Birds were caught and fumigated for lice in several localities across South China during 2012–2021 using standard mist nets (net size: 2×6 m; 2×12 m) following the methods outlined by Gustafsson et al. (2019b); see Ren et al. (2023) for exact collection localities. Hosts were identified using MacKinnon & Phillipps (2000) or Arlott (2017); host taxonomy has been updated to conform with Clements et al. (2021).

Lice were stored in a –80°C freezer at the Institute of Zoology, Guangdong Academy of Sciences (IZGAS), Guangdong, China. Voucher specimens (see below) were identified to genus by DRG level using the key of Gustafsson et al. (2019a). All specimens of Philopteroides from bulbuls were identified by DRG using the key of Gustafsson et al. (2022b); note that many specimens represent new species, following Ren et al. (2023), and could not be identified with this key.

Specimens of Philopteroides were obtained from seven of the 20 species of bulbuls occurring in China, representing 5 of the 7 genera of bulbuls in this country; most of these sequences were previously published by Ren et al. (2023). Additional sequences of Philopteroides specimens and other members of the Philopterus-complex, as well as some outgroup taxa, were obtained from GenBank, originating from Kolencik et al. (2022), Light et al. (2016) and Catanach et al. (2019). Only taxa for which both COI and EF-1α were available on GenBank were added. All specimens used in our analyses are listed in Table 1.

Selected lice (Table 1) were cut halfway through the pterothorax and extracted for DNA using the DNeasy Blood and Tissue Kit (Qiagen, Shanghai, China) following the manufacturer’s instructions except that extractions were left in 55°C water baths for 24 hours, and only 50μℓ were used for each elution. Exoskeletons were retrieved from the extraction fluid and slide mounted in Canada balsam as vouchers, following Palma (1978). Vouchers are deposited in the collection at IZGAS. Two gene loci were amplified and sequenced – a fragment of the mitochondrial cytochrome oxidase subunit I (COI, 379 bp) and a fragment of the nuclear gene elongation factor 1-alpha (EF-1α, 347 bp). PCR conditions followed those outlined by Bush et al. (2016), using primers L6625 and H7005 (Hafner et al., 1994) for COI, and EF1-For3 and EF1-Cho10 (Danforth and Ji, 1998) for EF-1α.

PCRs were performed using Cytiva PureTaq Ready-To-Go beads (GE Healthcare, Vienna, Austria), following the manufacturer’s instructions. Samples showing satisfactory bands on an electrophoresis gel were sent for sequencing using the same primers as for PCR to Tianyi Huiyuan Gene Technology, Co. Ltd. (Guangzhou, China). Sequences were assembled in Seqman Pro 7.1.0 (DNAStar Inc., Madison, Wisconsin) and checked manually to rule out mismatches between forward and reverse sequencing results for each gene and each individual.

Sequences were aligned separately in MEGA 11 using ClustalW and MUSCLE (Edgar, 2004; Larkin et al. 2007; Kumar et al. 2018). Substitution models for each gene were evaluated in MEGA 11; the best model for COI was GTR+G, and for EF-1α was TN93+G,. The 2 aligned and partitioned genes were imported into and concatenated by BEAST v1.10.4 (Suchard et al. 2018), with the default strict clock prior and a Yule speciation process prior, using random starting trees, with the options of linked trees, separated clock models for each gene, 4 Gamma Categories under the strict clock, and constant size of coalescence. Markov chain Monte Carlo (MCMC) tests were run for 1×10⁸ generations and sampled every 1000 generations. We used Tree Annotator v1.10.4 (Suchard et al. 2018) for tree integration and discarded the first 10,000,000 trees as “burnin.” The output tree from Tree Annotator was imported to FigTree v1.4.3 (http://tree.bio.ed.ac.uk/software/figtree) for figure illustration and edited in Adobe Illustrator 2021.

Slide-mounted voucher specimens were examined with a Nikon Eclipse Ni (Nikon Corporation, Tokyo, Japan), with a drawing tube attached for making illustrations. Drawings were scanned, then compiled and edited in GIMP (www.gimp.org). Measurements (all in mm) were made from live images in NIS-Elements (Nikon Corporation, Tokyo, Japan) for the following dimensions: AW = abdominal width (at segment V); HL = head length (at midline); HW = head width (at widest point of temples); PRW = prothoracic width; PTW = pterothoracic width; TL = total length (at midline).

