This is the second and more important instalment of a series of 15 CDs on the vocalisations of African birds; the first volume (four CDs covering 423 species) dealt with North-West Africa, Canary and Cape Verde islands and is reviewed, within a general introduction to the overall work, above. The geographical coverage of the second volume includes the islands of the Gulf of Guinea and continental Africa from Senegal to Congo-Kinshasa, east to the mountains of the Albertine Rift, and south to northern Angola and the Zambian border. It thus omits most of the Zambezian region (largely, though not entirely, covered by R Stjernstedt) and eastern Africa (soon to be treated by a major new publication), whereas the southern third of Africa is covered by G Gibbon'. Virtually all the species presented here are African residents or intra-African migrants, most Palaearctic species visiting the region having already been covered in the first volume, with the exception of a few of more eastern distribution (eg Thrush Nightingale Luscinia luscinia). This is without doubt the single most important source of Afrotropical bird vocalisations, and it will have an enormous effect on the efficiency of future field investigations and on our understanding of avian relationships.
C Chappuis started recording birds in West Africa in 1968. and although this collection has involved the collaboration of more than 130 recordists, the great majority of recordings were obtained by the author and compiler, a remarkable achievement. Covering 1,043 species, these 11 CDs supersede completely the earlier collection of 11 discs (33-rpm, with 450 species) that was published in collaboration with Alauda between 1974 and 1985. More than 200 species appear commercially for the first time, and the families that had already received much space in the Alauda series (eg Pycnonotidae) have been thoroughly revised, updated and augmented with many new recordings.
Individual species are not announced by voice but are given a specific track number: this both facilitates the rapid location of any desired recording and makes confusion impossible. In a few cases, however, two species or subspecies, are presented on the same track but are separated by a long silence (as long as that between tracks), whereas separate recordings of the same species or form follow each other more rapidly. Correctly locating cuts in this way requires a little practice, but generally should not lead to confusion. (I must add that some of my British (male) colleagues (pers comm, unnamed) tell me they regret the disappearance of the human voice to identify species, especially that of the French lady on the Alauda series, whose pronunciation of Latin names they found remarkably sexy...). One species, Grauer's Cuckoo-shrike Coracina graueri, appears out of sequence at the beginning of CD11 (among the Sylviidae); while this is mentioned in the text under CD9 (Campephagidae), a cross-reference under CD11 would have been useful.
The sound quality is generally good to excellent, when less so it is for obvious reasons and it is always preferable to have a poor recording of a species than none at all. Species appearing in the background are mentioned occasionally but not consistently (no doubt due partly to lack of space). In a very few cases a secondary species is actually more prominent than the main one, and this could be misleading. Thus the final cut of Cameroon Olive Greenbul Phyllastrephus poensis (CD9, track 72) is somewhat marred by the loud interference of a Pink-footed Puffback Dryoscopus angolensis (a double chukking noise, repeated then followed by a churr), whereas the usual song of the bulbul can be heard in the distance (also with that of Black-faced Rufous Warbler Bathmocercus rufus at the beginning). The second cut of Yellow-mantled Whydah Euplectes macrourus (CD14, track 92) is, to my ears, largely taken up by a song of Northern Double-collared Sunbird Nectarinia preussi. But these problems are exceptional.
The recordings presented and the informative booklet generally succeed well in the two main aims of the work, to provide the means of correctly identifying bird sounds, and to suggest ways in which these have a bearing on avian taxonomy. Most species' vocalisations occupy between 30 seconds and one minute, and although not large this often includes several cuts of different origins illustrating song, calls and dialectal variations; for example in just one minute we get an excellent presentation of the range of dialectal variation in the songs of Sulphur-breasted Bush-shrike Malaconotus sulfureopectus (CD13, track 65) from West to East Africa. All relevant information concerning location, time, recordist and much else is given in the accompanying booklet (I have the French version, but have not seen the English). Indeed this work is much more than a gigantic compilation of field recordings (3,200 cuts have been retained here from a selection of some 5,500), as the text includes a great deal of information on means of separating difficult species, and in some cases proposes identification keys for whole series of species. One example of this, which I find works particularly well, is the key to the vocalisations of the green turacos Tauraco spp.; shame on anyone who still manages to misidentify a calling turaco after this. Attention is drawn to the vocal distinctiveness of the members of the superspecies Guinea Turaco T.persa (with Schalow's T. schalowi standing apart from all others, cf. Dowsett-Lemaire & Dowsett 1988), whereas one needs to listen with care to separate the song of White-crested Turaco T. leucolophus from those of the superspecies Bannerman's T. bannermani / Red-crested T. erythrolopus. Although in the past T. leucolophus was not considered to be particularly closely related to the last, recent DNA investigation by Veron & Winney has placed T. leucolophus immediately next to T. erythrolopus on the 'neighbour-joining tree'; this is certainly supported by bioacoustic evidence. To take an example of Chappuis' didactic approach among the passerines, the text accompanying the extensive choice of motifs of the two forest orioles Western Black-headed Oriolus brachyrhynchus and Black-winged O. nigripennis is very helpful, as distinction between these is not always straightforward, not least because of their extensive repertoires. Thus two minutes and 30 seconds are devoted to O. brachyrhynchus and two minutes and 50 seconds to O. nigripennis, probably the two longest individual presentations in this volume.
