0. BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) 2 2 HOV i963 A ZOOLOGY VOL. 10 1963 BRITISH MUSEUM (NATURAL HISTORY) LONDON : 1963 DATES OF PUBLICATION OF THE PARTS No. i 3 January 1963 No. 2 ...... 14 March 1963 No. 3 29 March 1963 No. 4 ...... 29 March 1963 No. 5 28 May 1963 No. 6 31 May 1963 No. 7 9 July 1963 No. 8 16 July 1963 No. 9 ...... 20 May 1963 PRINTED IN GREAT BRITAIN BY THOMAS DE LA RUE & COMPANY LIMITED LONDON ZOOLOGY VOLUME 10 PAGE No. I. Ethiopian mites of the genus Androlaelaps Berlese s. lat. (Acari : Mesostigmata). By W. M. TILL i No. 2. The Francolins, a study in speciation. By B. P. HALL 105 No. 3. Taxonomy and biology of the genus Lebetus (Teleostei-Gobioidea) . By P. J. MILLER (PI. i) 205 No. 4. The freshwater gastropod molluscs of Western Aden Protectorate. By C. A. WRIGHT (Pis. 1-2) 257 No. 5. Observations on the chaetotaxy of the legs in free-living Gamasina (Acari : Mesostigmata). By G. OWEN EVANS 275 No. 6. A revision of the recent round herrings (Pisces : Dussumieriidae) . By P. J. P. WHITEHEAD 305 No. 7. The polychaete fauna of South Africa. Part 8 : new species and records from grab samples and dredgings. By J. H. DAY 381 No. 8. The freshwater gastropod mollusca of Angola. By C. A. WRIGHT (Pis. 1-16) 447 No. 9. New marine nematodes from off the coast of South Africa. By WILLIAM G. INGLIS 529 Index Volume 10 553 ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s.lat. (ACARI : MESOSTIGMATA) W. M. TILL BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 10 No. i LONDON : 1963 ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. (ACARI : MESOSTIGMATA) BY W. M. TILL British Museum (Natural History) Pp. 1-104 >' I 95 Text-Figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 10 No. i LONDON: 1963 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series, corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 10, No. i of the Zoological series. Trustees of the British Museum 1963 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued January, 1963 Price Thirty Jive shillings ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. (ACARI: MESOSTIGMATA). By W. M. TILL British Museum (Natural History], London CONTENTS INTRODUCTION EXTERNAL MORPHOLOGY CLASSIFICATION HOST-PARASITE LIST REFERENCES INTRODUCTION The earliest species of the genus Androlaelaps s. lat. (including Haemolaelaps} to be recorded from the Ethiopian region are those described by Berlese (1911-1918) and Hirst (1916). Radford (1939-1944) added several more species from this region and Zumpt & Patterson (1950, 1951) provided the first keys to the Ethiopian species of the genera Androlaelaps and Haemolaelaps. Since then many new species have been described by Zumpt and his co-workers, by Keegan (1956) and by Lavoipierre (1955, 1956). With the increased number of species, the existing keys have been found to be inadequate as they were based largely on superficial characters which are subject to intraspecific variation. It was also suspected that many of the new species might be synonyms of older, imperfectly described species, the types of which had never been thoroughly re-examined. The aim of the present investigation is therefore to give redescriptions of the known species, based wherever possible on type material, to try to evaluate the morphological characters observed in the different species, and to prepare a key to the females. Several new species have been included and a host-parasite list is given. This investigation has been made possible by grants from the British Museum (Natural History) and from the Wellcome Foundation, London. It is based chiefly on material in the collections of the British Museum (B.M.N.H.) and the South African Institute for Medical Research, Johannesburg (S.A.I.M.R.), the latter collection having been placed at my disposal by Dr. F. Zumpt. Material has also been received from Dr. P. L. G. Benoit, Musee Royal de 1'Afrique Centrale, Tervuren, Belgium (M.R.A.C.), Dr. R. Taufflieb, Institut de Recherches Scientifiques au Congo, Brazzaville (I.R.S.C.) and Miss J. B. Walker, East African Veterinary Research Organization, Kenya. Types and other specimens have been lent by Miss M. A. Johnson, Liverpool School of Tropical Medicine, Dr. M. M. J. Lavoipierre, ZOOL. 10, I I 4 W. M. TILL George Williams Hooper Foundation, California, Dr. C. D. Radford, Manchester, and by the Smithsonian Institute, United States National Museum, Washington. The work was carried out in the British Museum (Natural History) and I am grateful to the Trustees for providing me with the necessary facilities. My thanks are due to Dr. G. O. Evans for his guidance and many helpful suggestions, and to Dr. J. G. Sheals and Mr. D. Macfarlane for advice and assistance in various ways. EXTERNAL MORPHOLOGY Females GNATHOSOMA (figs. 1-5) : The gnathosoma is attached subterminally to the anterior end of the body. It consists of a cylindrical basis capituli (gnathosomal base) enclosing the pharynx and chelicerae and bearing antero-laterally a pair of seg- mented pedipalps. The dorsal wall of the basis capituli extends anteriorly as a membranous flap known as the tectum. In the genus Androlaelaps this is a trans- parent structure more or less rounded anteriorly and without ornamentation. On the ventral surface is a median longitudinal groove, the deutosternum, which normally bears six rows of anteriorly directed teeth. The number of teeth in each row varies from one to about six. Anteriorly the ventral wall is produced to form the hypostome and its processes, the horn-shaped corniculi (external malae) and the fimbriated hypostomal processes (internal malae). In most Androlaelaps species the corniculi are prominent and well sclerotized (figs, i & 3), but in A. sangsteri and A. heliosciuri they are reduced in size, weakly sclerotized and generally incon- spicuous (fig. 4). The hypostomal processes show a variety of modifications and are sometimes very difficult to see. In most Androlaelaps species they are similar to those of A. murinus (fig. i), the ventral processes having the form of a pair of triangular flaps, the outer margins of which are fringed. The depth of the fringe varies in different species and is relatively great in A. tachyoryctes (fig. 2). The hypostomal processes are unusually large and pilose in A. georychi (fig. 5), whilst in A. walker ae both the ventral and dorsal processes are modified to form prominent brush-like structures (fig. 3). In A. heliosciuri the inner margins of the processes are fringed (fig- 4)- The ventral surface of the gnathosoma bears four pairs of setae, the anterior rostral (g.s. i) and the posterior external and internal rostral setae (g.s. 2 and g.s. 3) on the hypostome, and the capitular setae (g.s. 4) on the basis capituli. The lengths FIGS. 1-5. Types of gnathosoma (ventral view) in females of the genus Androlaelaps s. lat. Fig. i. Androlaelaps murinus (Berlese) Fig. 2. Androlaelaps tachyoryctes (Radford) Fig. 3. Androlaelaps walkerae sp nov. Fig. 4. Androlaelaps heliosciuri sp. nov. Fig. 5. Androlaelaps georychi sp. nov. b.c., basis capituli; deut., deutosternum; ext. m., external mala; g.s., gnathosomal seta; hyp., hypostome; int. m., internal mala; la., labrum. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS RERLESE 5. lat. 6 W. M. TILL of these setae are relatively constant within the genus. The anterior rostral and capitular setae are usually subequal in length, except in the scapularis-group, and the posterior internal rostral seta is 2-3 times as long as the external seta. Lying dorsal to the hypostome and ventral to the chelicerae is a pointed, ciliated structure, the labrum. The chelicerae occupy the dorsal region of the gnathosoma. Each terminates in a fixed and a movable digit, both of which are normally dentate and well sclerotized. The fixed digit is usually shorter than the movable one, less strongly dentate, and bears a transparent, seta-like structure, the pilus dentilis, the form of which varies in different species. The structure of the chelicerae is peculiar in the scapularis-group, in that the movable digit is edentate, except for the terminal hook and at the most one rather blunt subterminal tooth, and bears an additional cutting process provided with large, outwardly directed teeth. The fixed digit is reduced, weakly sclerotized, and bears a flagellar pilus dentilis (fig. 170). In the majority of the Androlaelaps species the pilus dentilis is inflated distal to the point of insertion (subgenus 1 ' schnolaelaps sensu Fonseca), while in a few, for example A . hystrici, A . phoenicidi, A . casalis, A . walkerae, it is slightly inflated near its point of insertion, tapering progressively to a point (subgenus Atricholaelaps sensu Fonseca) . On the ventral surface of the chelicera, at the base of the movable digit, is a semicircle of arthrodial filaments. In most of the Androlaelaps species these are subequal in length or only slightly variable. In A. heliosciuri (fig. 81) two of the filaments are considerably longer than the others, and in A. murinus (fig. 109), and A. zumpti (fig. 195) there is one filament which is strikingly longer and thicker than the others. The arthrodial membrane between the chelae appears to form an extension which is usually a small, rounded, transparent flap, but which, in A . zumpti, is greatly elongated, extending to the tips of the chelae. The pedipalps are 5-segmented; the first free segment, or trochanter, bears 2 setae, the femur bears 5, the genu 6 and the tibia 14 setae. The tarsus, in addition to a number of simple setae, bears a two-tined seta-like structure at its base. IDIOSOMA: The idiosoma is usually ovoid in shape with no sharp demarcation between the dorsal and ventral surfaces. Almost the entire dorsum is covered by a single scutum which usually has recticular markings and in some species, e.g. those of the scapularis-group, may also have a granular appearance. The chaetotaxy of the dorsal shield follows the basic pattern described by Costa (1961, fig. i) for the genus Haemolaelaps. Most of the species have the pattern comprising 39 pairs of setae, although in some species, for example those of the mesopicos-group, one or more of the marginal pairs may be situated on the adjacent integument and seta PX$ may be lost. A. longipes has an additional pair of setae (pxi) on the posterior part of the plate (fig. 93), whilst A. centrocarpus and A. hirsti have extra pairs of setae (ax and PXI} on both the anterior and posterior parts of the shield. Many species show some degree of hypertrichy, a variable number of unpaired, accessory setae developing on the median part of the plate between the "/" series. In A . later ae and A . villosissimus the dorsal shield has a dense covering of setae which completely obscures the basic pattern, An intermediate condition occurs in ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE 5. lot. 7 A . walker ae in which the basic pattern is partially obscured by the addition of both unpaired and paired setae. The setae vary considerably in length and in the scapularis- and w^so^'cos-groups (excluding A. wilkini] and in A.galagus those on the central part of the shield are reduced to microsetae. The ventral surface of the idiosoma bears a sternal, a genital and an anal shield, all of which may show some degree of intraspecific variation. The sternal shield is usually broader than long, reticulate in most species but with a granular appearance in mites of the scapularis-group, and bears two pairs of pores and three pairs of setae. In most Androlaelaps species the sternal setae are more or less subequal in length, the second and third pairs being very slightly longer than the first pair. In A. arvicanthis, A. graingeri, A. marshalli and A. theseus the first pair is relatively short, being |-| as long as the second pair. Metasternal setae are always present. The presternal area also has a reticulate pattern and bears the tritosternum, a bipartite structure of uniform appearance throughout the genus. It consists of a short, undivided, basal portion and two ciliated arms, the laciniae. The genital shield consists of a relatively more strongly sclerotized posterior portion, varying in shape from the slender tapering form seen in A . capensis (fig. 20) and A. tateronis (fig. 165) to the greatly expanded, flask-shaped form of A. rhodesiensis (fig. 128), and a more weakly sclerotized anterior portion which forms a flap covering the genital orifice. The shield usually has a linear or reticulate pattern at least on the posterior portion. This pattern may be very weakly developed, as in the scapularis-group, or it may be very clearly defined and extend on to the anterior portion, as in A. dasymys (fig. 51). In some mites the shield also has a granular appearance and in all the species it bears only one pair of genital setae. The anal shield is ovate or triangular, with the apex directed posteriorly, and encloses the anal orifice. It bears a pair of paranal setae and a single postanal seta. The metapodal plates are small sclerotized structures lying posterior to the fourth pair of coxae. The peritreme extends from the stigma, between coxae III and IV, to a point over coxa I, and the peritremal plate extends posterior to the stigma for a distance equal to approximately twice the diameter of the stigma. The setae arising from the integument of the idiosoma may be simple or barbed and are variable in number, some species being very sparsely and others relatively densely haired, as for instance A . taterae and A . georychi. The marginal setae are defined as those lying between the dorsal shield and an arbitrary line commencing at the posterior tip of the anal shield and continuing anteriorly on both sides and equidistant from the lateral margin of the body. The names of the idiosomal shields have been abbreviated and their measure- ments taken as follows: DL = median length of dorsal shield DW = greatest width of dorsal shield SL = median length of sternal shield SW = width of sternal shield at level of second pair of setae GW X = greatest width of genital shield GW 2 = width of genital shield at level of genital setae 8 W. M. TILL AL = median length of anal shield AW = greatest width of anal shield. LEGS : The legs consist of six recognizable segments, namely, the coxa, trochanter, femur, genu, tibia and tarsus, each tarsus terminating in a caruncle with claws. The segments are usually smooth in outline, but in A. callosus some of them have prominent, heavily chitinized swellings (fig. 18). Leg II is usually stouter than the other legs and this feature is most pronounced in the Androlaelaps-group. The chaetotaxy of the legs follows a basic pattern which is subject to slight variations. The number of setae per segment found in the majority of the Androlaelaps species, and which I have taken to be the normal number for this genus, is given in the following table : Leg Coxa Trochanter Femur Genu Tibia Tarsus I 2 6 13 13 13 II ... 2 5 ii ii 10 16 III ... 