Ld. Harder et Sch. Barrett, POLLEN REMOVAL FROM TRISTYLOUS-PONTEDERIA-CORDATA - EFFECTS OF ANTHERPOSITION AND POLLINATOR SPECIALIZATION, Ecology, 74(4), 1993, pp. 1059-1072
Stamens produce male gametophytes and expose them to the pollination p
rocess, so that stamen characteristics should reflect selection on mal
e function. For animal-pollinated plants, restricting removal in favor
of using all available pollinators should generally promote pollen di
spersal, so that staminal characteristics should limit pollen removal
by individual pollinators. Here we describe the influences of anther p
osition on pollen removal from Pontederia cordata, a tristylous plant
with anthers at three distinct positions: exserted from the flower (lo
ng-level anthers), at the mouth of the flaring tubular perianth (mid-l
evel anthers), and near the perianth base (short-level anthers). To me
asure pollen removal, we presented unvisited flowers to freely foragin
g bees and counted the number of pollen grains left in flowers after 1
-4 visits. The pollinators included two generalist bumble bees (Bombus
impatiens and B. vagans: Apidae) and Melissodes apicata (Anthophorida
e), a specialist with morphological and behavioral adaptations for col
lecting pollen from P. cordata. Analysis of pollen removal incorporate
d the effects of bee species, anther position, pollen size, flower len
gth, and the number and cumulative duration of bee visits. Anther posi
tion and bee species significantly influenced pollen removal, although
the differences between bee species were not related to pollinator sp
ecialization. Long-level anthers dehisced early, lost an average of 39
% of their pollen during their first visit, and did not consistently c
ontact small pollinators, such as M. apicata. Short-level anthers rest
ricted pollen removal more (28% lost) than long-level anthers, but deh
isced more slowly and were subject to unpredictable pollen removal. Mi
d-level anthers combine the advantages of more exserted and inserted p
ositions without suffering the corresponding disadvantages as they res
tricted removal (24% lost), dehisced relatively early, and seemed to c
onsistently contact all pollinators. The overall benefits of positioni
ng anthers within the perianth mouth, the position occupied by mid-lev
el anthers in P. cordata, may explain the prevalence of this morpholog
y among bee-pollinated plants with flaring, tubular, monomorphic flowe
rs.