DURBAN MUSEUM NOVITATES

Transcription

1 DURBAN MUSEUM NOVITATES Issued by the Durban Museum, Durban, South Africa V o l. V I I I I s s u e d 31s t O c t o b e r, 1968 Part 14 GEOGRAPHICAL VARIATION IN THE YELLOW-BREASTED APALIS A P A L IS FLAVIDA OF AFRICA by W A L T E R J. LAW SON (Scientific Officer, Durban Museum, Durban) Introduction: Revisionary notes on the austral races of Apalis flavida (Strickland) have already been published b y the author (Lawson (1961)) and a further note, b y the same author, on this species describing a new subspecies appeared at a later date (Lawson (1962)). Subsequent to the publication of these notes on the southern races of A.flavida, papers on some of the taxa concerned from the hands of other workers appeared in the periodic literature. In the first of m y communications (Lawson (1961)) seven subspecies were dealt with in southern, and eastern Africa, these being A.f.flavida (Strickland): Ngamiland, A.f.neglecta (Alexander): Zam besi River, A.f.florisuga (Reichenow): eastern Cape Province, A.f.lucidigula Lawson: Farm "M alamala, near Newington, eastern Transvaal, A.f.tenerrima Grote: Mikindani, south-eastern Tanzania, A.f.golzi (Fischer & Reichenow): Great Arusha, Tanzania, and A.f. flavocincta (Sharpe): Athi River, Kenya. In the subsequent note (Lawson (1962)) a further subspecies, A.f.canora Lawson: Sumbu, Zambia, was added to the list. Both W hite (1962) and Irwin and Benson (1966) reject A.f. lucidigula and A.f.canora, placing A.f.lucidigula as a synonym of A.f.neglecta and A.f.canora as a synonym of A.f.tenerrima. Clancey (1967) also expressed doubt as to the validity of A.f.lucidigula but

2 200 Geographical Variation in the Yellow-breasted Apalis admitted A.f.canora as a valid race, postulating good size differences to separate it from tenerrima. In the light of the criticisms of some of the original findings of the present author (Lawson 1961, 1962) by other workers, it was decided that the question of geographical variation in A.flavida should be re-examined, utilising considerably more material than examined initially, and bearing in mind the comments and criticisms of other workers, none of whom, however, attempted a complete review of the species geographical variation. Over six hundred specimens were examined. Material w a s, assem bled. in the, Durban Museum from various other centres for the purpose of this, study. I am deeply indebted to the Directors of the following museums for the loan of material to supplement that in the Durban Museum: East London'M useum (through C. D. Quickelberge); Kaffrarian Museum, King W illiam s Town (through C. J. Skead); Natal Museum, Pietermaritzburg (Dr. J. A. Pringle); South African Museum, Cape Town (through Prof. J. M. W interbottom ); State Museum, W indhoek (through P. J. Buys); Transvaal Museum, Pretoria (through O. P. M. Prozesky); National Museum of Zambia, Livingstone (C. Cross); Museu Dr. Alvaro de Castro, Louren^o Marques (Prof. Dr. A. X avier da Cunha); National Museum of Rhodesia, Bulawayo (through M. P. Stuart Irwin); Instituto de Investigac;ao Cientifica de Angola, Sa da Bandeira (through Dr. A. A. da Rosa Pinto); British Museum (Natural History), London (through Mrs. B. P. Hall and I. C. J. Galbraith); Museum National d Histoire Naturelle, Paris (through Dr. J. Dorst); Zoologische Museum Alexander Koenig, Bonn (through H. E. W olters); Institut fur Spezielle Zoologie und Zoo- logisches Museum, Berlin (through Dr. G. Mauersberger); Academ y of Natural Sciences of Philadelphia (through Dr. J. B ond); American Museum of Natural History, New Y ork (through Dr. D. A m adon); Carnegie Museum, Pittsburgh (through Dr. K. C. Parkes); Field Museum of Natural History, Chicago (through M. A. Traylor); Los Angeles County Museum of Natural H istory (through Dr. K. E. Stager); Museum of Comparative Zoology, Cambridge (through Dr. R. A. Paynter); Peabody Museum of Natural History, New Haven (through Mrs. E. H. Stickney) and the United States National Museum,.Washington (through Dr. R. L. Zusi). Characters of the subspecies: In assessing the geographical variation numerous characters and combinations of characters are used to distinguish the various subspecies from one another.

3 by Walter J. Lawson 201 One of the principal characters used is variation in plumage coloration. In A.flavida this affects the yellow breast-band, both in its intensity and the degree of admixture of olive-green, as well as the extension of the yellow on to the white throat. The presence of white or yellow under tail-coverts is likewise important, as well as the amount of yellow on the outermost tail-feathers. The extent of grey on the head-top, whether narrowly restricted to the forehead, broadly dispersed towards the crown, or enveloping the entire crown and nape, is also utilised as a subspecific character. Variations in the green coloration of the mantle have not been used as a subspecific criterion in the case of the southern populations as these are slight and not considered to be sufficiently constant. The extent and development of the black "sp o t in the middle of the chest immediately below the yellow breast-band is also an ancillary character which is utilised, though care must be taken as faulty arrangement of the feathers when the skin is prepared, can either almost obliterate it or else extend it into a prominent bar across the lower breast. The yellow coloration of the breast-band and the admixture of olive-green to it are utilised in the south of the species range, as there it is a fairly constant character, but in eastern Africa it is most variable and is less reliable. In the East African subspecies the females generally have far less admixture of olivegreen to the yellow breast-band than do the males, and the amount of admixture varies. Two measurements have been taken from the specimens available, these being of the wing and tail. In addition, an index figure derived by dividing the wing-measurement b y that of the. tail for each specimen has been utilised. This figure reflects the relative sizes of the wing and tail and immediately reveals any alterations in the lengths of the wing and tail relative to one another. The colour differences between some of the subspecies are more pronounced than between others, and for this latter group the standard deviations of the measurements have been calculated and Student s t test applied to the means, in order to determine the probabilities of the samples to which the means pertain having been drawn from the same population. The 95 per cent, rule is applied for significance.' Only males have been measured. Females were also examined, but as young males and young females can, in some taxa, resemble adult females it was decided not to use female measurements as they could well be inaccurate due to the inclusion of data -from sub-adult birds.

4 202 Geographical Variation in the Yellow-breasted Apalis Discussion: The populations of this small warbler throughout its African range exhibit considerable geographical variation, and a thorough analysis of this has revealed the presence of no less than twelve valid subspecific groupings of populations. The variation in size and coloration exhibited b y the series examined has shown that the use of such variables in the delimiting of formal subspecific units has, however, to be approached with some caution. A great deal of variation exists in A.flavida and it would be most unwise to regard much minor and local variation as worthy of subspecific weighting. I have attempted to regard only the principal variants as subspecies and others as minor or local character trends, these often the result of intergradation between well-marked vicinal forms. Intergradation between the subspecies, as one would expect, is extensive and in cases where these differ widely the hybrids can assume distinctive facies of their own. In such cases I have preferred to regard them simply as intergrades and not give them formal subspecific rank, as indeed some workers might do, as they could hardly be regarded as incipient species in that form. Intergradation between races also affects the measurements recorded, as the measurements of the individual subspecies given must of necessity include some taken from birds collected in an area of intergradation, and in this way the differences between subspecies can be partly obscured. To leave out the intergrades from one s reckoning would leave a reviewer open to charges of falsifying the measurements. The most southerly populations range from the eastern Cape northwards through Pondoland to southern and interior Natal. To these populations the. name A.f.florisuga (Reichenow) is applied. This race is most distinctive and differs from all others in southern Africa in that the males.and females lack a black spot beneath the yellow breast-band. The males in all the other subspecies in southern ^frica have this black spot in varying degrees of prominence, though it is absent in the female. : The populations in Natal, from approximately Durban northwards to Zulularid, Mozambique, Malawi and Zambia were previously regarded as A.f.neglecta (Alexander) (Lawson 1961), but it has now becom e apparent that the allocation of all these populations to the latter is inaccurate. Traylor (1963), W hite (1962) and Irwin and Benson (1966) include the populations allocated to A.f.canora and A.f.lucidigula by Lawson (1961 & 1962) in the range of A.f. neglecta as well. A knowledge of distribution and subspeciation patterns in birds in the southern part of Africa should at once suggest that the possibility of populations of the eastern lowlands