Morphological terms used and their abbreviations used follow Clay (1951), Mey (1994, 2004), Gustafsson & Bush (2017); abbreviations include: aps = accessory postspiracular seta; as1 = anterior seta 1; lpmes = lateral posterior mesosomal seta; mms = marginal mesometanotal setae; mts1, 3 = marginal temporal setae 1, 3; os = ocular seta; pns = postnodal seta; ppss = pronotal post-spiracular seta; ps = paratergal seta; psps = principal postspiracular seta; pst1–2 = parameral setae 1–2; pts = posttemporal seta; s1–4 = sensilla 1–4 of dorsal postantennal head; sts = sternal seta; tps = tergal posterior seta; vms = vulval marginal setae; vss = vulval submarginal setae. These terms are indicated in the relevant figures.

Our analysis resulted in a tree in which the trabeculum-bearing genera (Philopterus-complex sensu Mey 2004) formed a monophyletic group, to the exclusion of closely related genera (Fig. 1). This clade is basally divided between the Mayriphilopterus Mey, 2004, parasitizing hosts in the Galbuliformes, and all other Philopterus-complex lice, parasitizing mainly hosts in the Passeriformes, but also the genera Clayiella (parasitizing motmots; Coraciiformes) and Vinceopterus (parasitizing trogons; Trogoniformes). Most of the relationships between genera within the Philopterus-complex are unresolved, and both Tyranniphilopterus, Philopteroides and Australophilopterus are recovered as paraphyletic.

Specimens of Philopteroides were placed in three larger clades, although the relationship between these clades were not clear (Fig. 1). Two of these clades are here described as new genera, Coronedaxnew genus and Stasiasticopterusnew genus. For convenience, these three genera together are here referred to as the “Philopteroides morpho-group”, which is not intended to indicate any close relationship between them. Four samples identified by Kolencik et al. (2022) as belonging to Philopteroides were placed outside the three main clades here, but as we have not examined these samples, their generic placement is unresolved; none of these three samples were identified to species level by Kolencik et al. (2022).

The first clade (Clade I; Fig. 1) comprises samples of Philopteroides from a variety of hosts, including honeyeaters (Meliphagidae), Australasian robins (Petroicidae), sunbirds and spiderhunters (Nectariniidae), and treecreepers (Climacteridae). Although the type species of Philopteroides was not included in this phylogeny, previous studies of specimens from honeyeaters (DRG, unpublished data) indicate that this group likely represents Philopteroides s. str.; alternatively, Philopteroides s. str. is not represented in any clade of this tree, and has not yet been sampled. The geographical range of these samples covers Asia and the Australo-Papuan region, but the clade is not divided geographically as the Asian samples are nested inside the Australo-Papuan samples. The Asian samples are all from Nectariniidae, but these do not form a monophyletic clade, as specimens from Aethopyga saturata are nested inside a clade of lice from various Australo-Papuan birds. Most of the relationships within this clade received no support, and the relationships between Philopteroides s. str. and its inferred closest relatives (Clayiella, Tyranniphilopterus s. lat) also received no support.

The second clade (Clade II; Fig. 1) comprises the specimens from bulbuls, which are here described as the genus Stasiasticopterus. This clade includes samples from both African and Asian bulbuls, but samples from the two continents do not form reciprocally monophyletic clades. Instead, samples from African-endemic host genera form two clades, and samples from Asia and the African representative of the genus Hypsipetes forms a third clade. The relationship among these three clades is unresolved. Specimens from the African Batis molitor and the Asian Chlamydochaera jeffreyi are placed as sister to Stasiasticopterus, but these specimens have not been examined.

The third clade (Clade III; Fig. 1) comprises the specimens from monarch flycatchers, which are here described as the genus Coronedax. The deepest divergence within this clade is between specimens from Africa and specimens from Asia, with specimens from Asian Terpsiphone incei being more closely related to specimens from Asian Hypothymis azurea than to specimens from other Terpsiphone hosts from Africa. This clade appears closely related to single specimens from another monarch flycatcher, Grallina cyanoleuca, and the petroicid Eugerygone rubra. As these specimens have not been examined, they are not here considered part of Coronedax.

