Second, floristic and life-form differences between
low and mid-elevation sites could drive pollination
system differences. For example, the frequencies of
Lauraceae (most of which are pollinated by small
diverse insects), Fagaceae (some wind pollinated,
notably Quercus spp.), and Ericaceae (largely bird
pollinated) are higher in montane forests, whereas
Arecaceae (often beetle pollinated) and Fabaceae
(often bee pollinated) decrease in importance in the
cloud forest relative to lowland wet forests. Simi-
larly, the incidence of epiphytes increases with ele-
vation, many of which are visited by hummingbirds
(Feinsinger et al. 1987; see Busby, "Hummingbird
Pollination," pp. 267-268).
Third, elevational shifts in the abundance of dif-
ferent insect groups and in the frequencies of floral
sexual systems (hermaphroditism, monoecy, and
dioecy) and breeding systems (pollen compatibility;
Sobrevilla and Arroyo 1982) could underlie shifts in
the frequency of pollination systems. Increases in the
frequency of dioecy, for example, might be accompa-
nied by increases in the pollinator types associated
with the small open flowers common in neotropical
dioecious trees, or vice versa.
Disturbance and pollination interactions. Natural dis-
turbances occur at a variety of scales; communities
are mosaics of patches that differ in age, structure, and
species composition. How a forest's natural vegeta-
tion heterogeneity affects other components of com-
munity structure is poorly understood, particularly
with respect to interactions among animals, or among
animals and plants. Research in Monteverde directly
addressed these questions (Feinsinger et al. 1987,
1988a, b, Linhart et al. 1987a, b). Hummingbird-plant
pollination interactions were studied in three "patch"
types that reflect the natural scales of disturbance:
forest, treefall gaps, and larger gaps cut to mimic natu-
ral small landslides (Feinsinger et al. 1987).
Hummingbird-pollinated flowers were visited and
pollinated similarly in the three patch types. Although
larger gaps had higher densities of a few plant spe-
cies, and more visits by flower-piercing birds, no evi-
dence for significant disturbance-related variation in
bird-plant interactions was found. Pollinators in dis-
turbances were nearly as predictable, carried similar
pollen loads, visited flowers of a given species at simi-
lar frequencies (with more variability in gaps), and
deposited similar pollen loads as pollinators in the
forest. The same patches were used to examine pat-
terns of nectar availability and consumption of nec-
tar by birds (Feinsinger et al. 1988b). They docu-
mented the availability and diversity of nectar, birds'
use of and demand for nectar, and diet breadths and
diet overlap. Some characteristics differed among the
patches but in different ways. Overall, responses to
the natural disturbance mosaic were subtle, varied,
and complex.
These studies of bird-plant interactions suggest a
strong distinction between the effects of natural,
small-scale disturbance and anthropogenic, large-
scale disturbance on interactions, characteristics of
populations, and community assemblages. Natural
disturbance caused neither the large fluctuations in
resource availability nor the shifts in species compo-
sition reported for large anthropogenic disturbances
in Monteverde (Feinsinger 1976, 1978) or Trinidad
and Tobago (Feinsinger et al. 1985). The studies also
suggest caution in assuming that large-scale anthro-
pogenic disturbances simply reflect natural distur-
bances at a larger scale.8.1.4. Pollinator Sharing: The Plant
Guild Perspective
Most plant species share pollinators. They may in-
teract neutrally, compete, or facilitate one another's
pollination; the outcome varies with local ecologi-
cal conditions (Rathcke 1983). Experimental inves-
tigations of potential ecological and evolutionary
outcomes of pollinator sharing have been based in
Monteverde.Neighborhood effects in bird-pollinated plants. Build-
ing on detailed information about hummingbirds'
interactions and guild structure (Feinsinger 1976,
1978), researchers led by Feinsinger examined plant-
plant interactions and population and community
patterns, providing one of the most comprehensive
sets of observations and experiments on plant-plant
interactions via pollinator sharing. One focus ad-
dressed how neighboring hummingbird-pollinated
plant species affect each others' pollination success.
Within each of the two bird-pollinated plant guilds,
most plant species occur with guildmates and over-
lap in flowering period. Stigmatic pollen loads of four
plant species were examined at several points in the
flowering season, in different floral neighborhoods
(Feinsinger et al. 1986). If neighbors affect pollination
via competition for pollinator service, larger numbers
of compatible grains should be delivered to stigmas
in neighborhoods with higher densities of flowers of
the same species, and lower numbers delivered when
neighbors include other species. Pollinator-sharing
did not have consistent competitive effects. Neighbor-
ing flowers had only sporadic effects on each others'
pollination success. In the few instances in which
there were effects, some were competitive and others
were facilitative.
A mechanism for competition exists (Feinsinger253 Plant-Animal Interactions