Following the taxonomy of Gill et al. (2021), and distributions reported by Fry et al. (2000), the pied boubous in eastern Africa are currently thought to comprise three species. The monotypic Ethiopian Boubou Laniarius aethiopicus (J. F. Gmelin, 1789) occurs in far eastern South Sudan north to Eritrea and east through Ethiopia to northwest Somalia and northern Kenya at Moyale.

Tropical Boubou Laniarius major (Hartlaub, 1848) is represented by three subspecies in the region covered here. L. m. major occurs in West and Central Africa east to South Sudan, Uganda, central and western Kenya, north-central Tanzania (Essimingor, Loliondo and Ngorongoro) south to Njombe, extreme north Malawi (where it intergrades with L. m. mossambicus) and west to north and north-west Zambia; L. m. ambiguus von Madarász, 1904, east of the Rift Valley in northern Tanzania (Mt. Kilimanjaro and Mt. Meru) and Kenya, north to Marsabit and Mt. Kulal; and L. m. mossambicus (Fischer & Reichenow, 1880) Zambia (except the north and north-west), Malawi (except the far north), eastern Botswana, Zimbabwe (except the far south-east) and Mozambique.

Finally, the monotypic East Coast Boubou Laniarius sublacteus (Cassin, 1851) is found in the lowlands of extreme southern coastal Somalia through coastal Kenya to north-east Tanzania (including the Usambara and North Pare Mountains) to Zanzibar.

However, the genus Laniarius, including the taxa mentioned above, has long confounded taxonomists and field workers (Harris & Franklin 2000), and it is only comparatively recently that genetic analyses have shed light on some of the less well understood relationships (Nguembock et al. 2008, Finch et al. 2016). A key finding has been that intraspecific polymorphism occurs in the genus, making some plumage traits unreliable indicators of genetic affinity and species-level taxonomy. Specifically, Nguembock et al. (2008) concluded that divergence may have taken similar form in separate lineages, resulting in distantly related taxa showing convergent morphology. While resolving some uncertainties, molecular studies also identified issues in need of further study, one of these being the taxonomy of pied boubous in the coastal forests of East Africa. Birds in this region were all considered as L. sublacteus (see above) until Nguembock et al. (2008) showed that, in fact, two visually identical species were involved: birds from the East Usambara and Rubeho Mountains of north-east Tanzania were referred to sublacteus Cassin, 1851 (type locality Eastern Africa?; see Discussion), and found to be possibly allied to extralimital Southern Boubou L. ferrugineus (J. F. Gmelin, 1788); and those in the Sokoke Forest, coastal Kenya, were found to be most closely related to L. major, and perhaps distantly allied to Turati's Boubou L. turatii (J. Verreaux, 1858: type locality Guinea-Bissau; see Bannerman 1939) from West Africa.

Due, however, to concern that the sample from Sokoke may have been contaminated, Nguembock et al. (2008) left this apparently distinct population unnamed. Subsequently, however, Finch et al. (2016) demonstrated that this population is genetically distinct from north Tanzanian birds, and instead was a close relative of a pied boubou from southern coastal Somalia named somaliensis Reichenow, 1905. Whilst somaliensis differs in appearance from closely related birds of coastal Kenya only by having white vs. black ‘shoulders’, it has incorrectly been regarded as a pied morph of the all-black Coastal Boubou L. nigerrimus (Reichenow, 1879) (= L. erlangeri; see Ash & Miskell 1998) with which it is sympatric in south-coastal Somalia. However, due to uncertainty over whether the type of sublacteus may be from the Sokoke Forest region, or from the Usambaras in north-east Tanzania, Finch et al. (2016) also declined to assign a name to birds in coastal Kenya. Therefore, while the pied boubou of coastal Kenya is specifically distinct from identical-looking birds in coastal Tanzania, it still bears the same name, sublacteus. Meanwhile birds from south-coastal Somalia (somaliensis), from the same genetic lineage as those in coastal Kenya, are currently not afforded recognition in any world checklist (Dickinson & Christidis 2014, del Hoyo & Collar 2016, Clements et al. 2019, Gill et al. 2021).

