Sucking lice (Phthiraptera: Anoplura) are obligate, permanent ectoparasites of eutherian mammals, parasitizing members of 12 of the 29 recognized mammalian orders and approximately 20% of all mammalian species. These host specific, blood-sucking insects are morphologically adapted for life on mammals: they are wingless, dorso-ventrally flattened, possess tibio-tarsal claws for clinging to host hair, and have piercing mouthparts for feeding. Although there are more than 540 described species of Anoplura and despite the potential economical and medical implications of sucking louse infestations, this study represents the first attempt to examine higher-level anopluran relationships using molecular data. In this study, we use molecular data to reconstruct the evolutionary history of 65 sucking louse taxa with phylogenetic analyses and compare the results to findings based on morphological data. We also estimate divergence times among anopluran taxa and compare our results to host (mammal) relationships.
The diversification time of sucking lice approximately 77 Ma is in agreement with mammalian evolutionary history: all modern mammal orders are hypothesized to have diverged by 75 Ma thus providing suitable habitat for the colonization and radiation of sucking lice. Despite the concordant timing of diversification events early in the association between anoplurans and mammals, there is substantial conflict between the host and parasite phylogenies. This conflict is likely the result of a complex history of host switching and extinction events that occurred throughout the evolutionary association between sucking lice and their mammalian hosts. It is unlikely that there are any ectoparasite groups (including lice) that tracked the early and rapid radiation of eutherian mammals.
Lice were obtained from 8 of the 15 sucking louse families (Additional File 1). Unfortunately, louse data from the remaining 7 families could not be obtained due to specimen rarity (some anopluran families have extremely narrow host ranges and some are monotypic, making collection almost impossible) or failure to amplify specimens in the laboratory. Because of PCR failure as well as availability of data from GenBank, some molecular data could not be collected. The genes 18S, EF-1α, and COI were not collected from 11, 7, and 2 specimens, respectively, and, except for one sample of Pedicinus pictus (Pedicinus pictus 2), none of the specimens analyzed were missing data from more than one molecular marker (Additional File 1).
All phylogenetic trees reconstructed in our analyses support a monophyletic Anoplura, sister to the chewing louse suborder Rhychophthirina (Figure 2). Several anopluran families (Hoplopleuridae and Polyplacidae), genera (Hoplopleura and Pterophthirus), and species (H. ferrisi) were not monophyletic. Phylogenetic constraints forcing the families and genera to be monophyletic were significantly worse than the best tree (ML Shimodaira-Hasegawa tests and examination of Bayesian suboptimal trees; P Lemurpediculus verruculosus) belonging to the anopluran families Pedicinidae, Pediculidae, and Pthiridae formed a highly supported monophyletic group (Figure 2).
The molecular clock was rejected in the combined 3-gene data set; thus, the most appropriate divergence dating techniques are those that relax a molecular clock . Similar to phylogenetic analyses, Bayes factors indicated that partitioned data sets are the preferred partitioning scheme, although analyses of partitioned and non-partitioned data sets produced similar results. Partitioned analyses (with model parameters unlinked across partitions) using all three calibrations resulted in a late Cretaceous origin of the Anoplura, and a time of basal diversification approximately 77 Ma (95% HPD 58-96 Ma; Figure 3 and Additional File 5). Upon initially parasitizing their eutherian hosts, the Anoplura segregated into two clades and then diversified rapidly soon after the Cretaceous-Paleogene (K-Pg) boundary approximately 65 Ma (Figure 3). In one clade (the top clade in Figure 2), sucking lice radiated to parasitize carnivores, artiodactyls, rodents, rabbits, and tree shrews, and in the other clade (the bottom clade in Figure 2), anoplurans diverged to colonize artiodactyls, shrews, rodents, and primates. There appears to be no evidence for parallel cladogenesis between sucking lice and their hosts early in their evolutionary history. Rather, it seems that these parasitic insects independently colonized diverse mammal groups possibly as these host lineages were radiating.