5 by Walter J. Lawson 203 of south-eastern Africa being genetically similar to those of the interior in parts of Angola, Congo and Zambia is unlikely, even though the eastern coastal populations are currently regarded as being identical with those of Zam bia by workers (viz. A.f.neglecta). The effect of any gene flow between populations so distantly separated in this sedentary species must be almost non-existent. Examination of the material suggests that tw o distinct groupings of populations comprise the A.f.neglecta of Lawson (1961), these being the populations of Natal, Zululand and Mozambique south of the Zambesi R iver on the one hand, and those of the upper and middle Zambesi River, Malawi, Zambia and eastern and western Rhodesia on the other. These two groupings are essentially similar in coloration but there is a distinct tendency for the interior populations to have the yellow breast-band clearer yellow, less suffused with olive-green, a feature which is not constant, however, and in having the black breast-spot somewhat smaller and more restricted to the edge of the yellow breast-band than do the coastal populations, which have a well pronounced black spot. In some specimens this latter almost attains the proportions of a bar of black under the yellow. Generally speaking, the wing-measurementr of the coastal populations average smaller than those of the interior, whereas the tails are generally longer. The differences in wing-measurements and the wing/tail index are significant, but differences in the tail- lengths are not (see Table I). The type-locality of A.f.neglecta is generally given as Zambesi R iver, but Mackworth-Praed and Grant (1963) restrict it to Zum bo, western Tete district, Mozambique. The populations of eastern and north-eastern Zambia, Malawi, the. middle and upper Zambesi R iver and eastern Rhodesia must therefore bear this name. No name exists for the coastal populations, a.nd for these Apalis flavida renata mihi is proposed hereunder. In series the differences between A.f.neglecta and A.f.renata are readily discernible. Examination of material of A.f.renata shows that there is some measure of individual variation within its ranks, as indeed there is within the populations grouped under A.f.neglecta, but it is preferred to regard these as minor variational trends within the formal subspecies. The populations of the eastern Transvaal, adjacent Swaziland and the southern and eastern Rhodesia lowveld differ from A.f.neglecta and A.f.renata in having the breast-band a clear, bright yellow, with no admixture of olive-green whatsoever. The black spot on the breast is smaller than in A.f.renata but essentially the same size as

6 204 Geographical Variation in the Yellow-breasted Apalis in A.f.neglecta. For these populations the name A.f.lucidigula Lawson is available. The tail of A.f.lucidigula is somewhat longer than in A.f.neglecta, but the principal difference between it and all other subspecies lies in the clearness of the yellow of the breast-band. The differences in the tail-measurements and the wing/tail index between A.f.neglecta and A.f.lucidigula are significant. In Lawson (1961) the range of A.f.lucidigula was extended from the limits given above to "the Zambesi River in the vicinity of Kariba Lake. However this study has shown that this is not so. The populations assigned to A.f.lucidigula do not com e into contact with those of the Kariba and middle Zambesi River areas as the country between these tw o areas is unsuitable for this Apalis and no gene flow exists between them. Nevertheless, the populations in the upper Zambesi River area do resemble A.f.lucidigula, especially in the clearness of the yellow breast-band. These birds are now considered to be intergrades between A.f.neglecta and A.f.flavida, and though superficially similar to A.f.lucidigula have originated from an entirely different source, and should not be referred to the latter. A.f.lucidigula is localised when compared with some of the wider- ranging subspecies, but it is doubtful whether the range is any more circumscribed than in A.f.florisuga, A.f.pugnax and A.f. viridiceps. In addition, the nature of the difference from the contiguous populations is such that it represents a distinct evolutionary tendency on the part of these populations, and has not originated from intergradation between any of the contiguous subspecies. A.f. lucidigula m ay of course represent the intergrade between A.f.flavida and A.f.renata, if one assumes that the range of A.f.flavida extended further east than at present under moister climatic conditions. There is a tendency for populations of some avian subspecies occurring in northern Botswana to extend eastwards, their influence being seen as far east as the interior of Inhambane district, coastal Mozambique. As already mentioned, the intergrades between A.f.flavida and A.f.neglecta are similar in coloration to A.f.lucidigula, but I prefer to regard the populations of this species with the blight yellow breast-band and no suffusion of green resident in the eastern Transvaal, eastern Swaziland and southern and eastern Rhodesian lowlands as a distinct subspecies, A.f.lucidigula. In Zululand and Swaziland A.f.lucidigula and A.f.renata occur on different sides of the Lebom bo Mountain range, A.f.lucidigula being on the west. However there is a certain amount of movement through the river valleys cutting through this range and this results in some degree of m ixing of the subspecies in this area.

7 by Walter J. Lawson 205 As shown above, the populations of the greater part of Zambia are regarded as A.f.neglecta. However, in the extreme north-western portions of the Western and Northern Provinces, and also in northeastern Angola and the south-eastern Congo occur birds which differ markedly and to which the name A.f.canora has been given. In A.f.neglecta the grey of the head is confined to the forehead, the crown and nape being green, i.e., concolorous with the mantle, whereas in A.f.canora, though the extent of grey is variable, there is a consistently greater development of this on the head. In most specimens the entire head-top and nape is grey, the latter always so. A.f.canora resembles A.f.flavida in this respect, the latter also having the extent of grey on the head-top somewhat variable. This variation in the amount of green extending distally on to the grey head-top is present in all localities, and does not represent a cline from one extreme to another as far as I have been able to ascertain. The tail of A.f.canora averages a little shorter than that of A.f.neglecta. This m ay be due to some influence from A.f.caniceps, which is a short-tailed race. Only the wing/tail index is significantly different between A.f.canora and A.f.neglecta. Irwin and Benson (1966) place A.f.canora as a synonym of A.f.tenerrima Grote, but it differs from this race in its larger wing- and tail-measurements as demonstrated by Clancey (1967). A.f. tenerrima likewise has the tendency to have the grey on the nape infused with green from the mantle, and the tw o subspecies are consequently rather similar to one another in coloration; all the measurements are significantly different. The nominate race of A.flavida comprises populations occurring in northern South-W est Africa, south-westem, southern and western Angola, south-eastwards to Ngamiland, northern Botswana and the adjacent areas of southern Barotseland, south-westem Zambia. The male of A.f.flavida has a very prominent black breast-spot and is further characterised b y having the yellow of the breast-band extending on to the throat and with a marked tendency for the grey of the forehead to invade the crown of the head. The sides of the body, flanks and under tail-coverts are washed with yellow, and the ventral surface of the tail exhibits prominent yellow edging to the inner webs of the rectrices. There is a certain degree of coloration variation exhibited in the populations of A.f.flavida, but I find that this is not geographically delimited. In eastern and south-easlem Tanzania and northern Mozambique occur populations named A.f.tenerrima by Grote, which differ from those of A.f.neglecta and A.f.renata in having the head-top

