This is the first major sound collection dealing with the Malagasy region. Madagascar has seen an upsurge of scientific and conservation activities since the political situation improved in the 1980s, and this has materialised in the production, inter alia, of two attractive and easy-to-use field and photographic guides of the birds (Sinclair & Langrand 1998, Morris & Hawkins 1998). Long gone are the days when access was restricted and, even in a popular reserve such as Perinet, one had to put up with policemen in civilian clothes following birders into the forest and listening to their conversations over dinner - a strange situation we experienced on our first visit in 1976. Now research and ecotourism are booming, understandably, in a region with such a fascinating flora and fauna.
The region boasts a very high level of endemism: on Madagascar alone, some 109 species are endemic, out of a total of 209 breeding species (Fishpool & Evans 2001). Despite this, and the ease of access in the last 20 years, documentation of bird vocalisations has been neglected. There are a couple of sound guides for the smaller islands, especially Herremans (2001) for the Comores and Rocamora & Sole (2001) for the Seychelles. But for Madagascar in particular no more than atmospheric 'soundscapes' had been available so far (e.g. Randrianary et al. 1997), with few species readily recognisable; an early publication by Roche (1971) included only 22 species. With a very large proportion of the local and migratory species now presented (327 species, thus 95% of the whole avifauna), this is a most welcome and much-awaited collection. The presentation of species is the same as in the previous two volumes of the series dealing with continental North and West-Central Africa (Chappuis 2000); the quality is good overall; the space devoted to each species varies from usually less than a minute, to exceptionally over three minutes (as in the case of the versatile Crested Drongo Dicrurus forficatus). A short accompanying text gives details of locality, month, and (a novelty) hints at the 'acoustic efficiency' on a scale of 0.1 - 1.0, i.e. tells us how useful the voice can be in species identification. For a few groups (e.g. the genus Coua, and the ground rollers Brachypteraciidae) keys to identification are given in an introductory section. The text could have been better proof-read as mis-spellings are frequent: Euritriorchis (for Eutriorchis), Corythornis vintsoide (for vintsioides), Zosterops kiriki throughout (for kirki), Oriola (for Oriolia), etc.
The tapes of couas contain some very interesting, and probably novel, material. The song of Green-capped Coua Coua (ruficeps) olivaceiceps is a striking, rapid, wu-kip-kip-kip-kipkif-kip-kip... (a dozen notes in c.2 seconds) followed by a low, muffled grou-grou-grou-grou-grou-grou. The high pitch and fast rhythm of the rising kip-kip series are unlike anything produced by other couas (as stressed by Morris & Hawkins 1998: 194). But the song of nominate Red-capped Coua C. r. ruficeps was practically unknown: F. Hawkins (pers. comm.), who spent two years in the western forests where this form is common, never consciously heard it; I never did either, in the two weeks I spent at Ampijoroa and Kirindy. P. Huguet recorded a song attributed to C. ruficeps (at Ampijoroa) consisting of a single, modulated, low-pitched wo-oue followed by a couple of low gro notes. This could not be more different from the song of C. olivaceiceps and lends support to the splitting of these forms as species. Surprisingly, Sinclair & Langrand (1998: 108) write that the two forms have a similar voice, but their description of the song does not correspond to either of these recordings (nor to the description of Mottis & Hawkins concerning C. r. olivaceiceps), so their statement is probably incorrect. Most couas also produce various low-pitched mewing growls, and they can be equally distinctive. Even though the text refers to the low call of Giant Coua C. gigas (in respect of that of Coquerel's Coua C. coquereli), it is not presented here, presumably through an oversight. In areas where C. gigas is common, this highly characteristic mewing growl is heard several times daily. It is almost impossible to transcribe it (Morris & Hawkins give it as weeerr-ouult), but it is unforgettable once heard and has high territorial importance. On one hot day at Kirindy Forest Station, I was playing tapes to a friend over lunch, which included this incredible growl: despite the time of day the local C. gigas immediately responded just behind us and nearly walked into the open restaurant, growling away.
