catchers, and Prong-billed Barbets) deposit most seeds
within 50-60 m of parent plants, although all are ca-
pable of moving seeds several hundreds of meters from
their source. Based on a computer model that combines
data on seed shadows, germination requirements, and
rates of gap formation, dispersal by any of the three bird
species results in a 16-36-fold increase in potential
lifetime reproductive output (total number of surviv-
ing offspring produced by the plants) over the case of
no dispersal. Differences in potential reproductive
output associated with dispersal by the three bird spe-
cies were estimated to be minimal (Murray 1988).
Another form of seed "dispersal" is important for
these pioneer plants—dispersal in time. Although
seeds of the three species normally fail to germinate
in the shady forest understory, where most are depos-
ited by birds, they do not die. Their seeds can remain
dormant on or in the soil, germinating later if and
when a gap forms overhead. The physiological mecha-
nisms that allow such enforced seed dormancy in-
clude responses to soil temperature fluctuations and
sensitivity to spectral characteristics of light, both of
which differ between gap and understory habitats
(Vazquez-Yanes and Orozco-Segovia 1984). When dis-
persal in time via seed dormancy was incorporated
into the model, reproductive output of all three plant
species was greatly increased. With the ability to re-
main dormant in the soil for just 2 years, for example,
plants could potentially produce 11-16 times as many
offspring as those with no capacity for enforced seed
dormancy. Plants whose seeds can remain dormant
up to 40 years can produce up to 611 times as many
offspring (Murray 1988).
These estimates of reproductive output did not
include the effects of seed mortality. The rates at
which otherwise viable seeds are removed from the
soil have important effects on the consequences of
dispersal by different birds. This work supports Howe
and Smallwood's (1982) "colonization hypothesis,"
as it demonstrates how seed dispersal enhances plant
fitness via colonization of discrete patches of habitat
that occur unpredictably in space.
Colonization of host trees by mistletoes. Most mistle-
toes are obligate parasites that colonize only a narrow
range of host plant species. Moreover, their seedlings
can establish only on a subset of the branches of host
plants (see Sargent, "Specialized Seed Dispersal,"
pp. 288-289). Mistletoes also provide an opportunity
to understand the consequences of dispersal by dif-
ferent animals: species more likely to frequent suit-
able host plants, and those that use the "correct" depo-
sition behavior (wiping the sticky, viscin-coated seeds
onto a branch or twig surface), are more likely to effect
successful mistletoe reproduction.Twelve species of mistletoes were investigated
by Sargent (1994; see Sargent, "Mistletoes," pp. 81-
82; "An Exceptional Mistletoe," p. 82; and "Special-
ized Seed Dispersal," pp. 288-289) in Monteverde.
They include members of three families (Viscaceae,
Loranthaceae, and Eremolepidaceae) that parasitize a
variety of host species. One mistletoe, Phomdendron
robustissimum (Viscaceae), has a narrow host range
(nearly obligate on the pioneer tree Sapium glandulo-
sum), which made it possible to census all individu-
als over a broad area (S. glandulosum occurs almost
exclusively in pastures). The movement patterns of
dispersers and the seed deposition patterns they were
likely to produce were examined as a function of fruit
abundance at three spatial scales (see Sargent,
"Mistletoes," pp. 81-82). Although the tanagers that
consume P. robustissimum fruits moved many seeds
over considerable distances, none provided highly ef-
ficient, directed dispersal service. They failed to perch
preferentially in the appropriate host species, and
they spent the greatest proportion of their time
perched in mistletoe plants themselves, where the
seeds would be unable to grow if deposited. This work
suggests that Howe and Smallwood's (1982) "coloni-
zation hypothesis" is a better model for understand-
ing the advantages of bird dispersal of mistletoe seeds
than is the "directed dispersal hypothesis." As with
the pioneer plants studied by Murray (1988), birds
transport seeds of mistletoes for considerable dis-
tances but do not appear to deposit them preferen-
tially in favorable sites for establishment.Other potential cases of directed dispersal. Habitat
preferences among other fruit-eating animals might
form the basis for nonrandom seed dispersal to par-
ticular habitats. The clearest demonstration to date
is Wenny and Levey's (1998) work on dispersal of
Ocotea endresiana (Lauraceae) by male Three-wattled
Bellbirds at Monteverde. During the breeding season,
these birds deposit most of the seeds they carry be-
neath the courtship display perches where they spend
most of their time. Wenny and Levey show that male
bellbirds preferentially locate their perches on the
edges of canopy gaps, and that survivorship of
O. endresiana seedlings is higher in gaps (due to lower
fungal attack) than at randomly chosen sites in the
forest understory. As a result, seedling survival was
higher for seeds dispersed by male bellbirds than for
those dispersed by other species.
Pioneer plants also might be dispersed preferen-
tially to treefall gaps if their dispersers prefer to feed
in such patches (Willson et al. 1982). Evidence for
such preferences has been mixed (Schemske and
Brokaw 1981, Hoppes 1987, Wunderle et al. 1987,
Levey 1988, Loiselle et al. 1996). In Monteverde, no265 Plant-Animal Interactions