8 206 Geographical Variation in the Yellow-breasted, Apalis entirely grey, and not only the forehead grey. As shown above, A.f.tenerrima and A.f.canora are alike in colour, but differ sharply in size; they are also not in contact. All the critical measurements of A.f.tenerrima, A.f.renata and A.f.neglecta are significantly different. The Type of A.f.tenerrima from Mikindani, in south-eastern Tanzania, was examined by me in the Durban Museum through the courtesy of Dr. G. Mauersberger. In the interior of Tanzania, excluding the low-lying eastern areas where A.f.tenerrima occurs, occur populations to which the name A.f.golzi (Fischer & Reichenow) is applicable. This taxon differs from A.f.tenerrima in having the under tail-coverts yellow, not white, and it is also larger in size of the wing and tail. It likewise differs from A.f.canora which has white, not yellow, under tail- coverts. The outermost rectrices of A.f.canora, as in A.f.tenerrima, are tipped terminally with yellow, whereas in A.f.golzi the entire visible portions of the feathers are yellow. The head-top and nape are entirely grey in A.f.golzi. This is a very stable character and only varies where the race concerned meets adjacent forms and inter- grades with them. Both A.f.canora and A.f.tenerrima are variable in this respect, as has already been noted. I have examined the Type of A.f.golzi, which was collected at Great Arusha, Tanzania. The only way in which it varies from the populations now bearing this name is in the coloration of the headtop, which has a few green feathers on the nape and crown. This is because the Type is from an area where there is some discernible intergradation between the populations now called A.f.golzi and those of A.f.pugnax. The name Apalis niassae was proposed b y Reichenow, in 1921, for populations of birds which we now regard as being a constituent part of the species A.flavida, having been named on the basis of specimens collected at Langenburg, north of Lake Malawi ( = Lake Nyasa). I have been fortunate enough to examine five specimens from northern Malawi (from Chinteche (two), Mzimba, D ow a bom a and Songwe) as well as the Type from Langenburg and three other specimens from the type-locality. The populations of Malawi are usually regarded as being A.f. neglecta, though the northern populations have been placed as A.f.golzi (Lawson 1961). All five of the above listed specimens from northern Malawi have the head-top and nape entirely grey, in which respect they resemble A.f.golzi, A.f.tenerrima and A.f.canora. H ow ever all have a few green feathers intruding into the proximal grey surface. One of the specimens from Chinteche (Transvaal

9 by Walter J. Lawson 207 Museum No ) has the under tail-coverts yellow, whereas the other from the same locality (T.M. No ) has white under tail- coverts. The specimen fiom Songwe, on the Malawi/Tanzania border, close to the type-locality of A.niassae (British Museum No ), has the under tail-coverts tinged with yellow, whereas the specimens from Dow a bom a (B.M. No ) and Mzimba (B.M. No ) have white under tail-coverts. The outermost rectrices of all these specimens are a very pale yellow, not the bright clear yellow as in A.f.golzi, yet not the olivegreen with yellow tip and yellow outer web as in A.f.neglecta, A./. canora and A.f.tenerrima. They fall within the size-range of A.f.golzi, A.f.canora and A.f.neglecta but are too large for A.f.tenerrima. All have the breast-band with only a small suffusion of olive-green, in which respect they differ from A.f.golzi and A.f.tenerrima which have a heavier suffusion of olive-green. All are more yellow-green on the mantle than in these races and the head-top is a paler grey. In his original description of A.niassae Reichenow noticed these last characters, as he wrote Der Apalis neglecta sehr ahnlich, aber das Grau des Kopfes heller, das Gelbgriin von Rucken und FliigeJn heller und gelber. The Type of A.niassae has the head-top and nape grey, with only very few green feathers over the hind-crown. The under tail-coverts are white and the rectrices, like the five other specimens dealt with above, are a very pale yellow. I have also examined a further three specimens from Langenburg collected b y Dr. F. Fiillebom, who collected the Type of niassae. Tw o specimens are females, both with entirely grey head-tops and no green admixture, one of them with white under tail-coverts, the other with yellow tinged under tail- coverts. The third specimen is a sub-adult bird, originally sexed as a male but later changed to a female. It has yellow under tail-coverts. Specimens resembling in part those discussed above are also found further south in Malawi. A specimen from Messenguese, Zambesi River, Mozambique (National Museum of Rhodesia N o. 9194) and another from Port Herald (N.M.R. No ) have the outer remiges considerably yellower than other specimens of A.f. neglecta in whose range they fall. The green on the inner web of the remiges is likewise yellower. Another specimen from 60 miles up the Zambesi River from Feira (N.M.R. No ) has the head- top and nape entirely grey, with a few green feathers intruded into the crown of the head. This specimen was collected very close to the type-locality of A.f.neglecta. Another from Port Herald (N.M.R. No ) also has the grey of the head-top extending on to the nape,

10 208 Geographical Variation in the Yellow-breasted, Apalis though there is a little more green on the crown than in the specimen from near Feira. Typically A.f.neglecta and A.f.renata never have the nape grey, always green. A female from Port Herald (B.M ) has the under tail-coverts yellow. A further eight specimens have been examined from Malawi; all have the grey of the head restricted to the forehead and crown, as has the last mentioned Port Herald specimen, and in this respect are typical of A.f.neglecta. However four of these specimens (male and female from Fort Johnston (B.M and 585), a female from Mzimba (B.M ) and a female from Namwera (B.M )) have the grey o f the head-top paler and the mantle a more yellow-green than typical A.f.neglecta, in which respect they resemble the specimens of A.niassae. The four other Malawi specimens are typical A.f.neglecta (being from Blantyre (B.M ), Mlanje bom a (B.M ), Sombani River, Mlanje (B.M ) and Fort Johnston (B.M )). These specimens all have the breast-band a clear yellow, not heavily suffused with olive-green as in A.f.golzi and A.f.tenerrima. The question now arises as to which name should be applied to these populations. The name A.f.niassae Reichenow is available. The five specimens from the northern areas of Malawi and the type series have a combination of characters which distinguish them from all the contiguous populations of A.flavida and five specimens of the nine others examined from Malawi have certain characters in com mon with the northern ones. These populations show characters which are clearly intermediate between the contiguous populations, namely of A.f.neglecta and A.f.golzi. It is doubtful whether the Malawi populations com e into contact with A.f.renata, A.f.tenerrima or A.f.canora. Initially (Lawson 1961) the tw o Chinteche specimens were placed as A.f.golzi, but I now consider that the Malawi specimens should not be afforded independent subspecific rank but should be allocated either to A.f.golzi or A.f.neglecta, depending on which they most closely resemble, as representing intergrading populations between the two. I do not believe that A.niassae is an earlier name for A.f.canora, even though the Type m ay resemble the norm of the stated race. The Type was collected from populations which are clearly intergrades between A.f.golzi and A.f.neglecta, and the name niassae should not be applied to the distinct populations occurring further west (A.f.canora), and with which the intergrades A.f. negleda^-a.f.golzi are not in contact. A further feature which distinguishes the populations of A.f.golzi from the austral populations is that both males and females have a black spot under the yellow breast-band. Intergrades between

