Four new species of Brueelia Kéler, 1936 (Phthiraptera: Ischnocera) from African hosts, with a redescription of Nirmus bicurvatus Piaget, 1880

1,2 Guangdong Key Laboratory of Animal Conservation and Resource Utilization, Guangdong Public Laboratory of Wild Animal Conservation and Utilization, Guangdong Institute of Applied Biological Resources, 105 Xingang West Road, Haizhu District, Guangzhou 510260, China. 3,5 Department of Biology and Wildlife Diseases, Faculty of Veterinary Hygiene and Ecology, University of Veterinary and Pharmaceutical Sciences, Palackeho tr. 1946/1, 612 42 Brno, Czech Republic. 4 Department of Veterinary Sciences, Faculty of Agriculture, Food and Natural Resources, Czech University of Life Sciences, Kamycka 129, 165 00 Prague 6, Czech Republic.


Introduction
The chewing lice of African songbirds have long been poorly known (Ledger 1980;Gustafsson & Bush 2015;Light et al. 2016;Takano et al. 2017Takano et al. , 2018. In particular, very few species of lice belonging to the Brueelia complex (sensu Gustafsson & Bush 2017) have ever been reported from Africa. In his summary of the chewing lice known from sub-Saharan Africa, Ledger (1980) listed only 3 species of Meropoecus Eichler, 1940, 36 species of Brueelia Kéler, 1936and 16 species of Sturnidoecus Eichler, 1944 (note that Ledger overlooked the many African species of Sturnidoecus described in Ansari 1968). Of these, only 25 species have been recorded from birds that breed in sub-Saharan Africa; the remaining lice in the list in Ledger (1980) concern species known from hosts that migrate to Africa from Eurasia. In at least some cases, this list includes species of lice that have never been recorded in sub-Saharan Africa, but which occur on hosts that migrate to this region.
In recent decades, many new records of lice in the Brueelia complex have been published from African countries (e.g., Balakrishnan & Sorenson 2006;Sychra et al. 2010aSychra et al. , 2010bNajer et al. 2012;Gustafsson & Bush 2015, 2017Bush et al. 2016;Light et al. 2016;Takano et al. 2017Takano et al. , 2018Gustafsson et al. 2018). However, in many of these reports, lice of the Brueelia complex have not been identifi ed to species level, and few new species of lice in this complex have been described from Africa since 1980 ( Fig. 1; Appendix). Data from several published phylogenies that include African lice of the Brueelia complex (Bush et al. 2016;Light et al. 2016;Takano et al. 2017Takano et al. , 2018 and from a morphological survey (D.R. Gustafsson and S.E. Bush, unpublished data) show that the diversity of lice in this complex is very high in Africa, and a vast number of species remain undescribed. The lack of descriptions, illustrations and identifi cation keys for African lice of the Brueelia complex limits our understanding of the biogeography and evolution of these lice and their host associations.
Here, we provide descriptions of four new species of Brueelia from Africa, as well as a redescription of Nirmus bicurvatus Piaget, 1880. All fi ve species treated here are from host species that are endemic to Africa, and all fi ve hosts belong to families that are very speciose in this region.

