8.16, bottom). The newly hatched larva (ca. 1 mm
long) burrows into the anther tube and eats the pol-
len as the bud grows. After 3 weeks, just prior to an-
thesis, the larva (now about 5-6 mm long) leaves the
anther tube and makes an exit hole (which looks like
the work of a nectar robber) in the corolla just above
the base of the flower (Fig. 8.16, top). Presumably, the
insect pupates in the soil; adult flies emerge from
pupae after 14-15 days.
Zygothrica neolinea belongs to a group of droso-
philid flies that are known from their massive mixed
species aggregations at patches of bracket fungi in the
forest (Grimaldi 1987). They gather by the thousands
to fight, feed on fungal spores, and mate. Some spe-
Figure 8.16. Centropogon solanifolius bud with latex
droplet over site of recent oviposition (bottom); and
male-phase flower with larval exit hole above the calyx
lobes (top).
cies oviposit in the fungi; others lay their eggs in flow-
ers. In some cases, the relationship between flower
species and fly seems to be very specific, with only
one floral host for a given species of fly; in others, the
association is more general (Pipkin et al. 1966). The
relationship between C. solanifolius and Z. neolinea
does not seem to be species-specific; Z. neolinea has
been reported from Passiflora and Aphelandra flow-
ers in Panama (Pipkin et al. 1966, Grimaldi 1987), and
larvae of other flies have been found in Centropogon
species (Pipkin et al. 1966).
Larval infestation may have important reproduc-
tive consequences for C, solanifolius at the individual
and the population levels. Larval pollen removal
shortens the duration of the male phase and hastens
the onset of the female phase, so relatively fewer flow-
ers are available to donate pollen. Pollen removed by
hummingbirds has the potential to reach an appropri-
ate stigma, but pollen consumed by larvae is unavail-
able for pollination. Larval infestation thus reduces
both the time that pollen is available and the total
amount of pollen in the system. These studies raise
other questions. How does the female-biased floral sex
ratio affect patterns of male and female fitness in the
population? How patchy is larval infestation across
space and time? How does the female fly locate her
host flowers, and once there, how does she know if a
bud already contains an egg? This phenomenon is an
area ripe for future research in Monteverde.DECEIT POLLINATION IN BEGONIA
Jon Agren & Douglas W. Schemsken many animal-pollinated tropical plants with
unisexual flowers (monoecious and dioecious
species), female flowers do not provide a reward
for pollinators; they are pollinated by deceit (Baker
1976, Renner and Feil 1993), For several of these spe-
cies, the rewardless female flowers mimic the floral
characters of the rewarding male flowers; pollinators
may be attracted to the female flowers by intersexual
mimicry (Baker 1976, Bawa 1980b, Little 1983, Willson
and Agren 1989),
Begonia is a large tropical genus of predominantly
monoecious species in which pollination by deceit is
common (Agren and Schemske 1991, Schemske etal.
1996). In several insect-pollinated Begonia species,
male and female flowers show a striking similarity,
The size and spectral properties of the stigmas in these
female flowers closely resemble those of the anthers
in the male flowers, and the number and size of pet-
als (or petallike sepals) do not differ between flowers
of different genders. The degree of similarity between
male and female flowers varies, which may reflect
among-species differences in characters that are im-
portant for the attraction of pollinators to the re-
wardless female flowers. For example, in insect-
pollinated B, fisheri and B, oaxacana, petal size and
number differ between flowers of different sex, but the
visitation rate to female flowers is high enough to re-
sult in high levels of fruit set (Schemske et al. 1996,
J. Agren and P. Schemske, unpubl. data).
Similarity between male and female flowers in
deceit pollination systems may reflect shared ances-
try (unisexual flowers are assumed to be derived from
hermaphroditic flowers), selection for similarity, or
both. The mimicry hypothesis implies that resem-279 Plant-Animal Interactions