A comprehensive approach towards the phylogeny and evolution of cervidae

Cervidae (deer) belong to Ruminantia together with Tragulidae (chevrotains), Antilocapridae (pronghorns), Moschidae (musk deer), Giraffidae (giraffes), and Bovidae (cattle, sheep, antelopes). After bovids, cervids are the second most diverse group of ruminants and large herbivores in general. Cervids are natively distributed in the Americas, Europe and Asia inhabiting a broad variety of habitats. Antlers, deciduous, osseous branched outgrowths of the frontal bone, are synapomorphic for all living member of Cervidae except for the Chinese water deer, Hydropotes inermis, in which they have presumably been lost. They are unique cranial appendages, which are shed and rebuilt at intervals. The antlerogenic process is controlled by a complex interaction of fluctuating levels of several hormones, most importantly testosterone. The oldest antler remains are recorded from the early Miocene. These have often been interpreted as non-deciduous appendages in the past, because of supposed permanent skin coverage and the lack of a burr. Extensive comparative morphological analyses of external features of the antler and of the abscission area showed that antlers of all extant and of eight Miocene cervid genera, including those of most of the earliest cervids Procervulus, Ligeromeryx, and Lagomeryx were deciduous. This insight is particularly important for the systematic classification of early Miocene species as Cervidae, because the absence of antler shedding and rebuilding would exclude them from the total group Cervidae. Cervid systematics has been puzzling researchers for over 150 years. The initial, gross (sub)classifications based on morphology and comparative anatomy are mostly supported by molecular data. In recent decades, molecular systematics has provided new input, but consensus could only be partially reached. Cervids are traditionally subdivided into two subfamilies, Cervinae, consisting of Muntiacini and Cervini, and Capreolinae, consisting of Alceini, Capreolini, Odocoileini, and Rangiferini. While the systematic relationships within Muntiacini and Cervini are, with very few exceptions, resolved, systematic relationships within Capreolinae are much more controversial. The position of Capreolini and Alceini is uncertain and there are many polyphylies within Odocoileini. The latter is the youngest clade of cervids with a quick diversification rate, which makes resolving the systematic relationships more difficult. Also, the rich fossil record of cervids has never been extensively phylogenetically tested concerning potential systematic relationships of fossil representatives to extant cervid taxa. The aim of this work was to investigate the systematic relationships of extant and fossil cervids using molecular and morphological characters and make implications about their evolutionary history based on the phylogenetic reconstructions. To achieve these objectives, molecular data were compiled primarily from public data bases such as GenBank. Some cervid species still lack molecular data because they are difficult to access in the wild. In order to complement the existing data, partial mitochondrial cytochrome b gene sequences from museum specimens of five previously unrecorded species were extracted. The resulting data set was so far the most extensive concerning the taxon sampling for cervids. All molecular data were combined to a data set including 17709 base pairs across five nuclear markers and the complete mitochondrial genome of 50 extant and one fossil cervid species. Several analyses using different data partitions, taxon sampling, partitioning schemes, and optimality criteria were undertaken. Divergence time estimates were undertaken on the molecular data sets using molecular clock node dating. Additionally, the most extensive morphological character matrix for such a broad cervid taxon sampling was compiled. In total 168 cranial and dental characters of 41 extant and 29 fossil cervid species were scored. Due to the highly conservative morphological features of cervids, solving systematic relationships was challenging for both extant and fossil species. However, it was possible to determine several characters useful to diagnose cervid subclades. The morphological and molecular data were combined for a total evidence approach and other phylogenetic testing. A tip dating analyses was also undertaken on the total evidence data set. Molecular topologies were partially in conflict with morphological topologies. Nuclear topologies suggested different systematic relationships for some taxa than the mitochondrial topologies. The combined molecular analyses provided robust topologies. The total evidence analyses were less robust and still contained large areas of unresolved taxa. Alternative approaches to combine both data sets and fossil and extant taxa were more useful. The two different divergence time estimate approaches provided differing but feasible results for each approach. The results showed that most of the Miocene cervids were considered to be more closely related to each other than to any other cervid. They were positioned either between the outgroup and all other cervids or as the sister taxon to Muntiacini. Two Miocene cervids were frequently placed within Muntiacini. Plio- and Pleistocene cervids could often be affiliated to Cervini, Odocoileini or Capreolini. Their morphology is more similar to extant cervids than the morphology of most Miocene cervids. The phylogenetic analyses of this work provide new insights into the evolutionary history of cervids. New systematic relationships were observed, some uncertainties persist and resolving phylogenetic relationships within certain taxa remain challenging. The analyses on the extensive data sets presented here concretised systematic problems and uncertainties, which can now better be targeted with phylogenetic approaches. Several fossil cervids could be successfully related to living representatives, confirming previously assumed affiliations based on qualitative comparisons and introducing new hypotheses. Cervid systematics remains an interesting and fascinating area of research with exciting new results to be expected in the future