11 by Walter J. Lawson 209 A.f.golzi and A.f.flavocincta, which latter has no black spot in either sex, m ay lack a black breast-spot, and immature birds also lack this character. The spot in the female m ay vary from prominent to incipient, but it is always present though smaller than the very prominent one which is displayed by the male. The breast-band of the female of A.f.golzi is a considerably clearer yellow than the male, which has a strong admixture of olive-green in the yellow. To the north of the range of A.f.golzi and A.f.tenerrima in Kenya, Somalia and southern Ethiopia occur tw o quite distinct gioups of populations. Those occurring in the highlands of Kenya have the grey on the head entirely confined to the forehead, with green on the nape and hind crown, a yellow-green mantle, a tail which is shorter in length than the wing, both sexes with a black breast-spot beneath the yellow breast-band, and with yellow under tail-coverts. In the eastern lowlands of Kenya north to Somalia, the Northern Frontier District of Kenya, and southern Ethiopia occur populations characterised b y having the head-top entirely green, any grey present being restricted to a very small patch above the lores, the mantle a dull greyish olive-green, the tail-length longer than that of the wing, neither male nor female with a black breast-spot, and with white under tail-coverts. The question of what names must be applied to these now arises. The Type of A.f.flavocincta no longer appears to be in existence (Mauersberger pers. comm.), and the original description, like the original descriptions of most of the subspecies of this apalis, does not help in defining the subspecific characters. Sharpe says,... unterschieden durch ihr weissliches Augenlid und durch die schmutzig olivengelben und dunkel aschfarbenen Ohrdecken. These features do not help at all and do not appear to be valid in any event. Confusion has existed as to which populations this name should be applied to. It is therefore necessary to examine all the important relevant literature in order to decide. The type-locality, Athi River, is sufficiently vague to be of little help, as this river flows through the ranges of both the subspecies in this area. Fischer and Reichenow (1884), in dealing with A.f.golzi and A.f.flavocincta, remarked that the only available specimen of the latter, namely the Type, was a young bird with a green head-top, no mention being made of any grey. Sharpe, in his original description, clearly did not regard the Type as being a young individual. However, Fischer and Reichenow were clearly looking at a green-headed bird for them to decide that it was a sub-adult, as the sub-adults of the highland Kenya populations and A.f.golzi have the head-top entirely green.

12 210 Geographical Variation in the Yellow-breasted Apalis The lowland Kenya populations are green-headed in both adult and juvenile. Neumann (1900) in remarking that Die echte flavocincta scheint das siidliche Somali Land zu bewohnen, quite clearly regarded flavocincta as being the green-headed, long-tailed subspecies. He also maintained that Reichenow s illustration of flavocincta in his Vogel Deutsch Ost-Afrikas was in fact A.f.golzi as it had a grey head. Reichenow (1905) cites tw o localities for A.flavocincta: the type-locality Athi River, and Kibuesi. Neumann (1906) likewise must have considered flavocincta to be green-headed, because in his original description of A.f.malensis he states, Am meisten der Apalis flavocincta Sharpe von Ukamba gleichend and gives its range as "U kam ba, vermutlich auch Teita und siidliches Somali Land. Zedlitz (1916), in naming A.f.neum anni of southern Somalia, quite clearly allies it with malensis and flavocincta. Both neumanni and malensis have all-green head-tops. He also seemed doubtful whether the Type of A.f.flavocincta was a young bird, and like Reichenow (1905) refers the specimen from Kibuesi to that race. Looking at the older literature, as summarised above, one finds that the name Aflavocincta is generally given to the green-headed, long-tailed birds, with the exception of the illustration by Reichenow, who obviously confused the totally green and grey-headed birds, as suggested by Neumann (1906) and Van Someren (1932). From the early 1920 s to the present tim e the literature has become somewhat confused on the allocation of names to the East African populations. Van Someren (1922) arranges his material of this warbler under tw o headings, chiefly on the length of the tail. He does not, however, give any measurements. Mine show that there are in fact three groupings if one uses the tail-lengths. It is clear, however, that the birds he calls A.f.flavocincta are essentially the Kenya highland birds, with only the front part of the head-top grey. His A.f.golzi, from the range and description he gives, include what we now regard as the populations comprising A.f.tenerrima. However, ih 1932 Van Someren again regards the partly grey-headed birds as A.f.flavocincta, which he considered as being a distinct species. His arguments are somewhat confused and do not get close to the true state of the geographical variation as I now see it. In retrospect, with some of. the same specimens which earlier workers had, as well as. recently collected material, all assembled before one, one can see exactly how the confusion arose. However the important point is that Van Someren (1922 and 1932) used the name A.flavocincta for the partly grey-headed populations from the Kenya highlands and not the totally green-headed ones. Sclater (1930), probably following Van Someren, likewise accorded the

13 by Walter J. Lawson 211 populations of the highlands of Kenya the name A.f.flavocincta. Friedmann (1937) however used this name for the totally greenheaded birds with no black pectoral spots. Jackson (1938) followed Van Someren and Sclater in giving the populations of the central highlands of Kenya the name A J.flavocinda. Mackworth-Praed and Grant (1955) did likewise. In defence of the earlier workers, though.they had far fewer specimens to work with, it must be conceded that they did examine the material available in many instances and did not depend entirely on the literature, which many authors of recent faunal works have relied on, I have examined the specimen from Kibuesi referred by Reichenow (1905) and Zedlitz (1916) to A.f.flavocincta:. It is a male collected by 0. Neumann on 13th December, It has a totally green head-top and the tail-feathers are similar to those of other green-headed birds from this area, in not being as' yellow as in the highland" birds, though the under tail-coverts are a very pale yellow. Kibuesi is in an area of intergradation between these tw o population groups. The fact that the type-locality is cited as the Athi R iver in Ukamba also suggests, from the geographical point of view, that the name A.f. flavocincta should be applied to the green-headed birds and not the partly grey-headed ones as proposed by Van Someren and subsequent authors (with the exception of Friedmann 1937). Dr. G. Mauers- berger, in attempting to trace the Type of flavocincta, informed me that in the catalogue of Type specimens in the Zoologisches Museum, Berlin, he found a note to the effect that it is probably a mounted bird (No ) from Kitui, in Ukamba. Unfortunately, he cannot find the specimen, though it is not impossible that it m ay yet be found. I have examined specimens from near Kitui and they are all green-headed with long tails. In consequence of these findings, I propose that the name A.f.flavocincta be used for the totally greenheaded, long-tailed populations with no black breast-spots, and not the Kenya highland birds. As the highland populations are innominate the name A.f.pugnax mihi is proposed below. A.f.golzi, of interior Tanzania, has the headrtop entirely g rey to the nape, the male and the female have a black breast-spot, the under tail-coverts are yellow, and the outerm ost rectrices are entirely yellow, the adjacent ones with the outer web yellow and the balance with a broad terminal yellow tip. The wing-length is longer than that of the tail. A.f.pugnax has the grey on the head confined to the forehead, the crown and nape being concolorous with.the mantle, i.e., a yellow-green, and the tail and black breast-spot characters are the

14 212 Geographical Variation in the Yellow-breasted Apalis same as A.f.golzi. The main difference between A.f.pugnax and A.f.golzi is the amount of grey on the head-top. However intergrades between the tw o exist and will be dealt with below. The wing-length is longer than that of the tail. In A.f.flavocincta the male and female lack the black breast-spot and generally have a considerable admixture of deep olive-green in the yellow breast-band. The head-top is concolorous with the mantle, being a greyish olive- green, not the bright yellow-green of A.f.golzi and A.f.pugnax which live in areas of higher rainfall. There is little or no grey on the forehead. The under tail-coverts are white, the outer rectrices are entirely yellow, the yellow however being rather pale, almost white. The remaining rectrices have a large terminal spot of pale yellow. The wing-length is shorter than that of the tail. This form ranges from central and northern areas of Uganda (not the southern and western portions in the vicinity of Lake Victoria), to the southern Sudan, southern Ethiopia, northern districts and lowlands of eastern Kenya, southern Somalia, and the extreme north-east of Tanzania. The name A.f.malensis Neumann: Schambala River, Male Country, southern Ethiopia, is now placed as a synonym of A J.flavocincta. Situated between the typical populations of A.f.golzi and A.f. pugnax occur others which are intermediate in regard to head-top coloration. These intermediates resemble the parental forms except in the coloration of the head-top, which in these birds is grey with an intrusion of a few green feathers, usually over the hind crown. It could be argued that these intermediate populations constitute A.f.aequatorialis Neumann: Angata Anyuk, near Mau, Kenya. Friedmann (.1937) records this taxon as ranging from south-western Kenya from near Mau and the Sotik district to around Lake Victoria. Van Someren (1922) records it as ranging from the north-west and southern shores of Lake Victoria to the K ivu district, Congo, and in South Ankole and Entebbe, Uganda. Jackson (1938), however, gives the distribution as the south-western part of Kenya from about Naivasha to the Loita Plains, the southern shores of Lake Victoria and the inland districts of Tanzania, ranging to Ankole and Kivu. The range Jackson gives approaches closest the distribution of the birds which show intermediaite features in head-top coloration. When the range of these intermediates is examined it is found that they could not possibly all be of similar origin. A solitary specimen from Iringa, Southern Highlands, Tanzania, is well south in the range of A.f.golzi and is probably an intergrade A.f.golzi^,A.f.tenerrim a. Two other specimens from Iringa are typically A.f.golzi. The intermediate Iringa specimen also has a relatively shorter tail, with the wing/tail index being Two specimens from Mwanza and Ukerewe Island,

