This enormous book (it weighs 5.5 kg, as our postman can testify) is the seventh in a line of handbooks which have served ornithologists and birdwatchers well in southern Africa for more than 60 years.
An important change overall between the Southern African Atlas (Harrison et al. 1997), earlier versions of Roberts and this one is in the radical modifications of the sequence of families and species, and some of the nomenclature. This move may not be popular with a lot of users, for obvious practical reasons (albatrosses and shearwaters are now the last of the non-passerines, appearing immediately before the pittas!). It may also be considered unwise, as so many of these changes are still tentative.
Of some 50 introductory pages, the bulk is occupied by a detailed explanation of generic characters - quite a useful innovation. Sections on environment and biogeography are, however, reduced to four pages; even though the word 'biome' is mentioned once, it is surprisingly not defined nor used anywhere in this book. A list of endemic or near endemic species presented on p. 15 could usefully have been situated in a pan-African context. There is no general analysis of important topics such as migration, breeding seasonality or conservation. This lack of perspective is to some extent attributable to the large number of authors involved and the unevenness of the specific treatments, but one feels nevertheless that the result as a whole falls rather short of expectations.
Those familiar with the Southern African Atlas will immediately recognise the maps (basically a reduced, monochrome version of those published in the Atlas) and several of the sections are very similar - especially distribution (shortened here), habitat (without the cumbersome vegetation cum geographical bar-graphs), movements and conservation. The degree of similarity of these sections depends to some extent on the author (there have been some changes since the Atlas). Other sections developed in more detail here concern identification (morphology, confusing species), voice, general habits, food, breeding behaviour and season, measurements, and geographical variation where appropriate. Compared to the shorter preceding versions of Roberts written by Maclean (1985, 1993), one major distinction and improvement is the large number of references cited throughout the text, with the use of numerical superscripts; each species account terminates with a reference list and full citations are given at the end of the book for each species (hence a certain amount of duplication). The index gives pages only for the species account and the literature citations, whereas one can find the page with the relevant colour plate listed directly next to the species heading (although this fact is not specified in the introduction). There is no cross-reference between the colour plate and the species text, nor (more annoyingly) between the text and the full literature citations.
These are all new since those of Maclean (1985, 1993). They are apparently not aimed at aiding identification, in the way that field guide plates are (we understand from Hugh Chittenden in litt, that a Roberts field guide is in the pipeline). Non-breeding and female plumages of sunbirds, weavers, bishops, whydahs and widowbirds are not shown, nor male plumage of Painted Snipe Rostratula benghalensis, or the important, distinctive immature plumage of species such as Starred Robin Pogonocichla stellata or White-backed Night Heron Gorsachius leuconotus. Flight patterns of waders, raptors, nightjars etc. are not usually shown either, and raptors appear in varying postures not directly comparable. If indeed the plates are presented here mainly for decorative reasons, one may regret the poor quality of some. Whilst the work of G. Arnott, A. Barlow, A. Clarkson, R. J. Cook and C. van Rooyen appears to us pleasant (although Arnott s colours are generally too pale and some of van Rooyen's raptors too heavily marked or coloured, cf. the red bars of African Cuckoo Hawk Aviceda cuculoides), other plates are not a great success. Two particular artists overemphasise glossiness (even the non glossy doves, white-eyes, plain cisticolas and starlings [e.g. Wattled Starling Creatophora cinerea] are shining in an unnatural way). Three artists have a style that exaggerates feather contours, which affects particularly ducks, francolins, crakes, cranes, one of the two heron plates, pigeons, turacos, cuckoos, bee-eaters, rollers, hornbills and many passerines. Plates by K. Newman (in Maclean 1985, 1993), especially of francolins, crakes, swifts and swallows, bee-eaters, kingfishers and starlings seem to us superior to most of those redone for this edition. K. Newman's raptors, waders, larks and robin-chats were also pretty good, and it is regrettable all these were redone.
We discuss below the various sections, standardised for most species (using, for illustration, the species names and spellings utilised in this new book).
Following the species' heading, there follows an often detailed Identification section, including juvenile plumage and confusing species. In the absence of colour plates showing these various
plumages, this section will best serve someone who has a bird in the hand.
