Repeated adaptive divergence of microhabitat specialization in avian feather lice

Abstract

Background: Repeated adaptive radiations are evident when phenotypic divergence occurs within lineages, but this divergence into different forms is convergent when compared across lineages. Classic examples of such repeated adaptive divergence occur in island (for example, Caribbean Anolis lizards) and lake systems (for example, African cichlids). Host-parasite systems in many respects are analogous to island systems, where host species represent isolated islands for parasites whose life cycle is highly tied to that of their hosts. Thus, host-parasite systems might exhibit interesting cases of repeated adaptive divergence as seen in island and lake systems.The feather lice of birds spend their entire life cycle on the body of the host and occupy distinct microhabitats on the host: head, wing, body and generalist. These microhabitat specialists show pronounced morphological differences corresponding to how they escape from host preening. We tested whether these different microhabitat specialists were a case of repeated adaptive divergence by constructing both morphological and molecular phylogenies for a diversity of avian feather lice, including many examples of head, wing, body and generalist forms.

Results: Morphological and molecular based phylogenies were highly incongruent, which could be explained by rampant convergence in morphology related to microhabitat specialization on the host. In many cases lice from different microhabitat specializations, but from the same group of birds, were sister taxa.

Conclusions: This pattern indicates a process of repeated adaptive divergence of these parasites within host group, but convergence when comparing parasites across host groups. These results suggest that host-parasite systems might be another case in which repeated adaptive radiations could be relatively common, but potentially overlooked, because morphological convergence can obscure evolutionary relationships.

Figure 1

Body forms of microhabitat specialists across diversity of avian feather lice included in this study. Host group and microhabitat are indicated.

Figure 2

Phylogeny based on combined maximum likelihood analyses of COI, EF-1α and Wingless gene sequences. The tree from gene-partitioned Bayesian analysis was identical in topology. Maximum likelihood bootstrap/Bayesian posterior probabilities shown for each node when greater than 50%. Values less than 50% indicated by -. Microhabitat specialization indicated by vertical bars. Brackets indicate terminal sister pairs of genera from the same host group but of different microhabitation specializations. Avian host group indicated using the first two to four letters of the host order or family (see Table 1).

Figure 3

Majority rule consensus of six equally parsimonious trees from analysis of 138 morphological characters. Numbers above branches indicate percentage of trees for which branch is recovered. Numbers below branches indicated bootstrap values from 1,000 bootstrap replicates. Microhabitat specialization indicated by vertical bars. Avian host group indicated using the first two to four letters of the host order or family (see Table 1).

Figure 4

Principal coordinate analysis of partition metrics comparing optimal molecular, morphological and combined trees. The scatter plot shows the results from the first two axes of the principal coordinates that explain 97% of the overall variation between the scores.

Figure 5

Null distribution over random trees for sister groupings of different ecomorphs on same host group. Cases in which lice of different microhabitat specializations, but from the same host group were terminal sister taxa were counted for each randomized tree. Actual number using the morphological (*) and molecular (**) trees indicated in relation to this null distribution.