No abstract

As in most pollinator-limited orchids lacking edible rewards, a population of C. reginae in southern Missouri showed a low conversion ratio of flowers into fruits (0.046-0.23) over two seasons. There was no relationship between the length of the secondary flowering stem, the number of foliage leaves on the same stem and the number of flowers (one or two) produced at the terminus of the stem. However, the size mattered based on the physical dimensions of pollinia-carrying insects vs. parameters of floral architecture. While a diverse range of floral visitors (Coleoptera, Diptera and Hymenoptera) to C. reginae were observed over three seasons, only six medium-sized bees (Anthophora, Apis and Megachile spp.) carried segments of massulate pollinia after three seasons of observation and collection. Pollinia were always deposited dorsally on the thorax. These bees had a mean width of 4.44 mm and depth of 3.41 mm whereas the rear exit length and width of the orchid measured 6.53 mm and 3.41 mm. respectively. In contrast, the more numerous but smaller bees (2.66 mm width and 2.16 mm depth) in the genera Augochlorella, Augochlora, Ceratina, Lasioglossum spp. etc., exited the flower via the same rear orifices without pressing against the dehiscent anthers. Larger bees (gynes of Bombus spp.) measuring 9.06 mm in width and 6.25 mm in depth, were too large to escape via the rear exits so they left the flower via the large, dorsal entrance (through which they first entered the labellum) never contacting either anther. As in the small-flowered C. plectrochilum, the larger flowered C. reginae receives many floral visitors but selects for pollinia-vectors of a discrete body size.

Fluorescence microscopy is used to compare frequencies of pollen tube penetration in in situ populations of Cypripedium 1.iardo/phicinum W.\X/. Smith et Farrer, Crpnpc'dzum flavum W.W. Smith, Cyprzpedtum iitofltafluJ)1 Dougl. ex !.indl., Cypripedium pari'iflorum Salisbury var. pubescens (Wildenow) ().'X. Knight, Cypripedtwn regincie Walter, and Cypripedium tibeticum Schltr. The average natural (insect-mediated) pollination rates measured over five seasons are wide ranging among the six species (0.08-0.74). However, the pollination rate of hand-manipulated populations (self and/or cross) is significantly greater than the rate of insect-mediated polimations in all species studied. A few pollen tubes in both self-and cross-pollinations display aberrant growth in the styles and/or ovaries, but their numbers are too small to suggest evidence of self-incompatibility. Pollen tubes germinate and grow up to the bases of styles within 48 h in C. bardolphianum, C. flavum, and C. tibeiicuni. Pollen tubes remain at the bases of the styles in C. rnonlanum for 5 d after pollination. In C. parviflorum, pollen tubes penetrate ovaries at 7 d. Pollen tube penetration of ovaries is observed within 15 d after hand pollination in all six species but remains incomplete at this time, with the greatest number of ovule penetrations observed in C. reginae (which has the shortest floral life span). Therefore, we suggest there are additional fsct()rs aside from low pollinator visitation for low conversion rates of flowers into fruits in these Cvpripediunz species. These include inadequate pollen loads deposited on receptive stigmas (pollen limitation), coupled with environmental stress and/or predation disrupting or destroying the slow processes of fertilization and/or fruit maturation.