Effects of Inflorescence Size on Pollination in Epilobium angustifolium

Abstract

In the field, inflorescences of Epilobium angustifolium L. were experimentally manipulated to contain above average (LARGE) or below average (SMALL) numbers of open flowers. We then estimated size-dependent fitness achieved per open flower from pollinator visitation rates, amount of pollen exported and pollen deposited. On average, rate of bee arrival (the principal pollinators) on LARGE (88 .+-. 7.9 S.E. arrivals/hour, N = 52) was 1.5 times that of SMALL (59 .+-. 5.5, N = 53), and bees visited 1.3 times as many flowers before they left (LARGE: 4.3 .+-. 0.2; SMALL: 3.2 .+-. 0.1). However, visitation rates to single flowers were similar for both size classes (LARGE: 23 .+-. 2.1 flowers/hour; SMALL: 24 .+-. 2.1); no difference with respect to sex of the flowers visited was found. In addition, both pollen export and pollen deposition were similar for the two size groups. We therefore conclude that for single flowers male function (pollen donation) and female function (pollen receipt) are both independent of inflorescence size. However, at the same time, flowers on LARGE inflorescences are visited by a larger number of individually different bees than flowers on SMALL, suggesting a higher diversity of mates for the flowers of LARGE. On LARGE inflorescences bees were also less likely to move directly from male to female flowers. With pollen carryover, however, this latter advantage may be counterbalanced by the generally longer visitation sequences on LARGE (4.3 .+-. 0.2 flowers/inflorescence/bee; SMALL: 3.2 .+-. 0.1), leading to a similar proportion of female flowers (incorrectly) visited after a visit to a male (LARGE: 67.6%, N = 146; SMALL: 61.6%, N = 144).