Deceptive pollination, reproductive success and niche evolution in the broughtonia alliance (orchidaceae)a phylogenetic approach

Resumo

Orchidaceae is a family particularly diverse in the tropics, where new species are continuously found. This dissertation aims to analyse the diversification of a monophyletic Antillean clade of orchids, the Broughtonia alliance, with special emphasis in the ecology and evolution of its reproductive strategies, as well as in the mechanisms that explain its diversity. This alliance is one of the few plant groups endemic to the Antilles, and despite of its relative low species richness it has an outstanding ecological diversity. As such evidence of diversity, Tetramicra riparia, a new species from eastern Cuba is described in this dissertation. Molecular analyses supported T. riparia as part of subgenus Tetramicra, which is more diverse in the neighbour island of Hispaniola. Thus, this species either colonized Cuba from Hispaniola or is a relict of previous geological periods during which parts of Cuba and Hispaniola were a single emerged land. This is in contrast with the rest of Cuban endemic species, which belong to the endemic sub-genera Auricola. Considering that orchid seeds are able to overcome geographic barriers, we tested for phylogenetic niche conservatism of the main bioclimatic requirements of this alliance in shaping its current distribution and diversification. For this, we inferred a time-calibrated phylogeny of the alliance, estimated its biogeographical history, and tested a set of models of trait evolution for the bioclimatic niche of the species. A partially conserved bioclimatic niche was detected, which probably constrained the distribution of most species and clades to sites within certain ranges of precipitation and temperature seasonality. This could explain the geographic structure of the estimated phylogeny, which is contrary to the pattern expected for a group of plants capable of over-sea dispersal by seed. This highlights the importance of intrinsic evolutionary constraints on the diversification and distribution of insular organisms. Also, to obtain a more complete scenario of the past and current ecological constrains on this group of orchids, we also examined whether or not the fruit set is limited by pollination activity, scarceness of resources, or both, as reported elsewhere among orchids. We found that Broughtonia lindenii, B. cubensis, and a group of Tetramicra species are severely limited by pollinator visitation, availability of resources or both. This was observed in other species of the alliance, including those that reward their pollinators with nectar. Also, the sexual reproduction of many of these species was secondarily limited by the activity of some herbivores consuming flowers and fruit. In fact, all the studied species are strictly pollinator-dependent for fruit initiation. Most species have long-lasting flowers or very long flowering periods, or both features. These traits probably enhance the chance of being visited by pollinators. Nevertheless, the existence of a mechanism of floral senescence after both pollen removal and deposition, even when could be limiting the chance of pollination, is here discussed as a contrivance saving resources that could be used in other on-going reproductive processes, but also as a possible way of improving cross-pollination and the reproductive success of other flowers or conspecific plants. In addition, we tested at a higher taxonomical level whether or not pollination success is driven by a group of traits related with floral display and flower accessibility. The study, conducted with most species of Tetramicra, used a confirmatory path analysis and showed that most attributes associated to floral display did not affect pollen exportation, but affected negatively pollen deposition, as well as pollination efficiency (i.e. the ratio between pollen exportation and deposition). In contrast, the variables more related to floral architecture affected only pollen exportation. This provides a phylogenetically contrasted example of the advantage of smaller floral displays for deceptive pollination strategies, which is discussed within and beyond the context of Orchidaceae. Finally, we assessed a group of so far un-tested evolutionary hypotheses regarding the evolution of vegetative and reproductive traits in orchids. Our results indicate that most traits in the Broughtonia alliance are highly labile. In addition, we tested the evolution of the traits under the assumption of a macroevolutionary adaptive landscape. Our results offer the first evidence of diversification following a landscape model in a group of insular plants. In the Broughtonia alliance the reproductive phenotypic convergence is outstanding and well-illustrated by the parallel adaption to hummingbird-pollination from three of the four genera of this alliance, which has happened independently from entomophilous ancestors in all the large islands and island banks except Bahamas. Thus, in addition to the bioclimatic constraints, the local biotic context of each Caribbean island has also shaped the evolution and diversification of the Broughtonia alliance. On the other hand, the deceptive pollination and the high degree of floral specialization of their members have homogenized the reproductive scenario of these orchids. Furthermore, the pervasive occurrence of rewardless flowers, and reproductive systems characterized by self-incompatibility or inbreeding depression, could represent additional sources of risk for the reproduction of these plants, many of which are currently severely endangered. This group, once elucidated the unknown relationship of its members with their fungi partners, could become a classic system to explore the role of the whole spectrum of abiotic and biotic aspects on the reproduction and diversification of insular plants.