Optimal inflorescence size for pollination in Prunella vulgaris.

Abstract

Using a patch of the inflorescences of Prunella vulgaris we observed visitation sequences of fourteen individual bees (Bombus vagans) to determine if the plant's characteristics have influenced the bee's behavior. Through statistical analysis of our data we show the following: the greater the number of flowers on an inflorescence, the higher will be the rate of visitation by Bombus vagans; the greater the number of flowers on an inflorescence, the greater will be the number of probes per visit. We derive a prediction of optimal inflorescence size for P. vulgaris from these data. Assuming one probe sufficiently pollinates P. vulgaris, our results predict an optimal inflorescence of five. The most commonly observed inflorescence size was six, the mean was 18.0 (s.d. = 12.4). However, we found that P. vulgaris is severely pollen limited; measurement of B. vagans tongues proved that P. vulgaris was not being pollinated by the bee because tongue length prevented the contact necessary with the stigmas to insure pollination. P. vulgaris may have an optimal inflorescence size for pollination by short-tongued bumble bees in a different part of its range. Consequently, our results provide a model for other P. vulgaris patches in southern states where Bombus tongue length is shorter and they are the main pollinator.