Material and methods
All material examined is deposited as slide-mounted specimens at the Natural History Museum, London (NHML). Specimens were examined in an Olympus CX31 microscope. Illustrations were drawn by hand, using a drawing tube fi tted to the microscope. Line drawings were scanned, collated and edited in GIMP (www.gimp.org). Grey lines in all illustrations denote the approximate extent of dark pigmentation; note that these patterns typically differ slightly between specimens of the same species, and sometimes between sides of the same specimen.
Measurements were made from photographs in Quick Photo Micro ver. 3.1 (Promicra, Prague, Czech Republic) and are given for the following dimensions (in mm): AW = abdominal width (at segment V) HL = head length (at midline) HW = head width (at temples) PRW = prothoracic width (at posterior end) PTW = pterothoracic width (at posterior end) TL = total length (at midline)  Ledger (1980), including the records published herein. The fi rst number represents the total number of records of host-louse associations reported for each country, disregarding that some records from different host species may refer to the same louse species. The second number represents the number of louse species identifi ed to species level. Countries with no records of lice of the Brueelia complex since 1980 are unmarked. References to lice in the Brueelia complex known from each country can be found in the Appendix. Note that as no precise collection locality is given on older slides, records for Sudan and South Sudan are here presented together. The numbers ʻ2/2ʼ in the Gulf of Guinea refer to São Tomé and Príncipe, which is not visible on the map. The islands of Réunion (France) and St. Helena (United Kingdom) are also not visible on the map; both have 1 record of a louse of the Brueelia complex in the time period, in both cases identifi ed to species level. In addition to the records shown here, one species was recorded from "Northern Africa" and one from "Congo", with no more detailed locality given on the slide labels.
An index of specifi city (IS) was calculated for all host families from which Brueelia are known in Africa, following Valim & Weckstein (2013). Two estimates were calculated: one based only on the IS of identifi ed species of Brueelia, and one on the IS of identifi ed species of Brueelia plus unidentifi ed species recorded from Africa (see Appendix). The latter IS was calculated under the assumption that all unidentifi ed species listed as ʻBrueelia sp.ʼ in the Appendix are accurately assigned to genus, which is likely not the case for many specimens.

Results
The index of specifi city was calculated for the genus Brueelia in Africa, following Valim & Weckstein (2013); from this index, two estimates of the diversity of this genus in Africa were calculated (Table  1). In Table 1, an IS of 1 means that all species of Brueelia are host-specifi c within this host family. In host families for which IS > 1, each host species is on average parasitized by more than one species of Brueelia; for instance, different host subspecies being parasitized by different species of Brueelia would give an IS > 1. An IS < 1 indicates that at least some Brueelia species parasitizing hosts in this family are not host-specifi c, but occur on more than one host species.
In almost all host families, the IS of Brueelia is 1, suggesting that Brueelia are generally host-specifi c. For nine host families, IS < 1; however, in most of these cases, the species of Brueelia parasitizing hosts in this family have been poorly described, and some of these cases may represent misidentifi cations. Only one host family, Corvidae, has an IS > 1.  Kéler, 1936 in Africa based on an index of host specifi city, following Valim & Weckstein (2013). Host taxonomy and species numbers follow Clements et al. (2018); louse numbers follow Gustafsson & Bush (2017) and Gustafsson et al. (2018). Host numbers in the African fauna includes residents, wintering migrants, introduced species and vagrants, following Lepage (2018). IS is the ʻindex of specifi cityʼ of Valim & Weckstein (2013), and represents #Brueelia species / host species of Brueelia. Estimates have been rounded off to the closest whole number. The second estimate includes the specimens listed as ʻBrueelia sp.ʼ in the Appendix, recorded from Africa after 1980 but not identifi ed to species. Species are included in this estimate under the assumption that all species listed as ʻBrueelia sp.ʼ in the Appendix belong to the genus Brueelia; most likely, this is incorrect in some cases, and the numbers given are thus probably overestimates.

Type species
Brueelia rossittensis Kéler, 1936: 257 (= Nirmus brachythorax Giebel, 1874 by original designation. Clay (1954) discussed the use of the post-spiracular sensillum in determining homology in the abdominal chaetotaxy of Ischnocera. She stated that in Brueelia, these sensilla are known from segments III-VII, whereas in all other groups of ischnoceran lice, they are never found posterior to segment V. Gustafsson & Bush (2017) included these sensilla in their illustrations, but neglected to discuss their importance in the text. Based on our investigation of several hundred species of lice in the Brueelia complex, it seems that these sensilla occur on segments II-III only in the following genera: Brueelia, Teinomordeus Gustafsson & Bush, 2017, Acronirmus Eichler, 1953and Sychraella Gustafsson & Bush, 2017. In all other genera of the Brueelia complex, these sensilla only occur on segments IV-V. However, they are typically very hard to see, especially in species with reduced tergopleurites. Gustafsson & Bush (2017) also neglected to explicitly state that it is the position of post-spiracular setae in relationship to this sensillum that determines whether they are psps or aps. Any setae positioned laterally to the sensillum are aps, whereas any setae situated immediately median to this sensillum are psps. Note that aps and psps on the fi rst abdominal segment bearing post-spiracular setae (often segment V or VI in Brueelia) may be similar in length. Moreover, in some species of, e.g., Olivinirmus Złotorzycka, 1964 there may be more than one psps per side on some segments. To our knowledge, no species in the Brueelia complex has more than one aps per side on any segment.