2 5 6 9 8 16 IV ... i 5 6 10 10 16 The number of tarsal setae listed does not include the two fine setae at the tip of the segment. Variations have been noted in the number of setae on the first and third tibiae and on the third and fourth genua. Tibia I has 14 setae in A. rhodesiensis and A. walker ae; tibia III has 9 setae in these two species and also in the species of the mesopicos-group, in A. centrocarpus , A. hirsti, A. longipes, A. phoeniculi, A. oliffi, A. tateronis, A. cricetomydis, A. congoensis, A. taterae and A. villosissimus. The last three species have ten setae on genu III and A. spatuliformis has nine setae on genu IV. The setae are variable in form and may be smooth or barbed, hair-like, spine-like, pointed or blunt. In the mites of the Androlaelaps-group one of the ventral setae on femur II is modified to form a stout spur-like process. Males The gnathosoma of the male is similar in structure to that of the female. The chelicera, however, has its movable digit modified to form a spermadactyl which serves as a copulatory organ. This may be relatively short and straight, as in A . casalis (fig. 29) , A . hystrici (fig. 88) and others, or relatively long and coiled as in A. taterae (fig. 164), A. patersoni (fig. 124) and A. spreo (fig. 145). The fixed digit bears a pilus dentilis, sometimes very difficult to detect, which resembles the pilus dentilis of the corresponding female. The dorsal shield has the same chaetotactic pattern as in the female, although the setae are often relatively longer. The ventral shields in most Androlaelaps species are fused to form a single holoventral plate which shows varying degrees of expansion behind the fourth pair of legs. The shield bears four pairs of sternal, one pair of genital, and the usual FIGS. 6-9. Androlaelaps casalis (Berlese). Venter (Fig. 6) and dorsum (Fig. 7) of deutonymph; dorsum (Fig. 8) and venter (Fig. 9) of protonymph. ETHIOPIAN MITES OF THE GENUS A NDROLA ELA PS BERLESE s. lat. TO W. M. TILL three anal setae. The number of setae in the preanal region varies between two or three pairs, in the species with a narrow plate, for example the mesopicos-group, and many, in those species with an expanded plate, for instance A. murinus and A . later ae. The usual number is five pairs. The anal shield is separated from the sternito-ventral shield in A. centrocarpus , A . hirsti and A . longipes (species with 40 or 41 pairs of setae on the dorsal shield) and also in A, tauffliebi. In all species the genital orifice is presternal. The chaetotaxy of the legs follows the same pattern as in the corresponding females although certain setae, for example in the mesopicos-group, may be modified to form thickened spines or spurs. Immature Stages According to Bregetova (1956) Haemolaelaps females are able to lay eggs with developed larvae or they may give birth to larvae or even to protonymphs. As far as the Ethiopian species are concerned only proto- and deutonymphs have been recorded up to now. The following description of the nymphal stages is based on specimens of A . casalis from deep poultry litter in England. The immature stages of the other species will not be described as so very little material is available. PROTONYMPH (figs. 8-9) : The gnathosoma resembles that of the female except that the palps have not yet acquired their full complement of setae. The trochanter bears one ventral seta, and the femur, genu and tibia bear 4, 5 and 12 setae respectively. The dorsum has a large anterior podonotal shield bearing eleven pairs of setae, a smaller, posterior, pygidial shield bearing eight pairs of setae, and three pairs of small intermediate platelets or mesonotal scutellae. The sternal shield, which bears three pairs of setae and two pairs of pores, does not extend beyond the middle of coxae IV. The peritremes are short and do not reach beyond coxae III. There is no genital shield or genital orifice. The legs have not yet acquired their full number of setae, those present being listed in the following table: Leg Coxa Trochanter Femur Genu Tibia Tarsus I ... 2 4 10 8 8 II ... 2 4 8 6 7 15 III ... 2 4 5 6 7 15 IV . i 4 4 6 7 15 DEUTONYMPH (figs. 6-7) : The deutonymph is larger than the protonymph and the palps and legs have the same number of setae as in the adult stages. The podonotal, mesonotal and pygidial shields have fused to form a single dorsal scutum which has a lateral incision on either side, behind seta 57. The number of setae on the shield is the same as in the adult. On the ventral surface the sternal shield extends to the level of the posterior border of coxa IV. It bears four pairs of setae (the sternals and metasternals) and three pairs of pores. The peritremes are long, reaching coxae I. Metapodal plates are present, and also several small areas of sclerotization on the median part of the venter, ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. n CLASSIFICATION It has long been recognized that the genera Androlaelaps and Haemolaelaps are very closely related to one another as well as to the genus Hypoaspis Canestrini. Both Androlaelaps and Haemolaelaps were originally described by Berlese as sub- genera of Laelaps Koch. Subsequently Haemolaelaps (Berlese, 1916) and then also Androlaelaps (Vitzthum, 1943; Zumpt, 1950; Zumpt & Patterson, 1950) were placed as subgenera of Hypoaspis, and more recently both have been given generic status (Strandtmann, 1949; Zumpt & Patterson, 1951; Keegan, 1956; and others). There is still some confusion, however, regarding the delimitation of the genus Haemolaelaps. The genera Atricholaelaps Ewing and Ischnolaelaps Fonseca were synonymized with Haemolaelaps by Strandtmann (1949), a decision which has been accepted by most other workers on this group. Fonseca (1959), however, recommended that Haemolaelaps be kept as a monotypic genus for H. marsupialis Berlese, the remaining species being allocated to the genus Atricholaelaps with its subgenus Ischnolaelaps. He based these divisions on the form of the pilus dentilis, a feature which he himself admits is sometimes difficult to interpret. Eubrachylae- laps Ewing (= Cyclolaelaps Ewing) and Zygolaelaps Tipton were added to the synonymy by Zumpt and his co-workers (1951, 1958). Cavilaelaps Fonseca was included in the synonymy by Zumpt & Patterson (1951), but was not listed by Zumpt & Till (1961). Its status is still doubtful as I have not been able to examine the type species. The genus Turkiella Zumpt & Till was synonymized with Androlaelaps by Keegan (1956). The genera Androlaelaps and Haemolaelaps resemble one another very closely in their dorsal and ventral sclerotization and chaetotaxy, in their leg chaetotaxy and in the structure of the chelicerae. In both genera the sternal shield of the female is broader than long, the pilus dentilis is inflated or long and slender, and the digits of the male chelicerae are weakly sclerotized and edentate, the fixed digit being shorter than the movable one. The only differences between Androlaelaps and Haemolaelaps are that in females of the former genus leg II is relatively more stout and one of the ventral setae on femur II is thickened to form a spur-like structure. These features, however, are not clear-cut. There is a gradation both in the thickness of the second leg and in the nature of its setae from the Haemolaelaps to the Androlaelaps species. Furthermore, males of the mesopicos-group of Haemolaelaps species have a stout spur-like seta in the position corresponding to the Androlaelaps spur, whilst the females have only a simple seta. The genus Hypoaspis differs from Androlaelaps and Haemolaelaps in that the length of the sternal shield of the female usually equals or exceeds the breadth ; the pilus dentilis is short, slender, scarcely visible, and the digits of the male chelicera are well sclerotized, equal in length, with dentate internal margins. The genus Gaeolaelaps 1 Tragardh, with Laelaps aculeifer Canestrini as type species, appears to bear the same relationship to Hypoaspis as Androlaelaps does to Haemolaelaps. It seems possible that these mites may all have arisen from a common free-living 1 The genus Gaeolaelaps was mentioned in Berlese's note book but was never described. Berlese referred to the genus Geolaelaps in 1924 (Redia 15 : 254), but without designating a type. Tragardh (1952, Ark, Zool. 4 ser, 2 : 66) states that Berlese (in MS) had designated L, aculeifer as the type. 12 W. M. TILL Hypoaspis-\ike ancestor, giving rise on the one hand to species with simple setae on leg II (Hypoaspis} and on the other to species having an enlarged second leg with thickened setae (Gaeolaelaps} . Modified, elongate, edentate male chelae could have developed in the mites of both branches which became adapted to a parasitic mode of life, giving rise to species of the Haemolaelaps and Androlaelaps types respectively. In deciding upon the status of these four genera there seemed to be three possibilities : 1. To give all four generic status. 2. On the basis of leg setae to combine Haemolaelaps with Hypoaspis and Gaeolaelaps with Androlaelaps. 3. On the basis of male chelae, female sternal plate and pilus dentilis, to combine Haemolaelaps with Androlaelaps and Gaeolaelaps with Hypoaspis. Of these three propositions the third seems to be the most appropriate and Haemolaelaps has consequently been synonymized with Androlaelaps. In this connection it is interesting to note that two species described recently (H. travisi Delfinado 1961 and H. vietnamensis Grochovskaya and Nguyen-Xuan-Hoe 1961) as having the Androlaelaps type of leg setation have been allocated to the genus Haemolaelaps by their respective authors. Bregetova (1956), in her diagnosis of the genus Androlaelaps, describes the chelicerae as being similar in structure in both sexes, the male having a spermadactyl added parallel to the movable digit. This is the type of chelicera found in males of the genus Hypoaspis. However, Berlese's (1887) drawing of A. hermaphrodita, the type species of Androlaelaps, shows that the male chelae are elongate and edentate. The Androlaelaps species described from the Ethiopian region are clearly co-generic with A. hermaphrodita, whereas those listed by Bregetova (.4. sardous Berlese, A. karawaiewi Berlese and A. pavlovskii Bregetova) are of the Hypoaspis type. In the key which I have given to the females of the genus Androlaelaps s. lot., couplets 2 to 5 refer to the species bearing the modified spur-like seta (Androlaelaps- group), whilst the remaining couplets refer to species previously assigned to the genus Haemolaelaps (Haemolaelaps-group). On the basis of morphological features, many of the species which I have included in the Haemolaelaps-group could be arranged in well-defined sub-groups. A. cryptomius, A. scapularis and A. tauffliebi, three species parasitic on mole-rats, form a distinct group having the movable cheliceral digit in two parts, microsetae on the dorsal shield, and " granular " dorsal and ventral idiosomal shields. The status of these mites may have to be reconsidered and the group regarded either as a subgenus of Androlaelaps or even as a distinct genus. A. taterae and A. villosissimus differ from all other Androlaelaps species in their dorsal chaetotaxy. A. centrocarpus, A. hirsti and A. longipes form a fairly well- defined group having 40-41 pairs of dorsal setae, nine setae on tibia III and a separate anal shield in the male. The mesopicos-group, comprising four species associated with birds, namely, ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 13 A. haydocki, A. mesopicos, A. steyni and A. wilkini, is characterized by the sinuous posterior setae, the position of the first pair of sternal setae on the presternal area and by the presence of nine setae on tibia III. The zulu-glasgowi-group comprises a number of apparently very closely allied species (A. dasymys s. lat., A. glasgowi, A. pachyptilae, A. patersoni, A. spreo, A. suncus and A. zulu) whose status is not quite clear, and it is possible that some of them are no more than subspecies or varieties. A. glasgowi represents a cosmopolitan species complex which has been recorded from a wide variety of mammalian and bird hosts (Strandtmann, 1949 ; Bregetova, 1956). According to Bregetova, this species has only 38 pairs of setae on the dorsal shield, setae 53 being absent. This has been confirmed in specimens from Japan and Astrakhan in the collection of the British Museum (Natural History) . American specimens which I have been able to examine, namely, the type specimen of A. glasgowi (lent by the Smithsonian Institute, Washington) and several specimens from the collection of the South African Institute for Medical Research, have 39 pairs of setae. Mites from Israel (Costa, 1961) and from Southern Africa also have 39 pairs of dorsal setae as well as a number of asymmetrically arranged setae on the posterior part of the dorsal shield. There are also slight morphological differences, chiefly in relative measurements, between mites from different host species (Strandtmann, 1949). A. zulu very closely resembles A. glasgowi, differing from it in having a larger genital shield and longer setae on the dorsal shield. In both species the pilus dentilis is strongly inflated proximally, the terminal portion being slender and recurved. Specimens previously identified by Zumpt & Till (1956) as forms of A. glasgowi in which the genital shield is bordered by three pairs of setae would appear to belong to A . zulu. The remaining members of this group differ from A. glasgowi and A. zulu, and from one another, in the form of the pilus dentilis, the relative lengths of the arthrodial filaments at the base of the movable digit of the chelicera and the dorsal setae, and the relative measurements of parts of the body. A. dasymys s. lat. is considered to include A. davisi and A. labuschagnei which appear to differ from A. dasymys s. sir. only in size, degree of sclerotization and relative setal lengths. A. dasymys was based on a single damaged specimen which lacks most of the dorsal setae and of which the pilus dentilis is not clearly visible. By courtesy of Dr. P. L. G. Benoit (Musee Royal de 1'Afrique Centrale) three mites from Rhabdomys pumilio have been examined which appear to be identical with A. dasymys and which are in rather better condition. These mites are relatively large and robust, the dorsal shield measuring 1025-1064^ by 700-750^. The type series of A . davisi consists of rather smaller, less heavily sclerotized mites, in which the dorsal shield measures 822-888{jL by 516-569^. Examination of further material from the type host of A . davisi shows the dorsal chaetotaxy of this species to be variable in that accessory setae may be present or absent. Mites from different host species show a variation in the relative lengths of the dorsal setae. The type specimens of A . labuschagnei, which unfortunately are not in very good condition, appear to be intermediate between A . davisi and A . dasymys s. str. I 4 W. M. TILL A. dasymys s. lat. has been collected from a variety of hosts, but appears to be restricted to the Muridae. The remaining species in the genus Androlaelaps are more difficult to arrange in groups. Whatever feature is used for the purpose of separation, the resulting groups are heterogeneous with respect to other features. Family LAELAPTIDAE Genus Androlaelaps Berlese Androlaelaps Berlese, 1903, Zool. Anz. 27 : 14; Zumpt, 1950, Parasitology 40 : 299; Zumpt & Patterson, /. ent. Soc. S. Afr. 1951, 14 : 69. Haemolaelaps Berlese, 1910, Redia 6 : 216 & 1916, Redia 12 : 170; Strandtmann, ig^g.J.Parasit. 35 : 3 2 5I Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 70; Bregetova, 1956, Acad. Sci., U.S.S.R. 61 : 84; Fonseca, 1959, Mem. Inst. ButantanS. Paulo (1957/58) 28 : 46 (syn. nov.). Atricholaelaps Ewing, 1929, Manual of External Parasites : 186. Eubrachylaelaps Ewing, 1929, Manual of External Parasites : 186. 1 schnolaelaps Fonseca, 1936, Mem. Inst. Butantan S. Paulo 10 : 19. Cyclolaelaps Ewing, 1933, Proc. U.S. Nat. Mus. 82 : 5. Turkiella Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 215. Zygolaelaps Tipton, 1957, / Parasit. 43 : 367. Type species: Laelaps (Iphis) hermaphrodita Berlese, 1887 KEY TO THE FEMALES 1. Femur II with a stout, blunt or thorn-like seta on its ventral surface (Androlaelaps- group) . . 2 - Femur II with only simple setae on its ventral surface (Haemolaelaps-group) . . 6 2. Dorsal shield with seta ax present. Genital shield small, slightly widened behind genital setae, then tapering posteriorly (figs. 173-174) .... A. theseus Zumpt - Dorsal shield with seta ax absent. Genital shield large, flask-shaped, broadly rounded posteriorly ............. 3 3. Dorsal shield with setae px2 and px$ present ....... 4 - Dorsal shield with setae px2 and px$ absent ....... 5 4. Tarsus II terminating in pointed setae only. Genital shield very closely approached to anal shield. Genital setae relatively short, less than half GW 2 (fig. 188). A. zuluensis Zumpt Tarsus II with three stout, blunt, terminal spines. Genital shield widely separated from anal shield. Genital setae relatively long, approximately equal to GW 2 (fig. 98). A. marshalli Berlese 5. Integument of idiosoma with about 21 pairs of relatively short, fine setae (figs. 70-71). A. graingeri Zumpt & Patterson - Integument of idiosoma with about 37 pairs of long, coarse setae (fig. 10). A. arvicanthis Radford 6. Hypostomal processes forming two pairs of prominent brushes, one ventral and one dorsal (fig. 3) . . . . . . . . . . .A. walkerae sp. nov. Hypostomal processes fimbriated or pilose, but not forming brushes ... 7 7. Dorsal shield with an overall covering of setae which completely obscures the normal chaetotactic pattern (fig. 161) . . . . . .A. villosissimus (Berlese) A. taterae (Zpt. & Patt.) Dorsal shield with chaetotactic pattern composed of not more than 41 pairs of setae, excluding the accessory setae on the postero-median part of the shield ... 8 8. Corniculi weakly sclerotized (fig. 4) or reduced, inconspicuous .... 