Phylogenetic Position of Pterocommatinae and Cavariella, and Implications for the Origins of Host Alternation in Aphids

Aphids are morphologically simple. Their numerous hypothesized convergent reductions, such as reduced siphunculi length in association with ant attendance, have made it difficult to define morphological synapomorphies that are necessary for phylogenetic studies. Thus, I used molecular characters both to reexamine the phylogenetic relationships of Cavariella and Pterocommatinae within Aphididae, and to further map host associations and life cycles onto these phylogenies to better understand the evolutionary lability of host alternation within Aphididae. Independent and combined analyses were performed under unweighted parsimony and maximum likelihood criteria for sequences of mitochondrial cytochrome oxidase II plus tRNA-Leucine plus partial cytochrome oxidase I (COII + trnL), and nuclear elongation factor-α (EF1α). Shimodaira-Hasegawa likelihood ratio tests were also employed to test for statistically significant differences between: (1) the tree topologies obtained from the analyses in this study; and (2) topologies supporting the traditional phylogenetic hypotheses based upon morphological data. These analyses recovered various relationships contradicting the current morphology-based phylogeny: 1) a highly supported sister relationship of Pterocommatinae to Cavariella; 2) paraphyly of Myzinae and Anuraphidinae, as well as paraphyly in some genera within Dactynotinae; and 3) support for the sister relationship of Pterocommatinae/Cavariella/Liosomaphis to the remaining macrosiphines. There was also evidence within Aphididae for an evolutionary rapid radiation and multiple origins of host-alternation. These results imply the need for further molecular analyses in resolving relationships within Macrosiphini, and for defining the morphological attributes that characterize these relationships