Tribe Pogonieae - The Pollination Biology of North American Orchids: North of Florida and Mexico - page 180

Virginia site, may have influenced selection for floral fragrance (Gregg 1991b ). Bees
learn to recognize scents more rapidly than colors or flower shapes: they are cor-
rectly recognized 97-100% of the time following only a single exposure (Kolterman
1969 ). Colors, on the other hand, are recognized at the 90% level only after three to
five visits (Menzel 1967 ) and shapes require at least 20 visits (Wehner 1967 ). An
absence of floral odor in the West Virginia population may explain why most associ-
ated bees have failed to discover the presence of a pollen reward (Gregg 1991b ).
The pollination of Cleistesiopsis , then, is not entirely by deceit. Bumblebees
regularly and actively collected its pollen, packing the grains into their hind tibial
corbiculae. Since bees discard unwanted pollen (Heinrich, personal communication
in Gregg 1991b ), we may infer that Cleistesiopsis pollen is transported back to the
nest as a food resource. However, the marked difference in the behavior of bees at
the two sites implies that Cleistesiopsis depends largely on a strategy of deceit in
West Virginia and more on a strategy of reward in North Carolina (Gregg 1991b ).
One explanation for a difference in reproductive strategies may be morphologi-
cal. West Virginia plants sometimes produced two flowers per stem or several simul-
taneously flowering stems from the same clone (Gregg 1991b ). No instances of the
former and very few of the latter were encountered in 7 years of observation in
North Carolina. Thus, geitonogamy would be more likely to occur in the West
Virginia population, and we have seen that artificial geitonogamous crosses reduce
the percentage of healthy embryos (Gregg 1989, 1991b ). Provision of a reward
might be expected to increase geitonogamous pollination and reduce the overall fit-
ness of the small West Virginia population, a population that produces only a few
hundred, scattered flowers per year. Ackerman ( 1986 ) regarded a reduction in gei-
tonogamy as one of the possible advantages of deceptive pollination.
Flower color, pattern, and sepal orientation can be highly variable in C. bifaria , and
this lack of uniformity may represent an additional adaptation to deceit pollination at
the West Virginia site (Gregg 1991b ). As noted already, Heinrich ( 1975 , 1979 a),
Boyden ( 1982 ), and Little ( 1983 ) believe that such variation increases the time required
for pollinating insects to learn to recognize the flowers. In North Carolina, pollinators
may show a strong preference for scented flowers, and the similar variation in flower
color and pattern observed here might be less important (Gregg 1991b ).
The mechanism by which these two highly interfertile species are maintained
has yet to be determined. A hypothesis that differences in column length (Table 9.4 )
play a role in reproductive isolation has not been substantiated by the pollinator data
presently available (Catling and Gregg 1992 ). However, phenological differences
may contribute. Gregg ( 1991b ) noted a difference in peak flowering times at the
North Carolina site: fresh flowers of C. bifaria were mostly gone by the time the
flower buds of C. divaricata began to open.
Fruiting Success and Limiting Factors
In the West Virginia, population of C. bifaria about 50-74% of open-pollinated
plants produced capsules over a 7-year period (Gregg 1989 ). This is a markedly
lower level of fruit set than that observed in hand pollinated plants (89-100%) and