Overall, the taxonomic treatment follows that of Birds of Africa volumes 1-6, although there are a number of divergences, for example among sunbirds, monarchine and platysteirine flycatchers. Thus Bates's Paradise Flycatcher Terpsiphone batesi, separated by C Erard (in Urban et al 1997) from Rufous-vented T. rufocinerea, is still presented as a race of the latter. In my experience and that of C Erard (op. cit.), the song of T. rufocinerea sensu stricto is very similar to that of the African Paradise Flycatcher T. viridis and unlike that of T. batesi; for this reason and others (including geographical contact), C Erard's treatment was probably the wiser. Unfortunately no-one has yet managed to get a tape-recording of the song of T. rufocinerea.
The importance of bioacoustics in the field of systematics has been increasingly evident in recent decades, as exemplified by the discovery and description of new species prompted by the study of vocal characters. Examples of relevance here include Eastern Green-tailed Bristlebill Bleda notata, Dorst's Cisticola Cisticola dorsti, Pale-crowned Cisticola C. cinnamomeus and several indigobirds Vidua spp. Many problems remain unresolved and Chappuis offers pointers to more potential splits. There is a great deal of interest here, but some of this is going to be difficult to unravel. The most striking example among non-passerines is that of Dusky Long-tailed Cuckoo Cercococcyx mechowi: recordings of the eastern and western song-types were first published separately (from Uganda by S Keith in 1971 and from the Nigeria / Cameroon border by C Chappuis in 1974) but are now presented together, and a sonogram demonstrating that these songs differ not only in their tempi but in their structural form was published in 1997. The eastern song type is consistent throughout a wide range from Uganda to Congo, Gabon and southern Cameroon (published recordings and pers obs); the geographical separation between the two cuckoo populations is probably around the Sanaga River, as the eastern form is very common in south and south-eastern Cameroon, whereas the western form appears somewhere north of the Sanaga River (from the Bakossi Mountains north, pers, obs.). However, C. mechowi from West to East Africa has always been considered a monotypic species; the three presently recognised species of Cercococcyx cuckoos (Barred Long-tailed C. montanus, Olive Long-tailedC. olivinus and C. mechowi) are difficult to separate on morphological characters, even for museum specialists, and the paucity of specimen material is not going to help resolve the problem of C. mechowi for some time. A similar difficulty arises with Little Rush Warbler Bradypterus baboecala, whose high-pitched and low-pitched songs are strikingly different, but cannot easily be attributed to particular morphological races. Even though Chappuis proposes to separate the form with the high-pitched song-type under the name of B. elgonensis (from Kenya / Uganda), part of another race, centralis (but only part of it) also possesses this distinctive song, reminiscent of insect-like stridulations. Thus centralis birds from northern Tanzania (cf. Zimmerman el al 1996), Rwanda (pers obs and tape-recordings) and eastern Congo-Kinshasa (birds in song collected and well described by Chapin) all sing like elgonensis, but somewhere between eastern Congo and Cameroon birds of (apparently) the same race suddenly change, since in Cameroon all centralis I have heard from the Haul Nyong to coastal locations (as is illustrated by a tape of it by C Chappuis) produce the low pitched song.