15 by Walter J. Lawson 213 Lake Victoria, have wing/tail indices of 1.19 and 1.15 respectively, which are well in excess of the figures for A.f.golzi and clearly show the influence of A.f.caniceps, though the green on the crown suggests some influence of A.f.pugnax as well. A specimen from Kigoma (wing/tail index 1.11) cannot possibly have any relationship with A.f.pugnax at all and may well represent an intermediate between A.f.golzi and A.f.canora. It can, therefore, be seen that a number of specimens, superficially alike, have at least four possible sources of origin. I have examined the Type of A.aequatorialis, a female collected by 0. Neumann on the 17th November, 1894, at Angata Anyuk. It has a pale grey head-top (as in A.f.caniceps), a short tail (which is slightly damaged), the under tail-coverts are virtually absent through damage, but one remaining feather is yellow, the outermost rectrices are pale yellow on the outer web and a rather greyish- yellow on the inner. The yellow breast-band has no visible black spot, but a search will reveal two or three black feathers present. The wing/tail index is This specimen is clearly an intergrade between A.f.caniceps and A.f.pugnax. I submit it would be incorrect to use the name aequatorialis for the populations I have now placed as A.f.pugnax. Firstly the Type of A.aequatorialis is far from typical of A.f.pugnax as defined above, and the populations from the vicinity of the type-locality are essentially intergrading ones. This Type is very close to A.f.caniceps in coloration and size, but the populations it represents are clearly intermediate between that race and A.f. pugnax. The name should be lodged in the synonym y of A.f.caniceps. In northern Somalia occur distinct populations to which the name A.f.viridiceps Hawker: Sheik W oofly, Somalia, is applied. As in A.f.flavocincta, neither male nor female has a black breast-spot, and the yellow breast-band is very heavily suffused with olive-green, generally more so than in flavocincta. In extreme cases there is probably less than 20% of the area of the breast-band yellow. The head-top is concolorous with the mantle, being a deep olive-green, any grey present being very limited and confined to the frons. A.f.viridiceps differs from A.f.flavocincta in that the outer rectrices are white, not yellow. The under tail-coverts are likewise white. Mr. I. C. J. Galbraith, of the British Museum (Natural History), kindly examined the Type of A.f.viridiceps and informs me that The H olotype, B.M. (N.H.) No , was collected b y R. M. Hawker and G. Cheetham at Sheik W oofly, Somaliland, on 28th November, It is a male, legs brown, bill black, irides light hazel. The wings measure 51 mm. (right) and 50 mm. (left), the tail 53 mm. (in moult). The head is also in moult and the plumage

16 214 Geographical Variation in the Yellow-breasted Apalis worn. The head is green like the mantle with grey restricted to the forehead; the under tail-coverts and the tips of the rectrices are white. I have been able to examine another specimen from Sheik W oofly, Somalia, collected at the same time as the Type (B.M ), which, incidently, has Type of Species written on the label. This confirms the identity of the populations placed as A.f.viridiceps. I have not been able to examine the Type of A.f.neumanni Zedlitz: Afgoi, Somalia, nor any specimens from the type-locality, but I have seen material from Jubaland (topotypical), Kismayu and Baduna. If A.f.neum anni were a valid subspecies then one might expect it to occur at at least one of these localities. Specimens from these areas of southern Somalia examined are typical of A.f.flavocincta in coloration. It is possible that the coastal birds may be smaller than the interior ones, which state of affairs exists for m any species in this part of East Africa. This variation in size, however, m ay be clinal and'over a rather short distance, and until such time as further material becomes available from this area I prefer to regard A.f.neum anni as a synonym of A.f.flavocincta. In W est Africa, from Portuguese Guinea eastward through southern Tchad, the Central African Republic and northern Congo to Uganda in the south and west in the vicinity of Lake Victoria occur very distinctive populations. In these the head-top and nape are grey, in which respect they differ from A.f.viridiceps and A.f. flavocincta. These populations represent nominate A.caniceps. The males and the females lack a black breast-spot, the tail is very much shorter than the wing, and there is generally far less suffusion of olive-green in the breast-band than in A.f.flavocincta, A.f.golzi or A.f.pugnax. The under tail-coverts are white, and all the rectrices are a deep olive-green, narrowly tipped with a very pale yellow- green, or almost white with an olive suffusion. The white under-parts seem to be somewhat greyish and not a clear white. I have examined the Type of Apalis uamensis Reichenow: Bosum, eastern Cameroons, and can find no characters b y which to separate this from A.f. caniceps, and this name is accordingly placed in the synonym y of that race. A further name applied to populations of this apalis remains to be dealt with. This is A.f.flavotorquata (Hartlaub): Magungo. This name is employed, as far as I can ascertain, b y only one worker, namely Jackson (1938). I have examined specimens from three of the localities cited by Jackson in the range he gives for A.f.flavotorquata, namely Entebbe, Kazi and Mpumu, and can find no features by

17 by Walter J. Lawson 215 which to distinguish them from examples of A.f.caniceps, of which I regard A.f.flavotorquata as a synonym. In addition to the range given above for A.f.caniceps, this race also extends south-west to Cabinda and the lower Congo and extreme northern Angola. I have only examined one specimen from Belas, Angola, attributable to A.f.caniceps, and find it identical to other examples examined. Because of its very short tail, A.f.caniceps' influence on the populations occurring in the extreme eastern Congo and extreme western Tanzania can be seen. Specimens from Baraka, at the northern end of Lake Tanganyika, have the short tail of A.f. caniceps but otherwise agree in coloration with A.f.canora. These particular specimens were formerly placed as the last named by Lawson (1962). I have no evidence that the range of A.f.caniceps entirely encircles the Congo forests, as in the south-eastern Congo A.f.canora occurs. Arising from the above discussions, the Yellow-breasted Apalis can be regarded as being composed of the following subspecies: (a) Apalis flavida flavida (Strickland) Drymoeca flavida Strickland, Contr. Orn., 1852, p. 148: Damara- land =N gam iland (vide Lawson (1961)). The male alone has a prominent black breast-spot which is wanting in the female. Differs from all other subspecies in having the yellow of the breast-band extending on to the throat. The top of the head and nape are grey, generally with an admixture of green on the nape and hind crown. The flanks are washed with yellow and the under tail-coverts are coloured likewise. The ventral surface of the outer rectrices have prominent yellow outer webs and terminal spots. Measurements: 61 3 $ wing (51.89); tail (49.83) m m.; wing/tail (1.044). Material: 87 (Botswana 21, South-W est Africa 13, Angola 45, Zambia 8). Localities: Botswana: Aha Hills, Cungena, Kabulabula, Lake Ngami, Mababe Flats, Maun, Ngoma, Nokaneng, Shorobe, Thama- lakane River; Zambia: Shangombo, Sinjembele; Angola: Belas, Bemfica, Benguella, Bibala,. Blaasbalkfontein, B ocoio, Cachoeiras, Calai, Cassinga, Chingoroi, Chipamba, Chitado, Chite, Cuto, Jau, Leba, Lungo, Mucuma, Mutundo, Pereira d Eca, Quipungo, Represa, Sa da Bandeira, Sambio, Techivinguiro, Teeba, Zuitala; South- West Africa: Elephants Vley, Grootfontein, Ondongua, Ovaquenyama, Rietfontein, Swartboois Drift. Kapaku, Mayara,