9 - Corniculi well sclerotized, normally developed, conspicuous (figs. 1-3, 5) . . 10 ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lai. 15 g. First pair of sternal setae inserted on anterior margin of sternal shield. Distance between genital and anal shields more than twice distance from anterior margin of anal shield to anterior margin of anus. Dorsal shield almost parallel-sided (figs. 131-132). A. sangsteri (Radford) First pair of sternal setae inserted on presternal area. Distance between genital and anal shields about half the distance from anterior margin of anal shield to anterior margin of anus. Dorsal shield ovoid (figs. 79-80) .... A. heliosciuri sp. nov. 10. Dorsal shield with seta /4 at least 5 times as long as seta J r (fig. 55). A. galagus (Lavoipierre) - Dorsal shield with setae /i and /4 subequal . . . . . . . n 11. Setae Z$ whip-like and sinuous (fig. 102). From birds (mesopicos-group) . . 12 - Setae Z$ normal. From mammals and birds . . . . . . . 15 12. Setae 24 and 22 very minute, their lengths about J the distance between their bases, or less. Sternal shield granular ........... 13 - Setae 14 and 22 longer, their lengths at least the distance between their bases. Sternal shield reticulate (fig. 182-183) ....... A. wilkini (Till) 13. Anterior seta of coxa III thick, spine-like (fig. 148) ... A. steyni (Till) - Anterior seta of coxa III normal, hair-like . . . . . . . . 14 14. Marginal setae of posterior half of dorsal shield simple (fig. 102) A mesopicos (Radford) Two pairs of marginal setae (82 and 53) of posterior half of dorsal shield whip-like, sinuous (fig. 78) A. haydocki (Till) 15. Some leg segments with pronounced chitinized swellings (fig. 18) . A. callosus (Berlese) Legs without swellings ........... iG 16. Dorsal shield with seta px i developed (fig. 83) ....... 17 Dorsal shield with seta pxi absent ......... 19 17. Seta ax present (figs. 31, 83) .......... 18 - Seta ax absent (fig. 93) ....... A. longipes (Bregetova) 1 8. Posterior dorsal shield without or with a few accessory setae restricted to the region /3 to /5 (fig. 83) . .... .A. hirsti (Keegan) Posterior dorsal shield with numerous accessory setae extending anterior to seta/i (fig. 31). A. centrocarpus (Berlese) 19. Tibia III with 9 setae ........... 20 Tibia III with 8 setae ........... 24 20. Genital shield greatly expanded, greatest width at least 1-6 times width at level of genital setae .............. 21 - Genital shield not, or only slightly expanded, greatest width not exceeding 1-3 times width at level of genital setae . . . . . . . . . . .23 21. Genu III with 10 setae ....... A. congoensis sp. nov. Genu III with 9 setae ........... 22 22. Lengths of setae 24 and z-z at least twice the distance between their bases (fig>. 129). Tibia 1 with 14 setae ........ A. rhodesiensis (Zpt. & Patt.) Lengths of setae 44 and z-z less than half the distance between their bases (fig. 43). Tibia I with 13 setae ........ A. cricetomydis sp. nov. 23. Tarsus II terminating in pointed setae only. Lengths of setae 24 and 22 approximately J, or less, the distance between their bases. Dorsal shield with 37 pairs of setae (fig. 126). A. phoeniculi (Zumpt & Till) Tarsus II with at least one of the terminal setae stout and blunt. Lengths of setae ^4 and Z2. approximately equal to the distance between their bases. Dorsal shield with 39 pairs of setae (figs. 115, 166) ....... A. tateronis (Radford) A. oliffl (Zumpt & Patterson) 24. Tarsus II with 3 stout, blunt terminal spines (fig. 157) ..... 25 Tarsus II terminating in pointed setae only ..... 26 25. Sternal shield granular. Caruncle of tarsus IV normal, its length being i times the greatest width of the tarsus . A. tachyoryctes (Radford) iC W. M. TILL - Sternal shield faintly reticulated. Caruncle of tarsus IV elongated, its length being at least twice the greatest width of the tarsus. .... A. capensis (Hirst) 26. Movable digit of chelicera in two parts, one part being dentate; fixed digit of chelicera weakly sclerotized, membranous in appearance (figs. 135, 170). Setae 24 and 22 very minute, their lengths not exceeding J the distance between their bases (fig. 134). (scapularis- group) . 27 - Movable digit of chelicera normal; fixed digit normally developed, well sclerotized. Setae 14 and z^ relatively longer, their lengths nearly half the distance between their bases, or longer ............. 29 27. Genital shield short, widely separated from anal shield (fig. 45) A. cryptomius (Radford) - Genital shield large, flask-shaped, extending almost to anal shield .... 28 28. Width of sternal shield at most 2 \ times its length. Arthrodial filaments at base of movable digit of chelicera subequal in length (figs. 133, 135) . . A. scapularis (Berlese) - Width of sternal shield 3 times its length. Four or 5 of the arthrodial filaments at the base of the movable digit of the chelicera conspicuously longer and stouter than the others (figs. 168, 170) A. tauffliebi sp. nov. 29. Genu IV with 9 setae A. spatuliformis (Lavoipierre) Genu IV with 10 setae ........... 30 30. Integument of venter, behind coxae IV, relatively densely haired, about 15 pairs, or more, of truly vental setae (figs. 59, 107) ......... 31 Integument of venter relatively sparsely haired, about 10 pairs, or less, of truly ventral setae .............. 32 31. Chelicera with arthrodial filaments subequal in length. Pilus dentilis moderately and more or less evenly inflated, elongate (fig. 61). Hypostomal processes very large and hairy (fig. 5) . ..... A. georychi sp. nov. - Chelicera with one very long arthrodial filament. Pilus dentilis strongly inflated basally, terminal portion slender, curved (fig. 109). Hypostomal processes normal (fig. i). A. murinus (Berlese) 32. Setae on dorsal shield very long and thick; seta Ji reaches almost to base of 74; lengths of setae 24 and 22 at least twice the distance between their bases (fig. 63). A. ghanensis sp. nov. Setae on dorsal shield relatively shorter and more slender; seta /i does not reach base of /3; lengths of setae 24 and 22 less than if the distance between their bases . . 33 33. Pilus dentilis slender (figs. 27, 91) ......... 34 Pilus dentilis inflated (figs. 187, 195) ......... 35 34. Genital shield widest anterior to first pair of flanking setae, then tapering posteriorly. Distance between genital and anal shields more than twice distance from anterior margin of anal shield to anus (fig. 89) A. hystrici (Zumpt & Till) Genital shield widest at level of second pair of flanking setae, broadly rounded posteriorly. Distance between genital and anal shields not exceeding distance from anterior margin of anal shield to anus (fig. 25) A. casalis (Berlese) 35. Anal shield i times longer than broad (fig. 193). . . A. zutnpti sp. nov. - Anal shield about as long as broad .... zulu-glasgowi complex a. Parasites of birds ........... b Parasites of mammals .......... d b. Tarsus IV with length to greatest width not greater than 5:1. Seta /i extends to a point approximately level with base of seta /2 (fig. 118). A. pachyptilae (Zumpt & Till) Tarsus IV with length to greatest width about 7:1. Seta /i extends well beyond base of /2, at least half way to base of /3 . . . . . . . c c. Pilus dentilis spatulate, terminating in a short, slightly bent, sharp point (fig. 122) A. patersoni (Zumpt & Till) Pilus dentilis with a larger proximal and a smaller distal inflation, the latter terminating in a very fine point (fig. 143) .... A. spreo (Zumpt & Till) ETHIOPIAN MITES OF THE GENUS ANDBOLAELAPS BERLESE s. la/. 17 d. Chelae 50-65^ long; 2 or 3 of the arthrodial filaments at the base of the movable digit distinctly longer than the others. Pilus dentilis as in fig. 53 A. dasymys (Radford) - Chelae 30-40(0. long; arthrodial filaments at base of movable digit usually subequal in length e e. Pilus dentilis with distal portion inflated (fig. 153). Tarsus IV with length to greatest width not more than 5:1 . . . . . A. suncus sp. nov. Pilus dentilis with distal portion slender, usually curved (figs. 67, 187). Tarsus IV with length to greatest width at least 6:1. . . . . . . f FIGS. lo-ii. Androlaelaps arvicanthis Radford, female. Venter (Fig. 10) ; dorsum (Fig. n). /. Setae of dorsal shield relatively short, lengths of 24 and Z2 at most slightly exceed the distance between their bases. Genital shield bordered by two pairs of setae (figs. 65- 66) A. glasgowi (Ewing) - Setae of dorsal shield relatively long, lengths of 34 and zz at least i times the distance between their bases. Genital shield bordered by three pairs of setae (figs. 185-186). A. zulu (Berlesc) Androlaelaps arvicanthis Radford Androlaelaps arvicanthis Radford, 1944, Parasitology 35 : 162, fig. 4. Turkiella arvicanthis, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 225. Androlaelaps equatoriensis Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 230, fig. 15. (synnov.). ZOOL. IO, I i8 W. M. TILL FEMALE (figs. 10-13) : Deutosternum with 6 rows of teeth, 4 to 5 in each row. Chelae 84(0, long, both digits dentate; pilus dentilis slender, arthrodial filaments at base of movable digit subequal in length. Dorsal shield relatively broad, with an overall reticular pattern, well sculptured, especially anteriorly. DL 1235-1 264^; DW= 950-1045^; DL/DW= 1-2-1-3. The shield bears only 37 pairs of setae, px2 and px% being absent. Seta ri distinctly longer than seta si ; lengths of setae 24 and Z2 about i| times the distance between their bases; seta Z$ about ij times as long as seta 75. Sternal shield with a distinct reticular pattern as well as a fine granulation, anterior and posterior margins very slightly concave. SL= 114-1 24|A; SW=257(x; SL/SW=o-4-o-5. First pair of sternal setae relatively short, not reaching posterior margin of sternal shield ; second and third pairs and metasternal setae considerably longer. Genital shield flask-shaped, with a reticular and granular pattern. GW X = 257-284^; GW 2 2igy.; GW 1 /GW 2 =i-2-i-3. Genital setae relatively long, approximately equal to GW 2 , and extending to bases of second pair of flanking setae. Anal shield with linear and granular markings, anterior margin straight or convex, anus situated at less than its length from the anterior margin. AL=i8o(j,; AW= 200-205^; AL/AW=o-9. Paranal setae level with posterior half of anus, about 4/5 as long as postanal seta. Integument of idiosoma with about 37 pairs of long ventral and marginal setae, many of which are barbed. Metapodal plates broadly oval, length to width about 4 =3- Chaetotaxy of legs normal, anterior and posterior setae on coxa I subequal in length. Inner ventral seta on tibia II slightly shorter and stouter than outer seta; inner ventral seta on genu II slightly shorter and stouter than outer seta, but longer than in A. marshalli. Tarsus II bears three stout, blunt, terminal setae. Length of tarsus IV about 7 times its width at the base. MALE (figs. 14-15): Deutosternum as in female; chelicera as illustrated. Chaetotaxy of dorsal shield as in female. DL= 1064-1092^; DW= 779-884^1; DL/DW=i-2-i-4. Holoventral shield reticulate, expanded behind coxae IV, bearing 28-29 setae, 15-16 being situated in the preanal region. Chaetotaxy of legs as in female, but inner ventral seta on tibia II stouter, resembling the condition in A. marshalli females. HOSTS AND LOCALITIES. Arvicanthis niloticus (Desmarest), Bunyoro, Uganda (types in B.M.N.H.). Tatera valida (Bocage), Bunyoro, Uganda (S.A.I.M.R.). Taiera species, Kerio Valley, Kenya and W. Nile district, Uganda (S.A.I.M.R.). Rattus natalensis (Smith), an Elephant Shrew (single record) and an unidentified rodent, Kerio Valley, Kenya (S.A.I.M.R.). Unidentified rodent, Yei, Sudan (paratype of A. equatoriensis Keegan in B.M.N.H.). Androlaelaps callosus (Berlese) Hypoaspis (Haemolaelaps) callosits Berlese, 1916, Redia 12 : 171. Cyclolaelaps lophuromius Radford, 1939, Parasitology 31 : 243, fig. i (syn. nov.}. ETHIOPIAN MITES OF THE GENUS AN DRO LA EL A PS BERLESE s. la/. 19 Hypoaspis (Haemolaelaps) lophuromius, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 70. Haemolaelaps lophuromius, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa, 10 : 238. Haemolaelaps sudanicus Zumpt & Till, 1954, ^ ev - Ecuat. Ent. Parasit. 2 : 213, fig. 3; Kee^an, 1956, /. Egypt, publ. tilth. Ass. 31 : 237, fig. 22 (syn. nov.). FEMALE (figs. 16-19): Deutosternum with 6 rows of 2 to 4 fine teeth. Chelae 40-46^1 long; pilus dentilis with basal 4/5 inflated, apex sharply recurved according 13 14 15 FIGS. 12-15. Androlaelaps arvicanthis Radford. Chelicera (Fig. 12) and ventral view of femur, genu and tibia of leg II (Fig. 13) of female. Chelicera (Fig. 14) and venter (Fig. 15) of male. to Keegan (1956) but indistinct in the specimens examined; one of the arthrodial filaments at the base of the movable digit longer than the others. Dorsal shield reticulate, relatively broad. DL=6oo-620(j.; DW= 445-486^; DL/DW=i'3-i'4. It bears 39 pairs of setae and one accessory seta at the level of /4. Anterior and marginal setae longer and stouter than those on central part of shield and provided with fine barbs. Lengths of setae 14 and Z2 approximately half the distance between their bases; seta Z$ about i| times as long as seta /5_ W. M. TILL Sternal shield with a reticular pattern which is most pronounced antero-laterally and indistinct on the central part of the shield. SL= 83-93^; SW= 134-145^; SL/SW=o-6-o-7. Sternal setae stout, first pair the same length as the metasternal setae, second and third pairs slightly longer. Genital shield granular and also with a reticulate pattern which is well marked FIGS. 16-19. Androlaelaps callosus (Berlese), female. Venter (Fig. 16); dorsum (Fig. 17) ; femur, genu, tibia and tarsus of leg IV (Fig. 18) ; chelicera (Fig. 19). posteriorly but less distinct anteriorly. GW X = 130^1; GW 2 = 118-127^. Length of genital setae less than half GW 2 . Anal shield approximately as wide as long, the length and width varying between 103^ and ii4(x. Anus situated at approximately its length from the anterior margin of the shield; paranal setae level with middle of anus, slightly shorter than postanal seta. Integument of idiosoma bears 25 pairs of ventral and marginal setae, the latter being barbed. Metapodal plates ovoid, their length to width about 2:1. Chaetotaxy of legs normal ; anterior seta on coxa I about 4/5 as long as posterior seta. Length of tarsus IV 4! times its width at the base. This species is charac- ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lal. zi terized by the strongly chitinized swellings on several of the leg segments, namely, femur, tibia and tarsus of leg I, femur II, tarsus III, tibia and tarsus of leg IV. MALE: Not known. HOSTS AND LOCALITIES: Lophuromys flavopimctatus Thomas (=Lophuromys zend) from Kenya (Berlese, 1916; Keegan, 1956), from Kampala, Uganda (syntype of C. lophuromius in B.M.N.H.), and from Torit, Sudan (type series of H. sudanicus in S.A.I.M.R.). Unidentified rodent from Uganda (B.M.N.H.). Keegan (1956) records this species from Tachyoryctes sp. and a mixed collection of Otomys and Arvicanthis species in the Rift Valley Province, Kenya. Androlaelaps capensis (Hirst) Haemolaelaps capensis Hirst, 1916, /. zool. Res. i : 79, figs, 12-14. Hypoaspis (Haemolaelaps} capensis, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 72. Haemolaelaps capensis, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa, 10 : 246. Haemolaelaps spinitarsus Berlese, 1918, Redia 13 : 123; Keegan, 1956, Trans. Amer. micr. Soc. 75 : 314, figs. 11-12; Zumpt & Till, 1958, /. ent. Soc. S. Afr. 21 : 267. Haemolaelaps mystromys Radford i942a, Parasitology 34 : 300, figs. 3-4 (syn. nov.). FEMALE (figs. 20-22) : Deutosternum with 6 rows of 3-4 small teeth. Chelae 56-64^ long; pilus dentilis slender; arthrodial filaments subequal in length. Dorsal shield ovoid, reticulate. DL=656-827[z; DW= 394-500^; DL/DW= 1-6-1-7. It bears 39 pairs of setae of moderate length, one or two unpaired accessory setae sometimes being present. Lengths of 24 and Z2 about 4/5 the distance between their bases; anterior and marginal setae a little longer; seta Z$ at least i \ times as long as /5. Sternal shield reticulate, merging anteriorly with presternal area, pattern weakly developed posteriorly. SL= 105-133^; SW= 142-150^; SL/SW= 0-7-0-9. Genital shield apparently smooth, rather slender, and may be slightly widened behind the genital setae. GW^ 102-128^; GW 2 = 94-106^; GW 1 /GW 2 =i-i-i-3. Length of genital setae about 2/3 GW 2 . Anal shield pear-shaped, 1-3-1-5 times longer