Remarks
Several of the species here belong to a group of pied Brueelia species found mainly on African hosts in the families Ploceidae, Estrildidae and Paridae. The only species of this group known from hosts outside Africa is Brueelia plocea (Lakshminarayana, 1968), from India. We have seen many additional species in this group, all from African hosts; however, suitable hosts in the same genera are found in South Asia. We here refer to this group as the "African pied Brueelia" group, to distinguish it from the New World ornatissima group, which have similar pigmentation patterns. This group comprises the following species: Brueelia plocea (Lakshminarayana, 1968); B. queleae Sychra & Barlev in Sychra et al., 2010a;B. cantans Sychra in Sychra et al., 2010b;B. aguilarae Gustafsson & Bush, 2017;B. mpumalangensis Gustafsson et al., 2018 The main characteristic of this group is the striking pigmentation pattern. This varies slightly between species, but typically includes having dark pigmentation on the anterior and posterior margins of sternites III-VI, the female tergopleurite IX + X, along the lateral margins of the abdomen, around the distal section of femora I-III and on the subgenital plates. The dark areas are generally at least dark brown, but may appear black in some species; both sternal and subgenital plates typically have distinct translucent fenestrae in both sexes.
In the phylogeny of Bush et al. (2016), members of this group (e.g., Brueelia queleae and Brueelia sp. (= B. mpumalangensis) ex Melaniparus niger) were placed in different parts of the tree, suggesting that they do not form a natural group; however, these placements received no statistical support. Apart from pigmentation patterns, the morphological characters of this group are also very diverse, suggesting that the division of this group in the phylogeny of Bush et al. (2016) may be correct.
Nevertheless, for the purposes of identifi cation and keying, we consider the ʻAfrican pied Brueeliaʼ group a useful grouping to help sort out the vast diversity of species of Brueelia on African hosts. As more species of Brueelia from African hosts become known, the relationships of the species in this informal group may have to be revised, and the group may be found to be artifi cial. We provide a key to the described species in this group below.

Key to the ʻAfrican pied Brueeliaʼ
Note that the dorsal abdominal setae in the original illustration of Brueelia plocea have been translocated to the ventral side (Lakshminarayana 1968). No dorsal setae are given in the original description (ibid.: table II), but multiple setae are illustrated on some segments; we interpret all setae on these segments except the sts as dorsal setae. The female of B. plocea is undescribed. 12. Subgenital plate with largely dark pigmentation apart from a central more or less T-shaped translucent fenestra (sections of this fenestra may be interrupted as in Fig. 29

Diagnosis
Brueelia pofadderensis sp. nov. is a fairly typical species of Brueelia from hosts in the genus Passer Brisson, 1760. This informal group of Brueelia parasitizing Passer spp. is characterized by long, slender heads with convex lateral margins of the preantennal area, elongated parameres and the presence of aps but not psps on the male tergopleurite V. We have seen several undescribed species of this group (D.R. Gustafsson and S.E. Bush, in prep.), but the only described species in this group is Brueelia cyclothorax (Burmeister, 1838) (including N. subtilis Nitzsch in Giebel, 1874 and B. obligata Eichler, 1954). No adequate illustrations or descriptions of B. cyclothorax have been published, but partial illustrations were published by Eichler (1954) and Złotorzycka (1964Złotorzycka ( , 1977. Note that the illustration in Eichler (1954) of the female abdomen mixes dorsal and ventral setae and characters on the same side. The frons in his illustrations are also incorrectly illustrated, as the hyaline margin has collapsed in his specimens.

Etymology
The specifi c epithet is derived from the type locality.