18 216 Geographical Variation in the Yellow-breasted Apalis Specimens examined from Zambia from Kazungula, Sesheke Boma and Namwala, and Sebungwe and Nampini, Rhodesia, are regarded as intergrades between A.f.flavida and A.f.neglecta. Range: Northern Botswana and northern South-W est Africa and adjacent areas of Barotseland, Zambia, north to Angola in the south-east, south and west, as far north as the Luanda district. T ype: In the Museum of Zoology, Cambridge. Not examined. (b) Apalis flavida florisuga (Reichenow) Euprinoides florisuga Reichenow, J ourn.fiir Ornith., 1898, p. 314: ' Kaf&rland King W illiam s Town district, Eastern Cape Province. Differs from all the other subspecies in southern Africa in that both the male and the female lack a black breast-spot. The yellow of the breast-band is clear and lacks any suffusion of olive-green as in A.f.renata. Measurements: 34 < < wing (48.88); tail (51.67) mm. wing/tail (0.946). Material: 38 (eastern Cape Province 34, Natal 4). Localities: Eastern Cape Provinces Alexandria Forest, Big Umgazi River, Committees, Debe Nek, Bathurst, East London, Grahams- town, Kei Bridge, Kei R oad, Mntafufa River, Need s Camp, Ngqeleni, Patensie, Pirie, Port St. John; Natal: Camperdown, Dargle, Karkloof, Pietermaritzburg. Range: Eastern Cape Province to Natal, where, north of Durban, it is found only in the interior forests, the coastal forests being populated b y the next subspecies. N ote: Specimens examined from East London, Kei R oad and Alexandria Forest have very small incipient black spots on the breast in three male specimens. (c) Apalis flavida renata, subsp. nov. T ype: cj, adult. Collected b y a Durban Museum Expedition at Mapinhane, Sul do Save, Mozambique (at 22 15'S., 35 07'E.) on 3rd June, In the collection of the Durban Museum, Reg. No Diagnosis: Differs from A.f.florisuga in that the male has a black breast-spot, which feature is absent in that subspecies, and from A.f.neglecta in having the yellow of the breast-band more heavily suffused with olive-green, and with the black breast-spot in the male very prominent, generally larger and more pyramidal in form. The wing is rather smaller but the tail is longer, resulting in a difference in the wing/tail index between the two subspecies. Separable from

19 by Walter J. Lawson 217 A.f.lucidigula b y having the yellow breast-band overlaid with olive-green and with a larger black breast-spot. The forehead and fore-crown of the head are grey, the hind crown and nape invariably olive-green, concolorous with the mantle. Measurements o f T ype: wing 50.0, tail 53.0 mm., wing/tail Measurements: 63 (Jc? wing (50.50), tail (50.33) mm., wing/tail (1.00). Material: 91 (Zululand 16, Mozambique 72, Transvaal 1, Swaziland 2). Localities: Zululand: Heatonville, Ingwavuma, Makane s Pont, Manaba, Maputa, Mapelane, Mposa, Ndumu, Penecuik, Pongola, Usutu R.; Mozambique: Beira, Bela Vista, Canganetole, Chimonzo, Chinhanguanine, Coguno, Inhaca, Inhaminga, Espungabera, Lacer- donia, Licuati, Lourengo Marques, Mabalane, Mabote, Manhica, Mapinhane, Maringua, Massinga, Mapai, Mopeia, Movene Maquese, Muanza, Mixixine, Panda, Pungwe/Gorongoza, Sofala Bay, Tete, Tinonganine, Umbeluzi; Transvaal: Inkomati River; Swaziland: Big Bend, Stegi. The influence of A.f.lucidigula can be seen in odd specimens from Panda and Massinga, an area known to reveal influences of interior races in many bird species. Range: Coastal Natal from approximately Durban northwards, Zululand east of the Lebom bo Range, and Mozambique to the Zambesi R iver with westward intrusions to the extreme eastern Transvaal and Swaziland up the m ajor river valleys. Note: Clancey (1967) has already laid down the characters of this new race. (d) Apalis flavida lucidigula Lawson Apalis flavida lucidigula Lawson, Durban Museum Novitates, vol. vi, 9, 1961, p. 123: Farm Malamala, near Newington, eastern Transvaal. Differs from A.f.renata and A.f.neglecta in the clear coloration of the yellow breast-band which has no admixture of olive-green. Only the male has a biack breast-spot, which is smaller than in A.f.renata. Resembles similarly coloured populations of the Zambesi R iver in the vicinity of Lake Kariba with which it is not in contact. Thiese latter populations are considered intergrades A.f.fla vid a ^ A.f. neglecta. Measurements: 59 (JcJ wing (51.27), tail (51.81) mm., wing/tail (0.991).

20 218 Geographical Variation in the Yellow-breasted Apalis M aterial: 64 (Zululand 11, Mozambique 9, Swaziland 8, Transvaal 20, Rhodesia 16). Localities: Zululand: Candover, Cecil Mack s Pass, Gwaliweni Forest, Ingwavuma, Kangazani Pan, Makane's Pont, Melmoth, U m folosi R iver; Mozambique: Umbeluzi River, Mamale, Maringua, Sabie; Swaziland: Big Bend, Horo, Ranches Ltd., Stegi; Transvaal: A com hoek, Elim, Klaserie, Olifants River, Koedoes River, Leydsdorp, Lydenburg, Motale River, Newington, Sand River, Umhlumi; Rhodesia: Birchenough Bridge, B ubye/lim popo River, Chikwiro Reserve, Holdenby, Mutema, Sabi/Lundi Rivers, Sijarira Mts., Turgwe River. One specimen from Blouberg and another from Klein Letaba, in the Zoutpansberg of the Transvaal, resemble A.f.neglecta in coloration. Range: Eastern and northern Transvaal south to Swaziland and western Zululand, north to eastern, south-eastern and southern Rhodesia in the lowlands, intruding into Mozambique and eastern Zululand through the river valleys. T ype: In the Durban Museum. Reg. No Examined. (e) Apalis flavida neglecta (Alexander) Chlorodyta neglecta Alexander, Bull.Brit.Orn.Club, vol. x, 1900, p. 17; south-east Africa, the Type from Zum bo, Mozambique. From A.f.renata separable in having a less heavy suffusion of olive-green in the yellow breast-band, a smaller black spot, a longer wing, a shorter tail and a larger wing/tail index, and from A.f. lucidigula in having a moderate amount of olive-green suffused in the yellow breast-band. Differs from A.f.canora in that the hind crown and nape are always green, concolorous with the mantle, and from A.f.tenerrima in its larger size of wing and tail, and the green hind crown and nape, which in the latter are grey. Measurements: 68 ($<$ wing (51.78), tail (49.78) mm., wing/tail (1.043). Material: 99 (Mozambique 11, Malawi 14, Rhodesia 29, Zambia 44, Botswana 1). Localities: Mozambique: Luenya River, Messenguese, Tambara Fort, Tsungu; Malawi: Blantyre, Chiromo, Fort Johnston, Mlanje, Malawe Hill, Mzimba, Namwera, Port Herald; Rhodesia: Chirundu, Devuli River, Haroni/Lusitu Rivers, H oldenby, Inyanga, Mt. Selinda, Mutema, Nuanetsi River, Sabi River, Sabi/Odzi Rivers, Sanyati River; Zambia: Chiawa, Chilanga, Chipongwe, Chiyengele, Kafue River, Kahatu, Kariba, Katondwe, Fort Rosebery, Livingstone, Luangwa River, Lusaka, Mankoya, Mwanya, Mwekera, Namwala Ngoma, Sandwe; Botswana: Kasane.

