Trait evolution in the adaptive radiation of Madagascan Bulbophyllum orchids

Adaptive radiation (AR) is the rapid diversification of a relatively large, monophyletic group of species that profoundly differ in the way they exploit different environments. AR is thus a key process in the origin of organismal diversity and arguably the most important one bridging ecology and evolution. Despite the considerable attention ARs have received so far, surprisingly little is known (especially in plants) about how their tempo and mode of lineage diversification is connected to the evolution of trait diversity, and as to whether there are particular temporal stages of trait evolution with regard to (i) macro-habitat; (ii) traits implicated in local adaptation and survival; and (iii) traits controlling reproductive isolation. To fill this gap, this research proposal aims to investigate the AR of Bulbophyllum orchids in Madagascar and surrounding islands (Comores, Seychelles, Mascarenes) and with a special focus on evaluating whether trait diversification happened in temporal stages. To this aim, we will first generate a time-calibrated, highly resolved, and well-sampled multi- gene phylogeny of Madagascan Bulbophyllum (= 157 spp., = 75% of the regional total) based on DNA sequence data. Second, we will gather new datasets of four trait complexes, which likely had a major role in the groups radiation, using state-of-the-art comparative methods in a multi-disciplinary way: (1) macro-ecological habitat; (2) alternative photosynthetic pathways [water saving Crassulacean Acid Metabolism (CAM) vs. conventional C3 photosynthesis], possibly enabling micro-habitat adaptation and survival; and two probably reproductively selected floral traits, i.e. (3) flower morphology/shape and depending on live material available (4) floral scent. Finally, we will use phylogenetic comparative methods to explicitly fit various macro-evolutionary models of lineage diversification, trait evolution, and trait-dependent diversification to the trait data and/or the molecular phylogeny. Here, it will be of particular interest to determine both the time-dependence of diversification for each trait category to test whether trait diversification occurred in the temporal sequence predicted by the stages model of adaptive radiation (i.e. first: macro-habitat; second: photosynthetic pathway; third: flower morphology; and perhaps fourth: floral scent). Overall, these analyses will allow us gaining novel insights into the tempo and mode of multiple trait evolution in shaping the diversity of Madagascars most species-rich orchid group, and thus should contribute to a better understanding of a striking plant adaptive radiation in a one of the worlds most important but also threatened biodiversity hotspots.

Background and aims - The term 'adaptive radiation' (AR) implies the rapid diversification of a relatively large, monophyletic group of species that profoundly differ in the way they exploit different environments. However, only limited evidence is available (especially in plants) to test whether their tempo and mode of lineage diversification is connected to the evolution of trait disparity. For instance, the stages model suggests a temporal sequence of trait evolution with regard to (i) macro-habitat; (ii) specific traits (e.g., physiological) implicated in local adaptation and survival; and (iii) traits controlling reproductive isolation (e.g., flower morphology). To fill this gap, our research aimed to investigate the AR and trait evolution of Bulbophyllum in Madagascar (c. 210 spp.), with focus on evaluating whether trait diversification in this monophyletic lineage of tropical orchids happened in temporal stages. Methods - We generated the most complete multi-gene phylogeny of Malagasy Bulbophyllum to date (179 spp.; c. 85.2% of the total), and gathered new datasets comprising variation in climatic niche space ('macro-habitat'), photosynthetic pathways (conventional C3 vs. Crassulacean Acid Metabolism/CAM), three-dimensional (3D) flower shape, and floral scent. We then fitted various macro-evolutionary models of species diversification, trait evolution, and trait-dependent diversification to the trait data and/or the time-calibrated molecular phylogeny. Due to insufficient data, floral scent has so far been excluded from these analyses. Key Results - In contrast to predictions raised by the stages model, our results provide no support for a temporal order of trait categories driving the radiation of Bulbophyllum in Madagascar since the early Late Miocene ( 12.70 million years ago, Ma). Instead, we find that the great majority of species still retain the ancestral ('Central Highland') niche as well as the ancestral photosynthetic pathway (C3), while their flower shapes evolved predominantly under a small number of 'optimal' selective regimes, likely mediated by the group's main pollinators (small Dipteran flies). Conclusions - Our results provide no evidence for a trait-diversification coupling, as to be expected for traits fitting constrained models of evolution. Instead, they demonstrate a clade-age-dependent process of diversification with (near) constant-rate accumulation of species over time. Nevertheless, we identified a major incidence of accelerated niche evolution, coinciding with the evolution of CAM in a small subclade (c. 30 spp.) that likely paved the way for the group's expansion into the lowlands (c. 7.36 Ma), i.e., from the sub-humid Central Highlands over the humid 'Eastern Coastal' to the seasonally dry Northwest 'Sambirano' rainforests. As such, Malagasy Bulbophyllum still qualifies as an 'adaptive radiation'; however, tempo and mode of lineage diversification seem to be niche-neutral and uncoupled from 3D flower shape evolution in Madagascar's most species-rich orchid group.