Description
Head slenderly rounded, dome-shaped (Fig. 4), lateral margins of preantennal area convex, frons slightly convex. Marginal carina narrow, with shallowly undulating median margins, displaced and widened at osculum. Ventral anterior plate large, with fl at to shallowly concave anterior margin. Head chaetotaxy and pigmentation patterns as in Fig. 4; head sensilla and pts not visible in examined specimens. Preand postocular nodi of similar size. Marginal temporal carina slender, with undulating median margin. Gular plate broad, lanceolate. Thoracic and abdominal segments and pigmentation patterns as in Figs 2-3, 37-38.

Female
Thoracic and abdominal chaetotaxy as in Fig. 3. Subgenital plate rounded trapezoidal, with moderately wide connection to cross piece (Fig. 8). Vulval margin convergent to rounded median point, in specimen from type host subspecies with 2-3 short, slender vms and 3 short, thorn-like vss on each side; 3 short, slender vos on each side of subgenital plate; distal 1 vos median to vss. In material from P. m. vicinus, with 4 short, slender vms and 3-5 short, thorn-like vss on each side of vulval margin, and 2-3 short, slender vos on each side of subgenital plate, with distal 1 vos situated median to vss.

Remarks
Specimens from the Transvaal host subspecies, Passer melanurus vicinus, differ from the specimens from the type host subspecies by having proportionately shorter and more rounded preantennal heads, in size and in the female genital chaetotaxy; the latter character may be different only due to the small number of specimens examined. The male genitalia and abdominal chaetotaxy are essentially similar between the two populations. We presently do not consider these differences substantial enough to warrant the erection of a new taxon for the specimens from Transvaal, but note that in some other cases, different host subspecies are parasitized by different species of Brueelia (D.R. Gustafsson and S.E. Bush, in prep.).

Diagnosis
Brueelia semiscalaris sp. nov. belongs to the "African pied Brueelia" group (see above), but is the palest described member of this group. Within this group, B. semiscalaris sp. nov. is most similar to B. aguilarae Gustafsson & Bush, 2017, with which it shares the following characters: head relatively slender (Fig. 11), with fl attened frons and only slightly convex lateral margins of the preantennal area; mesosome with nearly parallel lateral margins and somewhat angular postero-lateral corners (Fig. 13); aps absent on male tergopleurite IV (Fig. 9); tps absent on male tergopleurites V-VI (Fig. 9); ps present on female abdominal segment IV (Fig. 10).
Brueelia semiscalaris sp. nov. can be separated from B. aguilarae by the following characters: tps present on male tergopleurite VII in B. semiscalaris sp. nov. (Fig. 9), but absent in B. aguilarae; parameres broad and roughly oval in B. semiscalaris sp. nov. (Fig. 14), but slender and elongated in B. aguilarae; proximal mesosome with convex lateral margins in B. aguilarae, but with concave lateral margins in B. semiscalaris sp. nov. (Fig. 13); gonopore about as long as wide and penile arms reaching beyond distal margin of mesosome in B. semiscalaris sp. nov. (Fig. 13) (Fig. 15), but with median point in B. aguilarae.

Etymology
The specifi c epithet is derived from the Latin ʻsemiʼ for ʻpartial, incompleteʼ and ʻscalarisʼ for ʻladderʼ, referring to the pigmentation pattern of the abdomen.

Description
Head rounded trapezoidal (Fig. 11), lateral margins of preantennal area slightly convex proximally and slightly concave distally, frons concave. Marginal carina moderate, deeply displaced and much widened at osculum, with almost even median margin. Ventral anterior plate large, with deeply concave anterior margin. Head chaetotaxy and pigmentation patterns as in Fig. 11. Preantennal nodi elongated. Preocular nodi much larger than postocular nodi. Marginal temporal carina moderate in width, with undulating median margin. Gular plate slender, lanceolate. Thoracic and abdominal segments and pigmentation patterns as in Figs 9-10, 39-40.

Female
Thoracic and abdominal chaetotaxy as in Fig. 10; ps present on abdominal segment IV. Subgenital plate rounded triangular, with broad connection to cross piece and unique pigmentation pattern (Fig. 15). Vulval margin fl attened, median section somewhat concave, with 6-7 short, slender vms and 7-8 short, thorn-like vss on each side; 3 short, slender vos on each side of subgenital plate; distal 1 vos just anterior

Type locality
Original material from the Leiden Museum, but no type locality given by Piaget (1880).

