D. M. (eds). Ecolog?a de un Bosque 
Tropical. Ciclos estacionales y cambios a largo plazo. 
Smithsonian Press, pp. 393-402. 
Snow, D. W. 1965. A possible selective factor in the evolution 
of fruiting seasons in tropical forest. - Oikos 15: 274-281. 
Snow, D. W. 1971. Evolutionary aspects of fruit-eating by 
birds. - Ibis 113: 194-202. 
Sokal, R. R. and Rohlf, F. J. 1995. Biometry. The principals 
and  practice of statistics in biological  research,  3rd  ed. 
- WH Freeman and Company. 
Sosa, M. and Soriano, P. J. 1996. Resource availability, diet 
and reproduction in Glossophaga longirostris (Mammalia: 
Chiroptera) in an arid zone of the Venezuelan Andes. - J. 
Trop. Ecol. 12: 805-818. 
Stiles, F. G. 1980. Patterns of fruit presentation and seed 
dispersal in bird-disseminated woody plants in the eastern 
deciduous forest. - Am. Nat. 116: 668-670. 
Stiles, E. W. and White, D. W. 1986. Seed deposition patterns: 
influences of season, nutrients, and vegetation structure. 
- In: Estrada, A. and Fleming, T. H. (eds). Frugivores and 
seed dispersal. Dr. W. Junk, Dordrecht, pp. 45-54. 
Tebbs, M. C. 1990. Revision of Piper (Piperaceae) in the New 
World.  2.  The  taxonomy  of Piper section  Churumayu. 
- Bull. Br. Mus. Nat. Hist. 20: 193-236. 
Terborgh, J. 1983. Ecological relationships in the Manu pri- 
mate community. - In: Terborgh, J. (ed.). Five New World 
primates. A study in comparative ecology. Princeton Univ. 
Press, pp. 190-210. 
Terborgh, J. 1986. Keystone plant resources in the tropical 
forest. - In: Soule, M. E. (ed.). Conservation biology. 
Academic Press, pp. 330-344. 
Thies, W., Kalko, E. K. V. and Schnitzler, H.-U. 1998. The 
roles of echolocation and olfaction in two neotropical 
fruit-eating bats, Carollia perspicillata and C. castanea, 
feeding on Piper. - Behav. Ecol. Sociobiol. 42: 397- 
409. 
Thomas, D. W., Cloutier, D., Provencher, M. et al. 1988. The 
shape of bird- and bat-generated seed shadows around a 
tropical fruiting tree. - Biotropica 20: 347-348. 
Thompson, J. N. and Willson, M. F. 1979. Evolution of 
temperate fruit/bird  interactions:  phenological  strategies. 
- Evolution 33: 973-982. 
van Schaik, C. P., Terborgh, J. and Wright, S. J. 1993. The 
phenology of tropical forests: adaptive significance and 
consequences for primary consumers. - Annu. Rev. Ecol. 
Syst. 24: 353-377. 
Wenny, D. 2000. Seed dispersal of a high quality fruit 
by   specialized   frugivores:   high   quality   seed   dispersal? 
- Biotropica 32: 327-337. 
Wheelwright, N. T. 1983. Fruits and the ecology of resplen- 
dent Quetzals. - Auk 100: 286-301. 
Wheelwright, N. T. 1985. Competition for dispersers, and the 
timing of flowering and fruiting in a guild of tropical trees. 
- Oikos 44: 465-477. 
Wheelwright, N. T. 1991. How long do fruit-eating birds 
stay in the plants where they feed? - Biotropica 23: 29- 
40. 
Windsor, D. M., Rand, A. S. and Rand, W. M. 1990. Carac- 
ter?sticas de la precipitaci?n de la isla de Barro Colorado. 
- In: Leigh, E. G., J.r., Rand, A. S. and Windsor, D. M. 
(eds). Ecolog?a de un Bosque Tropical. Ciclos estacionales 
y cambios a largo plazo. Smithsonian Press, Washington, 
D.C, pp. 53-71. 
Wolda, H. 1978. Fluctuations in abundance of tropical insects. 
- Am. Nat. 112: 1017-1045. 
Worthington, A. H. 1990. Comportamiento de forrajeo de dos 
especies de saltarines en respuesta a la escasez estacional de 
frutos. - In: Leigh, E. G., J.r., Rand, A. S. and Windsor, 
D. M. (eds). Ecolog?a de un Bosque Tropical. Ciclos 
estacionales y cambios a largo plazo. Smithsonian Press, 
Washington, D.C, pp. 285-304. 
Wright, S. J. 1996. Phenological responses to seasonality in 
tropical forest plants. - In: Smith, A. P., Mulkey, S. S. and 
Chazdon, R. L. (eds). Tropical forest plant ecophysiology. 
Chapman & Hafl, pp. 440-460. 
Wright, S. J. and van Schaik, C. P. 1994. Light and the 
phenology of tropical trees. - Am. Nat. 143: 192-199. 
Wright, S. J. and Calder?n, O. 1995. Phylogenetic patterns 
among tropical flowering phenologies. - J. Ecol. 83: 937- 
948. 
Wright, S. J., Carrasco, C, Calder?n, O. et al. 1999. The El 
Ni?o Southern Oscillation, variable fruit production, and 
famine in a tropical forest. - Ecology 80: 1632-1647. 
376 OIKOS 104:2 (2004)