21 by Walter J. Lawson 219 Range: Middle and upper Zambesi R iver valley, south to the eastern regions of Rhodesia, and north to Malawi and south-eastern, eastern and southern Zambia. N ote: Material from the low-lying areas of eastern Rhodesia generally has the tail longer than the wing, whereas birds from the higher areas of eastern Rhodesia have the wing longer than the tail. Sketch Map to Show Distribution of APALIS FLAVIDA 1. A.f.flavida 7. A.f.tenerrima 2. A.f.florisuga 8. A.f.golzi 3. A.f.renata 9. A.f.pugnax 4. A.f.lucidigula 10. A.f.flavocincta 5. A.f.neglecta 11. A.f.viridiceps 6. A.f.canora 12. A.f.caniceps Localities from which specimens have been examined. H ortors

22 220 Geographical Variation in the Yellow-breasted Apalis (/) Apalis flavida canora Lawson Apalis flavida canora Lawson, Bull.Brit.Orn.Club, vol. lxxxii, 7, 1962, p. 134: Sumbu, Zambia. Differs from A.f.neglecta in that the grey of head-top extends to the nape. The head-top m ay be entirely grey or with a few green feathers over the nape, but the nape is always grey and never entirely green as in A.f.neglecta. The tail of A.f.canora averages a little shorter than that of A.f.neglecta and the wing/tail index is larger. Separable from A.f.tenerrima, with which it is not in contact, b y its considerably larger size. Measurements: 19 $ $ wing (52.63), tail (48.94) mm., wing/tail (1.073). M aterial: 33 (Zambia 23, Angola 10). Localities: Zambia: Bulaya, Chocka, Kasempa, Kilwa Island, K abom po River, Liuwa Plain, Lungwevungu River, Luwenge, Mankoya, Manyinga River, Museshya, Mayau, Mutanda, Mundwiji Plain, Mwekera, Mwinilunga, Nchelenge, Sesheke Bom a; Angola: Calunga River, Cazela, Cuito-Canavale, Lagoa Muibilo, Longa, Manbi River. Range: Extreme north-western portions of the Western and Northern Provinces of Zambia, eastern Angola, and south-eastern Congo. T ype: In the National Museum of Rhodesia. Reg. No Examined. (g) Apalis flavida tenerrima Grote Apalis flavida tenerrima Grote, Ornith.Monatsber., 1933, p. 119: Mikindani, south-eastern Tanzania. Distinguishable from A j.neglecta in having the head-top and nape entirely grey, with a small amount of admixture of green in the nape in some specimens, but this latter never entirely green. Smaller in size of wing and tail. From A.f.canora separable b y its smaller wing and tail, and similarly so from A.f.renata; also in the presence of grey on hind crown and nape, which are always green in A.f.renata. Differs from A.f.golzi in having the under tail-coverts white, not yellow, and the rectrices narrowly, not broadly, tipped with yellow; only the male has a black breast-spot; from A.f. flavocincta in having a black breast-spot in the male and a shorter tail and larger wing/tail index. The yellow breast-band has less suffusion of olive-green.

23 by Walter J. Lawson 221 Measurements: 22 (JcJ wing (49.23), tail (44.27) mm., wing/tail (1.112). Material: 36 (Mozambique 7, Tanzania 29). Localities: Mozambique: Alto Ligonha, Alto Molocue, Boror, Mcutuche, Nampula; Tanzania: Dar-es-Salaam, Kitaya, Luwipa River, Magogoni, Mahenge, Mtwasa, Nchingidi, Songea, Ulanga, Zanzibar. Range: Northern M ofam bique, north of the Zambesi River to south-eastern and eastern Tanzania. Replaced by and intergrades with A.f.flavocincta in extreme north-eastern Tanzania. T ype: In the Zoologisches Museum, Berlin. Reg. No Examined. (h) Apalis flavida golzi (Fischer & Reichenow) Euprinoides golzi Fischer & Reichenow, Journ.fiir Ornith., 1884, p. 182, Great Arusha, Tanzania. Apalis niassae Reichenow, Journ.fiir Ornith., 1921, p. 264: Langenburg, northern Malawi. Male and female have a black breast-spot, a feature only found in one other subspecies, namely A.f.pugnax. Also resembles this latter race in having the under tail-coverts yellow and the outermost rectrices entirely yellow, but differs from it in that the head-top and nape are entirely grey, with no admixture of green on the last named. Intergrades between the present race and A.f.pugnax have an intrusion of a small amount of green on to the grey of the nape and hind crown. Differs from A.f.flavocincta which has no black spot in either male or female, and which form also has a considerably longer tail and a smaller wing/tail index; likewise from A.f.tenerrima, A.f.neglecta and A.f.canora all of which have white, not yellow, under tail-coverts; and from A.f.caniceps which form lacks the black breast-spot in both male and female, has a shorter tail and larger wing/tail index. The male has a heavier suffusion of olive-green in the yellow of the breast-band than does the female, and the black spot in the female is also smaller than in the male. Measurements: 20 (JcJ wing (52.7), tail (48.93) mm., wing/tail (1.079). Material: 36, all from Tanzania. Localities: Tanzania: Babati, Bussisi, D oinyo Narok, Dongo, Guyame, Igawa, Insel Sirwa, Iringa, Ketum beine Mt., Kibrori, Kjafum a, Lake Chaya, Loietai, Loliondo, Mara River, Marangu, Monduli, Naberera, Rukwa, Tum ba, Wallafluss.

24 222 Geographical Variation in the Yellow-breasted Apalis Range: Throughout mainland Tanzania, except for extreme east where A.f.tenerrima occurs and the extreme northern boundary where it intergrades with A.f.pugnax (see localities under that race), extreme north-western and western boundaries where it intergrades with A.f.caniceps (see localities under A.f.caniceps) and Malawi where it intergrades with A.f.neglecta. T ype: In the Zoologisches Museum, Berlin. Reg. No Exam ined; of A.niassae, in Zoologisches Museum, Berlin. Examined. (i) Apalis flavida pugnax, subsp. nov. T ype: <$, adult. Collected b y P. A. Clancey at Nanyuki, lower Mount Kenya, Kenya (at 0 01' N., 37 05' E.) on 20th March, 1958, at 6,000 feet altitude. In I he collection of the Durban Museum, Reg. No Diagnosis: Differs from A.f.golzi in having the head-top and nape entirely green, concolorous with the mantle, the grey being confined entirely to the forehead, otherwise similar; from A.f.flavocincta in that both the sexes have a black spot on the breast and the tail is shorter and the wing/tail index is larger. Separable from A.f.caniceps in that both the sexes have a black breast-spot and the tail is longer and the wing/tail index is smaller. The under tail-coverts are yellow, the outermost rectrices entirely yellow, the rest with a large terminal spot of yellow. The male has a heavy admixture of olive-green in the yellow breast-band, that of the female clearer yellow. The black spot is more prominent in the male than the female. Measurements o f T ype: wing 55.0, tail 52.5 mm., wing/tail Measurements: 27 $ $ wing (54.50), tail (52.00) mm., wing/tail (1.048). M aterial: 36, all from the highlands of Kenya. Localities: Kenya: Eldom a Ravine, Embu, Guaso Nyiro, Karrat, Karura Forest, K yam bu Forest, Kijabe, Mau, Mt. Kenya, Msara, Mwiega, Nairobi, Nanyuki, Ngong, Samburu. Intergrades with A.f.golzi at the following localities: D oinyo Narok, Kibigori, K igcm a, Lake Elmenteita, Loietai, Marangu, Monduli, Moschi, Naivasha, Ngaptuk, Nakuru, 01 Mesuti, Umbugwe. Range: The highland areas of central Kenya. (j) Apalis flavida flavocincta (Sharpe) Euprinoides flavocinctus Sharpe, Journ.fiir Ornith., 1882, p. 346: A thi River, Kenya. (T ype may be from Kitui, near Athi River, in Ukamba, Kenya). Apalis malensis Neumann, in Reichenow s Vog.Afr., vol. iii, 1905, p. 612: Schambala River, Male Country, southern Ethiopia.