Currently, the morphology-based classification of Anoplura detailed by Kim and Ludwig  and with modifications by Durden and Musser  is followed by most researchers studying sucking lice. The molecular phylogeny reported here (Figure 2) agrees with aspects of this morphology-based classification such as the distinct familial lineages of Anoplura associated with pinnipeds (anopluran family Echinophthiriidae), bovids (Linognathidae), bovids and suids (Haematopinidae), hominids (Pediculidae and Pthiridae) and cercopithecids (Pedicinidae). The differences between the two phylogenies (compare Figures 1 and 2) are intriguing; however, it is possible that morphological features may support some of the molecular-based relationships proposed here. For example, the molecular phylogeny places the hoplopleurid genus Pterophthirus within the genus Hoplopleura (Figure 2). Morphologically, the only difference between these two genera is the extension, to varying degrees, of the second pair of paratergal plates on the abdomen. Perhaps the varying extensions of the second pair or paratergal plates evolved more than once within Hoplopleura and, as such, it may not warrant the recognition of Pterophthirus as a distinct genus . The wide separation between hoplopleurid genera Hoplopleura (including Pterophthirus) and Ancistroplax in the molecular phylogeny (Figure 2) also has morphological ramifications. Kim and Ludwig  recognized two subfamilies within the Hoplopleuridae, the only anopluran family for which they recognized subfamilies. Members of the subfamily Hoplopleurinae (genera Hoplopleura, Pterophthirus, and Paradoxophthirus) have a large continuous sternite on abdomonal segment 2 that physically connects with the corresponding paratergal plates, situated laterally. However, in members of the subfamily Haematopinoidinae (genera Ancistroplax, Haematopinoides, and Schizophthirus), the abdominal segment two sternite is clearly divided medially resulting in two separate plates. It is feasible that this morphological difference actually defines two distinct families rather than subfamilies as supported by the large separation between the two clades in Figure 2. As such, it would be beneficial for future researchers to include other relevant hoplopleurid genera in their molecular phylogenetic reconstructions of Anoplura evolutionary history to determine if the two Hoplopleura subfamilies remain genetically distinct.
The most obvious differences between the morphological (Figure 1) and molecular (Figure 2) anopluran phylogenies involve the family Polyplacidae, which is monophyletic based on morphological data but paraphyletic based on molecular data. Interestingly, the morphological definition of the Polyplacidae is quite variable. Notwithstanding the features they share with all other anopluran families, the only morphological characters that are common to all members of the Polyplacidae, as currently recognized, are the presence of 5 antennal segments, 6 pairs of spiracles on the abdomen, small forelegs, and the absence of a notal pit on the thorax . However, none of these characters are synapomorphies for Polyplacidae. Statements reflecting the morphological variability of Polyplacidae in current descriptions include: "antennae...usually sexually dimorphic," "thorax with mesothoracic phragma usually present," "abdomen with paratergites usually highly developed...and occasionally represented by small sclerites or completely lacking," "tergal and sternal plates usually highly developed and at times reduced or lacking," "male...with variously shaped basal apodeme, parameres and pseudopenis," and "female with...spermatheca usually indistinct" . This extreme morphological variability within the Polyplacidae may actually encompass more than one family as suggested by the separate polyplacid lineages shown in Figure 2. The molecular data suggest that rigorous taxonomic reassessment of what is currently treated as Polyplacidae is warranted. In fact, all five of the separate polyplacid lineages shown in Figure 2 correspond with distinct morphological characters that could be used to define separate families and other suprageneric taxa if future taxonomic reevaluation supports such action.
Without complete taxonomic sampling at the family level, it is difficult to compare Kim's  morphological hypothesis of anopluran relationships (Figure 1) to the molecular phylogeny (Figure 2). A few additional differences, however, are apparent. For one, morphological data support a sister relationships between Pedicinidae and Pthiridae, and these two louse families are closely related to Pecaroecidae, Haematopinidae, and Hybophthiridae, all to the exclusion of Pediculidae . Molecular data, however, support monophyly of all non-polyplacid primate lice (families Pedicinidae, Pediculidae, and Pthiridae) with a relatively distant relationship to the Haematopinidae (Figure 2). Monophyly of these three primate louse families has been found in previous molecular studies [28, 29]; however, molecular data from the families Hybophthiridae, and Pecaroecidae will be needed for a more rigorous comparison to the morphological study of Kim . Interestingly, Kim  noted a relatively close relationship among the families Hoplopleuridae, Linognathidae, and Polyplacidae (Figure 1). The molecular data presented herein also support a close relationship among these three families (or at least specific clades within Hoplopleuridae and Polyplacidae; Figure 2) and it will be interesting to see if these relationships hold with additional investigations. Although a more comprehensive morphological study is currently underway (Smith and Light, unpubl. data), additional molecular data and better taxon sampling will be necessary to properly compare morphology and molecules and it is likely that these data will both agree that substantial taxonomic revision of Anoplura will be necessary. 041b061a72