Occasionally, the motif presented here is not the most characteristic of the species: the song of the Brown Mesite Mesitornis unicolor consists usually of a long series of alternating modulated whistles chuu-wee, chuu-wee, chuu-wee... (chooil-woop, whooil-woop in Morris & Hawkins 1998). The bird recorded here is undoubtedly this species, but produces only a half-heart-series of single notes, consisting just of the chuu component (descending in pitch), a tired bird apparently. The identification key to the ground rollers is misleading in respect of Short-legged Ground Roller Brachypteracias leptosomus: this bird gives long series of hollow hoop notes at intervals of c.1 per second (pers. obs., see also Morris & Hawkins 1998: 218), and not (usually) short series with three or four notes' at intervals of one in 2 - 3 seconds. When camping at Masoala and Mantadia, I heard this species call at dawn every day for long periods of 20 - 30 minutes, always at the rate of one note / second. The notes of that recorded here sound the right pitch, but are unusually spaced out, and the bird rests briefly between bouts of 3 - 4 notes: was this perhaps recorded in the afternoon? It sounds curiously unmotivated. The tape of Madagascar Red Owl Tyto soumagnei comes from Mantadia: the bird was not seen but assumed to be this species (and not Barn Owl T. alba) from the forest habitat. But we saw T. alba on the edge of forest at Mantadia, and it would be wrong to assume that these two owls are completely separated in their habitat requirements. After comparing the tape to one of authenticated T. soumagnei (by F. Hawkins), I think the recording by P. Huguet is probably of the same species, as the screech sounds rather more muffled than in T. alba. But this may also be influenced by the conditions of the recording, so some doubt must remain.
Indeed, the one quibble I have with this collection is that the two main authors (aided by a limited number of collaborators) do not seem to have quite the level of familiarity with the local avifauna that C. Chappuis evidently had with African species. The number of species included is impressive, but several misidentifications have crept in, and the repertoire of some of the commoner species could have been illustrated better. The warblers in particular have posed problems. The authors rightly stress the difficulty of separating Stripe-throated Jery Neomixis striatigula from Rand's Warbler Randia pseudozosterops as the general style of song and timbre are similar; but what they claim as R. pseudozosterops is none other than N. striatigula again, and the real R. pseudozosterops appears as Green Jery Neomixis viridis. The differences between the first two are well explained by Morris & Hawkins (1998: 272, 280): the song of N. striatigula is more complex, a long series of trilled buzzy notes that rise and then descend the scale. That of R. pseudozosterops is significantly shorter and purer, with one main phrase of pure whistles, slightly descending; the lack of blurred notes and the lack of an ascending component should eliminate confusion. The song of N. viridis, on the other hand, should never have been confused with that of R. pseudozosterops, as it consists of very high-pitched, 'seeping' notes, ending in some clicking notes (Morris & Hawkins 1998: 276 and pers. obs.).
Another error concerns the genus Newtonia: the recording of Red-tailed Newtonia N. fanovanae is not of that species but of Common Newtonia N. brunneicauda. It is possible the recordists were influenced by the illustration of N. fanovanae in Sinclair & Langrand (1998: 141), which merely shows a N. brunneicauda with a rufous tail stuck on - the artist evidently never saw a specimen of the real N. fanovanae to draw from. Compare this to the plate drawn by M. Andrews in Morris & Hawkins (1998: 277) and everyone can realise that we are dealing with a completely different bird. N. fanovanae is even more distinctive than the latter drawing suggests, as it is rather elegant, with a conspicuously long, thin bill; it feeds in the midstorey, flitting through the foliage with its red tail slightly open in the manner of an Erythrocercus flycatcher. M. Andrews' drawing is in fact a little too schematic; the silver-grey cap blends into the pale grey cheeks and does not contrast so vividly with the white throat as shown (all based on pers. obs. in Andohahela Forest, November 1999). Tapes of the real N. fanovanae are in existence and circulate widely among bird tour leaders, thus such an error could have been avoided. F. Hawkins kindly lent us his tape-recording while we were working in Madagascar, and we