25 by Walter J. Lawson 223 Apalis flavida neumanni Zedlitz, Journ.fiir Ornith., 1916, p. 89: Afgoi, southern Somalia. Neither the male nor the female has a black spot; the only other subspecies with this characteristic are A.f.viridiceps and A.f.caniceps, and A.f.florisuga of southern Africa. Differs from A.f.viridiceps in that this form has the yellow breast-band very heavily suffused with olive-green and the under tail-coverts and outer rectrices are white and not pale yellow, and from A.f.caniceps in having a considerably longer tail and a smaller wing/tail index. The head-top and nape are grey in A.f.caniceps, green in A.f.flavocincta. Measurements: 45 $<$ wing (50.49), tail (55.29) mm., wing/tail (0.914). Material: 74 (Tanzania 1, Kenya 45, Somalia 6, Ethiopia 18, Uganda 3, Sudan 1). Localities: Tanzania: Tanga; Kenya: Bura, Campi-ya-bibi, Embu, Barsaloi, Isiolo, Kacheliba, Kibwezi, Lali, Makindu, Maungu, Mbuyuni, Moyale, Lodermoru, N. Uaso Nyiro, Nzui River, Tana River, Tharaka, Tsavo Stn., W am ba, W aso; Somalia: Baduna, Jubaland, Kismayu; Ethiopia: Amar Koshi, Lake Rudolf, Lake Stephanie, Laku Bali, Y avello; Uganda: Karam oja, Mt. M oroto; Sudan: Loelli (flavocinctat^caniceps intergrade). Range: Central and northern regions of Uganda (not the southern and western portions in the vicinity of Lake Victoria), southern Sudan, southern Ethiopia to northern Kenya and the lowlands of eastern and south-eastern Kenya, southern Somalia and extreme north-eastern Tanzania. Specimens from Usambara Mts., northeastern Tanzania, are flavocinctar^j.enerrima intergrades. Type: Formerly in the Zoologisches Museum, Berlin. N ot traced. (,k) Apalis flavida viridiceps Hawker Apalis viridiceps Hawker, Bull.Brit.Orn.Club,. 7, p. 55, 1898: Sheik W oofly, Somalia. Neither the male nor the female has a black breast-spot. The head-top and the mantle are entirely green with little or no grey on the forehead. The yellow breast-band is very heavily suffused with olive-green. Differs from A.f.flavocincta in that the outer rectrices are white, not yellow; under tail-coverts also white. Measurements: 7 g g wing (50.64), tail (59.29) mm., wing/tail (0.847). Material: 14, all from Somalia.

26 224 Geographical Variation in the Yellow-breasted Apalis Localities: Burao, Sheik W oofly. Range: Northern Somalia. T ype: In the British Museum (Natural History) Reg. No , not examined, but description kindly provided b y Mr. I. C. J. Galbraith. (I) Apalis flavida caniceps (Cassin) Camaroptera caniceps Cassin, Proc.Philad.Acad.Nat.Sci., 1859, p. 38: Camma River, Gabon. Tricholais flavotorquata Hartlaub, Proc.Zool.Soc.Lond., 1880, p. 624: Magungo. Apalis aequatorialis Neumann, Journ./iir Ornith., 1900, p. 307: Angata Anyuk, near Mau, Kenya. Apalis uamensis Reichenow, Journ./iir Ornith., 1921, p. 264: Bosum, eastern Cameroon. Differs from all the other subspecies in having a very short tail. The head-top and nape are entirely grey, and there is no black spot on the breast in either male or female. The under tail-coverts are white, and the rectrices a deep olive-green, narrowly tipped with very pale yellow-green or almost white with a faint olive suffusion. The white under-parts are washed with grey. Measurements: 11 $ $ wing (52.00), tail (40.30) mm., wing/tail (1.299). M aterial: 19 (Kenya 1, Uganda 5, Cameroon 2, Nigeria 3, Ghana 1, Tchad 1, Portuguese Guinea 2, Angola 1). Localities: Kenya: Angata Anyuk; Uganda: Busoga, Entebbe, Mabira, Mpumu; Cameroon: Bosum ; Nigeria: Abeokuta, Port Harcourt, Shonga; Ghana: Gambaga; Tchad: Bamingui; Portuguese Guinea: Bissau; Angola: Belas. Range: W est Africa through southern Tchad, the Central African Republic to the northern Congo and Uganda south and west of Lake Victoria; also in the west, south to the Cabinda enclave, southern Congo forests and in extreme northern Angola. The influence of this race is apparent in specimens from Kazi, Uganda (caniceps^golzi intergrades) and Baraka, northern Lake Tanganyika. T ype: Not examined. Of A.uamensis in the Zoologisches Museum, Berlin. No Exam ined; of A.aequatorialis in Zoologisches Museum, Beilin. Reg. No Examined.

27 by Walter J. Lawson 225 Wing renata neglecta tenerrima canora neglecta lucidigula renata tenerrima t P Tail renata neglecta tenerrima canora neglecta lucidigula renata tenerrima Wing/Tail renata neglecta tenerrima canora neglecta lucidigula renata tenerrima "] ^ Y *i Y ^ y <{ > < J ^ Y T A B L E 1. Probabilities of means using Student's t Test. X Mean t Student s 't' value s Standard Deviation P Probability n Degrees of Freedom ,001,001, <., ,001

28 226 Geographical Variation in the Yellow-breasted Apalis BIBLIOGRAPHY Chapin, J. P Bull.Amer.M us.nat.hist., vol. lxxva, pp Clancey, P. A Durban M us.novit., vol. viii, 9, pp Fischer, G. A. and Reichenow, A Journ.fiir Ornith., p Friedmann, H Bull.U.S.Nat.M us., No. 153, pp Irwin, M. P. S. and Benson, C. W Arnoldia (Rhodesia), vol. ii, No. 32, pp Jackson, F. J The Birds of Kenya Colony and Uganda Protectorate, vol. ii, pp Lawson, W. J Durban Mus.Novit., vol. vi, 9, pp Lawson, W. J Bull.Brit.Orn.Club, vol. lxxxii, 7, pp Mackworth-Praed, C. W. and Grant, C. H. B Birds of Eastern and North-Eastern Africa, Series 1, vol. ii, pp Neumann, O Journ.fiir Ornith., vol. xlviii, p Neumann, O Journ.fiir Ornith, vol. liv, p Reichenow, A Die Vogel Afrikas, vol. iii, pp Schouteden, A Ann.Mus.Roy.Congo Beige, C.Zoologie, Series iv, vol. iv, pp Sclater, W. L Systema Avium Aethiopicarum, part 2, pp Traylor, M. A Publ.cult.Co.Diam.Angola, No. 61, p van Someren, V. G. L Novitates Zoologicae, vol. xxix, 1, pp van Someren, V. G. L Novitates Zoologicae, vol. xxxvii, pp W hite, C. M. N Occ.Papers Nat.Mus.S.Rhod., vol. xxvib. Zedlitz, O. G Journ.fiir Ornith. vol. lxiv, pp