Description
Head rounded trapezoidal (Fig. 18), lateral margins of preantennal area slightly convex proximally, but slightly concave distally, frons broadly fl attened to slightly concave. Marginal carina slender, deeply displaced and much widened at osculum, and with median margin slightly undulating. Ventral anterior plate oblong, with rounded posterior margin. Head chaetotaxy and pigmentation pattern as in Fig. 18. Preantennal nodi elongated. Preocular nodi larger than postocular nodi. Marginal temporal carina slender, with slightly undulating median margin. Gular plate slenderly lanceolate. Thoracic and abdominal segments and pigmentation patterns as in Figs 16-17.

Male
Male subgenital plate does not reach terminal end of abdomen (Fig. 16). Thoracic and abdominal chaetotaxy as in Fig. 16. Basal apodeme broad, with concave lateral margins (Fig. 19). Proximal mesosome broad, convergent to median point, and with lateral extensions, making entire structure somewhat arrow-shaped (Fig. 20). Mesosomal lobes long and broad, with almost parallel lateral margins, rounded postero-lateral corners, and extensive rugose area along distal margin. Gonopore roughly semioval, about as long as wide. Penile arms long, reaching beyond distal margin of mesosome. Parameres long and broad, distal section elongated, with pst1-2 as in

Remarks
The presence of psps on the female tergopleurites IV-V is unusual in Brueelia. No examples of females with psps on these segments were included in the list of variation in abdominal chaetotaxy of Brueelia published by Gustafsson & Bush (2017: table 3). We know of no other species of Brueelia in which psps are present on the female tergopleurite IV, but psps are present on the female tergopleurite V in several species of Brueelia found on icterid hosts (Cicchino & Castro 1996). Carriker (1963) illustrated setae on female tergopleurites III-VII in the position of psps in B. mirabile Carriker, 1963, but did not illustrate any sts. It is therefore doubtful whether these setae represent psps or sts translocated to the dorsal side; we have not examined Carriker's material. In females of most genera in the Brueelia complex, psps are absent on tergopleurites IV-V, and the general absence of these setae in Brueelia (except in B. bicurvata) is unusual for the complex (Gustafsson & Bush 2017:  commonly found on female tergopleurites IV-V in many other genera of the Brueelia complex (see Gustafsson & Bush 2017), it is possible that this represents the ancestral condition in the Brueelia complex, and that this placement as sister to the remaining Brueelia is correct.
Notably, these setae are present in both Formicaphagus Carriker, 1957(see Price & Clayton 1996 and Formicaricola Carriker, 1957(see Price & Clayton 1995, two of the genera most closely related to the Brueelia complex in the phylogeny of Bush et al. (2016). In the closest relative, Neopsittaconirmus Conci, 1942, the distribution of psps varies, e.g., present on III-VII in N. albus (Le Souëf & Bullen, 1902) (Price & Emerson 1978), but present only on IV-V in many species (Guimarães 1974) and only on IV in N. gracilis Guimarães, 1974 (see also Sychra 2005).

Type locality
Kenya.

Diagnosis
Brueelia terpsichore sp. nov. belongs to the informal ʻAfrican pied Brueeliaʼ group (see above). Within this group, B. terpsichore sp. nov. is not particularly similar to any species. The extensive dark pigmentation of the sternites and subgenital plates of both sexes, the abdominal chaetotaxy, and the head shape suggests that B. aguilarae may be the closest relative of B. terpsichore sp. nov. These two species can be separated by the following characters: tps present on male tergopleurites VI-VIII in B. terpsichore sp. nov. (Fig. 23), but absent in B. aguilarae; male abdominal segment IV with 2 ps on each side in B. terpsichore sp. nov. (Fig. 23), but with 1 ps on each side in B. aguilarae; aps absent on male tergopleurite V in B. terpsichore sp. nov. (Fig. 23), but present in B. aguilarae; proximal mesosome broadly rounded in B. aguilarae, but trapezoidal in B. terpsichore sp. nov. (Fig. 27); translucent fenestra of female subgenital plate T-shaped in B. aguilarae, but divided into smaller, isolated fenestrae in B. terpsichore sp. nov. (Figs 29, 41; note that there is some variation between specimens in this species); vulval margin gently rounded in B. terpsichore sp. nov. (Fig. 29), but convergent to median point in B. aguilarae.