made further (identical) recordings at Andohahela. The voice is sweet and very distinctive, consisting of two types of repeated whistles: either a series of identical fui-fui-fui-fui-fui-fui... at the rate of c.5 / second (sweep-sweep-sweep- in Morris & Hawkins); or a series of double whistles fitiu-fitiu-fitiu-fitiu-fitiu... at the rate of 3 double notes / second (pitchi-pitchi-pitchi... in Morris & Hawkins). N. fanovanae is restricted to a few localities of true lowland forest, usually below 500 m. Although Morris & Hawkins mention it to be 'probably present in the Perinet...area' there are no confirmed records. The locality on the CD attributed to this species, Anjozorobe, is at 900-1,450 m, and from this fact alone, the occurrence of N. fanovanae there is remotely unlikely. Even though the voices of the three eastern newtonias are easy for human ears to distinguish, it may be relevant to mention here that the birds do not make such a ready distinction. As we had the opportunity of visiting a little-known section of the low-altitude forest in the east of Masoala National Park, I made special efforts to search for N. fanovanae with F. Hawkins' tape. I failed to find it but observed with some surprise that both Dark Newtonia Newtonia amphichroa and Common Newtonia N. brunneicauda occasionally reacted to the playback of the song of N. fanovanae, N. amphichroa sometimes very strongly, coming to within 1 m of the source of playback. The same reactions were obtained in Mantadia National Park, where N. fanovanae does not occur (as confirmed by F. Hawkins pers comm.). The song of N. amphichroa is similar in timbre but is much less stereotyped and more rhythmically varied; the song of N. brunneicauda is of a different timbre altogether, rather dry and consisting most often of a repetition of a double note tchika-tchika-tchika-tchika... at the rate of 3 - 4 double notes/second (and reminiscent to my ears of the metallic voice of Grey-backed Camaroptera Camaroptera brachyura). Thus the songs of N. fanovanae and N. brunneicauda are close in rhythm but not in timbre, and the reverse is true for the other species pair. There is little doubt that most (if not all) of the unverified reports of N. fanovanae from the Perinet / Mantadia region and elsewhere are misidentifications, some prompted by these surprising reactions on the part of congeneric species: even if the tape played is of the right, much sought-after species, one can actually call up another, and get very confused in the poor light of the understorey (especially if the only field guide available is that of Sinclair & Langrand 1998). On the other hand, N. amphichroa and N. brunneicauda, which are widely sympatric, do not appear to react to each other's songs. Possibly, N. brunneicauda and N. amphichroa have not 'learned' to separate the song of their rarer congener as they are also so rarely in contact with it (especially the high-altitude N. amphichroa), and some basic structural features common to all are sufficient to elicit a response. In any case, special care must be taken with both sound and visual identification aids as N. fanovanae is on the globally threatened list of species, and it is important to get the distribution right.
Of the other families, one should also mention that the recording of another very localised (and endangered) Malagasy endemic, Red-shouldered Vanga Calicalicus rufocarpalis, should be attributed instead to Red-tailed Vanga C. madagascariensis. What is presented (a low pitched 2-note warble fu-fui) is one of the dialectal song variants of the latter, which indeed does occur just north of Toliara (pers. obs., F. Hawkins pers. comm.), whence came this recording. The typical song of C. rufocarpalis consists of a two-note whistle fu-feeeeeeee (or fu-fiiiiiiie in French onomatopoeia), the second note very much louder and longer than the first, so that at a distance one may hear just the second component. This long whistle is somewhat reminiscent of that of Hook-billed Vanga Vanga curvirostris. I also have doubts about two of the Berniera (ex-Phyllastrephus) species: although I have never come across Dusky Tetraka B. tenebrosa, the calls presented do not sound like the tape F. Hawkins played to me, but more like those of Spectacled Tetraka B. zosterops (with which it can easily be confused, especially as the illustration of B. tenebrosa in Sinclair& Langrand is too pale, cf. Morris& Hawkins). And the nasal, tchac, tchac or tchep, tchep calls given under B. zosterops sound far too low-pitched for this species: they correspond instead to the typical calls of Long-billed Tetraka B. madagascariensis.
