SEXUAL REPRODUCTION IN SAXIFRAGA-OPPOSITIFOLIA L AND SAXIFRAGA-BIFLORA ALL (SAXIFRAGACEAE) IN THE ALPS

Plant life in high alpine or arctic areas is strongly limited by abiot ic rather than biotic environmental factors. In climate warming, invad ing plants from lower altitudes may outcompete indigenous alpine speci es, which might only survive by colonizing new habitats. Because diasp ores are most effective for long-distance dispersal, colonizing specie s mostly depend on reproduction by seeds. The breeding systems of Saxi fraga oppositifolia and Saxifraga biflora were investigated in the fie ld. Saxifraga oppositifolia was found to be mostly cross-pollinated, a nd S. bipora showed a mixed mating system. Relative seed set (seed:ovu le ratio) was higher in S, biflora than in S. oppositifolia. Saxifraga biflora also showed a higher relative reproductive success (RRS) in n aturally cross-pollinated plants, because of high fruit set, and in ge itonogamous selfing compared with S, oppositifolia. Germination percen tage was lower in S, biflora than in S. oppositifolia, but only a few seeds germinated in both species. Variation of germinability could not be attributed to the pollen source of the various pollination treatme nts. Saxifraga oppositifolia is among the earliest plants to flower in spring, at a time when low temperatures Limit pollinator activities. Since pollen transfer is essential in this preferentially outbreeding species, S. oppositifolia is a typical ''pollen risker'' Late-flowerin g species such as S, biflora have numerous pollinators at their dispos ition and usually display a high selfing rate. Early snowfall, however , can destroy the whole seed crop. In this context, S. biflora resembl es a ''seed risker.'' Because of a high outcrossing rate, ''pollen ris kers'' are expected to have a high genetic variability, which is regar ded as preadaptation to stochasticity in tundra habitats and to climat e fluctuations. A possible future climate warming will reduce the chan ces that ''seed riskers'' lose their whole seed yield, which could be advantageous for colonizing new habitats. This demonstrates the need o f learning about species-specific reproductive strategies in order to make predictions on the vulnerability of tundra plants in a changing c limate.