The section on Voice differs from Maclean (1993) mainly in having dropped the spectrographic illustrations. We find this regrettable, especially as interpretation of these was aided by onomatopoeic descriptions above the sonograms, thus even the non-specialist could understand them. Vocal descriptions rarely include comparisons between similar taxa: thus the fact that the songs of Eurasian and African Reed Warblers Acrocephalus scirpaceus and A. baeticatus are indistinguishable is not mentioned. Unfortunately, no reference is made to the extensive audio records published for the region by Guy Gibbon (e.g. Gibbon 1991).
The Atlas of Southern African Birds (Harrison et al. 1997) lacked a gazetteer of localities, and this serious drawback is still present here; moreover, the maps astonishingly have no coordinates nor a scale. The distribution texts are often very brief and do not usually proffer precise range limits for species with localised distributions, thus there is rarely any interpretation of the maps. For vagrants and several rare migrants, maps are presented for the first time in this work but it is often difficult to relate them to the text. For instance, the text on Acrocephalus scirpaceus clearly states that there are no records from Zimbabwe, yet the map shows two crosses in that country. The map of Cardinal Quelea Quelea cardinalis shows two crosses, but as mentioned in the text only one is acceptable, records from the Caprivi area having long been known to be erroneous (see Varden 1999). For Green Tinkerbird Pogoniulus simplex there are two records and one question mark on the map, but only one accepted record (and one query) in the text. The map of Angola Swallow Hirundo angolensis has as many as four plots, but the author of the text makes it clear there is only one known acceptable record (confirmed by C. N. Spottiswoode pers. comm., who did not see the map pre-publication). This is likely to lead to misinterpretations. All of these new maps (we are informed in the Introduction) originate from a multimedia work, produced by Gibbon (2002), and a field guide (Sinclair et al. 2002), neither of which provides any information on the sources of records or the criteria used to judge their accuracy.
Some of the species whose occurrence was rejected or treated as unconfirmed in the Atlas (Harrison et al. 1997) now appear to be accepted, without any justification being offered: one example is Eleonora's Falcon Falco eleonorae, based on the same records as in the Atlas (and by the same author), even though these have not been revised by any rarities committee, and one at least had been rejected by authorities in Zimbabwe (Hustler et al. 1990). A large proportion of references in support of records comes from unpublished sources, and often no more than the observer's name is mentioned. In some cases, e.g. Greater Frigatebird Fregata minor, one of only three inland records of this exceptional vagrant is given without any details: how can these records be interpreted by the scrupulous reader, in the absence of any published South African Rarities Committee report since 1997? Under Madagascar Cuckoo Cuculus rochii, the authors fail to mention that several records in southern Africa are based on singing birds, one of which was tape-recorded and published (by Gibbon 1991, under the wrong species Lesser Cuckoo C. poliocephalus). The possibility that a small population may well breed in southern Africa should to have been drawn to the reader's attention.
Population & Demography
This section contains much of interest on threats to survival and records of longevity, and presents many regional estimates of density. It could usefully have included a note of caution regarding these last, where figures are often based on survey techniques of doubtful accuracy, given the problems of censusing tropical avifaunas when birds are not individually colour-marked.
Movements & Migrations
For some species this is a detailed and fascinating section, supported by references and with much scope for further research. Accuracy of treatment varies somewhat between authors, and extreme dates of presence or mean dates of passage for strict migrants are not always given.
The section on Habitat is usually rather brief; an effort could have been made, perhaps, to define such expressions as 'savanna woodland,' 'open woodland,' 'mesic woodland' in terms of structure, e.g. tree height, percentage cover etc. Reference could have been made to the standard definitions used in botanical works, the most appropriate being White (1983: 44-55).
For most species, General Habits are described in detail, e.g. social structure, ecological niche and predatory behaviour.
One of the more extensive sections concerns Foraging & Food, and one can find much of interest here, especially when this information is based on local sources. Occasionally, much space is given to long lists of food types recorded in other parts of Africa, and in the case of very marginal species this may seem rather a waste of space: thus under Schalow's Turaco Tauraco schalowi (occurring only marginally in the extreme north of the region), some 50 of 58 species of fruit eaten in the montane forests of northern Malawi are listed, English and scientific names (from Dowsett-Lemaire 1988). Moreover, the dry riparian forests of the Zambezi Valley of southern Africa have virtually nothing in common with the vegetation studied in Malawi; the name of one fig species (Ficus polita) has been changed (to F. bizanae, a close relative), even though the latter tree does not occur in Malawi. Surely one line summarising fruit types and sizes would have sufficed?