The structure of the male genitalia of specimens and published illustrations of lice in the Philopteroides morpho-group fall into three categories (Figs 2–17; Table 2).

Group one includes species in which the mesosome is prominent, roughly rectangular, and with a gonopore that has elongated projections distally; moreover, the parameres are long, less restricted in their flexibility, and lack prominent apical setae (Figs 2–5). This group corresponds to the specimens in Clade I in Fig. 1, and represent Philopteroides s. str.

Group two includes species in which the mesosome is much reduced ventrally, but may be visible as a plate dorsally, and with a prominent gonopore of varying structure, that may project distally and may be associated with rugose median projections; moreover, the parameres are more intensely sclerotized, restricted to be highly convergent distally, and lack apical setae (Figs 13–17). This group corresponds to Clade II in Fig. 1, and represent the new genus Stasiasticopterus.

Group three includes species in which the mesosome is reduced to at most a thickening of the distal margin of the basal apodeme, and the gonopore lacks distal projections; moreover, the parameres are long and slender, less restricted in their flexibility, and have prominent apical setae (Fig. 12). This group corresponds to the specimens in Clade III in Fig. 1, and represent the new genus Coronedax.

Comparisons of other morphological characters are inconclusive. Potentially, the dorsal preantennal suture reaches the lateral margin of the preantennal head at the site of anterior seta 1 only in Clade II species (Stasiasticopterus), but this character is not illustrated in sufficient detail in all described species to evaluate. The ventral chaetotaxy may separate Coronedax from other genera, but this is also not conclusive. Female specimens cannot presently be identified to genus level, until the variation of in morphology of the dorsal preantennal suture and the ventral abdominal chaetotaxy have been examined in more detail.

Species and genus delimitation in the Philopterus-complex are notoriously difficult. For instance, whereas patterns of abdominal chaetotaxy can often be used to delimit taxa in the Brueelia-complex (e.g., Gustafsson & Bush 2017), Philopterus-complex lice almost uniformly have setal rows on all tergopleurites and sternal plates (e.g., Mey 2004; Najer et al. 2016, 2020b; Gustafsson et al. 2022a). The numbers of setae in these rows may be different between species, but often constitute ranges with individual variations that may overlap between species; at the genus level, few useful chaetotaxy characters are known. Similar homogeneity can be found in many other character sets, such as the overall structure of the head and the female genitalia.

Other morphological elements, such as the male genitalia, are often much reduced, so that characters can be difficult to compare. For instance, reduction in parameres is found both in Mayriphilopterus (see Mey 2004) and Philopteroides s. lat (see Tandan 1955; Najer et al. 2016; Figs 6–8), and possibly elsewhere. Similarly, the female subgenital plate is almost uniformly reduced to the same general shape in many genera of the Philopterus-complex, and female vulval chaetotaxy is both homogeneous and reduced compare to that of many other louse groups (see e.g., Gustafsson & Bush 2017; Gustafsson et al. 2020).

Moreover, several characters seem to have evolved convergently several times. For instance, the secondary sclerotization of the hyaline margin is known from several distantly related genera (cf. Mey 2004 with Kolencik et al. 2022) and may even occur on subgroups within genera that otherwise lack this character (e.g., Philopterus species on swallows; Gustafsson et al. 2022a). Several distantly related groups in the phylogeny of Kolencik et al. (2022) also have similar preantennal structure (Mey 2004); notably, Mayriphilopterus Mey, 2004, was placed as a sister group to the rest of the Philopterus-complex in the phylogeny of Kolencik et al. (2022) as well as in our phylogeny (Fig. 1), but the preantennal structure in this genus is the same as in Philopterus, which is deeply nested inside the complex (Mey 2004; Kolencik et al. 2022).

In parallel to this, published descriptions and illustrations of many species in the Philopterus-complex are often inadequate to establish which genus they belong to. For instance, of the eleven species of Philopterus s. lat described by Złotorzycka (1964), rough outlines of the male genitalia are given only for four, none of which are detailed enough to be identifiable or placeable in the classification of Mey (2004). Only the dorsal anterior plate is illustrated for all species, but the utility of this character for species delimitation has never been evaluated; moreover, its utility for genus-level classification is probably very limited. Poorly described species like these constitute a large part of the species in the Philopterus-complex, making assessments of taxon limits difficult without reexamining type specimens.