than broad. AL= 114-134^; AW=76-ioo[z. Anus situated at approximately its length from the anterior margin. Paranal setae level with middle of anus, about 4/5 as long as postanal seta. Integument of idiosoma bears 23 pairs of ventral and marginal setae, some of the latter having fine barbs. Metapodal plates elongate, length to width about 4:1. Chaetotaxy of legs normal. Anterior seta on coxa I about 4/5 as long as posterior seta. Tarsus II has 3 stout, blunt, terminal spines and 3 stout ventral spines. Caruncle of tarsus IV relatively long, its length being 2f times the greatest width of the tarsus. Length of tarsus IV about 8 times its width at the base. MALE (figs. 23-24): Chelicera as figured, with a slender pilus dentilis. Dorsal shield reticulate, chaetotaxy as in female. DL=570[j.; DW=323[x; DL/DW=i-8. Holoventral shield reticulate, with 23 setae. Integument bears about 23 pairs of marginal and ventral setae, Chaetotaxy of legs as in female, but femur II with a stout spur, 22 W. M. TILL FIGS. 20-24. Androlaelaps capensis (Hirst). Venter (Fig. 20), dorsum (Fig. 21) and chelicera (Fig. 22) of female, Venter (Fig. 23) and chelicera (Fig. 24) of male, ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 23 HOSTS AND LOCALITIES. Cryptomys hottentotus (Lesson), Grahamstown, Cape Province (type series in B.M.N.H.) and several localities in Natal and the Transvaal (S.A.I.M.R.). Cryptomys holosericeus (Wagner), Kruger National Park, Transvaal (S.A.I.M.R.). Georychus capensis (Pallas), Knysna, Cape Province (S.A.I.M.R.). Mystromys albicaudatus (Smith), Albany, Cape Province (types of H. mystromys in B.M.N.H.). Androlaelaps casalis (Berlese) Iphis casalis Berlese, 1887, Acari, Myr. Scorp. Ital. Fasc. 38 N.8, figs. 3-5. Haemolaelaps casalis, Bregetova, 1956, Acad. Sci. U.S.S.R. 61 : 89 & 99, figs. 143-4. Hypoaspis oculatus Oudemans, 1915, Arch. Naturgesch. (A) 81 : 134. Haemolaelaps molestus Oudemans, 1929, Ent. Ber. 8 (169) : 13; Willmann, 1939, Abh. Nat. Ver. Bremen 31 : 173, figs. 5-7. Atricholaelaps megaventralis Strandtmann, 1947, Proc. ent. Soc. Wash. 49 : 112. Hypoaspis freemani Hughes, 1948, Mites associated with stored food products: 129, figs. 173-179; Strandtmann, 1956, /. Kansas ent. Soc. 29 : 138. Haemolaelaps haemorrhagicus Asanuma, 1952, Misc. Rep. Res. Inst. Nat. Res. No. 25 : 87, fig. i. (For further references see Strandtmann & Wharton, 1958, Manual of Mesostigmatid Mites: 34)- FEMALE (figs. 25-27) : Deutosternum with 6 rows each of 5 fine teeth. Chelae 34-40^ long ; pilus dentilis slender ; arthrodial filaments subequal in length. Dorsal shield oval, reticulate. DL=62o-7iojj.; DW= 376-429^; DL/DW 1-5-1-7. It bears 39 pairs of symmetrically arranged setae, as well as a variable number of unpaired accessory setae between the / series. Setae 34 and Z2 are longer than half the distance between their bases; the longer posterior setae are barbed; seta Z$ is ii-i|- times as long as /5. Sternal shield reticulate; SL= 86-96^; SW= 124-140^; SL/SW=o-7-o-8. First pair of sternal setae about as long as the metasternal setae, but a little coarser; second and third pairs of sternal setae slightly longer than first pair. Genital shield expanded and broadly rounded posteriorly, with a well-defined reticulate pattern which extends at least as far as the anterior flap of the shield. GW 1 =i43-i66[jL; GW 2 = 96-105^; GW 1 /GW 2 =i-4-i-6. Genital setae short, less than half GW 2 . Anal shield reticulate, approximately as long as broad. AL=9O-H4fx; AW= 96-ii4fji. Anus situated at about its length from the anterior margin. Paranal setae inserted on a level between middle and posterior end of anus, and approximately the same length as the postanal seta. Integument of idiosoma bears about 21 pairs of ventral and marginal setae. Metapodal plates elongate, length 4-7 times the width. Chaetotaxy of legs normal. Posterior seta on coxa I slightly longer than anterior seta. Three ventral and 3 apical setae on tarsus II are thickened, but do not form blunt spines. Setae generally rather fine and of moderate length. Length of tarsus IV about 6 times width at base. MALE (figs. 28-29) : Chelicerae as figured, the fixed digit bearing a slender pilus dentilis. Chaetotaxy of dorsal shield as in the female. DL= 466-485^; DW= W. M. TILL 27 28 29 FIGS. 25-29. Androlaelaps casalis (Berlese). Venter (Fig. 25), dorsum (Fig. 26) and chelicera (Fig. 27) of female. Venter (Fig. 28) and chelicera (Fig. 29) of male. ETHIOPIAN MITES OE THE GENUS ANDROLAELAPS BERLESE s. hit. 25 284-322(0.; DL/DW 1-5-1-6. Holoventral shield reticulate, expanded behind coxae IV, bearing 23 setae. Chaetotaxy of legs as in female. HOSTS AND LOCALITIES : Cosmopolitan, from a wide variety of bird and mammalian hosts, and also found in straw, hay, detritus, etc. Recorded from the Ethiopian region from Rattus rattus (Linnaeus) and Xerus inauris (Zimmermann), Transvaal, and from Sciurus carolinensis Gmelin, Cape Province (S.A.I.M.R.). Androlaelaps centrocarpus (Berlese) Haemolaelaps centrocarpus Berlese, 1911, Redia 7 1432; Costa, 1961, Bull. Brit. Mus. (nat. Hist.) Zool. 8 : 27, figs. 39~44- Haemolaelaps mauritanicus Hirst, 1925, Bull. Com. Etud. Hist. Sci. : 98, fig. 3 (syn. nov.). FEMALE (figs. 30-32) : Deutosternum with 6 rows of 2-3 teeth. Chelae 6o[o, long; pilus dentilis slightly inflated, tapering towards the tip ; arthrodial filaments subequal in length. Dorsal shield reticulate, widest at level of setae r6, then tapering gradually posteriorly. DL= 836-911(0.; DW= 485-523(0.; DL/DW= 1-7-1 -8. The shield bears additional paired setae in the ax and pxi positions, as well as numerous unpaired setae on the postero-median part and extending anteriorly as far as setae 25. The setae are long and subequal in length, except for /5, which are shorter. Lengths of setae 24 and 22 approximately i| times the distance between their bases; seta Z5 at least i| times as long as seta /5. Some marginal setae have fine barbs. Sternal shield reticulate. SL= 105-130(0.; SW= 180-190(0.; SL/SW=o-6-O7. Metasternal setae relatively short, approximately half as long as the sternal setae. Genital shield granular, with a weakly defined reticular pattern posteriorly; slightly widened behind the genital setae, then tapering abruptly. GW X = 143-162(0. ; GW 2 = 130-143(0.; GW 1 /GW 2 =i-i. Genital setae fairly long, about 4/5 GW 2 , reaching to base of second pair of flanking setae. Anal shield pear-shaped, anterior margin straight. AL= 120-145(1; AW= 110-124(0.; AL/AW=i-i-i-3. Anus situated at less than its length from the anterior margin. Paranal setae level approximately with middle of anus, very slightly longer than postanal seta. Integument of idiosoma bears numerous long ventral and marginal setae, many of the latter being barbed. Metapodal plates elongate, length to width about 6 : i. Chaetotaxy of legs normal except that tibia III bears 9 setae. Posterior seta of coxa I thicker than anterior seta, but approximately the same length. Femur I has 2 short, stout, spine-like setae on its dorsal surface, and trochanter I has a stout, spine-like dorsal seta. Leg II has a ventral spine-like seta on the trochanter, femur and genu, and 2 on the tibia. Tarsus II has 3 spine-like ventral setae, and 3 stout, blunt, apical setae. The upper dorsal seta on trochanter IV is considerably thicker than the other setae on this segment. Length of tarsus IV nearly 10 times width at base. MALE (figs. 33-34) : Chelicerae as figured, with a pilus dentilis resembling that of the female. Dorsal shield as in female. And shield separated from sternito- ventral shield, the latter strongly reticulate and bearing 7 pairs of setae, 2 pairs being placed on the area posterior to coxae IV. Chaetotaxy of legs as in female. 26 W. M. TILL 34 33 FIGS. 30-34. Androlaelaps centrocarpus (Berlese). Venter (Fig. 30), dorsum (Fig. 31) and chelicera (Fig. 32) of female. Venter (Fig. 33) and chelicera (Fig. 34) of male. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE 5. lat. 27 HOSTS AND LOCALITIES. Berlese recorded this species from " Mures " in British Somaliland. The above description is based on females from Taterillm species, Kerio Valley, Kenya (S.A.I.M.R.) and on males from Gerbillus gerbillus Olivier, Israel (received by courtesy of Dr. M. Costa). FIGS. 35-37. Androlaelaps congoensis sp. nov., female. Venter (Fig. 35); dorsum (Fig. 36); chelicera (Fig. 37). NOTE : The specimens described above have been compared with drawings of the type made by Dr. G. O. Evans. Haemolaelaps mauritanicus Hirst, from Gerbillus pyramidum Geoffrey, Mauritania, is most probably a synonym of A. centrocarpus. It is slightly smaller, its setae are relatively finer, and the ventral setation is more symmetrical than in A. centrocarpus. 28 W. M. TILL Androlaelaps congoensis sp. nov. FEMALE (figs. 35-37): Deutosternum with 6 rows of 3-5 very fine teeth and an ill-defined 7th row with 4 minute teeth. Chelae 48^ long, pilus dentilis small, slightly inflated, terminating in a sharp point ; arthrodial filaments short, subequal in length. 38 39 40 FIGS. 38-41. Androlaelaps congoensis sp. nov., venter (Fig. 38) and chelicera (Fig. 39) of male. Androlaelaps cricetomydis sp. nov.. venter (Fig. 40) and chelicera (Fig. 41) of male. Dorsal shield reticulate. 01^=703-722^; DW= 456-504^; DL/DW=i-4~i'6. The shield bears the usual 39 pairs of setae as well as 11-13 unpaired accessory setae between the / series. Lengths of setae 24 and Z2 approximately equal to the distance between their bases ; seta Z5 twice as long as seta 75. Setae all apparently simple except seta Z$ which has very fine barbs. Sternal shield reticulate, merging with presternal area, SL (from level of first ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BEKLESE 5. la/. 29 sternal setae) = 105(0,; SW 114-120(0.; SL/SW^o-g. Sternal and metasternal setae subequal in length. Genital shield patterned posteriorly, greatly expanded. GW X = 205-219(0.; GW 2 = 1 14-133(0.; GW 1 /GW 2 =i-6-l'8. Genital setae relatively short, slightly more than half GW 2 . Anal shield broader than long. AL= 109(0.; AW= 12410.; AL/AW=o-9. Anus situated at approximately its length from the anterior margin of the shield. Paranal setae near posterior margin of anus, extending beyond base of postanal seta and about 3/5 as long as postanal seta. Integument of idiosoma bears 21 pairs of ventral and marginal setae, all apparently simple. Metapodal plates elongate, length about 4 times the breadth. Genu III bears 10 setae and tibia III has 9. Anterior seta on coxa I is about 2/3 as long as posterior seta. Posterior ventral seta on femur IV extremely long, about 3/5 the total length of the femur. Tarsus IV long and slender, its length about 7 times its width at the base. MALE (figs. 38-39) : Deutosternum with 4 rows of very fine teeth, the 5th and 6th rows ill defined. Fixed digit of chelicera bears a short pilus dentilis, as in the female; spermadactyl about Sojo. long. Dorsal shield 560(0, long, 330(0, wide. Holoventral shield bears 23 setae. Chaeto- taxy of legs as in female. HOSTS AND LOCALITIES: Holotype ($), allotype (<^) and 7 female paratypes (1933.11.14.309-318 part) collected by F. W. ]. Cox from the nest of Spermestes cucullatus Swainson at Mulungu, Belgian Congo, 5 November, 1919. This species appears to be very common in birds' nests and has also been found in the nests of a serin finch, weaver, oriole, sparrow, bunting and wagtail, all from Zomba, Nyasaland. The specimens are all in the collection of the British Museum (Natural History) . Androlaelaps cricetomydis sp. nov. FEMALE (figs. 42-44) : Deutosternum bears 6 rows of 3-4 small teeth. Chelae about 45(0. long; pilus dentilis inflated; arthrodial filaments subequal in length. Dorsal shield reticulate, widest between the posterior setae Si and 82. DL= 850-883(0.; DW= 560-599(0.; DL/DW=i-5. The shield bears 39 pairs of setae, accessory setae being absent ; setae on central part of shield relatively short, lengths of 54 and Z2 being less than half the distance between their bases. Marginal setae longer, coarser and barbed; Si more than twice as long as /i, Z$ about i% times as long as /5. Sternal shield reticulate, heavily sclerotized. SL= 128-138(0.; SW= 190-195(0.; SL/SW=o-7. Genital shield greatly expanded behind the genital setae, posterior margin straight or slightly concave. The reticulate pattern extends anterior to the genital setae where it becomes less well defined. GW X = 238-247(0.; GW 2 = 128-138^; GW X /GW 2 = 1-8-1-9. Genital setae relatively short, less than half GW 2 . Anal shield slightly broader than long. AL= 133-144(0.; AW= 152(0.. Anus 3 o W. M. TILL situated at less than its length from the anterior margin. Paranal setae level with posterior margin of anus, about as long as postanal seta. Integument of idiosoma bears about 23 pairs of marginal and ventral setae. The 3 pairs of ventral setae which flank the genital shield are apparently simple, the remaining ventral and marginal setae are barbed. Metapodal plates approximately twice as long as broad. 42 43 FIGS. 42-44. Androlaelaps cricetomydis sp. nov. Venter (Fig. 42), dorsum (Fig. 43) and chelicera (Fig. 44) of female. Chaetotaxy of legs normal, except that tibia III bears 9 instead of 8 setae. Tarsus II bears 3 stout, ventral setae; terminal setae not particularly stout. Anterior and posterior setae on coxa I approximately equal in length. Length of tarsus IV about 6|- times its width at the base. MALE (figs. 40-41) : Deutosternum bears 6 rows of 2-5 teeth ; pilus dentilis inflated basally. Dorsal shield 650^ long, 437^ wide ; chaetotaxy similar to that of female. Holoventral shield reticulate, expanded behind coxae IV and bearing 23 setae, 5 pairs lying in the region between coxae IV and the anus. Chaetotaxy of legs similar to that of the female. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 31 HOST AND LOCALITY: The mites, received from Mr. H. J. Disney, were found in the food store of Cricetomys gambianus Waterhouse, Malengo Highlands, Tanganyika, 30 August, 1960. Holotype female (1962.6.12.1), allotype male (1962.6.12.2) and ten female paratypes (1962.6.12.3-7) in the collection of the British Museum (Natural History). 45 46 FIGS. 45-47. Androlaelaps cryptomius (Radford), female. Venter (Fig. 45); dorsum (Fig. 46); chelicera (Fig. 47). Androlaelaps cryptomius (Radford) Ischnolaelaps cryptomius Radford, 1939, Parasitology 31 : 248, fig. 5. Hypoaspis (Haemolaelaps) cryptomius, Zumpt & Patterson, 1951, J. ent. Soc. S. Afr. 14 : 71. Haemolaelaps cryptomius, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 241. Liponyssus lawrencei Zumpt & Patterson, 1951,7. ent. Soc. S. Afr. 14 : 89, fig. 10; Zumpt & Till, 1953, S. Afr. J . Med. Sci. 18 : 8 (syn. nov.). FEMALE (figs. 45-47) : Deutosternum provided with 6 rows of teeth, 2 large and 2 small teeth in the first row, i large tooth in the 2nd to 5th rows, and 2 small teeth in the sixth row. Anterior rostral seta about i| times as long as capitular seta. Chelae small, about 24^ long. Fixed digit weakly sclerotized, bearing a long, slender pilus dentilis; movable digit in two parts, one part edentate except for terminal hook, the other part weakly sclerotized but with large teeth. Arthrodial filaments not distinct in the specimens examined. W. M. TILL Dorsal shield widest at the level of setae ?5, with a reticulate and granular pattern. DL= 508-530^; DW= 350-385^; DL/DW=i-4-i-5. The shield bears 39 pairs of setae. Those on the central part are short and fine, the lengths of setae i<\ and Z2 being less than 1/5 the distance between their bases. The anterior and marginal setae are longer, and seta Z$ is approximately 2.\ times as long as seta /5- Sternal shield granular, with some reticulation near the lateral margins ; presternal area reticulate. SL=7o~75[j.; SW= 130-140^ ; SL/SW=o-5-o-6. First pair of 49 FIGS. 48-50. Androlaelaps cryptomius (Radford), venter of male (Fig. 48). Androlaelaps dasymys (Radford), venter (Fig. 49) and chelicera (Fig. 50) of male. sternal setae 4/5 as long as second and third pairs : metasternal setae about half as long as first pair of setae. Genital shield relatively short, granular, very slightly widened behind the genital setae, with a weakly defined reticular pattern. GW X = 96-104^; GW 2 = 84-88^. Genital setae relatively long, about 2/3 GW 2 . Anal shield approximately as long as broad, length and width varying between 94(x and 104^. Anus situated very close to anterior margin. Paranal setae level approximately with middle of anus, slightly longer than postanal seta, but not reaching the base of the latter. 