Meanwhile, the vocalisations of the pied boubou complex have yet to be studied in light of the genetic findings. The well-known duets in Laniarius spp. are typically loud and distinctive, involving repeated simple motifs, with each sex's contribution highly synchronised (Fry et al. 2000, Harris & Franklin 2000). Duet parts of each sex may be overlapping or antiphonal, and serve to defend a territory. Consistent duetting patterns across all groups involve either: (a) lower or higher bell-like notes from males, and a snarl-like note (hereafter snarl), by females, or (b) slow, monotone whistles and croaking notes by males with snarls by females (pers. obs.). Across the wide geographical area covered here, these duets are sufficiently variable that regional patterns may be detected that provide further insight into pied boubou taxonomy, especially where polymorphism may obscure cryptic diversity. In assigning birds to vocal groups based on their duets, I seek to reconcile these with genetic data and published distributions, to better define the range limits of taxon groupings. Given persisting disagreement, I also suggest a possible revised nomenclature and taxonomy for the group.

Methods

Recordings of pied boubou vocalisations were sourced via correspondence with observers and online at Xeno-canto ( www.xeno-canto.org), Macaulay Library ( www.macaulaylibrary.org) and AVoCet ( www.avocet.zoology.msu.edu). Additional recordings at the British Library but not available online were not consulted. Recordings were analysed from across all of East Africa, but with a focus on coastal southern Somalia south to Malawi and northern Mozambique. Vocalisations of some extralimital taxa were also analysed, for context. The vocalisations of duetting pairs formed the principal basis of the analysis, to recognise distinct vocal groups and, if possible, allocate taxa to them. While each sex mimics the voice of the other, and is therefore capable of producing duet phrases in solo renditions, there appears to be no evidence suggesting that the sexes reverse their roles during actual duets, and the attribution of sexual identity throughout this paper is inferred based on extensive review of audio recordings, video footage and personal field experience over a period of 30 years in Kenya and Tanzania. In describing repertoires of each group, only the main songs and calls are considered, these being heard 75% or more of the time in the field (pers. obs.). Examples are presented as sonograms, created using Syrinx software (Burt 2006). For my analyses, birds at Sokoke Forest, Kenya are treated as L. somaliensis, in line with genetic and morphometric congruence (Finch et al. 2016).

Results

Three hundred and sixty-five audio recordings of pied boubou taxa were analysed. Duets recorded from across East Africa were grouped based on various shared characteristics and the presence or absence of certain notes in different populations. For example, snarls comprise the only duet vocalisations by females in some populations while, among the more variable notes delivered by males, some are also specific to certain groups (Table 1). The analysis of vocalisations using these criteria suggests the presence of four main vocal groups in East Africa, with one split into two subgroups (Fig. 6), and a fifth group comprising intergrades, as detailed below. Following each group name, the data in parentheses note the taxa assigned to it and the number of recordings of my overall sample that pertain to each group.

Tropical group (major and ambiguus; n = 170 [northern subgroup], n = 39 [southern subgroup]; see Figs. 6–7). Distribution.—Recordings from West Africa (Cameroon, Ghana, Ivory Coast), east to Nyankunde, DR Congo, to Uganda and Kenya (Murchison Falls, Mt. Elgon, Tugen Hills, Mt. Kenya, the Aberdares, Nairobi, Chyulu Hills) south through northern and western Tanzania and the Albertine Rift (Kilimanjaro, Arusha, Iringa and Mufindi, Sitebe, Kigoma, Minziro Forest, Kigali, Bujumbura, Goma) to northern Zambia (Kitwe, Mayau, Salujinga) and southern DR Congo (Lubumbashi).

Basic duet.—Two bell-like notes on different pitches; the female initiates with a high note, followed by a shorter and lower note by the male, which typically overlaps with the end of the female note, e.g. tee-loo, repeated (Fig. 1a). These form the basis for more excited duets involving multiple notes by each bird (and which may also be initiated by the male?), usually at a faster pace but always on different pitches, e.g. loo-tee-loo-tee-loo or too-too-too-too-lee.

Variations.—During aggressive territorial displays, the length of the bell-like notes shortens and speed of delivery increases, resulting in duets of multiple too-too-too note series without higher notes (Fig. 1b). During particularly aggressive duets the female may also switch to using a snarl, to accompany the bell-like notes of the male, e.g. too-grrrr-too or too-too-too-grrrr, with the snarl typically delivered at the same time as the male's notes. However, this is not common, with most duets comprising solely bell-like notes. Bell-like notes in duets may be repeated singly or as multiples and, while male and female sometimes overlap, duets are usually antiphonal.