Breeding, behaviour and season, is another important section, but again treatment of key aspects such as seasonality varies greatly between authors. There has been no systematic attempt at analysing breeding seasonality in terms of number of clutches laid each month (a point raised by Dowsett & Dowsett-Lemaire 2001 regarding the Atlas); yet in a country with so many observers (and a nest records scheme) one would have welcomed sub-regional quantitative analyses in a part of the continent where seasonality varies greatly.
There follow: Conservation (with attention drawn to species that have fragile populations), Moult (seasonality, with some data from outside southern Africa) and Geographical Variation (including succinct descriptions of the races recognised), this last supplemented by a Taxonomic Note in many of those cases in which opposing views have been expressed. Measurements are presented for various parameters, e.g. wing-length, mass (where weights are available), usually based on southern African material.
Each species account ends with an extremely useful list of relevant, numbered References. Here we regret that a significant proportion of the information is taken from general handbooks (e.g. The Birds of Africa, Brown et al. 1982 et seq.), where the original source is often not given or is unclear. It is not always evident that current literature in the main journals has been checked: for instance one Brown-hooded Kingfisher Halcyon albiventris is said to have moved more than 200 km in Malawi, based on Benson & Benson (1977), but a subsequent publication by the same authors in Ostrich (Benson & Benson 1979; see Dowsett et al. 1997: 156) has been overlooked, which corrected this record and showed it to be no more than a local retrap.
So far we have dealt with points that, allowing for some shortcomings, represent a very great advance in documenting the southern African avifauna. But, for a work that is going to be the standard for several decades, it is regrettable that the nomenclature at subfamily level presents many problems. Readers may acquire the impression there has been splitting of hitherto accepted species for the sake of splitting, that is to say where even the slightest doubt has been suggested, subspecies are raised in rank to species. In several cases the evidence cited points in fact in the opposite direction. One striking example is that of Bennett's Woodpecker Campethera bennettii and Speckle-throated Woodpecker C. scriptoricauda: even though Benson (1952) showed pretty clearly that these intergrade where they meet and should be considered conspecific, a view also adopted by Short (1973), they are split here, despite there having been no study anywhere questioning the work of those authorities. Green-backed Camaroptera Camaroptera brachyura and Grey-backed C. brevicaudata are treated as separate species, although several authors have shown there to be much intergradation between green-backed and grey-backed birds (see references in Dowsett & Dowsett-Lemaire 1993: 362), in what is essentially an almost continuous distribution. The splitting of the Rufous-winged Cisticola Cisticola galactotes complex is based on a study of only a few of the populations, which did not take into account the considerable variation in vocal dialects and other characters. That the scientific index is based only on the genus complicates the search for some species' accounts (although generic names used in the previous Roberts and not in this work are cross-referenced).
More serious are numerous errors of citation in the taxonomic sections. We regret having to draw attention to them here, but an offer to check this aspect of the book long before the publication date was not taken up.
These errors take several forms. The wrong citation may be given (e.g. the species heading for Orange-breasted Bush-shrike Telophorus sulfureopectus is in fact that for the race similis), or it may be missing (e.g. for Souza's Shrike Lanius souzae tacitus). Dates are sometimes wrong: Chinspot Batis Batis molitor was named by Kttster in 1836, not 1850 (the second edition of the work concerned). Dryoscopus cubla (Latham 1801) pre-dates Shaw 1809. There are errors of page number (e.g. Southern Boubou Laniarius ferrugineus savensis Pinto 1963 was named in Mem. Inst. Invest. Cient. Moc. 5: 47 (p. 19 is that of a reprint), or of volume number (Black-throated Wattle-eye Platysteira peltata cryptoleuca Oberholser 1905, Proc. U.S. Nat. Mus. was vol. 28: 913, not vol. 29).
Some names have been attributed to the wrong authors; for example, Cardinal Woodpecker Dendropicos fuscescens camacupae was described by Bowen in 1930, and not by Roberts. Natal Francolin Pternistis natalensis neavei (Mackworth-Praed 1920) is not a Roberts name. Rockrunner Achaetops pycnopygius was named by P. L. Sclater alone, not together with Strickland. Yellow-throated Longclaw Macronyx capensis colletti Schou was proposed by a German called Schou and not by the Belgian Schouteden (Clancey et al. 1987: 29). European Roller Coracias garrulus semenowi was named in 1902 by Loudon& Tschusi (full family name: Tschusi zu Schmidhoffen), and not Tschudi (who died in 1889).
