Does a supergeneralist plant locally specialize on the most effective pollinators?

Umbellifers, a plant family that includes carrots, have striking inflorescences on stalks. Botanists have thought umbellifers have been generalists. They attract many different types of insect to their flowers. But is this assumption true? Zych and colleagues have looked at recent studies that show only a few insects are the most effective pollinators. The scientists wondered if populations of umbellifers had different pollinator assemblages. They also asked if a variation in a pollinator assemblage was associated with different floral traits. If there was, how did it affect reproduction?

Angelica sylvestris

The focus was on Angelica sylvestris, wild angelica. This is a common European species visited by a taxonomically diverse insect assemblage. The botanists followed three populations, located along a 700-km transect. They watched the plants over three growth seasons. The team identified insect visitors, and surveyed pollen loads present on the insect bodies to see how effective they were. They followed insect activity on plants, tracking umbels, nectar and scent composition. They also performed some transplantation experiment.

Zych and colleagues found the different Angelica populations had varying nectar and scent profiles. Despite the similar taxonomic composition of insect visitor assemblages, the populations were effectively pollinated by different insects. In some places flies, in other places beetles. This suggests local adaptations to the most effective pollinators. Yet, analyses of body pollen loads and behaviour on umbels demonstrated functional equivalency of the visitor morphogroups. This is probably related to the fact that A. sylvestris bears few ovules per flower. When the scientists transplanted the plants, something odd happened. Or maybe failed to happen. The insects did not exhibit a preference towards local genotypes.

Zych et al. conclude that Angelica sylvestris is morphologically well adapted to ecological generalization. They find little evidence that the surveyed populations represent distinct pollination ecotypes. Instead, they argue, the observed variation in floral characters can be interpreted as ‘adaptive wandering’.