33 ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s.lat. Integument of idiosoma bears about 24 pairs of ventral and marginal setae. Metapodal plates elongate, their length to width about 8 : i. Chaetotaxy of legs normal. Posterior seta on coxa I very slightly longer and thicker than anterior seta. The leg setae are all simple and are not modified to form thickened or blunt spines. Length of tarsus IV 5-6 times its width at the base. 51 52 FIGS. 51-53. Androlaelaps dasymys (Radford), female (from Mystromys albicaudatus) Venter (Fig. 51); dorsum (Fig. 52); chelicera (Fig. 53*2). Pilus dentilis of specimens from Tatera afra (Fig. 536) and Otomys sp. (Fig. 53c). MALE (fig. 48) : Structure of the chelicerae not very distinct in the only specimen available, but there appears to be a flagellar pilus dentilis as in the female. Dorsal shield similar to that of the female, but setae on central part relatively longer. DL=475pi.; DW = 284^; DL/DW=i-7. Holoventral shield expanded behind coxae IV and bears 21 setae. The shield has an overall reticulate pattern, and also a granular appearance especially anterior to coxae IV. Chaetotaxy of legs as in female. HOSTS AND LOCALITIES: Georychus capensis (Pallas) from Wynberg, Cape Province ZOOL. IO, I 3 34 W. M. TILL (types in Dr. Radford's collection and M.R.A.C.), and from Knysna and Drostdy, Cape Province (S.A.I.M.R.). Androlaelaps dasymys (Radford) (s. lat.) [schnolaelaps dasymys Radford, 1939, Parasitology 31 : 245, fig. 2. Hypoaspis (Haemolaelaps) dasymys, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 73. Haemolaelaps dasymys, Zumpt & Till, 1953, Ann. Inst. Med. trap. Lisboa 10 : 247. Hypoaspis (Haemolaelaps) labuschagnei Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 85, fig. 7 (syn. nov.). Haemolaelaps davisi Zumpt & Till, 1956, Z. Parasitenk 17 : 287, figs. 5-8 (syn. nov.). FEMALE (figs. 51-53) : Deutosternum with 6 rows of 3-4 teeth, one tooth in each row being larger than the others in some specimens. Chelae 50-65^ long; pilus dentilis as figured, its terminal portion appearing either swollen or ribbon-like, depending on the orientation; two or three of the arthrodial filaments distinctly longer than the others. Dorsal shield reticulate, with 39 pairs of setae; unpaired accessory setae may be present (2-6) or absent. Lengths of setae 24 and Z2 at least 4/5 the distance between their bases, up to 1-4 times this distance in specimens from Tatera afra, Rattus rattus, Parotomys species, and some specimens from Otomys species. Marginal setae coarse, barbed; seta Z$ 1-6-2-5 times as long as seta /5. DL= 822-1 064^; DW=5i6-750|^; DL/DW=i-4-i-7. Sternal shield reticulate, sharply demarcated from presternal area. SL= 123-163^.; SW=i54-200[x; SL/SW=o-8-i-o. Metasternal setae a little shorter than first pair of sternal setae, second and third pairs of sternal setae slightly longer than first pair. Genital shield with a distinct reticulate pattern extending almost to the anterior extremity. GWj= 170-209^; GW 2 =ii4-i65[x; GW 1 /GW 2 =i-3-i-5. Length of genital seta slightly more than half GW 2 . Anal shield reticulate anteriorly and laterally, approximately as long as broad. AL=i27-200[z; AW=i27-i95[x. Anus situated at a little less than its length from the anterior margin ; paranal setae level with middle of anus, 4/5 as long as postanal seta and extending beyond its base. Integument of idiosoma bears 18-28 pairs of ventral and marginal setae, the latter being longer, coarser, and provided with fine barbs. Metapodal plates vary in shape from roughly elliptical to almost round, their length 1-4-2-5 times their width. Chaetotaxy of legs normal. Posterior seta on coxa I very slightly longer and thicker than anterior seta. Tarsus II bears three stout ventral setae but no blunt spines ; trochanter IV bears a stout seta on its dorsal surface. Length of tarsus IV about 7 times its width at the base. MALE (figs. 49-50) : Chelicera as figured ; dorsal shield similar to that of female. DL=684-788[x; DW 418-420^; DL/DW= 1-6-1-9. Holoventral shield reticulate, expanded behind coxae IV, bearing 23 setae. Chaetotaxy of legs as in female. HOSTS AND LOCALITIES: Dasymys incomtus (Sundevall), Kakumiro, Uganda (syntype of /. dasymys in B.M.N.H.). ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lal. 35 56 57 FIGS. 54-58. Androlaelaps galagus (Lavoipierre). Venter (Fig. 54), dorsum (Fig. 55) and chelicera (Fig. 56) of female. Venter (Fig. 57) and chelicera (Fig. 58) of male. 36 W. M. TILL Mystromys albicaudatus (Smith) from Basutoland (type series of H. davisi, the figured specimens) and from Aliwal North, Cape Province (S.A.I.M.R.). Rattus rattus (Linnaeus) from the Transvaal (type series of H. labuschagnei in S.A.I.M.R.). Rattus natalensis (Smith) from Natal, Cape Province, Transvaal, Orange Free State (S.A.I.M.R.). Rattus chrysophilus (De Winton) and Rattus namaquensis (Smith) (Zumpt & Till, 1961; material not re-examined). Otomys irroratus (Brants) and Otomys species from several localities in South Africa (S.A.I.M.R.). Otomys saundersiae Roberts (Zumpt & Till, 1961; material not re-examined). Otomys sloggetti (Thomas), Basutoland (S.A.I.M.R.). Parotomys brantsi (Smith), Port Nolloth, Cape Province (S.A.I.M.R.). Parotomys littledalei Thomas, van Rhynsdorp, Cape Province (S.A.I.M.R.). Tatera afra (Gray), Basutoland and Transvaal (S.A.I.M.R.). Rhabdomys pumilio (Sparrman), Ngorongoro, Tanganyika (M.R.A.C.). Lophuromys sikapusi (Temminck), Cameroons (Taufflieb & Mouchet, 1959). Androlaelaps galagus (Lavoipierre) Haemolaelaps galagus Lavoipierre, 1955, Ann. trap. Med. Parasit. 49 : 304, figs. 5-6. FEMALE (figs. 54-56) : Deutosternum with 6 rows each of 2 small teeth. Chelae 54-56^1 long; pilus dentilis inflated, elongate, bent near tip and terminating in a fine point; arthrodial filaments subequal in length. Dorsal shield reticulate, bearing 39 pairs of setae. Those on the central part of the shield are very short and fine ; they are extremely minute and difficult to detect in the type specimen examined (by courtesy of Dr. M. M. J. Lavoipierre), but are more conspicuous in the specimens from Cryptomys sp., on which the drawings are based. Some of the setae are extremely long and barbed, namely, /4, 3-5 and 54-5; seta /4 is at least 5 times as long as /i. DL= 910-960^; DW= 605-709^; DL/DW=i-3-i-5. Sternal shield with a reticulate pattern which is most marked antero-laterally. SL=i45(j,; SW= 205-2 14^; SL/SW=o-6-o-7- First sternal seta i| times as long as metasternal seta, second and third sternal setae a little longer than first. Genital shield sculptured posteriorly, greatly expanded behind the genital setae. GW 1= 247-271 {j.; GW 2 =i52-i70[z; GW 1 /GW 2 = 1-5-17. Anal shield approximately as wide as long. AL=i43-i62(x; AW= 152-170^. Anus situated at approximately its length from the anterior margin. Parana! setae inserted on a level between middle and posterior end of anus, about f as long as postanal seta. Integument of idiosoma bears 26 pairs of marginal and ventral setae, many of which are barbed. Metapodal plates variable in shape, length to width about 2:1. Chaetotaxy of legs normal ; anterior and posterior setae on coxa I approximately the same length. Trochanter IV has 2 thick spine-like setae on its anterior surface, and femur IV has a stout ventral seta and a stout dorsal spine. Antero-dorsal seta ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 37 59 FIGS. 59-61, Androlaelaps georychi sp. nov., female. (Fig. 60); chelicera (Fig. 61). 60 Venter (Fig. 59) ; dorsum on this segment barbed, its length exceeding the width of the femur. Length of tarsus IV about 6 times its width at the base. MALE (figs. 57-58) : Chelicera with an inflated pilus dentilis which tapers distally. Dorsal shield as in female, but setae on central part relatively longer, and 5 accessory setae present between /3 and //j.. DL=694[j.; DW=456[z; DL/DW=i-5- Holo- ventral shield reticulate, expanded behind coxae IV, bearing 20-21 setae. Chaeto- taxy of legs as in female; one ventral seta on femur, genu, tibia and tarsus of leg II stout and spine-like. HOSTS AND LOCALITIES: Galago alleni Waterhouse, Barombikang Forest, British Cameroons (type). 38 W. M. TILL Cricetomys sp., burrow, Kivu, Belgian Congo (M.R.A.C.). Cricetomys gambianus Waterhouse, Yaounde, Cameroons (Taufflieb & Mouchet, 1959)- Androlaelaps georychi sp. nov. FEMALE (figs. 59-61) : Deutosternum with 6 rows of 3-4 very fine teeth, a larger median tooth in rows 2 to 4. Hypostomal processes very hairy and large, extending to anterior margin of palp femur. Chelae about 58 [i long; pilus dentilis elongate (about 40(ji long), moderately inflated; arthrodial filaments subequal in length. Dorsal shield reticulate, ovoid in shape, widest at level of seta sj. DL= 964-988^ ; DW=6i8[x; DL/DW=i-6. The shield bears the usual 39 pairs of setae, which are more or less subequal in length, and 5-6 accessory setae in the postero-median region. Lengths of setae 24 and Z2 about i-| times the distance between their bases; seta ^5 i| times as long as seta 75. Sternal shield reticulate; SL= 143-152^; SW= 185-200^; SL/SW= 0-7-0-8. Genital shield with a reticulate pattern which extends over the anterior flap; greatest width of shield 1-4-1-5 times width at level of genital setae (GW 2 ). Length of genital setae about 4/5 GW 2 . Anal shield approximately as broad as long. AL= 143-1 62 [z; AW =148-162^. Anus situated at less than its length from the anterior margin of the shield. Paranal setae near posterior margin of anus, extending to a point level with the posterior tip of the anal shield, their length about f that of the postanal seta. Integument of idiosoma bears numerous ventral and marginal setae which all appear to be barbed, except those adjacent to the genital shield. Metapodal plates short, ovoid, length to width not greater than 2:1. Chaetotaxy of legs normal. Length of tarsus IV about 8 times its greatest width. Caruncle IV elongated, its length more than twice the width of tarsus IV and about i^ times the length of caruncle III. HOST AND LOCALITY: Four females from Georychus capensis (Pallas), Knysna, Cape Province, 24 April, 1954. Holotype and one paratype in the collection of the S.A.I.M.R., Johannesburg, two paratypes (1962.6.12.8-9) in the collection of the British Museum (Natural History). Androlaelaps ghanensis sp. nov. FEMALE (figs. 62-64) : Deutosternum with 6 rows of 2-4 teeth. Chelae about 62 [x long; pilus dentilis moderately inflated; arthrodial filaments subequal in length. Dorsal shield reticulate, 817-903^ long, 570-580^ wide at the broadest point, DL/DW=i-4-i-6. The shield bears 39 pairs of setae and one unpaired accessory seta at the level of setae 74- All the setae except ii and ri are long and thick. Lengths of setae 24 and Z2 about twice the distance between their bases; seta Z$ about 1 1 times as long as seta 75 ; seta Ji reaches almost to the base of seta 74- Sternal shield reticulate, approximately as long as broad; SL=i7i-i8o(x; ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 39 SW=i8o[x. Metasternal setae about 2/3 as long as first pair of sternal setae, second and third pairs of sternal setae slightly longer. Genital shield widened near its posterior end and with a distinct pattern extending to its anterior margin. GW X = 143-152^ ; GW 2 =io5p.; GW 1 /GW 2 =i-4. Genital setae about as long as first pair of sternal setae, slightly shorter than GW 2 . 63 Venter (Fig. 62) ; dorsum 62 FIGS. 6264. Androlaelaps ghanensis sp. nov., female. (Fig. 63); chelicera (Fig. 64). Anal shield as long as broad or slightly longer. AL= 157-171^; AL/AW=i-o-i-i. Anus situated at approximately its length from the anterior margin of the shield. Paranal setae level with posterior margin of anus, about 3/5 as long as postanal seta. Integument of idiosoma bears about 27 pairs of ventral and marginal setae; the marginal and outermost ventral setae are all barbed. Metapodal plates oval in shape, length to width about 5 : 2. 4 o W. M. TILL Chaetotaxy of legs normal. Trochanter IV bears two very stout spines, one on the lower ventral and one on the upper dorsal surface. Femur IV bears a stout sword-like basal spine. Length of tarsus IV about 6 times the width at the base. HOST AND LOCALITY: Four females from a dormouse (Claviglis spec.) in Ghana, i January, 1911. Holotype (1962.6.12.10) and three paratypes (1962.6.12.11-13) in the collection of the British Museum (Natural History). Androlaelaps glasgowi (Ewing) Laelaps glasgowi Ewing, 1925, Proc. ent. Soc. Wash. 27 : 6. Haemolaelaps glasgowi, Strandtmann, 1949, /. Parasit. 35 : 343 ; figs. ; Zumpt & Till, 1956, Z. Parasitenk. 17 : 282, figs. 1-2. Laelaps californicus Ewing, 1925, Proc. ent. Soc. Wash. 27 : 5. Laelaps virginianus Ewing, 1925, Proc. ent. Soc. Wash. 27 : 6. Hypoaspis microti Oudemans, 1926; Ent. Ber. 7 : 101; Bregetova, 1952, Zool. Zh. 31 : 867. Haemolaelaps mohrae Oudemans, 1928, Ent. Ber. 7 : 374. Haemolaelaps cricetophilus Vitzthum, 1930, Zool. Jb. Jena, Abt. Syst. 60 : 417, figs. Laelaps stegemani Hefley, 1935, /. Kansas ent. Soc. 8 : 22. 1 ' schnolaelaps rhabdomys Radford, 1939, Parasitology 31 : 249, fig. 6 (syn. nov.). Haemolaelaps scalopi Keegan, 1946, Trans. Amer. Micr. Soc. 65 : 71. Atricholaelaps sigmodoni Strandtmann, 1946, J. Parasit. 32 : 164. Atricholaelaps strandtmanni Fox, 1947, Ann. ent. Soc. Amer. 40: 580, figs. Hypoaspis (Haemolaelaps) eos Zumpt & Patterson, 1951, J. ent. Soc. S. Afr. 14 : 79, fig. 2. (For further references see Strandtmann & Wharton, 1958, Manual of Mesostigmatid Mites : 37) FEMALE (figs. 65-67) : Deutosternum with 6 rows of 3-5 teeth. Chelae about 36(0, long; pilus dentilis strongly inflated basally, distal portion slender, curved or recurved; arthrodial filaments usually subequal in length, slightly variable in populations from Otomys irroratus. Dorsal shield reticulate, with a " double " margin. DL= 613-653^; DW= 394-436^; DL/DW= 1-5-1-6. The shield bears 39 pairs of setae and 2-3 unpaired accessory setae between setae /2 and 74- Lengths of setae 24 and 22 slightly exceed the distance between their bases; seta Z$ about twice as long as 75- Sternal shield reticulate and granular. SL 96-113^; SW= 130-155^; SL/SW 0-75. Genital shield slightly widened behind genital setae, its reticular pattern extending over the anterior flap. GW 1 =i23-i4O[x; GW 2 =96-io5{ J t; GW 1 /GW 2 = 1-2-1-4. Genital setae relatively short, about half GW 2 . Anal shield approximately as long as broad. AL= 90-109^; AW= 100-105^. Anus situated at less than its length from the anterior margin ; paranal setae level with middle of anus, about 2/3 as long as postanal seta. Integument of idiosoma bears 22 pairs of ventral and marginal setae, the latter being a little coarser than the former and provided with fine barbs. Two pairs of ventral setae flank the genital shield. Metapodal plates 3 or 4 times longer than broad. Chaetotaxy of legs normal ; posterior seta on coxa I slightly thicker than anterior seta, but subequal in length. Length of tarsus IV at least 6 times its greatest width. MALE (figs. 68-69) : Chelicera as illustrated, pilus dentilis inflated basally. Dorsal shield reticulate, chaetotaxy as in female. DL=428[x; DW=257[x; DL/DW=i-7. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE ,. //. 65 69 68 FIGS. 65-69. Androlaelaps glasgowi (Ewing). Venter (Fig. 65), dorsum (Fig. 66) and chelicera (Fig. 67) of female. Venter (Fig. 68^ and chelicera (Fig. 69) of male. 42 W. M. TILL Holoventral shield reticulate, expanded behind coxae IV and bearing 23 setae. Integument with about 14 pairs of ventral and marginal setae. Chaetotaxy of legs as in female. HOSTS AND LOCALITIES: Rhabdomys pumilio (Sparrman) and Otomys irroratus (Brants) from several localities in Southern Africa (S.A.I.M.R. ; B.M.N.H; types of /. rhabdomys in Dr. C. D. Radford's collection). FIGS. 70-73. Androlaelaps graingeri Zumpt & Patterson, female. Venter (Fig. 70); dorsum (Fig. 71); chelicera (Fig. 72); femur, genu, tibia of leg II (Fig. 73). Androlaelaps graingeri Zumpt & Patterson Androlaelaps graingeri Zumpt & Patterson, 1952, /. ent. Soc. S. Afr. 15 : 162, fig. 2. Turkiella graingeri, Zumpt & Till, 1953, Ann. Inst. Med. trap. Lisboa 10 : 224. FEMALE (figs. 70-73): Gnathosoma ventral in position in all the specimens examined, only the free segments of the palps extending beyond the anterior margin of the idiosoma. Deutosternum with 6 rows of 3-4 teeth, only one tooth visible in the 5th row. Chelae 74-76^ long, both digits dentate, movable digit with a pronounced terminal hook which arches over the fixed digit ; pilus dentilis slender, arthrodial filaments subequal in length. Dorsal shield subcircular, with a " double " margin, granular and weakly reticulate. DL=iooo-io50[x; DW= 825-900^; DL/DW=i-2. Chaetotaxy of dorsal shield similar to that of A . arvicanthis, i.e. setae px2 and px$ missing. Seta ri considerably longer than seta ii ; seta Z$ about twice as long as seta /5 ; setae 4-*5> 22, /i-/4 ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 43 rubbed off in all the specimens examined, but are probably subequal with 23, as they are in the male. Setae generally finer than in A. arvicanthis and apparently without barbs. Sternal shield with an overall granulation and very weak reticular pattern. SL=8o-96[ji; SW= 197-228^; SL/SW=o-4-o-5. First sternal seta relatively short, 2nd and 3rd sternal and the metasternal setae considerably longer. 