Aim The use of intuitive or explicit biogeographical reconstruction techniques to estimate the timing of species radiations on the Hawaiian Islands, while often incorporating evidence on the timing of island formation, may be misleading if the effects of extinction are overlooked. Using phylogenies of the bark louse genus Ptycta (Psocidae), we compared results from biogeographical character mapping with those from molecular dating analyses. These results were used to evaluate the 'progression model', which predicts an oldest-toyoungest pattern of island colonization.Location The main Hawaiian Islands (Kauai, Oahu, Maui, Lanai, Molokai and Hawaii).Methods With a data set including 101 Hawaiian Ptycta specimens and 18 outgroup species, we constructed molecular phylogenies based on sequences of the nuclear gene wingless and mitochondrial genes 12S, 16S and cytochrome c oxidase subunit I (COI). Maximum-likelihood, maximum-parsimony and Bayesian phylogenetic analyses were used. We used beast to estimate a timecalibrated tree under a relaxed clock model. ResultsOur analyses suggested a single colonization event, followed by extensive movement of Ptycta among islands and species radiations within and between islands. This monophyletic radiation is composed of two wellsupported clades that are also supported by two synapomorphic characters of the male genitalia.Main conclusions Very different biogeographical patterns are inferred by molecular dating versus approaches based on character mapping. Simple biogeographical reconstruction over the molecular phylogeny supports a pattern of youngest-to-oldest island colonization, the reverse of that predicted by the 'progression model', and implies that this radiation could be as young as the most recent islands of Maui (1.4 Ma) or Hawaii (0.5 Ma). Molecular dating, however, infers a pattern consistent with oldest-to-youngest island colonization, and suggests a lineage age of 7.1 Ma. Extinction on the oldest island of Kauai may account for the differences in results between the two analyses.

Although diurnal birds of prey have historically been placed in a single order due to a number of morphological characters, recent molecular phylogenies have suggested that this is a case of convergence rather than homology, with hawks (Accipitridae) and falcons (Falconidae) forming two distantly related groups within birds. The feather lice of birds have often been used as a model for comparing host and parasite phylogenies, and in some cases there is significant congruence between the two. Thus, studying the phylogeny of the lice of diurnal raptors may be of particular interest with respect to the independent evolution of hawks vs. falcons. Using one mitochondrial gene and three nuclear genes, we inferred a phylogeny for the feather louse genus Degeeriella (which are all obligate raptor ectoparasites) and related genera. This phylogeny indicated that Degeeriella is polyphyletic, with lice from falcons vs. hawks forming two distinct clades. Falcon lice were sister to lice from African woodpeckers, whereas Capraiella, a genus of lice from rollers lice, was embedded within Degeeriella from hawks. This phylogeny showed significant geographical structure, with host geography playing a larger role than host taxonomy in explaining louse phylogeny, particularly within clades of closely related lice. However, the louse phylogeny does reflect host phylogeny at a broad scale; for example, lice from the hawk genus Accipiter form a distinct clade. © 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, 114, 837–847.

Megascops is the most species-rich owl genus in the New World, with 21 species currently recognized. Phylogenetic relationships within this genus are notoriously difficult to establish due to the considerable plumage similarity among species and polymorphism within species. Previous studies have suggested that the widespread lowland Amazonian M. watsonii might include more than one species, and that the Atlantic Forest endemic M. atricapilla is closely related to the M. watsonii complex, but these relationships are as yet poorly understood. A recently published phylogeny of Megascops demonstrated that M. watsonii is paraphyletic with respect to M. atricapilla and that genetic divergences among some populations of M. watsonii are equal to or surpass the degree of differentiation between some M. watsonii and M. atricapilla. To shed light on the taxonomic status of these species and populations within them, we conducted a multi-character study based on molecular, morphological, and vocal characters. We sequenced three mitochondrial (cytb, CO1 and ND2) and three nuclear genes (BF5, CHD and MUSK) for 49 specimens, covering most of the geographic ranges of M. watsonii and M. atricapilla, and used these sequences to estimate phylogenies under alternative Bayesian, Maximum Likelihood, and multilocus coalescent species tree approaches. We studied 252 specimens and vocal parameters from 83 recordings belonging to 65 individuals, distributed throughout the ranges of M. watsonii and M. atricapilla. We used Discriminant Function Analysis (DFA) to analyze both morphometric and vocal data, and a pairwise diagnostic test to evaluate the significance of vocal differences between distinct genetic lineages. Phylogenetic analyses consistently recovered six statistically well-supported clades whose relationships are not entirely in agreement with currently recognized species limits in M. watsonii and M. atricapilla. Morphometric analyses did not detect significant differences among clades. High plumage variation among individuals within clades was usually associated with the presence of two or more color morphs. By contrast, vocal analyses detected significant differentiation among some clades but considerable overlap among others, with some lineages (particularly the most widespread one) exhibiting significant regional variation. The combined results allow for a redefinition of species limits in both M. watsonii and M. atricapilla, with the recognition of four additional species, two of which we describe here as new. We estimated most cladogenesis in the Megascops atricapilla-M. watsonii complex as having taken place during the Plio-Pleistocene, with the development of the modern Amazonian and São Francisco drainages and the expansion and retraction of forest biomes during interglacial and glacial periods as likely events accounting for this relatively recent burst of diversification.