Conservation of fragmented wild orchid populations. A study based on the rewarding Gymnadenia conopsea and the rewardless Orchis mascula

Habitat fragmentation is one of the most important threats to biodiversity. Plant populations are confined to very small habitat fragments in a hostile landscape matrix resulting in small population sizes. Orchids are very environmentally demanding, making them extremely vulnerable to even small changes in land use and the consequences of habitat fragmentation. Orchid species often depend on a specific pollinating genus for pollination, which makes them extremely vulnerable to disruptions of plant-pollinator interactions by habitat fragmentation. Also, many orchid species attract pollinators by deceit, and don t provide a floral reward. This can be expected to influence mating patterns and gene flow. Our general objective was to determine the viability of fragmented orchid populations, both of a rewarding (Gymnadenia conopsea) as well as a rewardless orchid species (Orchis mascula). The study area is the valley of the river Viroin, 75km to the south of Brussels. Since the second half of the 19th century, abandonment of the traditional land use has resulted in the loss of large areas of calcareous grassland. Moreover, the remaining grasslands are often small and spatially isolated. In a subset of populations we studied the relationship between the fitness of the populations and fragmentation characteristics for both species. We also measured microvariation in the surrounding environment, to test whether the viability of orchid populations is influenced by environmental characteristics. For G. conopsea, the effects of habitat fragmentation where quantified throughout a six year period. Next, we also calculated population genetic variables based on AFLP markers for a subset of populations of O. mascula and we quantified between and within population gene flow in this subset of populations through performing a full parentage analysis. We also evaluated three common techniques to assess reproductive success in orchids (embryo counting, tetrazolium test, in vitro germination). Finally, an extensive account on the ecology of the rewarding species G. conopsea was provided. Reproductive success of G. conopsea and O. mascula was not significantly influenced by environmental characteristics. Percentage fruit set was positively related to population size in G. conopsea, but no such effect was observed in O. mascula. On the other hand, seed viability in G. conopsea remained low, irrespective of population size, while a large population size positively influenced seed viability in O. mascula. Our results also suggest that in large populations where pollinators are expected to be more abundant, percentage fruit set increases in the rewarding G. conopsea, but at the same time, geitonogamous pollination reduces seed viability in these populations. Seed viability of the rewardless O. mascula seems to benefit from increased pollinator availability in larger populations, whereas the low percentages of fruit set in O. mascula in large populations may be explained by the relatively low attractiveness of rewardless orchid species for pollinators. Connectivity did not influence reproductive success of O. mascula. For G. conopsea, on the contrary, the isolation-effect varied among years, most likely due to the influence of varying weather conditions on the behaviour of its pollinators. Fragmented populations of O. mascula were not strongly genetically differentiated from each other and the exchange of alleles between populations occurs mostly through pollen flow, whereas seed dispersal was limited to the first few meters from the mother plant. The results of the three most often applied methods to measure reproductive success in orchids (embryo counting, tetrazolium test and in vitro germination) were not correlated to each other. The counting is the easiest procedure, although it overestimates the seed viability in comparison with the tetrazolium test and germination experiments. The results from the tetrazolium test overestimated viability as compared to the germination experiment. Based on the results in this PhD work, we can conclude that both rewarding G. conopsea and rewardless O. mascula suffer from the detrimental consequences of habitat fragmentation, each in their specific manner. As far as conservation management is concerned, each species requires a different approach. Increasing connectivity will sustain benefits for the rewarding species, and increasing population size will be particularly beneficial for the rewardless species.nrpages: 201status: publishe