74 75 FIGS. 74-76. Androlaelaps graingeri Zumpt & Patterson, venter of male (Fig. 74). Androlaelaps haydocki (Till), male. Venter (Fig. 75) ; dorsal view of femur and genu of leg I (Fig. 76). Genital shield broadly rounded posteriorly but not greatly expanded behind the genital setae, granular, with a weak reticular pattern. GW X = 2 15-236^; GW 2 = i88-200{x; GW 1 /GW 2 =i-i 1-2. Genital setae relatively long, slightly less than GW 2 , reaching base of second pair of flanking setae. Anal shield reticulate, with a pair of lateral, granular patches; approximately as long as broad. AL= 162-184^; AW= 158-175^1. Paranal setae near posterior margin of anus, about 4/5 as long as postanal seta. 44 W. M. TILL Integument of idiosoma with 21-22 pairs of fine setae. Metapodal plates subcircular. Chaetotaxy of legs normal. Anterior and posterior setae of coxa I approximately equal in length. Ventral setae on genu and tibia II similar to those in A . arvicanthis. Length of tarsus IV about 6| times the width at the base. 77 78 FIGS. 77-78. Androlaelaps haydocki (Till), female. Venter (Fig. 77); dorsum (Fig. 78). MALE (fig. 74) : Gnathosoma as in female; spermadactyl about 115^ long. Dorsal shield with chaetotactic pattern similar to that of female; DL=846[x; DW=646[x; DL/DW=i - 3. Holoventral shield granular with a weak reticular pattern, not expanded behind 4th pair of coxae, bearing 21 setae, of which 4 pairs lie in the preanal region. Metapodal plates oval, length about 2| times width. Chaetotaxy of legs as in female. HOST AND LOCALITY: Unidentified rodent, Kerio Valley, Kenya (type series, S.A.I.M.R.). ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 45 Androlaelaps haydocki (Till) Haemolaelaps haydocki Till, 1959, /. ent. Soc. S. Afr. 22 : 426, figs. Androlaelaps haydocki is very closely allied to A. mesopicos (Radford) and may prove to be only a subspecies or variety. FEMALE (figs. 77-78): The female differs from A. mesopicos in that the dorsal shield bears 38 pairs of setae, and the marginal setae 57, S2, and 53 are long and 79 80 FIGS. 79-81. Androlaelaps heliosciuri sp. nov., female. Venter (Fig. 79); dorsum (Fig. 80); chelicera (Fig. 81). sinuous. The genital shield is slightly more expanded posteriorly (GW 1 /GW 2 = 1-3-1-5), and the anal shield is slightly more elongate (AL/AW=i-2-i-4). MALE (figs 75-76) : The male differs from A . mesopicos in having a long dorsal seta on genu I. HOSTS AND LOCALITIES : Thripias namaquus (Lichtenstein) , Luanshya (type series) and Muliashi Area, Northern Rhodesia; Debeete, Bechuanaland ; Mabelikwa, N. 46 W. M. TILL Transvaal ; Mapalane, Mozambique (S.A.I.M.R. and B.M.N.H.). Campethera abingoni (Smith), Luanshya, N. Rhodesia (B.M.N.H.). Androlaelaps heliosciuri sp. nov. FEMALE (figs. 79-81) : Deutosternum with 6 rows of 2-3 small teeth; corniculi very weakly sclerotized. Chelae about 4O(x long, edentate except for terminal hook on movable digit; pilus dentilis slender; two of the arthrodial filaments considerably longer and thicker than the others. Dorsal shield reticulate, widest in posterior half at level of seta 52. DL 618- 637^; DW=385-4O4(u.; DL/DW=i-6. All 39 pairs of setae are present, but rz, r\ and r6 are off the shield. Lengths of setae 24 and 22 not greater than half the distance between their bases; seta Z$ about 3 times as long as seta /5. Sternal shield granular, with a faint reticular pattern as well. SL=76-89[j.; SW=i28-i33[A; SL/SW=o-6-o-7. First pair of sternal setae situated off the plate, slightly shorter than 2nd and 3rd pairs ; metasternal setae as long as 2nd and 3rd pairs of sternal setae. Genital shield reticulate, the pattern extending over the anterior flap. It is closely approached to the anal shield and widest at the level of the 2nd pair of flanking setae. GW 1= 142-152^; GW 2 =9o-95^; GW 1 /GW 2 = 1-5-17. Genital setae relatively short, less than half GW 2 . Anal shield broader than long, anterior margin straight. AL=ioi-iiO[ji; AW 1 20-123 [j.; AL/ AW 0-8-0-9. Paranal setae situated between middle and posterior end of anus, about 2/3 as long as postanal seta. Integument of idiosoma bears about 34 pairs of ventral and marginal setae, the former being fine and simple, the latter coarser and barbed. Metapodal plates long and slender, length to width at least 6 : i. Chaetotaxy of legs normal. Femur IV bears a stout antero-dorsal seta. Length of tarsus IV about 4 times the width at the base. HOST AND LOCALITY: Eight females from Heliosciurus gambianus, (Ogilby), Amani, Tanganyika, 3 February, 1957. Holotype and three paratypes in the collec- tion of the South African Institute for Medical Research, four paratypes (1962.6.12. 14-17) in the collection of the British Museum (Natural History). Androlaelaps hirsti (Keegan) Haemolaelaps hirsti Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 242, figs. 26-30; Costa, 1961, Bull. Brit. Mus. (nat. Hist.) Zool. 8 : 14, figs. 17-20. FEMALE (figs. 82-84) : Deutosternum with 6 rows each of 4 small teeth, except the last row which has 3 teeth. Chelae about 80 [x long ; pilus dentilis inflated basally, tapering distally; arthrodial filaments subequal in length. Dorsal shield ovoid, with a " double " margin, widest about the middle, tapering posteriorly, granular in appearance with faint reticulations. DL= 1092-1188^; DW= 703-817^; DL/DW=i-4-i-6. The shield bears 41 pairs of setae, one pair being present in the ax and one pair in the pxi position. One of the specimens ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 47 examined has an unpaired accessory seta between setae /4 ; the remaining specimens have no accessory setae. The setae on the central part of the shield are relatively long, the lengths of 24 and Z2 exceeding the distance between their bases. The posterior terminal setae Z$ are approximately the same length as setae 14 and at least 2\ times as long as setae 75. 82 FIGS. 82-84. Androlaelaps hirsti (Keegan), female. Venter (Fig. 82); dorsum (Fig. 83); chelicera (Fig. 84). Sternal shield granular with a faint reticulate pattern, presternal area reticulate and slightly granular. SL= 162-180^; SW= 225-257^; SL/SW=o-6-o-8. Sternal setae long, first pair reaching posterior margin of shield, third pair reaching middle of coxae IV ; metasternal seta about half as long as second sternal seta. Genital shield granular, not expanded posteriorly; GW 2 = 162-180^. Genital setae relatively long, reaching base of second pair of flanking setae, their length nearly equal to GW 2 . Anal shield as long as, or slightly longer than, broad. AL= 175-190^; AW= 48 W. M. TILL i62-i75(x; AL/AW=i-o-i-2. Anus less than its length from the anterior margin of the shield. Paranal setae inserted on a level between middle and posterior margin of anus, extending beyond base of postanal seta, and slightly longer than this seta. Integument of idiosoma with numerous long marginal and ventral setae, many of the former being barbed. Metapodal plates oval, length about twice breadth. 85 FIGS. 85-88. Androlaelaps hirsti (Keegan), venter (Fig. 85) and chelicera (Fig. 86) of male. Androlaelaps hystrici (Zumpt & Till), venter (Fig. 87) and chelicera (Fig. 88) of male. Tibia III has 9 setae instead of the usual 8. Ventral setae on coxa I nearly equal in length but posterior seta much thicker. Two short, stout, spur-like setae on femur I ; one stout ventral seta on genu II, two on tibia II, 3 stout ventral and 3 stout apical setae on tarsus II, two of the apical ones being relatively blunt. Length of tarsus IV is 7-8 times its width at the base. MALE (figs. 85-86) : Fixed digit of chelicera with a long slender pilus dentilis. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 49 Deutosternum with 6 rows of 2 or 3 teeth. Dorsal shield as in the female, but the setae are relatively longer. Anal shield separated from the sternito-ventral shield which bears 7 pairs of setae, 2 pairs being situated on the part posterior to coxae IV. Chaetotaxy of legs as in the female. HOSTS AND LOCALITIES : Unidentified rat from Yubo, Equatoria, Sudan (Paratype in B.M.N.H.). The above description is based on specimens from Gerbillus Pyramidum Geoffroy from Israel, received by courtesy of Dr. M. Costa. r5 ' /'/ T\ ' r6 \ FIGS. 89-91. Androlaelaps hystrici (Zumpt & Till), female. (Fig. 90); chelicera (Fig. 91). Venter (Fig. 89) ; dorsum Androlaelaps hystrici (Zumpt & Till) Haemolaelaps hystrici Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 235, figs. 12-16; Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 239, fig. 23. FEMALE (figs. 89-91): Deutosternum with 6 rows each of 3 teeth. Chelae 34^ long ; pilus dentilis slender ; arthrodial filaments subequal in length. Anterior rostral seta at least i| times as long as capitular seta. Dorsal shield ovoid, posterior margin almost straight. DL= 600-662^; DW= ZOOL. IO, I 4 5 o W. M. TILL 290-362^; DL/DW =1-8-2-3. Only 36 pairs of setae are present on the shield, r^-rb being inserted on the adjacent integument. Setae simple, of moderate length ; lengths of 24 and Z2 about f the distance between their bases ; seta Z$ at least 1 1 times as long as seta 75. Sternal shield reticulate. SL= 93-1 14(1; SW= 103-134^; SL/SW=o-8-i-o. Setae relatively short, first seta reaching base of second. Genital shield slightly granular, with a few indistinct traces of a reticulate pattern, very slightly widened behind genital setae, then tapering. GW-,^ 93-1 03^; GW 2 = 83-93^. Length of genital setae about half GW 2 . Anal shield pear-shaped. AL= 93-134^; AW =93-103^; AL/AW= 1-0-1-3. Anus situated at approximately its length from the anterior margin. Paranal setae inserted on a level between middle and posterior end of anus and approximately as long as the postanal seta. Integument of idiosoma bears about 22 pairs of ventral and marginal setae, of which two pairs border the genital shield. Metapodal plates extremely narrow, length to width about 10 : i. Chaetotaxy of legs normal. Posterior seta on coxa I slightly longer than anterior seta. Legs relatively slender; length of tarsus IV 8 times the width at the base. MALE (figs. 87-88) : Fixed digit of chelicera with a slender pilus dentilis. Dorsal shield ovoid, 5oo[x long and 263^ wide ; DL/DW=i-9- Chaetotaxy as in the female, except that r^-r6 are inserted on the shield. Holoventral shield narrow, bearing 19 setae. Metapodal plates broader than in the female. Chaetotaxy of legs as in female, but some of the setae relatively stouter, for example, femur II has two broad, spine-like ventral setae, tibia II has one very broad, spear-like ventral seta, tarsus II has two broad, spine-like, ventral setae, but the apical setae are slender. HOSTS AND LOCALITIES: Hystrix cristata Linnaeus from Karamoja district, Uganda (type series in S.A.I.M.R. and B.M.N.H.) and from Njoro, Rift Valley Province, Kenya (Keegan, 1956). Keegan (1956) records this species from Arvicanthis niloticus (Desmarest) and Rattus rattus (Linnaeus) in Egypt, and also an additional specimen, which should probably be referred to this species, from Crocidura manni Peters, Kano, Nigeria. Androlaelaps longipes (Bregetova) Haemolaelaps longipes Bregetova, 1952, Zoo/. Zh. 31 : 867, figs.; Costa, 1961, Bull. Brit. Mus. (nat. Hist.) Zoo/. 8 : 21, figs. 28-32. Haemolaelaps namrui Radford, 1954, Fieldiana, Zoo/. 34 : 310, figs. 56-57 (syn. nov.). Haemolaelaps aegyptius Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 246, figs. 37-38. FEMALE (figs. 92-94) : Deutosternum with 6 rows of 3-4 small teeth. Chelae about 55jj, long; pilus dentilis slightly inflated basally, distal 2/3 slender; arthrodial filaments subequal in length. Dorsal shield ovoid, with a reticulate pattern. ~DL=yo^[i ; DW=475[x ; DL/DW= 1-5. The shield bears 40 pairs of setae, one pair being present in the PXI position on the posterior half of the shield. There is also an unpaired seta near the mid-line ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 51 between setae /3 and /4. The setae on the central part of the shield are relatively long, the lengths of 14 and Z2 being approximately equal to the distance between their bases. The anterior and marginal setae are longer, some of the longest setae having fine barbs. Seta Z5 is more than twice as long as seta 75. Sternal shield reticulate. SL=g^; SW=i47{ji; SL/SW=o7. Sternal setae long, the first pair reaching almost to the posterior margin of the shield, second and 92 FIGS. 92-94. Androlaelaps longipes (Bregetova), female. (Fig. 93); chelicera (Fig. 94). 93 Venter (Fig. 92) ; dorsum third pairs slightly longer, the 3rd pair extending well beyond the bases of the metasternal setae. Metasternal seta less than half as long as 3rd sternal seta. Genital shield with a reticulate pattern, slightly widened behind genital setae, then tapering posteriorly. GW^ 125(1,; GW 2 =i20fi.. Genital setae relatively long, about 9/10 GW 2 , extending to base of second pair of flanking setae. Anal shield slightly longer than wide, anterior margin straight. Length (to base of postanal seta) is 103(1,, greatest width is 98^. Anus approximately equidistant between anterior margin of shield and postanal seta. Paranal setae situated between middle and posterior margin of anus, about the same length as the postanal seta. 52 W. M. TILL Integument of idiosoma bears about 37 pairs of apparently simple ventral and marginal setae, of which 2 pairs flank the genital shield. Metapodal plates elongate, 7 times as long as broad. Chaetotaxy of legs normal except that tibia III has 9 setae. Posterior seta on coxa I is 1 1 times as long and thick as anterior seta. Stout ventral setae are present on the femur, genu, tibia and tarsus of leg II, three of the terminal setae on tarsus II are stout and blunt. The dorsal seta on trochanter IV is much thicker than the 95 97 FIGS. 95-97. Androlaelaps longipes (Bregetova), venter (Fig. 95) and chelicera (Fig. 96) of male. Androlaelaps marshalli Berlese, venter of male (Fig. 97). other setae on this segment. Length of tarsus IV about 8 times the width at the base. MALE (figs. 95-96): Fixed digit of chelicera (in A. longipes from the U.S.S.R.) bears a large, tapering pilus dentilis. Dorsal shield similar to that of female. Holo ventral shield divided into a sternito- ventral shield and a separate anal shield. The sternito-ventral shield is moderately expanded behind the 4th pair of coxae and bears 7-8 pairs of setae, 2 or 3 pairs being situated on the part posterior to ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s.lat. 53 coxae IV. The ventral shield is more reduced in the Russian specimen than in the type specimen of A. namrui, on which the description and drawings are based. Integument of idiosoma bears about 31 pairs of setae. Chaetotaxy of legs as in the female. HOSTS AND LOCALITIES: Meriones rex Yerbury & Thomas, Ta'izz, Yemen (types of A. namrui in Dr. C. D. Radford's collection). 