Still among the Bradypterus, Grimes in Urban et al treated the form bangwaensis (of the mountains of Cameroon and eastern Nigeria) as a race of the more widespread Evergreen Forest Warbler B. lopezi, even though both forms coexist at Mt Manenguba; in 1989 Bob Dowsett and myself presented a case for the specific distinctiveness of Bangwa Forest Warbler B. bangwaensis, based on morphological characters (bangwaensis' rich coloration is closer to that of Cinnamon Bracken Warbler B. cinnamomeus than lopezi), partial sympatry with B. lopezi, and vocal characters (in respect only of call notes). C Chappuis presents here some of my recordings of songs from eastern Nigeria, but unfortunately not the call notes. As shown by sonograms' song motifs of lopezi and bangwaensis are similar in all their main characters, but the call notes are very different. Those of lopezi are presented on CD10 (tracks 49-50), so it is a pity those of B.bangwaensis have been omitted (just one low churr can be heard in the background of the second cut on track 51), although they have been tape recorded. Further field work in western Cameroon has demonstrated that these vocal characters are consistent throughout the range of bangwaensis; where the two species meet (Mt Manenguba) they occupy different niches, with bangwaensis at forest edges and lopezi within primary forest.
In many cases much more field work will be necessary to investigate, prove or disprove some of the ideas of species separation proposed here. One example is that of Rufous-naped Lark Mirafra africana: some unusual song types recorded on the Téké Plateau in Gabon (CD8, track 91) have induced the compiler to suggest that the local race (malbranti) may represent an entirely new species. But, we are dealing with a lark of very wide distribution in Africa and presenting a large range of dialectal forms across the continent. Individual birds already show quite an array of different motifs: in the one individual recorded by P Christy at Léconi, the first motif is certainly very unusual for a M. africana, but by the time we get to the third motif (by the same individual), we already have something more typical of the species elsewhere. Another suggestion to treat with care concerns the race leoninus of Cameroon Sombre Bulbul Andropadus curvirostris; I agree that the vocal dialect of this form (CD9, track 53) is distinctive (from that of nominate curvirostris recorded in south-west Cameroon, track 52), but dialectal variation in the songs of this bulbul is extensive, with different populations in various forest blocks in Cameroon and Congo having their own particular motifs (and these all belong to the same race).
Conversely, suggestions for reconsidering some geographical forms (treated as different species in recent volumes of Birds of Africa) within the same species are made for a number of superspecies; examples include White-bearded Greenbul Criniger olivaceus ndussumensis, Black-collared Apalis Apalis pulchra / ruwenzorii and Yellow-bellied Eremomela Eremomela icteroyigialis / salvadorii. Bioacoustic evidence is clearly in favour of lumping here; this is also supported by playback experiments, i.e. positive reactions from one race to the songs of the other. Indeed, before a good tape-recording of Criniger o. olivaceus became available, the best way of locating this discreet species in the forests of Côte d'lvoire was by playing a tape of C. o. ndussumensis (C Carter pers comm). The Apalises, A. p. pulchra and A.p. ruwenzorii possess a variety of song-types, of varying tempi, but the timbre of voice is identical in both geographical forms and fast or slow motifs in one form can be matched to fast or slow motifs in the other; thus the last (fast) song of Apalis p. pulchra (CD11, track 20) is identical to the first song-type of A.p. ruwenzorii (CD11, track 21), and their direct juxtaposition through astute editing makes this more convincing than any amount of text.
In a work of this magnitude, inevitably a few errors of transcription have crept in. Thus the song of Papyrus Yellow Warbler Chloropeta gracilirostris (CD10, track 56) is correctly identified but cannot come from the Nyika Plateau in north-east Zambia (a high montane area without swamps where the species is quite unknown); the same recording was previously published by R Stjernstedt' and was taped at the mouth of the Luapula River. If the call note of Apalis pulchra ruwenzorii (CD11, track 21) was indeed taped in Kenya, then it must be referred to that of nominate pulchra (ruwenzorii being endemic to the Albertine Rift). The second cut of Yellow Longbill Macrosphenus flavicans (CD11, track 55) from eastern Nigeria sounds like a song of Kemp's Longbill M. kempi to me; the recordist (R Demey pers comm) confirms that he did not see the bird and his tape comes from an area of overlap between the two Macrosphenus. Of the many M. flavicans I have heard in Cameroon, Equatorial Guinea and Congo, I have never found one that produced anything similar to a song of M. kempi, and observations in the Korup area of western Cameroon (also an area of sympatry) confirm that each species retains its own individual, characteristic song (Rodewald et al). This contradicts the suggestion made here that vocal barriers may break down in areas of sympatry. If Many-coloured Bush-shrike Malaconotus multicolor must lie split from Black-fronted Bush-shrike M. nigrifrons, then the fourth cut of Malaconotus multicolor (CD13, track 66), from Misaka Forest in Zambia, cannot be attributed to multicolor but only to nigrifrons. Birds from the Zambian plateau all belong to one race (manningi) and their main song-type consists of two whistles (the second longer and higher pitched, as on track 66), reproduced again under M. nigrifrons (CD13, track 67, second cut, from Mayau, Zambia).