Any investigation into taxon limits in the Philopterus-complex thus must be seen against a background of very low morphological variation in some characters, and convergence or reduction in others. Even if potentially distinct morphological characters are found in a specimen, it is often difficult to assess what this specimen may be related to, due to the lack of adequate comparative illustrations (see Gustafsson et al. 2022a). More than for many other recently revised groups of Ischnocera, the classification of the Philopterus-complex may benefit from the use of genetic data to supplement what few morphological characters can be found.

Here, we have detailed one such investigation, in which genetic and morphological characters come together to separate three distinct group within the Philopterus-complex. These three groups are “cryptic” in the sense that most morphological characters except the male genitalia are either homogeneous (e.g., most chaetotaxy), reduced (e.g., female subgenital plate, some elements of the male genitalia), or seemingly convergent (e.g., sclerotization of frons, preantennal structure) in all three groups. Due to the limited number of species of each genus available for analysis, and the lack of sequence data for most described species, the placement of many species is uncertain, and the variation in e.g., the parameres within some groups is unknown (e.g., the lack of parameres in some Philopteroides; Figs 6–8). It is not clear whether females of these groups can at all be separated, depending on where e.g., Philopteroides beckeri and Ph. pilgrimi are placed (see above).

Moreover, species within some of these genera are remarkably homogeneous. Ren et al. (2023) showed that specimens they considered congeneric in some cases belonged to several distinct genetic clades, and the new species described here, Coronedax longiceps, can be separated from the only other known species in the genus by a range of morphological characters, each of which is rather minor when seen in isolation. Genetic data for C. terpsiphoni has not been published, but genetic data from other African members of this group, previously identified as morphologically close to C. terpsiphoni (Light et al. 2016), is distinct. It seems likely that other members of these three genera, and of the Philopterus-complex as a whole, will be similarly homogeneous morphologically, but clearly separable genetically. This echoes the findings in several other studies on various groups of lice, where morphologically homogeneous specimens have been separated by long branches, and sometimes even been paraphyletic (e.g., Johnson et al. 2003, 2021; Gustafsson & Olsson 2012; Martinů et al. 2015; Escalante et al. 2016).

It should be noted that the type species of Philopteroides has not been examined genetically, and the voucher specimens of the sequences published by Kolencik et al. (2022) have not been examined by us. We have previously examined Philopteroides morpho-group specimens from over two dozen species of honeyeaters (DRG, unpublished data), which have all belonged to one of three morpho-types with regards to the male genitalia (Figs 26–28). The first morphotype (Fig. 26) is the same, or close to, that of the type species of Philopteroides, but notably with reduced parameres compared to described species (Figs 2–5). The second morphotype (Fig. 27) resembles the paramere-less group within Philopteroides (Figs 6–8) but apparently have stout lateral setae instead of parameres; it is not clear whether these setae represent lpmes or pst2, or some other seta, nor if parameres are actually present but much reduced in length. The third morpho-group (Fig. 28) is without parameres and lateral setae, but resemble other species of Philopteroides as defined here regarding the overall shape of the genitalia (e.g., Fig. 8).

There thus appears to be a wide diversity of Philopteroides species from honeyeaters, almost none of which are described. The only described species, Philopteroides mitsusui (Uchida, 1948), is poorly known, and its genitalia were not illustrated or described by Uchida (1948). All specimens we have examined fall into the variation of what we here consider Philopteroides, and seem to straddle the variation among the known species with regards to the presence or absence of parameres. A more extensive examination of lice on honeyeaters is needed to establish whether they form a monophyletic group, as implied by the genetic data (Kolencik et al. 2022; Fig. 1). By comparison, the lice in the Brueelia-complex parasitizing honeyeaters are known to fall into at least three genera (Valim & Palma 2015; Mey 2017).

Here, we tentatively consider the specimens in Clade I (Fig. 1) to represent Philopteroides s. str. However, more studies are needed to establish whether this is the case, or if these specimens actually represent a fourth, unnamed, genus. In any case, based on morphology of the male genitalia, the type species of Philopteroides does not fall within either of the genera described as new here.