Etymology
The specifi c epithet is in honor of the Greek muse Terpsichore, goddess of dance; ultimately from the Greek ʻterpoʼ for ʻI delightʼ and ʻkhorosʼ for ʻdanceʼ. This refers to the peculiar lekking behaviour of the type host, which includes the construction of a small stage on which the male dances by jumping high into the air and singing to attract females (Andersson 1989(Andersson , 1991.

Description
Head rounded trapezoidal (Fig. 25), lateral margins of preantennal area slightly convex or almost straight, frons broadly fl attened. Marginal carina broad, deeply displaced and much widened at osculum, with undulating median margin. Ventral anterior plate not visible. Head chaetotaxy and pigmentation pattern as in Fig. 25. Preantennal nodi large, bulging. Pre-and postocular nodi large. Marginal temporal carina broad, with distinctly undulating median margin. Gular plate slender, lanceolate. Thoracic and abdominal segments and pigmentation patterns as in Figs 23-24, 41-42.

Female
Thoracic and abdominal chaetotaxy as in Fig. 24. Pigmentation pattern of subgenital plate differing slightly between specimens; two variants shown in Fig. 29 (separated by grey line), one with anterolateral fenestra connected to antero-median fenestra and one with these fenestrae unconnected. Anteromedian fenestra generally pale brown in pigmentation, may extend posteriorly to approach posteromedian fenestra (not shown). Exact shape of postero-lateral fenestrae and translucent lateral borders of distal subgenital plate also differ between specimens. Subgenital plate rounded triangular, with moderate connection to cross piece. Vulval margin gently rounded, with 3-5 short, slender vms and 4-6 short, thorn-like vss on each side; 2-3 short, slender vos on each side of subgenital plate; distal 1 vos median to vss. Material from both host species with same vulval chaetotaxy. Measurements ex Euplectes jacksoni

Remarks
No signifi cant differences have been found between specimens from the two host species, except that the single examined male from Euplectes progne delamerei lacks tps on tergopleurite VI. These are present in all specimens from the type host; however, the number varies between 1 and 2. As only one male from E. p. delamerei was examined, we presently do not attach any signifi cance to this difference, as the psps on this segment is also missing on one side. The absence of these setae may thus be due to individual variation.

Diagnosis
Brueelia sima sp. nov. is part of the informal ʻAfrican pied Brueeliaʼ group (see above). Within this group, B. sima sp. nov. does not appear to be particularly similar to any other species, but the head shape and proportions are most reminiscent of those in B. cantans Sychra in Sychra et al., 2010. These two species can be separated by the following characters: in B. cantans, aps and tps are present on male tergopleurites V-VI, but they are absent in B. sima sp. nov. (Fig. 30); multiple tps are present on male tergopleurite VII in B. cantans, but only a single tps is present on this segment in males of B. sima sp. nov. (Fig. 30); the mesosome has angular postero-lateral corners in B. cantans, but has rounded postero-lateral corners in B. sima sp. nov. (Fig. 34); parameres less elongated in B. sima sp. nov. (Fig. 35) than in B. cantans; the female subgenital plate with an anterior transversal fenestra, interrupted medianly, and with a large central fenestra separated from the anterior transversal fenestra in B. sima sp. nov. (Figs 36, 43), but with all these fenestrae continuous in B. cantans.

Etymology
The specifi c epithet is derived from the Latin ʻsimusʼ for ʻsnub-nosedʼ, referring to the relatively short and broad preantennal area of this species.