A similar confusion has arisen with Long-billed Pipit Anthus similis (including nyassae). C Chappuis proposes to split the Zambian race nyassae from similis (without specifying how many of the other 15 races on the continent should be included in one or the other), on the basis mainly of the number of notes (3-4) per song, even though this character is recognised as being highly variable within individuals and populations. A cut from north of Lilongwe (Malawi) appears correctly under nyassae (CD9, cut 2 of track 29), but the first cut presented under similis (from west of Lilongwe, track 28) also refers to the race nyassae, which occurs throughout miombo woodland in Malawi and neighbouring countries (cf. D J Pearson in Keith et al, p, 224). In fact, all populations of similis sensu lato give songs of 3-4 notes, and it is unrealistic to try and separate forms based on the number of notes per song. We continue to believe (cf. Dowsett & Dowsett-Lemaire) that the best treatment of this complicated species to date is that of D J Pearson (op. cit.).
The recording of Eurasian Lesser Cuckoo Cuculus poliocephalus (CD7, track 7) from the Transvaal first appeared on Gibbon under that name, but it is the typical four-note song of Madagascar Lesser Cuckoo C. rochii, and the error was corrected in 1992 (Hockey el al; see also Becking 1988, who did much to clarify the status of the two small cuckoos C. poliocephalus and C. rochii in Africa). This bird (one assumes it was perhaps the same individual) advertised a territory in the Transvaal for two successive seasons from November to February. The omission of this and other records (including a specimen) of C. rochii in southern Africa from the Atlas of Southern African Birds was due to a printers error (C Spottiswoode in litt, Spotliswoode & Allan). There have been other reports of C. rochii singing in Africa during the local summer (including Zambia and Zimbabwe), thus it appears a few Malagasy birds do not return to Madagascar for the breeding season. Thus far no genuine C. poliocephalus (whose song is very different) has ever been heard to sing in Africa.
The song and calls of Forest Swallow Hirundo fuliginosa are quite different from those of any saw-wings Psalidoprocne sp., and it appears that the recording from Mt Cameroon (CD9, track 12) is none other than Mount Cameroon Saw-wing P.fuliginosa (listen to track 5, P. fuliginosa, and track 12 in succession). H. fulginosa can be very difficult to separate visually from some Psalidoprocne, but its voice completely lacks the whining quality of saw-wing calls: its most common call is a light vit, vit (reminiscent of Barn Swallow Hirundo rustica), excited or alarmed birds also give a double pritchi, and the song (given very rarely) is a fast, musical warble of Hirundo style (pers obs). Among the bulbuls, I was puzzled by the recording of Toro Olive Greenbul Phyllastrephus hypochloris (CD9, track 70) from western Kenya. I have no experience of the species, but to my ears this sounds like one of the motifs of Little Greenbul Andropadus virens: extraordinary convergence or confusion? Zimmerman et al have had similar doubts and conclude that the voice is 'not known with certainty'; L D C Fishpool (pers comm) who recently saw and heard the species in Uganda confirms that the voice is unlike that of A. virens. This problem appears worthy of further investigation. Among the sunbirds, the last cut under Collared Sunbird Anthreptes (Hedydipna) collaris (CD12, track 77) is not that species but Yellow-chinned A. rectirostris, an editorial slip I believe. This recording first appeared in the Alauda series under Lemon-bellied Crombec Sylvietta denti (third cut), but was later re-identified as Anthreptes rectirostris (pers obs). These loud, down-slurred whistles, often alternating (pee-peeuw), are very characteristic of this sunbird and facilitate identification of the species when it is calling (unseen) from the canopy. Loud call notes in sunbirds are often more useful in species identification than their more subdued songs. The loud, descending series of 4-7 whistles in Blue-throated Brown Sunbird Nectarinia (Cyanomitra) cyanolaema (well presented here, CD12, track 81) are equally diagnostic, just as the pit, pit flight calls of Johanna's Nectarinia (Cinnyris) johannae (CD13, track 14) betray the species. To finish with the sunbirds, it seems also that the advertising calls of Cameroon Blue-headed Nectarinia oritis (loud, descending tjee-tjee-tjee-tjee, very similar to those of N. cyanolaema, but usually mixed with series of double notes te-tjee-te-tjee-tetjee-te-tjee, which excludes possible confusion with the latter) have been presented under Ursula's Mouse-coloured Sunbird N. Ursulae instead (CD13, track 18, first cut).
