Botanic Gardens Trust, Sydney, Australia

Testing speciation models

Integrating phylogenetic, phenetic, population genetic and bioclimatic analyses with landscape history for testing speciation models

Dr Maurizio Rossetto - Principal Research Scientist and Manager, Evolutionary Ecology

The various models of speciation are categorised primarily on the role and nature of geographic isolation during the origination of isolating barriers between differentiating populations. In particular, allopatric speciation differs from other models in specifying a period during which differentiating populations are completely geographically isolated from one another prior to, and during, the evolution of intrinsic isolating barriers. In contrast, the parapatric and sympatric models specify a more elaborate process whereby a range of selective barriers must have a stronger impact than localised gene flow.

Telopea is a particularly interesting genus for studying speciation modes. It includes five species mostly separated by large distributional gaps with only one minor area of overlap south of the Sydney Basin. Within this area, a disjunct putative population of T. oreades is found in sympatry with a T. mongaensis population at Monga NP, and the distribution of the latter species potentially overlaps with T. speciosissima. Phylogenetically, the four recognised mainland taxa are further differentiated into two clades, T. speciosissima and T. aspera to the north, and T. oreades and T. mongaensis to the south.

Our study shows that the geographical location of chloroplast DNA haplotypes (cpDNA) reflected latitudinal partitioning of genetic variation rather than taxonomic boundaries. This could either be the result of incomplete lineage sorting among sister taxa, or hybridisation. Although the cpDNA data alone does not provide a definite answer, a contrasting topology obtained with nDNA suggests current or historical hybridisation as a possible source of shared haplotypes. In order to obtain a finer scale resolution of species-level genetic divergence, a population-level genotypic analysis was conducted using nuclear microsatellites (SSRs).

The SSR data corroborated the latitudinal patterns of genetic differentiation verified that those correspond to strong genetic differentiation. Five latitudinally, morphologically and genetically differentiated lineages were identified in correspondence with a range of recognised landscape barriers. T. aspera and T. speciosissima are separated by the Hunter Valley (an important floristic boundary), the northern and southern T. speciosissima populations are separated by the Shoalhaven catchment and, with the exception of the T. oreades population at Monga NP, T. mongaensis and T. oreades are separated by a ‘southern corridor’ previously identified in the distribution other southern taxa.

In a pattern consistent with allopatric speciation, where taxa are distributed along geographical (latitudinal and altitudinal) gradients, it would be expected that where overlaps / secondary contact occur there should potentially be evidence of hybridisation. As a result we explored in more depth the hybridisation hypothesis at the Monga NP site where two sister taxa, T. oreades and T. mongaensis, occur in sympatry. Morphological observations were not conducive to the expectation of a hybrid swarm, however the combination of the chloroplast and nuclear datasets suggests that the T. oreades population at Monga NP became isolated from the rest of the species and is now being assimilated by the local populations of T. mongaensis.

This study represents a rare empirical analysis identifying a range of taxa at different stages of allopatric divergence. We have integrated a range of analytical approaches to circumscribe five genetically distinct lineages, within two major clades of relatively recent descent. They are strongly latitudinally differentiated at the population level, and can introgress (some times extensively) when overlaps in distribution occur.

Nuclear microsatellite-based assignment results

Nuclear microsatellite-based assignment results for the number of subdivided populations (K=6) found to be most significant using a Bayesian analysis method. Single bars represent individuals and the colour of the bar represents assignment percentage to one of the six populations identified. Population identities: 1-19 T. speciosissima (17-19: southern populations); 20-23 T. aspera; 24-31 T. oreades (31: Monga NP population); 32-36 T. mongaensis.





T. speciosissima
T. speciosissima Photo: M Rossetto