The relationships between the three genera in the Philopteroides morpho-group are ambiguous. Philopteroides has been divided into two species groups, the “mitsusui” and “beckeri” groups (Valim & Palma 2013), which differ mainly by the shape of the head and preantennal area; the mitsusui species group includes the type species of Philopteroides. Valim & Palma (2013) placed all species except two in the mitsutsui species group, and Najer et al. (2016) added two more species supposedly belonging to the beckeri species group; Gustafsson et al. (2022b) stated that all the species they treated belonged to the mitsusuit species-group. These placements of species described after 2013 were based on the perceived differences in head shape between the mitsusui and beckeri species-groups. Here, the mitsusui species-group likely falls within the genus Philopteroides (see above), but the placement of the beckeri species group is unknown.

Light et al. (2016) included specimens here placed in both Stasiasticopterus and Coronedax, which formed a monophyletic group in their COI+EF-1α dataset; however, no other Philopterus-complex species were included, and the relationship between these two groups could thus not be evaluated. Similarly, Najer et al. (2020a) included only specimens of what is here called Stasiasticopterus, which all formed a single clade. In the more taxon-rich phylogeny of Kolencik et al. (2022) recovered two clades of lice in the Philopteroides morpho-group, which they identified as representing the mitsusui and beckeri species-groups. The majority of the clade they identified as the beckeri species-group is the one here described as Stasiasticopterus, which may not represent the beckeri species-group, as the species from bulbuls lack the morphological characters associated with this group (cf. Mey 2004; Valim & Palma 2013; Gustafsson et al. 2022b). Possibly, the undescribed species from platysteirid and turdid hosts placed as sister to Stasiasticopterus (Fig. 1; Kolencik et al. 2022) may represent the beckeri species group, but the structure of the male genitalia in Ph. beckeri (Fig. 10) may be more similar to those of Coronedax (Figs 11, 12).

The relative position of the three genera in the Philopteroides morpho-group has varied between analyses (Figs 29–32). In the phylogenies of Kolencik et al. (2022; COI, EF-1α) and Ren et al. (2023; COI, hyp, TMEDE6) the groups here called Philopteroides and Coronedax were placed as sisters, with Stasiasticopterus more distantly related to both. Here, based on COI and EF-1α but with a denser taxon sampling in these groups (particularly in Stasiasticopterus), we find no statistical support for the placement of any of the three genera in the Philopteroides morpho-group within the Philopterus-complex (Fig. 1). More molecular markers are needed to resolve the deeper nodes of the Philopterus-complex tree. Moreover, more morphological data is needed to evaluate whether the specimens marked “Philopteroides?” in Fig. 1 belong to Stasiasticopterus or Coronedax, respectively, or if the morphological differences are sufficient to consider these separate, undescribed, genera.

Both new genera proposed are here considered to be limited to one host family each: Monarchidae for Coronedax, and Pycnonotidae for Stasiasticopterus. This parallels the known distribution of some other Philopterus-complex genera (Table 4). As more species of the Philopteroides morpho-group are described and examined in detail, it seems likely that further cases of host family-specific lice in this group will be discovered. This would also parallel the situation in the Brueelia-complex, which is distributed across more or less the same bird groups as the Philopterus-complex, and which comprises a mixture of host family specialists and more widely distributed genera (Gustafsson & Bush 2017).

Notably, in both the Brueelia- and Philopterus-complexes there appear to be a concentration of more specialist genera occurring on hosts that are limited to the Australo-Papuan and Indo-Malayan regions, but a lowered diversity in the Neotropics and in more boreal areas (cf. Mey 2004; Gustafsson & Bush 2017; Table 4). This likely mirrors the extensive early radiation of passeriform birds in Gondwana and what is today the Australo-Papuan region (e.g., Barker et al. 2004; Jønsson et al. 2008, 2016; Aggerbeck et al. 2014). The estimated ages for the radiation out of this area for both the Passerida and the Corvides (~40–45 Mya; Barker et al. 2004; Jønsson & Fjeldså 2006) predates the estimated age of the Brueelia-complex (~30 Mya; Johnson et al. 2018), but no estimation of the age of the Philopterus-complex has been published. While estimating the radiation dates of ischnoceran lice reliably is difficult, due to the lack of fossil evidence, it is possible that much of the currently known diversity of ischnoceran lice on passeriform hosts post-dates the time when these hosts left their inferred region of origin. The more widely distributed radiations within each louse complex may thus represent the lice that were present on the host lineages that left the Australo-Papuan region, whereas the more specialist louse genera may represent those that parasitized hosts that remained behind. Presumably, as more species of Philopterus-complex lice are described, these patterns may become clearer.

