Abstract
The aims of the study were to reveal: (1) evolutionary lineages of the diploid taxa and their genetic differentiation; (2) the past distribution of the rediscovered 'Mediterranean diploid'; and (3) possible migration routes of diploids in the Mediterranean. A combined approach involving sequencing of two plastid regions (trnL-trnF and rpl32-trnL), nrDNA ITS, rDNA FISH analyses, climatic niche characterization and spatio-temporal modelling was used.
Among the examined diploid species, only two well-differentiated evolutionary lineages were recognized: Anthoxanthum gracile and A. alpinum. The other taxa - A. aristatum, A. ovatum, A. maderense and the 'Mediterranean diploid' - form a rather intermixed group based on the examined molecular data.
In situ rDNA localization enabled identification of the ancestral Anthoxanthum karyotype, shared by A. gracile and two taxa from the crown group. For the studied taxa, ancestral location probabilities for six discrete geographical regions in the Mediterranean were proposed and likely scenarios of gradual expansion from them were suggested.
Modelling past and present distributions shows that the 'Mediterranean diploid' has already been occurring in the same localities for 120 000 years. Highly congruent results were obtained and dated the origin and first diversification of Anthoxanthum to the Miocene.
The later divergence probably took place in the Pleistocene and started polyploid evolution within the genus. The most recent diversification event is still occurring, and incomplete lineage sorting prevents full diversification of taxa at the molecular level, despite clear separation based on climatic niches.
The 'Mediterranean diploid' is hypothesized to be a possible relic of the most recent common ancestor of Anthoxanthum due to their sharing of ancestral features.