Of seedeaters, the song of Broad-tailed Whydah Vidua obtusa from Francistown in Botswana (CD15, track 64) contains imitations of Melba Finch Pytilia melba (whereas V. obtusa normally imitates Orange-winged Pylilia P. afra) and should be referred instead to Long-tailed Paradise Whydah V. paradisaea (as confirmed by R B Payne in litt, who also points out that only V. paradisaea has a flight display, which the recorded bird was observed to have). Moreover, V. obtusa is not known from this far south in Botswana (cf. Penry). The song of Pin-tailed Widow Vidua macroura is normally, non-imitative, and I cannot identify the 'imitations' at the end of the track (as mentioned in the text) nor can R B Payne (in litt); perhaps the author meant 'motifs' rather than imitations. Finally, the second cut (song) of Stripe-breasted Seed-eater Serinus reichardi from Harare (CD15, track 75) is unlikely to be that species, as it remains completely unknown from Zimbabwe; it compares well to the song of Streaky-headed Seed-eater S. gularis recorded in South Africa by G Gibbon.
These few misidentifications or editing errors in no way detract from the enormous value of this work; on the contrary, because this important collection represents a landmark in African bioacoustic publications, one that will (and must) be widely used in the field and which will remain unsurpassed for many years to come, it is important that observers be aware of any slips. What, indeed, remains to be done after this? Species for which no tape-recordings appear to exist are listed under II in the accompanying booklet and number c70 for this volume. A few, however, have been published elsewhere (for example Oberlaender's Ground Thrush Zoothera oberlaenderi was published by Keith & Gunn, Miombo Double-collared Sunbird Nectarinia manoensis appears in both Stjernstedt and Gibbon and Scarlet-tufted Malachite Sunbird N. johnstoni in Stjernstedt). Prigogine's Nightjar Caprimulgus prigoginei has apparently been forgotten from the list of undocumented species. There is a very possible recording of its song made at the type-locality (Itombwe) by T Butynski, matched by others from Congo and Cameroon by myself, and it is a pity that this has been omitted, although confirmation of its identity is still required; a wider distribution of this tape could help solve the problem. Among,the 1,040 or so species of volume 2, the tape material can be improved for a number of species or forms: when will the songs of Black-eared Ground Thrush Zoothera camaronensis and Grey Ground Thrush Z. princei be unravelled at last? The display noise of Green-breasted Pitta Pitta angolensis reichenowi (described in the literature as identical to that of African Pitta P. a. angolensis) is not yet available for direct comparison with that of the nominate form. One would welcome a longer selection of motifs for some noisy species with a varied repertoire (eg some of the puffbacks Dryoscopus spp,). The quiet turaco-type song of Violet Turaco Musophaga violacea and Lady Ross's M. rossae are not presented (though they are less frequent than the characteristic chorus calls illustrated here), These remarks are no criticism, but suggestions of where gaps remain and further research can be directed. Some of this probably exists already in other ornithologists' unpublished recordings. New material also continues to appear: in the months preceding and following the publication of this series, several field workers obtained first recordings of some rare or local species, including Yellow-footed Honeyguide Melignomon eisentrauti. Baumann's Greenbul Phyllastrephus baumanni, Emerald Starling Coccycolius iris and Black-chinned Weaver Ploceus nigrimentum. However, for the time being and many years to come this magnificent publication will be unequalled and C Chappuis must be warmly congratulated on his own tremendous contribution, as well as the successful and arduous compilation of the work of others.
I am grateful to R J Dowsett and R B Payne for comments on a draft of this review.