Pollen collection does not guarantee
pollen delivery
The need for a pollen transfer agent is
clearly a factor that can limit the fertility of self-incompatible flowering
plant species. On several previous occasions we have discussed some of the more
spectacular pollination mechanisms. However there is still much to be
discovered about more ‘conventional’ pollination systems. This is well
illustrated by the very thorough research carried out by Adler and Irvin, University of Georgia, USA (pp. 141–150) using the
bee-pollinated perennial vine, Gelsemium sempervivens. There are two
floral morphs: thrum (anthers protrude beyond the corolla tube, stigma
concealed within the tube) and pin (stigma protrudes, anthers are within the
tube). The authors observed visits from several types of bee, and focused their
investigation on five of these: Bombus, Habropoda, Osmia, Apis and Xylocopa.
The latter often ‘robs’ flowers by collecting nectar through slits made in the
base of the corolla. Capture of bees followed by counting Gelsemium
pollen grains revealed a >10-fold range between species in the amount of
pollen carried. Then came an elegantly designed study
of pollen transfer (or transfer of fluorescent particles as a pollen
surrogate). Captured bees were carefully wiped clean of adhering grains and
then allowed to forage for pollen/fluorescent particles, after which they were
offered emasculated flowers as targets for pollen/particle transfer. These
investigations showed clearly that the amount of pollen transferred to a
receptive stigma was not tightly related to the amount of pollen collected: the
bee species carrying the most pollen were not necessarily the most effective
pollinators. Overall, Xylocopa was the least effective pollinator, even
when it did not rob the flowers; Bombus, Osmia and Habropoda
were the most effective. Surprisingly, floral morphology did not affect
pollination efficiency in this species.
Professor J. A. Bryant
University of Exeter, UK
j.a.bryant{at}exeter.ac.uk