There is inconsistent use of parentheses where authors have or have not used the current genus in the original description: Mosque Swallow Hirundo senegalensis monteiri Hartlaub 1862 was named in the genus Hirundo, so no parentheses should be placed around the author's name, whereas Cape Glossy Starling Lamprotornis nitens culminator (Clancey & Holliday 1951) was named in Lamprocolius, thus parentheses are required.
Regarding type localities, that for Acrocephalus baeticatus cinnamomeus is north of Lake Albert-Edward, which is the old name for Lake Edward, not Lake Albert. Errors of date here involve Cape Crow Corvus capensis amongst others (the date of type locality restriction being 1954, not 1951).
The International Code of Zoological Nomenclature has not always been respected in the spelling used. Some examples: Black Saw-wing Psalidoprocne holomelas must not be changed to P. holomelaena (see David & Gosselin 2002a: 41, no. 197), nor holomeleana (as in the Taxonomic Note). In Black-faced Babbler Turdoides melanops querulus, because Turdoides has to be treated as feminine (David & Gosselin 2002b: 281), querula should be used. Among other inconsistencies in the adjectival gender endings of races, several concern the masculine genus Tchagra, which should all end in -us, in agreement. In Wailing Cisticola Cisticola lais monticolus, -cola is the correct form, as this is a noun phrase (see David & Gosselin 2002a: 34, no. 137). Giant Kingfisher Megaceryle maxima is correct (not maximus), as the genus is feminine (even though Ceryle is masculine; David& Gosselin 2002b: 268). Red Phalarope Phalaropus fulicarius is correct, not fulicaria (David & Gosselin (2002a: 17). Yellow-mantled Widowbird Euplectes macrourus should be macroura (see David & Gosselin 2002a: 42, no. 209). Sabota Lark Calendulauda sabota vesey-fitzgeraldi is incorrect (hyphens are not permitted in zoological names). Red-capped Lark Calandrella cinerea niveni was named for Mrs C. K. Niven (not Mr Niven), thus must be corrected to nivenae (as already noted by Clancey et al. 1987: 19).
An attempt has been made to explain the meaning of the scientific names, an interesting innovation that has long existed in Australia (e.g. Cayley 1971). However, a number of mistakes have been introduced, of which we mention a few here. In Grey Penduline-tit Anthoscopus caroli 'Caroli' is not a corruption of Charles, but based on the original Latin Carolus, caroli meaning 'of Charles' (Andersson). Wire-tailed Swallow Hirundo smithii was named after Lt.-Col. Charles Hamilton Smith (1776-1859), and not Andrew Smith. Groundscraper Thrush Psophocichla litsitsirupa is not based on a name meaning 'ground scraper,' but is onomatopoeic (Cole 1984). The origin of the name Karoo Scrub-robin Cercotrichas coryphaeus has nothing to do with the goddess Diana; it is masculine, Le Coriphee being the leader of a group.
Although taxonomic treatment at the subspecific level is always to a great extent subjective, undoubted errors are to be found in the new Roberts. Levaillant's Cisticola Cisticola tinniens brookei Herremans et al. 1999 is pre-dated by C. t. elegans (Hartlaub & Finsch 1870), also from the Cape (its identity confirmed by Lynes 1930). Eastern Clapper Lark Mirafra fasciolata deserti (Roberts 1926) is pre-dated by Mirafra fasciolata damarensis Sharpe 1875, and so becomes a synonym.
Although these are points that will concern few readers, nomenclatural accuracy is important in serious ornithology, and such errors are likely to be perpetuated in an apparently authoritative book such as this.
To conclude, this is the most important single work to come from the one part of the African continent which has a long history of ornithological study, an abundant literature and numerous observers in the field, both professional and amateur. It also contains a large proportion of the entire Afrotropical avifauna, and so the book has an importance far beyond the confines of southern Africa. The greatest value of this book is to be found in the detailed accounts of well-studied endemics and other species characteristic of the subregion, with an abundance of referenced details, especially on food and breeding behaviour. From this review it will be gathered that we are disappointed that it does not do full justice to the knowledge available. Whilst for many species it does provide a very useful bibliography largely lacking in earlier Roberts, it is left to the cautious reader to make fuller use of this resource.
Concerning the taxonomic errors mentioned above, a 15-page .pdf is available on request from Dowsett@aol.com. This was made available to the editors of Roberts, and we understand that some (though not all) will have been corrected in a reprint.