Several of the host families of the Philopteroides morpho-group genera are distributed across a large portion of the Old World tropics, from Australia through South Asia to much of Africa (Clements et al. 2021). The number of species for which DNA sequences or morphological data are available is still very limited; for instance, with the exception of lice from sunbirds and allies (Nectariniidae), the known geographical distribution of Philopteroides s. str. as defined here is largely limited to the Australo-Papuan region (Table 3). Presumably this will change as more species are described, as some host families from which Philopteroides s. str. are known occur more widely (e.g., Rhipiduridae).

However, even among the few species that are included in our analyses, a few potentially significant patterns can be seen. For instance, bulbuls are divided into an African and a largely Asian clade, but members of the latter clade have subsequently recolonized Africa several times (Shakya & Sheldon 2017). Ren et al. (2023), using only Asian louse species and different genetic markers, found some louse clades to contain only specimens parasitizing one host radiation, but at least one clade contained lice from a mixture of host radiations; moreover, although not all relationships between their four clades were resolved, there was no indication of a basal division following the division in the host phylogeny. In our analysis, Stasiasticopterus is divided into three clades, but the relationship between these is unresolved. Two clades (S2 and S3 in Fig. 1) comprise only specimens from Africa, which are placed in the “Africa” clade of Shakya & Sheldon (2017). The third clade (S1 in Fig. 1) contains all species of Stasiasticopterus from Asian hosts, as well as a single specimen from an African host; notably, this African host is one of those that have colonized Africa from Asia, and is nested deeply inside the “Asia” clade of Shakya & Sheldon (2017).

Fewer specimens of Coronedax are available for analysis. However, Asian and African species of Coronedax appear to be separated into different clades, despite both clades containing specimens parasitizing Terpsiphone spp. All the hosts of the included species of Coronedax are in Clade A of Monarchidae (sensu Andersen et al. 2015); the host genera of these species are sister taxa (Fabre et al. 2012; Andersen et al. 2015). The discrepancy between host associations and louse phylogeny may be an artefact of the few species included, but may also be a result of the complicated biogeographical history of these flycatchers (Fabre et al. 2012). The sister group of Coronedax as defined here is also from monarch flycatchers, but specimens from a petroicid host is also closely related to these, suggesting that the true host range of Coronedax may be wider than suggested here.

The partial overlap in available gene sequences between different data sets makes reconciliation of patterns in different studies difficult. However, at least in Coronedax and Stasiasticopterus, it would appear that host biogeography may have helped shape the radiation patterns in lice. Biogeographical patterns are less obvious in Philopteroides s. str., as most specimens included in our analysis are derived from the Australo-Papuan region. The few samples from outside this region in our analysis are all from nectariniid hosts from China and Africa, but the exact placement of these specimens within Philopteroides received little support.

Based on published phylogenies (Najer et al. 2020a; Kolencik et al. 2022; Ren et al. 2023; Fig. 1), a substantial amount of genus-level diversity remains to be discovered within the Philopterus-complex. Several genera appear to be paraphyletic, and even within presumably monophyletic genera, significant morphological differences may exist. However, these differences may be limited to single character sets, particularly the male genitalia (Figs 2–8). This suggests that much of the diversity of the Philopterus-complex may be hidden, especially in cases where only females are known. Descriptions of new taxa in this complex should be accompanied by detailed illustrations of dorsal and ventral features of the male genitalia, so that they may be placed accurately in future revisions of the group.

Species-level circumscription and identification may be even more difficult, although detailed illustrations of male genitalia and other characters may help differentiating species (Gustafsson et al. 2022a). In some cases, traditionally illustrated characters such as the head shape, shape of the dorsal anterior plate, and setal counts may be useful, but the utility of these characters on larger scales needs evaluation. Najer et al. (2020b) considered specimens from several different host species conspecific, even if there was some variation in e.g., the shape of the dorsal anterior plate. Ultimately, at least COI sequences may be necessary to supplement morphological data for species descriptions in the Philopterus-complex.