Unique notes.—Multiple too-too-too-too notes delivered during aggressive displays are absent in other groups. In these displays, multiple pairs (up to four; pers. obs.) may engage in a lazy but coordinated routine of bounding from branch to branch on a circular track through the subcanopy of a large tree, while snarls are layered over the multiple bell-like phrases.

TABLE 1

Attributes of duet phrases by pied boubous Laniarius spp. in East Africa. Presence or absence of distinguishing characters in duet vocalisations are highlighted. The Tropical group is divided into northern and southern subgroups, with differences between them shown in Fig. 1. * = examples from Zanzibar, where the drawn out flute-like whistle is also used; ** = examples from the Zambezi and Luangwa Valleys, where the croak note and drawn-out flute-like whistle are also used; green = main differences in duet characteristics between the East Coast and other groups; yellow = duet characteristics unique to both the Somali group and southern Tropical subgroup; orange = duet characteristics common to both the Ethiopian and East Coast groups; blue = possible absence of a female snarl in the repertoire of the Somali group. Attribution of sexual identity is inferred based on extensive review of audio recordings, video footage and personal field experience in Kenya and Tanzania.

Figure 1.

Sonograms of the Tropical group showing (a) variation in the basic tee-loo high-low duet (one motif repeated once), (b) rapid, multiple bell-note too-too-too calls of Tropical Boubou Laniarius major and L. m. ambiguus from West Africa to Kenya (single bird), and (c) varying-pitch notes with gonolek-like quality in duets of L. major from south-west Uganda, Zambia, and the Mufindi and Ngorongoro regions of Tanzania. For recording credits see Appendix.

Subgroups.—The Tropical group divides into two subgroups, northern and southern, based on differences in duet structure (Fig. 1) Thus, duets by L. major populations from the southern fringes of the Congo Basin, north to the Albertine Rift and east to the Crater Highlands of Tanzania differ from those of L. major (and ambiguus) elsewhere by the inclusion of a varying-pitch note by either the male (Fig. 1c, examples 2, 3, 5 and 6 from left to right) or female (Fig. 1c, examples 1 and 4) in tandem with bell-like notes by the other sex. This varying-pitch note (which is most similar to that used by the Gabela Bushshrike L. amboimensis of western Angola) affords a ‘plonking’ quality to duet phrases, a characteristic perhaps best known among the gonoleks, e.g. too-k’Yonk-too-k’Yonk-too. Emphasis is on the low element of this note in the male, and on the high element in the female (like the Somali group).

General comments.—Almost all duets comprise bell-like notes with snarls, by the female, perhaps incorporated into c.20–30% of duets on average. Birds in this group make rich, mellow and reverberating notes characteristic of the evergreen forests of interior East Africa.

Figure 2.

Sonograms of the Ethiopian group showing (a) variation in the basic male-initiated too-lee low-high duet, and (b) various snarl notes incorporated into duets or delivered alone by females in the repertoire of Ethiopian Boubou Laniarius aethiopicus. Recordings are all from Ethiopia unless labelled otherwise. For recording credits see Appendix.

Figure 3.

Sonograms of the East Coast group showing (a) variation in the basic croak-snarl-croak duet, and (b–d) the repeated, drawn-out double-monotone whistle that can also be incorporated into a duet (d; Zanzibar) by the addition of a female snarl in the repertoire of East Coast Boubou Laniarius sublacteus and L. m. mossambicus. For recording credits see Appendix.

Figure 4.

Sonograms of the Somali group showing (a) variation in the basic, yet highly distinctive, two bell-note, female-initiated duet of Juba Boubou Laniarius somaliensis from the Kenya coast (single motif repeated once) and (b) the three-note male-initiated version. For recording credits see Appendix.

Figure 5.

Sonograms of pied boubous Laniarius spp. showing two examples of duets from Zambia considered to represent mixed pairings and / or intergrades, with characteristics of both the Tropical (southern subgroup) and East Coast groups. For recording credits see Appendix.

Discussion

Vocalisations and taxonomy.—My results are broadly congruent with those of published DNA analyses. The distributions mapped in Fig. 7 integrate the genetic results from Nguembock et al. (2008) and Finch et al. (2016) with the findings reported here, also taking the literature and previously mapped distributions into account. The vocal evidence presented here fully supports the three key findings of genetic studies by Nguembock et al. (2008) and Finch et al. (2016).