Finally, the Yellow-bellied Sunbird-Asity Neodrepanis hypoxantha sounds more like Common Sunbird-Asity N. coruscans, and the occurrence of the former has not yet been confirmed in the Mantadia / Andasibe region (the locality given here). The typical calls of the former are very high-pitched, detached, clipped metallic notes, and are not presented here (see also the sonograms published in Hawkins et al. 1997). The two species have been much confused: even the photographs taken in 1973 and 1976 on a hill c.l 00 m higher than Perinet-Analamazaotra (at Maromizaha, in fact) and mentioned by A. D. Forbes-Watson (in Collar & Stuart 1985: 363) as belonging to N. hypoxantha have been re-identified as N. coruscans (F. Hawkins pers. comm.). When R. J. Dowsett and I visited Maromizaha, we noticed that the local guides readily claim that N. hypoxantha occurs there when all we could see when we had good views was N. coruscans. Even though the locality of Maromizaha is mentioned by Morris & Hawkins (1998) as a possible spot for N. hypoxantha, this was in fact based on unverified rumours (F. Hawkins pers. comm.) and the species was not included in the local Important Bird Area biome list (IBA MG054 in Fishpool & Evans 2001).
Sometimes the sex of the singing bird mentioned in the text seems wrong: as far as we know, it is the female (not the male) that sings in the Greater Painted-snipe Rostratula benghalensis. As for Crossley's Babbler Mystacornis crossleyi, whose songs include two types of long whistle (either rising or nearly monotonous), the rising whistles are attributed here to females and the longer, monotonous ones to males. But I have seen males produce both types, and as Morris & Hawkins mention that the female plumage may be found in sexually mature males, no doubt all songs are given by males.
There is apparently a recording of the western race (or morph) of Tylas Vanga Tylas eduardi albigularis from Andasibé: if this is correct, that would be the first record of the western form in that eastern area (F. Hawkins pers. comm.); the voice does not appear to differ much from that of typical T. eduardi. The western race is only at all regular in some of the western forests, especially Andranomena (pers. obs., Hawkins 1995, Morris & Hawkins 1998); but as occasionally birds of both forms occur together, albigularis may be more a morph than a race.
Inevitably, a large proportion of the rarer migrants or vagrants have been tape-recorded outside the region. This sort of geographical shortcut may be a problem when dealing with local taxonomic forms. Thus the song and call of Little Bittern Ixobrychus minutus come from the nominate race in Europe, not from the endemic Malagasy race I. m. podiceps. The voice of Sacred Ibis Threskiornis aethiopicus was taken from birds introduced in France: although this is not specified, these birds almost certainly did not come from the endemic Malagasy race bernieri, sometimes considered a separate species. Similarly, recordings of Roseate Tern Sterna dougallii and House Sparrow Passer domesticus also come from France (thus dealing with different subspecies). As Little Ringed Plover Charadrius dubius is no more than a vagrant to parts of the region, presenting the display song of a breeding pair in France may seem superfluous. And was it necessary to include Java Sparrow Padda oryzivora, as this introduced finch became extinct in the region long ago? Although Sinclair & Langrand (1998) mention it has survived on Mauritius, it was in fact last seen there in 1892 (Diamond 1987). At least some of the space taken up by vagrants might have been more usefully devoted to improved coverage of the numerous endemic species.
Despite the few misgivings explained above (and worth spending a little time on, as they often concern threatened species), I strongly recommend this collection to anyone interested in the avifauna of the region. A second edition might, in time, include corrections and a few additions - such as Madagascar Serpent Eagle Eutriorchis astur, Slender-billed Flufftail Sarothrura watersi, the distinct song type of Neomixis striatigula pallidior (a very common bird in the dry south-west) and Bernier's Vanga Oriolia bernieri. Tapes of these exist already and should be made available in future collections. Apparently other CDs of Malagasy birds are in preparation, but for the moment this series is unsurpassed and deserves a wide distribution.
I am grateful to F. Hawkins and R. Safford for comments on a draft of this review.