The effect of pollinator activity on gene flow in colonies of Viola were examined by measuring pollinator flight distances, the frequency of interplant flights and percent pollination under different plant spacing patterns. Pollinator flight distances were directly proportional to spacing parameters while the frequency of interplant flights and percent pollination were inversely proportional to spacing parameters. These findings show that gene flow is reduced by pollinator activity over a wide r
ange of spacing parameters but in populations with low spacing means highly localized gene exchange can occur within the colony. Isolation of colonies may be expected under these circumstaces and cleistogamy may be the optimal breeding system. However, chasmogamous flowers may be important both in promoting with-in-colony gene exchange and long distance between-colony gene exchange corresponding to the sexual functions proposed in several recent models. Viola colonies appear to be semi-isolated demes with pollinator service which can bring adaptive genes to high localized frequencies, but which maintains low frequency, long-distance gene dispersal. This pattern corresponds to the "Shifting Balance" view of evolution.
The orchid Leporella fimbriata is pollinated by pseudocopulation with winged males of the ant Myrmecia urens. This recently studied interaction provides a unique opportunity to examine the two current hypotheses concerning the apparent rarity of ant pollination systems worldwide. The first hypothesis requires a series of specialized growth forms and floral characteristics regarded as adaptations to ant pollination. L. fimbriata does not possess them. The second considers the pollenicidal effects
of secretions from the metapleural gland of ants. These glands are absent in M. urens males and it may be that the occurrence of ant pollination requires the absence of metapleural glands in the vector.
The effects of the surface secretions of eight species of ants on three types of pollen were bioassayed by exposure to the integument of undisturbed, living individuals for 20 min. Ant species included Atta texana which cultures fungi by means of various types of secretions. The frequency of grains showing membrane dysfunction, and therefore reduced viability, was quantified by means of a fluorochromatic test. Comparisons of treated and control samples showed that in 46 out of 50 bioassays there
was a reduction in pollen viability following exposure to ants, 38 being statistically significant. Variation in the outcome of bioassays showed differential potency among ant species and differential vulnerability among pollen types. Ant pollination may be uncommon because surface secretions, often from the metapleural glands, cause membrane dysfunction in pollen. Ant species without metapleural glands may be pollinators, but ant pollinated plants may have pollen resistant to the secretion.
Ant metapleural glands secrete surface antibiotics that affect pollen as well as bacteria and fungi. This may be one reason why ant pollination is rare. It is predicted that pollination by ants is possible only in the presence of certain ant and/or plant traits. Two traits are investigated; first, absence of the metapleural glands, and second, the presence of stigmatic secretions that insulate pollen from the ant integument. The pollinator of the orchid Leporella fimbriata is the ant Myrmecia ur
ens. Only one caste is involved, the winged males, and they differ significantly from the queen and worker castes in that they do not possess metapleural glands. This paper reports experiments which test for differential effects on pollen between the males and other castes and evaluates the importance of stigmatic secretions. The results show that the absence of metapleural glands makes no difference as all three castes have strong disruptive effect on pollen artificially applied to the integument. However, during pollination the orchid secures the pollen mass to the ant surface by stigmatic secretions and normal pollen function, fruit production and seed set occur. It appears that both ant and plant traits are pre-adaptive having evolved for functions other than ant pollination.