98 99 FIGS. 98-100. Androlaelaps marshalli Berlese, female. Venter (Fig. 98); dorsum (Fig. 99) ; ventral view of femur, genu and tibia of leg II (Fig. 100). Bregetova (1952) recorded A. longipes from Meriones species in the U.S.S.R. (material in B.M.N.H.). Keegan's specimens (A . aegyptius) were collected from a variety of hosts at several localities in Egypt, and Costa (1961) has found this species chiefly on Meriones species in Israel. Androlaelaps marshalli Berlese Laelaps (Androlaelaps) marshalli Berlese, 1911, Redia 7 : 429. Androlaelaps marshalli, Keegan, 1956, /. Egypt, publ. Hthl Ass. 31 : 227, fig. 13; Costa, 1961, Bull. Brit. Mus. (nat. Hist.) Zool. 8 : 45, figs. 75~79- 54 W. M. TILL Hypoaspis (Androlaelaps) africanus Zumpt, 1950, Parasitology 40 : 30, fig. 2. Turkiella africanus, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 216. Hypoaspis (Androlaelaps) africanoides Zumpt & Patterson, 1950, S. Afr. J. med. Sci, 15 : 72, figs. 2-3. Androlaelaps oudemansi Radford, 1944, Parasitology 35 : 161, fig. 2 (syn. nov.). Turkiella oudemansi, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 224. FEMALE (figs. 98-100) : Deutosternum with 6 rows of teeth, 3 to 6 small, finely pointed teeth in each row. Chelae similar to those of A. arvicanthis, 70-90^ long; pilus dentilis slender; arthrodial filaments subequal in length. Dorsal shield with a " double " margin and an overall reticular pattern. DL= 863-1050(1,; DW= 675-763^; DL/DW= 1-2-1-5. The shield bears 39 pairs of setae as well as additional unpaired setae between /3 and 74- Many of the setae are barbed. Lengths of setae 24 and Z2 slightly exceed the distance between their bases; seta Z$ only slightly longer than 75. Sternal shield with a well-defined reticular pattern and with concave anterior and posterior margins. SL= 88-123^; SW=i8o-i93(x; SL/SW= 0-5-0-6. First sternal seta reaches second pore, second and third sternal and metasternal setae distinctly longer. Genital shield broadly rounded posteriorly, with an overall recticular pattern. GW 1 =228-250(A; GW 2 =i5o-i75[z; GW 1 /GW 2 =i-4-i-6. Genital setae relatively long, approximately equal to GW 2 , and extending almost to the base of the second pair of flanking setae. Anal shield with a reticular pattern ; anus at slightly less than its length from the anterior margin; paranal setae situated near posterior half of anus, only slightly shorter than postanal seta. AL=i23-i66[i,; AW 150-175^; AL/AW=o-8-i-o. Integument of idiosoma bears 30-55 pairs of long ventral and marginal setae, most of which are barbed. Metapodal plates oval, length about twice the breadth. Chaetotaxy of legs normal, except that the type of A . oudemansi and some of the other specimens assigned to this species have 9 setae on tibia III instead of the usual eight. Anterior and posterior setae on coxa I approximately equal in length. Thorn-like seta on femur II relatively large; inner ventral seta on genu II short and spine-like. Inner ventral seta on tibia II a little shorter and considerably stouter than the outer seta in the types of A. oudemansi and A. africanoides, whereas in specimens previously identified as A. africanus the differences between these two setae are less striking, as in A. arvicanthis. Tarsus II bears 3 stout, blunt, terminal setae. Length of tarsus IV about 8 times its width at the base. MALE (fig. 97) : Deutosternum as in female ; chelicera similar to that of A . arvi- canthis. Dorsal shield similar to that of the female. DL= 890-980^; DW= 570-6o8(j,; DL/DW=i-6. Holoventral shield strongly reticulate, expanded behind fourth pair of coxae, and bears 23-31 setae. The number of setae between coxae IV and the anus is 10-19, but there may be more on one side than on the other. Chaetotaxy of legs as in the female. HOSTS AND LOCALITIES: Recorded in Southern Africa from the following hosts (Zumpt & Till, 1961) : Tatera afra (Gray), Gerbillus paeba Smith, Desmodillus auricularis (Smith), Rattus paedulcus (Sundevall), Rattus natalensis (Smith), Rattus chrysophilus (De Winton), ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 55 Rattus namaquensis (Smith), Rhabdomys pumilio (Sparrman), Lemniscomys griselda (Thomas), Saccostomus campestris Peters, Steatomys pratensis Peters, Pedetes capensis (Forster), Cryptomys hottentotus (Lesson) and Oryctolagus cuniculus (Linnaeus). Specimens from the following East African hosts have been examined: Arvicanthis niloticus (Desmarest), Bunyoro, Uganda (type of A. oudemansi in B.M.N.H.). FIGS. 101-103. Androlaelaps mesopicos (Radford), female. Venter (Fig. 101); dorsum (Fig. 102) ; tip of tarsus IV, ventral view (Fig. 103). Tatera valida (Bocage), Bunyoro, Uganda (S.A.I.M.R.). Tatera nigricauda Peters, Lali Hills, Kenya (S.A.I.M.R.). Tatera species, W. Nile District, Uganda (S.A.I.M.R.). Taterillus emini (Thomas), W. Nile District, Uganda (S.A.I.M.R.). Tachyoryctes splendens (Riippell), Muguga, Kenya (S.A.I. M.R.). Unidentified rodent, Kerio Valley, Kenya (S.A.I.M.R.). 5 6 W. M. TILL A. marshalli has also been recorded from an unidentified rodent in the Sudan and from gerbils in Egypt (Keegan, 1956) and from gerbils in Israel (Costa, 1961). Two series of mites which have been provisionally assigned to A. marshalli have been obtained from carnivores, namely, Suricata suricatta (Schreber) from Queens- town, Cape Province (S.A.I.M.R.), and Ictonyx striatus (Perry) from Muguga, Kenya (S.A.I.M.R.). The drawings of A . marshalli are based on paratypes of A . africanoides from Tatera afra (Gray), S. Rhodesia. Androlaelaps mesopicos (Kadlord) Haemolaelaps mesopicos Radford, 1942, Parasitology 34 : 192, figs. 14-19; Till, 1959, /. ent. Soc. S. Afr. 22 : 423, figs. Hypoaspis (Haemolaelaps) mesopicos, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 72. A series of mites from Mesopicos griseocephalus , Natal, has been assigned to this species (Till, 1959). Although differing slightly from the types, especially in the male sex, these mites probably represent no more than a form or variety of A. mesopicos. FEMALE (figs. 101-103) : Deutosternum provided with 6 rows of 2-4 small teeth. Chelae about 40^ long; pilus dentilis small, inflated; arthrodial filaments subequal in length. Dorsal shield oval in shape, truncate posteriorly, with a reticulate pattern. DL= 700-735 [x; DW= 420-438^; DL/DW= 1-6-1-7. Not more than 37 pairs of setae can be distinguished on the shield, seta 53 being absent and seta r6 arising from the integument adjacent to the shield. Seta p%3 is present in the type but absent in the specimens from Natal. The marginal setae are variable in size, the longest of the S setae being more than 3 times as long as the / setae. Seta Z$ is sinuous and whip-like and about 20 times as long as seta /5. The setae on the median part of the shield (i2-i$, J series, z series, Zi-Z4, s2-s6) are very minute, subequal, the lengths of setae 24 and Z2 being less than a quarter the distance between their bases. Sternal shield granular, presternal area weakly reticulate. SL=84-io5(x; SW= 152-158^; SL/SW=o-6-o-7. First pair of sternal setae slightly shorter than second and third pairs, and lying on the presternal area. Metasternal setae as long as the 2nd and 3rd sternal setae. Genital shield slender, weakly reticulate posteriorly; GWj 96-144^; GW 2 = 80-96^; GW!/GW 2 = 1-1-1-3. Genital setae relatively short, less than half GW 2 . Anal shield pear-shaped with a reticular pattern antero-laterally, relatively shorter and broader in the specimens from Natal. AL= 123-150^; AW= 96-105^; AL/AW=i-2-i-6. Paranal setae level approximately with middle of anus, not more than 2/3 as long as postanal seta. Integument of idiosoma bears about 23 pairs of marginal and ventral setae, which all appear to be simple, except for the long, sinuous, posterior terminal pair. Metapodal plates slender, slightly curved. Tibia HI bears 9 setae and tarsi II, III and IV each bear two stout, blunt, ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 57 terminal spines. The anterior and posterior setae on coxa I are approximately equal in length and thickness. Length of tarsus IV about 4 times the width at the base. MALE (figs. 104-106): Fixed digit of chelicera with a small setiform structure near its base. Dorsal shield reticulate, and also granular in the specimen from Natal ; shape and chaetotaxy as in the female, the marginal setae, however, being 104 105 FIGS. 104-106. Androlaelaps mesopicos (Radford), male. Venter of type specimen from Uganda (Fig. 104); venter (Fig. 105) and chelicera (Fig. 106) of specimen from Natal. considerably longer. In the type, setae 82 and 5$, and in the Natal specimen 84 also, are long and sinuous. DL=6i8-646[ji; DW=35o-36o[i.; DL/DW=i -6-1-7. Holoventral shield bears 19 setae. Sternal region similar in shape to sternal shield of female, with a granular appearance and faint reticular markings ; reticular pattern more distinct on remaining part of shield. Posterior to coxae IV the shield is narrow with irregular margins. Integument of idiosoma bears about 16 pairs of marginal and ventral setae, the 58 W. M. TILL posterior pair being long and sinuous as in the female. In the specimen from Natal the longest marginal setae are approximately twice as long as in the type specimen. Metapodal plates as in the female. The legs bear the same number of setae as in the female, but there are modifications in the form of some of the setae. On femur II, the lower medio-ventral seta is modified to form a thickened spur. Tarsi II, III and IV each bear one stout, terminal spine, and tarsus II has, in addition, two blunt ventral spurs. 107 FIGS. 107-109. Androlaelaps murinus (Berlese), female. Venter (Fig. 107); dorsum (Fig. 1 08); chelicera (Fig. 109). HOSTS AND LOCALITIES: Mesopicos griseocephalus ruwenzori Sharpe, Kabale, Uganda (types in B.M.N.H.). Mesopicos g. griseocephalus (Boddaert), Pietermaritzburg, Natal (S.A.I.M.R. and B.M.N.H.). Androlaelaps murinus (Berlese) Laelaps (Haemolaelaps] murinus Berlese, 1911, Redia 7 : 432. Haemolaelaps murinus, Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 233, figs. 16-17. Ischnolaelaps arvicanthis Radford, 1939, Parasitology 31 : 246, fig. 3. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 59 FEMALE (figs. 107-109) : Deutosternum provided with 6 rows of 3-4 teeth. Chelae 52-54(1 long, movable digit with a pronounced terminal hook which arches over the fixed digit ; pilus dentilis strongly inflated basally, terminal portion slender, curved ; one of the arthrodial filaments strikingly longer than the others. Dorsal shield ovoid; DL= 860-912(1; DW= 595-648(1; DL/DW=i-4-i-5. The shield bears 39 pairs of setae and 2-3 unpaired accessory setae. Lengths of setae 110 112 FIGS. 110-113. Androlaelaps murinus (Berlese), venter (Fig. no) and chelicera (Fig. in) of male. Androlaelaps oliffi (Zumpt & Patterson), venter (Fig. 112) and chelicera (Fig. 113) of male. 24 and Z2 a little less than the distance between their bases; seta Z$ about twice as long as seta 75. Sternal shield strongly reticulate; SL= 120-140(1; SW= 166-184(1; SL/SW= 07-0-8. Metasternal setae shorter than first pair of sternal setae. Genital shield slightly widened posteriorly and having a reticulate pattern which extends over the anterior flap. GW^ 184-210(1; GW 2 = 128-140(1; GW l /GW 2 = 1-4-1-5. Length of genital setae about half GW 2 . Anal shield approximately as long as broad; AL= 145-150(1; AW= 145-16051. 60 W. M. TILL Anus a little less than its length from the anterior margin; paranal setae inserted slightly behind middle of anus, a little shorter than the postanal seta. Integument of idiosoma bears about 60 pairs of setae of which about 15 pairs are truly ventral in position. The marginal setae tend to be stouter than the ventral ones and have fine barbs. Chaetotaxy of legs normal. Anterior seta on coxa I 3/4 as long as posterior seta ; dorsal seta on trochanter IV stout and spine-like ; two of the dorsal setae on femur IV spine-like. Length of tarsus IV about 7 times its width at the base. MALE (figs, iio-iu): The chelicerae have the structure indicated in the figure. Holoventral shield reticulate, expanded behind coxae IV, bearing 31-32 setae. Integument of idiosoma bears numerous setae. Chaetotaxy of dorsal shield and legs as in the female. DL=665[x; DW=437[x; DL/DW 1-5. HOSTS AND LOCALITIES: Arvicanthis niloticus (Desmarest) from Uganda (type of /. arvicanthis in B.M.N.H.) and from Kenya (Keegan, 1956). Otomys species, Uganda (B.M.N.H.). Otomys irroratus (Brants) and Rattus natalensis (Smith) from Southern Rhodesia (S.A.I.M.R.). Oenomys hypoxanthus (Pucheran) from the Belgian Congo (B.M.N.H.) and from the Cameroons (I.R.S.C.). Dasymys incomtus (Sundevall) from Kampala, Uganda (S.A.I.M.R.) and from the Rift Valley Province, Kenya (Keegan, 1956). Rhabdomys pumilio (Sparrman), Lophuromys flavopunctatus Thomas and Lemnis- comys striatus (Linnaeus) from the Rift Valley Province, Kenya (Keegan, 1956). Androlaelaps oliffi (Zumpt & Patterson) Hypoaspis (Haemolaelaps) oliffi Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 81, fig. 3. Haemolaelaps oliffi, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 243. FEMALE: (figs. 114-116) : Deutosternum with 6 rows of 2-4 teeth. Chela 38-40^ long; pilus dentilis slightly inflated; arthrodial filaments subequal in length. Dorsal shield ovoid, reticulate. DL= 620-644^; DW=368-4iO(x; DL/DW= I '5~ 1 '7- The shield bears the usual 39 pairs of setae; lengths of setae 24 and Z2 slightly exceed the distance between their bases; seta Z$ is if times as long as seta 75. Sternal shield reticulate; SL=96-io5fx; SW= 130-1 35^; SL/SW=o-7-o-8. Metasternal setae slightly shorter than first pair of sternal setae. Genital shield granular and with a reticulate pattern which extends over the anterior flap. The shield is not expanded behind the genital setae, GWj and GW 2 both varying between 127^ and i35[x. Length of genital setae about 2/3 GW 2 . Anal shield approximately as long as broad, or slightly longer; AL= 123^1 ; AW=ii4-n8(jL. Anus situated at about i| times its length from the anterior margin. Paranal setae level approximately with middle of anus, almost reaching base of postanal seta. Postanal seta about the same length as the paranal setae, but slightly thicker. Integument of idiosoma bears 15 pairs of setae, of which 3 pairs border the ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 61 genital shield; some marginal setae are barbed. Metapodal plates slender, length to width about 7:1. Chaetotaxy of legs normal except that tibia III bears 9 setae. Posterior seta on coxa I nearly twice as long and thick as anterior seta. Anterior seta on coxa II and both setae on coxa III rather stout. Three of the apical setae on tarsus II are stout and one of these is blunt. Length of tarsus IV about 5| times the width at the base. I.' V! I'./ M 'U FIGS. 114-116. Androlaelaps oliffi (Zumpt & Patterson), female, dorsum (Fig. 115); chelicera (Fig. 116). 115 Venter (Fig. 114); MALE (figs. 112-113): Chelicera as figured; dorsal shield reticulate, chaetotaxy as in the female ; DL = 428-437/4 ; DW = 276/4 ; DL/DW = 1-6. Holoventral shield reticulate, bearing 23 setae ; these, excluding the anals, are longer and stouter than those of the integument. Integument bears about 16 pairs of ventral and marginal setae ; metapodal plates elongate. Chaetotaxy of legs as in the female. HOSTS AND LOCALITIES : Widely distributed in Southern Africa, mainly on Tatera afra (Gray) and Gerbillus paeba Smith, and occasionally found on Saccostomus campestris Peters and Desmodillus auricularis (Smith) (S.A.I.M.R., B.M.N.H. and M.R.A.C.). Found also in Tanganyika on Tatera afra (Gray) (B.M.N.H.). 62 W. M. TILL Androlaelaps pachyptilae (Zumpt & Till) Haemolaelaps pachyptilae Zumpt & Till, 1956, Z. Parasitenk. 17 : 285, fig. 3. FEMALE (figs. 117-119): Deutosternum with 6 rows each of 3-4 small teeth. Chelae 40[x long; pilus dentilis inflated basally, distal half slender and curved; arthrodial filaments subequal in length. Dorsal shield ovoid, reticulate, widest between setae rj and 57. DL= 604-691^; 117 FIGS. 117-119. Androlaelaps pachyptilae (Zumpt & Till), female, dorsum (Fig. 118); chelicera (Fig. 119). 118 Venter (Fig. 117); DW= 385-446^; DL/DW 1-5-1-6. The shield bears 39 pairs of setae. Those on the central part of the shield are relatively short, the lengths of setae 24 and zz being about half the distance between their bases. The anterior and marginal setae are considerably longer; seta Z$ is barbed and is about 2\ times as long as seta /5. Sternal shield with a reticulate pattern, which is more pronounced antero- laterally, the central part of the shield having a granular appearance. Presternal ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 63 area strongly reticulate. SL=79-ioo[x; SW= 140-158^; SL/SW=o-5-07. First sternal seta nearly i| times as long as metasternal seta. Genital shield short, slightly widened behind genital setae and then tapering posteriorly; a reticulate pattern is present posteriorly but is not very sharply defined. GW 1 =ioo-ii4 f x; GW 2 =g2-g6yi; GWJGWg^ 1-1-1-2. Genital setae the same length as the ventral setae which flank the shield, about half GW 2 . Anal shield slightly wider than long. AL= 105-1 14^; AW= 114-123^; AL/AW= 0-9. Anus situated at less than its length from the anterior margin. Paranal setae level with middle of anus, not more than 2/3 as long as postanal seta. Length of postanal seta almost equal to its distance from the anterior margin of the shield. Integument of idiosoma bears 16 pairs of ventral and marginal setae, two pairs bordering the genital shield and a third pair lying between the genital and anal shields. The marginal setae are mostly barbed, the ventral ones simple. Metapodal plates elongate, length to width about 4:1. Chaetotaxy of legs normal. Posterior seta on coxa I slightly longer than anterior seta. Length of tarsus IV about 5 times the width at the base. MALE: Not known. HOST AND LOCALITY: Pachyptila desolata (Gmelin) from Heard Island (type series in S.A.I. M.R.). Androlaelaps pater soni (Zumpt & Till) Haemolaelaps patersoni Zumpt & Till, 1956, Z. Parasitenk. 