Pollen flow and pollinator movements in fragmented Primula elatior populations

Primula elatior, an obligate outcrossing, insect-pollinated, long-lived perennial herb, is a widespread species in Belgium, but occurs in fragmented ancient deciduous forests. In the Brussels Capital Region, it is rare and its populations are usually small, located in small, isolated forest fragments. In such conditions we may expect genetic erosion, in particular in small populations. Gene flow is an important factor that has an influence on preventing genetic erosion. For many plant species, seed dispersal is restricted and gene flow mainly occurs by pollen. When the pollen vectors are insects, both the pollinator species present and their behaviour determine the efficiency of pollen transport and the distances over which pollen grains are exchanged. Therefore, it is important to characterise gene flow by pollen.

Using two highly polymorphic microsatellite loci we investigated (1) within- and among-population genetic variation and structure, including seedling and adult life stages, and (2) realized pollen flow within two small populations using paternity analyses. Pollinator movements and pollen dispersal were also investigated within four populations using fluorescent powdered dyes as pollen analogues, and the influence of individual plant size (recipient and dye source) and population size was tested. This indirect method was compared with the results obtained using paternity analyses (direct method).

There was a high genetic variation for seedlings and adults as expected for an obligate outcrossing species. Differentiation between populations of both age classes was not significant or low, which implies moderate or high levels of recent / historical gene flow. Within populations no inbreeding could be detected.

Dye dispersal distances showed a leptokurtic distribution, resulting in deposition of the majority of pollen grains close to the source plant and only incidental dispersal over larger distances. The maximal distance of dye transfer was 314 m. Floral display of the recipient individuals was only significant in two populations when distance to dye source was held constant.
When small and large dye sources were compared within populations, this was only significant for one population and there was a difference in response in distance to dye source. When within-population patterns of large and small dye sources were compared between populations there was an influence of population size.
For the paternity analyses with only two highly polymorphic microsatellite loci, it was possible to determine a potential father for almost half of the seedlings with 95% confidence in both populations.

The estimates of pollen dispersal found for dye movement were in agreement with the effective pollen flow based on paternity analyses.