Phylogenetic Analysis of Petunia sensu Jussieu (Solanaceae) using Chloroplast DNA RFLP

Branches in the trees for Petunia and Calibrachoa

It is said that taxonomists come mostly in one of two types: lumpers or splitters. We can see this at several taxonomic levels, including that of the genus, as described by Ando et al., an international team based in (pp. 289�297). They have worked on the �old� broad Petunia genus which, since the late 1980s, has been divided into two genera, Petunia and Calibrachoa. Following this split, several more Petunia species have been transferred into Calibrachoa. But, a leading question is whether or not the evidence to support these changes is strong enough. According to the authors, it is, although the two genera are certainly close. This conclusion is based on a molecular genetic analysis using RFLPs (restriction fragment length polymorphisms) in chloroplast DNA, a technique that has been used to establish relationships in several other angiosperm groups. In addition to the two genera in question, Nicotiana langsdorffii was used to provide data from a closely related genus in the same family. A total of 212 RFLPs was detected; 89 of these were shared by at least two taxa and were therefore used in the construction of a phylogenetic tree; 85 of the RFLPs could be unambiguously ascribed to gain or loss of a specific restriction site and these were used to construct a distance matrix. The results of the two techniques for assigning relatedness gave remarkably similar results. The most obvious feature is the clarity of the division between the two genera. Further, Calibrachoa is a more genetically diverse genus than Petunia and itsspecies fall into two clear groups; one contains C. pygmaea and the type species C. parviflora, while the second contains all the other species. This grouping supports the one suggested by more traditional taxonomic methods. However, the molecular analysis does not resolve unequivocally the inter-specific relationships in Petunia. Consequently, the authors suggest that further molecular data should now be obtained.

 

Professor J. A. Bryant
University of Exeter, UK
j.a.bryant{at}exeter.ac.uk