A COMPARATIVE STUDY ON FLORAL TRAITS IN TWO MEDITERRANEAN ORCHIDS WITH A DIFFERENT POLLINATION STRATEGIE

The extraordinary diversity of living forms is the direct consequence of the rising of reproductive barriers which interrupt genic flow between populations causing their differentiation and so an independent evolution. Most species of flowering plants offer rewards to their pollinators, but some base their pollination on deception (1). Deception is typical of Orchidaceae and some authors define this strategy as the key element of the diversification of this family (2). Using a morphological and functional approach on eight orchids species as sample of food (3) and sex (4) deception strategies, we predicted that stigmatic cavity and pollinia will covary, since they are both involved at different levels as a pre-zygotic barrier; their relation should be stronger in sexual deceptive orchids than in those that use food deception. We also investigated (i) the variation of each floral part within each species and each deception strategy; (ii) if pollinia and stigmatic cavity are interrelated and finally (iii) if a precise morphology of stigmatic cavity and pollinia is promoted by natural selection. In Sardinia there are 16 endemic species among 68 taxa of orchids, within the subtribe Orchidinae, the genera Anacamptis, Himantoglossum and Orchis are used examples of food deception and Ophrys as an example of sexual deception. We examined 160 flowers over the duration of this study. Means and standard deviations of floral traits were run. Correlations between pollinia length and height and width of stigmatic cavities are performed (Fig. 1). We correlated these values both at species and strategy levels. Positive correlations are generally expressed in both strategies. We found that phylogentically closely related species have pollinia length very similar. Hence, we suggest that this is due to adaptation to the similar morphology of strictly phylogenetically related pollinator or because the studied species can be affected by phylogenetic constraints. The comparative analysis of the various floral traits underlines how the each sterile portion is variable within in each species. This confirms the idea that in deceptive orchids the variation of sterile floral traits can be relevant in the deception process. In contrast, a stabilizing selection is proposed to explain the low variability in fertile traits