Figure 6.

Map showing the locations of audio recordings of pied boubous (Laniarius spp.) in East Africa and adjacent regions, allocated to the four distinct groups based on duet type (purple, yellow, green and red markers) and one subgroup (blue and yellow markers) recognised herein. Red and yellow markers denote audio recordings with elements from two different groups, indicating some intergradation.

Figure 7.

Map showing the distribution of the species and subspecies in the pied boubou complex (Laniarius spp.) in East Africa, as inferred from vocal analyses in this study and previous genetic studies (Nguembock et al. 2008, Finch et al. 2016).

Further attesting to the significance of the vocal evidence presented here concerning the two coastal forms, are the prophetic words of Sclater & Moreau (1933) who were well aware of these vocal differences and their implications. They clearly described the voice of somaliensis, from just south of the Kenya / Tanzania border on the coastal plain at Tanga (at the southern limit of its distribution), and that of sublacteus in the Usambara foothills, only 60 km inland:

This finding is also in line with common regional species distribution patterns, in that the Tanga-Pangani region of Tanzania comprises a significant biogeographical divide on the east coast of Africa. This marks the southern limit of species typical of the Somali biome, such as Golden Palm Weaver Ploceus bojeri and Ethiopian Swallow Hirundo aethiopica, as well as the northern limit of species more typical of woodland of the southern tropics such as Böhm's Bee-eater Merops boehmi and Piping Cisticola Cisticola fulvicapilla.

While the evidence presented here supports the recommendation of Nguembock et al. (2008) that species status is warranted for sublacteus, it also indicates, unexpectedly, that mossambicus (currently treated within L. major by all authors and not adequately sampled by Nguembock et al. 2008) should be united with it. Based on vocalisations, these two taxa, along with the similar-sounding extralimital limpopoensis, appear to comprise three subspecies within one discrete genetic lineage. Finch et al. (2016) suggested that mossambicus may be specifically distinct from sublacteus. That conclusion is not supported here, although vocal evidence does argue strongly that mossambicus should be transferred to sublacteus from major.

In Zambia, Dowsett & Dowsett-Lemaire (1980) and Dowsett et al. (2008) noted the unique croak note of birds in the Livingstone area (assigned to mossambicus), and which is absent in birds from northern Zambia (which they attributed to major), a treatment implying they considered the differences in voice to be dialectical. While there is no vocal evidence of intergradation between these taxa over most of Zambia, both specimen evidence and vocalisations are, however, supportive of it in a small area of northern Malawi and northern Zambia (Nyika National Park) to southern Tanzania (Ufipa to Mbeya) (Fry et al. 2000). Some duets from this region are characterised by elements from both taxa (Fig. 5), while a specimen from the Mbeya region was recognised as an intergrade and described as L. hybridus Neumann, 1899.

Also unexpected, and not uncovered by previous molecular studies, vocalisations suggest the presence of an unrecognised biogeographic division within major. Consistent use of a varying-pitch note in the duets of birds in the south of its range, entirely absent to the north, is suggestive of two discrete lineages. The location of the divide is in south-west Uganda, which also marks that between many species pairs with fragmented subtropical distributions (e.g., Northern Melaenornis edolioides and Southern Black Flycatchers M. pammelaena, Lesser Lamprotornis chloropterus and Miombo Blue-eared Starlings L. elisabeth, Northern Ptilopsis leucotis and Southern White-faced Owls P. granti, or Dark-eyed Melaniparus leucomelas and Pale-eyed Black Tits M. guineensis). The division between southern major and ambiguus is similarly evident in the Crater Highlands of northern Tanzania, where birds with a varying-pitch duet meet and possibly intergrade with taxon ambiguus of areas east of the Rift Valley, at Essimingor to Mt. Meru. This is suggested by photographs of birds that show the white median coverts of both forms, but with white proximal greater coverts as in major (black in ambiguus), and all-black secondaries as in ambiguus (two are white in major; e.g., see ML 249878891 and ML 291083661). The presence of a varying-pitch note in duets of southern major raises the possibility of a relatively close affinity with coastal somaliensis, which also has a varying-pitch note. Indeed, some recordings of major from northern Zambia are not dissimilar by ear to duets of somaliensis.