Remarks
The abdomen of the single examined female is partially disrupted due to mounting; however, one side of every segment is undistorted. We have here illustrated the abdomen tentatively, based on the undistorted sides. However, the gonapophysal setae on segments VIII-X are absent on both sides of the specimen and are not illustrated here. These setae are present in all other species of the Brueelia complex, and when more specimens of B. sima sp. nov. are examined, it is likely that these setae will be found.

Discussion
The genetic data published by Bush et al. (2016), Light et al. (2016) and Takano et al. (2017Takano et al. ( , 2018 suggest that many of the Brueelia complex louse records from Africa represent distinct species, most of which appear to be host-specifi c. However, the vast majority of the species of chewing lice on African birds remain undescribed. We estimate that the potential Brueelia fauna of African passeriforms and piciforms comprises well over 1000 species (Table 1). The actual number of Brueelia recorded from Africa (including this report) is below 50 species. It is thus fair to say that the Brueelia fauna of Africa is almost entirely unknown, with less than 5% of the potential species described so far. These numbers are even worse when other genera in the Brueelia complex are included (data not shown).
For instance, 305 bird species that could potentially be parasitized by lice of the Brueelia complex (i.e., Passeriformes, Trogoniformes, Meropidae, Picidae) are listed as resident in South Africa by Sinclair et al. (2014). Chewing lice of the Brueelia complex have been reported from 51 of these potential host species since 1980, but only 18 of these reports are identifi ed to species level (Appendix). Assuming that most of the lice of the Brueelia complex in South Africa are host-specifi c, and that most of the resident host species are infested with at least one louse species of the Brueelia complex, this suggests that the number of such species in South Africa may be underestimated by a factor of ten. The number of species of the Brueelia complex in all of Africa is most likely even more badly underestimated. The entirety of sub-Saharan West Africa has only 31 identifi ed records of lice in the Brueelia complex (Appendix), but the region is home to over 560 species of potential host species (Borrow & Demey 2014). Note that these numbers do not include migrants to these regions, and the number of potential host species is therefore much higher if non-resident bird species are included.
In addition, many passeriform birds are parasitized by more than one species of chewing lice of the Brueelia complex (Gustafsson & Bush 2017). Moreover, there are indications that some host species are infested with different species of lice in the same genus in different parts of its range (e.g., Brueelia spp. on different subspecies of Plocepasser mahali (Smith, 1836); Gustafsson & Bush, in prep.). In some cases, multiple Brueelia species occur on the same host even in the absence of different host subspecies (e.g., B. zohrae Ansari, 1956 andB. moreli Ansari, 1957, both on Ptilostomus afer (Linnaeus, 1766)). By contrast, some species of lice in the Brueelia complex occur on multiple host species (e.g., Balakrishnan & Sorenson 2006;Bush et al. 2016), sometimes including host species from different families. Notably, four of the new genera described by Gustafsson & Bush (2017)  To a large extent, the lack of identifi cations to species level in these reports is due to the fact that very few chewing lice in the Brueelia complex have even been described from Africa. There is thus no framework of descriptions and illustrations with which to compare collected specimens. Here, we describe four new species of Brueelia from African hosts and redescribe one species. Three of these species belong to an apparently widely distributed informal group, characterised by their striking pigmentation patterns; a key to this informal group is provided. We hope that, together with those recently published by Gustafsson & Bush (2015, 2017 and Gustafsson et al. (2018), the descriptions, illustrations and key characters presented here will make future identifi cations and descriptions of African species of Brueelia easier, and help us better understand the biodiversity and evolutionary history of the Brueelia complex. Appendix (continued on next 15 pages). Species of chewing lice in the Brueelia complex (sensu Gustafsson & Bush 2017) reported from Africa since Ledger (1980). Louse taxonomy follows Gustafsson & Bush (2017); however, many of the specimens identifi ed only to genus level have not been examined, and may in fact belong to a different genus than listed below. For brevity, louse subgenera have been excluded, and host subspecies are included only if reported in the original publication. Host taxonomy follows Clements et al. (2018). Almost all specimens from Sudan were collected by Richard Meinertzhagen, who was stationed in Kenya. It thus seems likely that at least some of these specimens were collected in what is today South Sudan; however, the only locality information given on the slides is "Sudan".