Nomenclature.—The name East Coast Boubou has been applied to pied boubous of the entire east coast of Kenya and Tanzania by most authors, under the assumption these birds were all one species and referable to sublacteus (e.g., Britton 1980). More recently, the same name has again been used in conjunction with sublacteus, particularly for birds in north-east Tanzania (i.e. not those birds genetically matching somaliensis in coastal Kenya; Nguembock et al. 2008). The vocal data presented here support the genetic discovery that north-east Tanzanian birds are not closely related to those in coastal Kenya, and I propose that the English name East Coast Boubou should also include the taxa mossambicus Fischer & Reichenow, 1880, and limpopoensis Roberts, 1922 (Fig. 7). While Clements et al. (2019) currently use Zanzibar Boubou for L. sublacteus, East Coast Boubou is preferred here to reflect the extensive distribution of subspecies sublacteus and mossambicus in coastal Tanzania from the Saadani–Pangani region southwards.

Of importance to the taxonomy and nomenclature of birds of coastal Kenya (see below) is justification for their referral to somaliensis here. Finch et al. (2016) opted not to assign a name to birds on the Kenya coast although the genetic data pointed to conspecificity with somaliensis, because of uncertainty over the type locality of the form sublacteus, labelled ‘Eastern Africa’ (Cassin 1851, vide Grant & Mackworth-Praed 1944). With the origin and precise type locality of sublacteus not considered traceable, several locations in coastal Kenya were posited, apparently arbitrarily (Grant & Mackworth-Praed 1944, 1947). Irrespective of exactly where, it seems likely that coastal Kenya was suggested simply because birds with the phenotype of sublacteus had been collected there.

The improbability that the type locality of sublacteus is in Kenya can be inferred from details of the 12,500+ bird specimen collection belonging to François Massena, Second Duke of Rivoli, which the Academy of Natural Sciences of Philadelphia acquired in 1844. From this collection, ten specimens were described as new species, three with type localities designated as Zanzibar (Cassin 1851), so at least some material originated there. Collection of birds from what is now Kenya did not commence until Baron von der Decken's expedition in 1859 (D. A. Turner in litt. 2021) and the first specimens known from mainland Tanzania were taken during the expedition of Captain J. H. Speke in 1860 (N. Baker in litt. 2021). Therefore, sublacteus can have originated only from Zanzibar, which European naturalists visited from the 1820s onwards. It therefore seems justified to refer birds on the Kenya coast to somaliensis, and while a genetic study of the sublacteus type specimen is planned (B. Finch in litt. 2021), the treatment proposed here seems appropriate in the absence of contradictory DNA evidence.

Meanwhile, somaliensis was afforded the English name Juba Pied Shrike by van Someren (1932), reflecting its type locality ‘Unterlauf des Ganale’ (i.e. ‘lower course of the Ganale’; Reichenow 1905), generally thought to be the Juba River. To draw much-needed awareness to the conservation plight of the lower Juba’s riverine forests, which are currently experiencing severe levels of deforestation, the name Juba Boubou is proposed here for L. somaliensis.

Conclusion

This study, based on vocal differences, supports the findings of Nguembock et al. (2008) and Finch et al. (2016) that two distinct and unrelated species of pied boubous should be recognised on the East African coast, with the English name East Coast Boubou for L. sublacteus, as already in use, and Juba Boubou, as recommended here, for L. somaliensis. Vocal data also suggest that white-winged mossambicus is better aligned with black-winged sublacteus than with white-winged major, again supporting statements by Nguembock et al. (2008) that plumage criteria are unreliable indicators of taxonomic affinity. Meanwhile, vocal evidence also suggests southern populations of Tropical Boubou L. m. major are worthy of further taxonomic investigation, based on consistent vocal differences from northern populations.

Future work on this complex in East Africa should seek to ascertain several outstanding details raised by the treatment proposed here. First, DNA comparisons between sublacteus and mossambicus are needed to test the hypothesis of conspecificity, or if mossambicus is better considered specifically. Second, audio recordings of the white-shouldered morph of somaliensis in south-coastal Somalia should be obtained and compared with those of the black-shouldered morph of somaliensis in coastal Kenya, to test previous genetic findings of conspecificity. Third, DNA comparisons of major from the southern (e.g. Zambia to western Tanzania) and northern (west Kenya to Cameroon) parts of its range should test my vocal data, that these birds may comprise two separate lineages.