17 : 286, fig. 4. FEMALE (figs. 120-122) : Deutosternum with 6 rows of 3-6 very fine teeth. Chelae 42-46(0, long; pilus dentilis spatulate, terminating in a short, fine, slightly bent point ; arthrodial filaments subequal in length. Dorsal shield bears 39 pairs of setae and a variable number (5-12) of unpaired accessory setae which may extend anterior to setae 25. Lengths of setae 24 and 22 approximately equal to the distance between their bases; seta Z$ at least twice as long as seta 75. DL= 683-75 3^; DW =420-481^; DL/DW-i-5-1-6. Sternal shield reticulate; SL= 100-114^; SW= 130-149^; SL/SW= 0-7-0-8. Metasternal setae a little shorter than first pair of sternal setae. Genital shield moderately expanded posteriorly, with a reticulate pattern ex- tending over the anterior flap. GW X = 135-154^; GW 2 = 96-114^; GWJGW^ I '3~ 1 '5- Length of genital setae at least half GW 2 . Anal shield with an overall reticulate pattern; AL= 114-131^; AW 127-131^; AL/AW=o-9~i-o. Anus situated at approximately its length from the anterior margin of the shield. Paranal setae level with middle of anus, about 4/5 as long as postanal seta. Integument bears about 30 pairs of setae. Outer metapodal plates elongate, length to width 4-5 : i. Chaetotaxy of legs normal. Anterior and posterior setae on coxa I subequal in length and thickness. Length of tarsus IV about 7 times the width at the base. MALE (figs. 123-124) : The chelicerae have the structure shown in the figure, the fixed digit bearing a pilus dentilis similar to that of the female. 6 4 W. M. TILL 124 122 123 FIGS. 120-124. A ndrolaelaps patersoni (Zumpt & Till). Venter (Fig. 120), dorsum (Fig. 121) and chelicera (Fig. 122) of female. Venter (Fig. 123) and chelicera (Fig. 124) of male. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 65 Chaetotaxy of dorsal shield as in the female, but the setae are relatively longer. DL= 569-586^; DW=324-359(jt; DL/DW=i5. Holoventral shield reticulate, expanded behind coxae IV and bearing 23 setae. The integument bears about 15 pairs of setae. Chaetotaxy of legs as in the female. HOST AND LOCALITY: Gyps coprotheres (Forster) from Rustenburg, Transvaal (type and other material, S.A.I.M.R.). 125 126 FIGS. 125-127. Androlaelaps phoeniculi (Zumpt & Till), female. Venter (Fig. 125); dorsum (Fig. 126); chelicera (Fig. 127). Androlaelaps phoeniculi (Zumpt & Till) Haemolaelaps phoeniculi Zumpt & Till, 1954; R ev - Ecuat. Ent. Parasit. 2 : 209, figs. 1-2. FEMALE (figs. 125-127) : Deutosternum with 6 rows of 1-3 teeth. Chelae 30-32(0, long; pilus dentilis slender; arthrodial filaments subequal in length. Dorsal shield reticulate, truncate posteriorly, bearing 37 pairs of setae, two pairs being absent from the anterior part of the shield. Lengths of setae 24 and Z2 do not exceed | the distance between their bases ; the setae at the anterior end of the shield are longer, ii being at least twice as long as 24. Posterior terminal seta Z5 barbed, 5-6 times as long as seta /5. DL=5i7-579[j.; DW= 290-340^; DL/DW= i -6-1-8. " ZOOL. IO, I 66 W. M. TILL Sternal shield similar to that of the mesopicos-group, but relatively broader. SL=62~72[A; SW= 134-156(0.; SL/SW= 0-4-0-5. Anterior and lateral parts of the shield granular and reticulate, remaining part of shield apparently smooth ; presternal area reticulate. First pair of sternal setae and metasternal setae subequal in length, second and third pairs slightly longer. Genital shield extends almost to the anal shield and is apparently without any distinct pattern. It is slightly widened behind the genital setae, with its lateral margins subparallel and its posterior border almost straight. GW 1 =ioo-ii4(j.; GW z =88-g6[L', GW 1 /GW 2 =i-o-i-3. Genital setae as long as first pair of sternal setae, about 1/3 GW 2 . Anal shield distinctly longer than broad, anterior margin straight or slightly convex. AL=93-ii4[jt.; AW=72-83[z; AL/AW= 1-1-1-4. Anus situated at less than half its length from the anterior margin of the shield. Postanal seta about 1 1 times as long as paranal setae, the latter being level approximately with the middle of the anus. Integument of idiosoma bears about 29 pairs of ventral and marginal setae; the ventral setae are short, fine and simple, the marginal ones coarser and mostly barbed. Metapodal plates elongate, slightly curved, length to width about 8 : i. Tibia III with 9 setae ; anterior seta on coxa I at most i| times as long as posterior seta. The leg setae are simple and there are no stout, blunt spurs. Legs short and stumpy; length of tarsus IV about 4 times the width at the base. MALE (not figured) : Length 530^, width 340^. Holoventral shield narrow behind coxae IV, with a constriction anterior to the anal region. It bears 15 setae, the first pair lying on the presternal area as in the female. Metapodal plates drop- shaped. HOST AND LOCALITIES: Phoeniculus purpureus Miller from Potchefstroom, Transvaal (type series), from Tsessebe, Bechuanaland, and from Northern Rhodesia (S.A.I.M.R.). Androlaelaps rhodesiensis (Zumpt & Patterson) Hypoaspis (Haemolaelaps} rhodesiensis Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 82, fig. 4. Haemolaelaps rhodesiensis, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 243. FEMALE (figs. 128-130) : Deutosternum provided with 6 rows of 3-5 small teeth. Chelae 62-70^ long ; pilus dentilis slightly inflated, ribbon-like ; arthrodial filaments subequal in length. Dorsal shield broadly oval, reticulate, widest about level of seta Si. DL= 825-912^; DW= 586-700^; DL/DW=i-3-i-4. The shield bears the usual 39 pairs of setae, as well as two accessory setae at the level of 74. The setae are all very long, some having fine barbs; lengths of setae 24 and Z2 exceed twice the distance between their bases; seta /5 is about 4/5 as long as seta Z$. Sternal shield reticulate; SL=i 14-130^; SW= 170-193^; SL/SW= 0-6-0-7. Metasternal seta about as long as first sternal seta, second and third sternal setae slightly longer than the first. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 67 Genital shield with a reticulate pattern extending to the anterior flap, greatly expanded behind the fourth pair of coxae. GWj= 254-300^; GW 2 = 130-140^; GWj/GW 2 = 2-0-2-2. Genital setae relatively short, their length about 2/3 GW 2 . Anal shield triangular, broader than long. AL= 150-160^; AW= 170-184^; AL/AW= 0-8-0-9. Anus slightly more than its length from the anterior margin. Paranal setae inserted slightly behind middle of anus, 2/3 to 4/5 as long as the postanal seta. 128 U9 FIGS. 128-130. Androlaelaps rhodesiensis (Zumpt & Patterson), female. Venter (Fig. 128); dorsum (Fig. 129); chelicera (Fig. 130). Integument of idiosoma bears 15 pairs of long ventral and marginal setae, of which 3 pairs flank the genital shield; most of the setae have fine barbs. Metapodal plates roughly oval, length about twice the width. Tibia I has 14 and tibia III has 9 setae. Anterior seta on coxa I about 4/5 as 68 W. M. TILL long as posterior seta. Most of the leg setae are barbed; setae on some segments stouter than on others, but there are no blunt spines. Length of tarsus IV about 7 times the width at the base. MALE: Not known. HOSTS AND LOCALITIES: Saccostomus campestris Peters from several localities in Southern Rhodesia (S.A.I. M.R.). Rattus rattus (Linnaeus) and Rattus morio (Trouessart) from Brazzaville, French Congo (Zumpt & Till, 1961). Steatomys pratensis Peters from Southern Rhodesia (one female only, Zumpt & Till, 1961). 131 132 FIGS. 131-132. Androlaelaps sangsteri (Radford) female. Venter (Fig. 131); dorsum (Fig. 132). Androlaelaps sangsteri (Radford) Liponyssus sangsteri Radford, 1942, Parasitology 34 : 190, fig. n; Zumpt & Till, 1953, 5. Afr. J. Med. Sci. 18 : 7. Haemolaelaps sangsteri, Strandtmann & Wharton, 1958, Manual of Mesostigmatid Mites : 44. FEMALE: (figs. 131-132) : Deutosternum provided with 6 rows of teeth, 4 teeth in the first row, 2 in the second and one in each of the remaining rows. Corniculi reduced, comparatively small and inconspicuous. Chelicerae missing in the type, the only available specimen. Dorsal shield narrow with lateral margins subparallel, granular in appearance, with faint reticulations. DL=5i8pi; DW=257pi; DL/DW=2-o. The number of ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 69 setae is reduced, only 34 pairs being distinguishable with certainty. Those on the central part of the shield are short, the lengths of setae 24 and 22 not exceeding half the distance between their bases. The anterior and marginal setae are longer; seta Z5 is at least 4 times as long as seta /5 and has fine barbs. Sternal shield granular, anterior margin slightly convex, posterior margin straight, SL=96(ji; SW=i32(x; SL/SW=o7. Sternal setae subequal in length, relatively short, not reaching bases of following setae. Metasternal setae about as long as sternal setae. FIGS. 133-135. Androlaelaps scapularis (Berlese), female. Venter (Fig. 133); dorsum (Fig. 134); chelicera (Fig. 135). Genital shield granular, with an indistinct reticulate pattern; widest near its posterior border, between the first and second pairs of flanking setae ; anterior part of shield rather sharply pointed. GW 1 =ioopi; GW 2 =72[x; GWJGWg 1-4. Length of genital setae about half GW 2 . Anal shield triangular, approximately as long as broad. AL=82[j.; AW=79jji; Parana! setae level with middle of anus, about 4/5 as long as postanal seta. Integument of idiosoma bears 26 pairs of ventral and marginal setae, the longer ones near the posterior end of the body being barbed. Metapodal plates not visible. Legs short and stumpy, their chaetotaxy following the typical Androlaelaps yo W. M. TILL pattern ; all the setae simple and hair-like. Length of tarsus IV nearly 4 times the width at the base. MALE : Not known. HOST AND LOCALITY : Protoxerus stangeri (Waterhouse) from Bwamba Valley, Toro, Uganda (type in B.M.N.H.). Androlaelaps scapularis (Berlese) Hypoaspis (Haemolaelaps) scapularis Berlese, 1918, Redia 13 : 124; Keegan, i956a, Trans. Amer. micr. Soc. 75 : 314, figs. 13-14. 1 'schnolaelaps bathyergus Radford, 1939, Parasitology 31 : 250, figs. 7-8 (syn. nov.). Haemolaelaps eloffi Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 226, figs. 6-9 (syn. nov.}. Haemolaelaps natalensis Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 230, figs. 10-11 (syn. nov.}. FEMALE (figs. 133-135) : Deutosternum provided with 6 rows of 2-4 teeth, the teeth in rows 5 and 6 being very small and inconspicuous ; anterior rostral seta at least 1 1 times as long as capitular seta. Chelae very small, 18-2 2 fz long; fixed digit weakly sclerotized, bearing a slender pilus dentilis ; movable digit in two parts, one part edentate except for the terminal hook and a small subterminal tooth, the other part dentate but more weakly sclerotized. Arthrodial filaments subequal in length. Dorsal shield broadly oval, granular, with a faint reticulate pattern. DL=445- 570(x; DW=28o-390[x; DL/DW=i-3-i-6. The shield bears 39 pairs of setae. The centrally situated ones are very small, the lengths of 24 and 22 being less than 1/5 the distance between their bases; the anterior and marginal setae are longer; seta Z5 is at least 2.\ times as long as seta 75. Sternal shield granular, presternal area reticulate. SL=^i-82[i; SW= 103-133^; SL/SW= 0-4-07. Sternal setae subequal in length, first pair slightly shorter than the others ; metasternal setae about half as long as third pair of sternal setae. Genital shield flask-shaped, moderately widened behind the genital setae, granular in appearance, with a faint reticulate pattern. GW^ i 03-162^; GW 2 = 75-105 pi; GWj/GW2= 1-2-1-5. Genital setae relatively short, not more than half GW 2 . Anal shield triangular, distinctly wider than long. AL=52-94fx; AW=83-ii4(j.; AL/AW=o-5~o-7. Paranal setae level approximately with middle of anus, or slightly anterior, about i times as long as postanal seta. Integument of idiosoma bears about 20 pairs of simple ventral and marginal setae, of which 3 pairs border the genital shield. Metapodal plates elongate, length to width about 6 : i. Chaetotaxy of legs normal. Posterior seta on coxa I is i|-i| times as long as the anterior seta. Anterior seta on coxa III about twice as long and thick as posterior seta. Length of tarsus IV about 4 times the width at the base. MALE (fig. 136) : Fixed digit of chelicera with a slender pilus dentilis as in the female. Dorsal shield as in the female except that the centrally situated setae are relatively longer. Holoventral shield expanded behind coxae IV, bearing 23 setae, 5 pairs lying in the region between coxae IV and the anus. Integument of iosoma with about 13 pairs of setae. Chaetotaxy of legs as in female. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE 5. lot. 71 HOSTS AND LOCALITIES: Bathyergus suillus (Schreber) from Strandtfontein, Cape Town (types in Dr. C. D. Radford's collection and in M.R.A.C.), and from Tulbagh District, Western Cape Province (S.A.I.M.R.). Cryptomys hottentotus (Lesson) from Bloemfontein, Orange Free State, and Pietermaritzburg, Natal (S.A.I.M.R. and B.M.N.H.). 136 FIGS. 136-138. Androlaelaps scapularis (Berlese), venter of male (Fig. 136). Androlaelaps spatuliformis (Lavoipierre), venter (Fig. 137) and chelicera (Fig. 138) of male. Cryptomys holosericeus (Wagner) from Pietermaritzburg, Natal (S.A.I.M.R.). NOTE: The specimens from Bathyergus suillus are larger than those from the Cryptomys species, but there are no reliable features which could be used to separate these groups. The types of /. bathyergus, H. eloffi and H. natalensis have been compared with drawings of the type of H. scapularis Berlese made by Dr. F. Gorirossi (Keegan, 1956) and it is believed that they are all conspecific. The deep posterior concavity of the sternal shield, indicated by Zumpt & Till (1953) in their drawing of H. eloffi, appears to be the result of over-clearing. 72 W. M. TILL Androlaelaps spatuliformis (Lavoipierre) Haemolaelaps spatuliformis Lavoipierre, 1956, Ann. trop. Med. Parasit. 50 : 293, fig. 2. FEMALE (figs. 139-140) : Deutosternum provided with at least 4 rows of 2-3 small teeth, 5th and 6th rows very indistinct. Chelae about 40^ long; pilus dentilis and arthrodial filaments not clearly visible in the one specimen available for examination. 139 140 FIGS. 139-140. Androlaelaps spatuliformis (Lavoipierre), female, dorsum (Fig. 140). Venter (Fig. 139); Dorsal shield oval; DL=525(x; DW=36o(j.; DL/DW=i-4- Several of the setae are missing from the specimen examined, but the chaetotaxy appears to follow the 38- or 39-pair pattern. Seta px$ cannot be distinguished in this specimen. Seta Z5 is more than twice as long as seta /5. Sternal shield has a few very indistinct reticulations; presternal area weakly ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 73 reticulate. Sternal and metasternal setae subequal in length; first pair of sternal setae inserted on presternal area. SL= 94-96^; SW= 140-1 43^; SL/SW=O7. Genital shield spatulate in shape, very slightly widened behind genital setae, then tapering gradually; GW 1 =ioo[j.; GW 2 =97;j.. Genital setae relatively short, about i GW 2 . Anal shield slightly longer than broad; AL=ioo-ii2[x; AW=9i^.. Paranal setae inserted near posterior margin of anus, about f as long as postanal seta. Integument of idiosoma bears about 19 pairs of ventral and marginal setae. Metapodal plates provided with a pronounced medial protrusion. Chaetotaxy of legs normal except for genu IV which has only 9 instead of 10 setae. Posterior seta on coxa I about i| times as long as anterior seta. MALE (figs. 137-138) : Structure of chelicera indicated in fig. 138 ; pilus dentilis not visible. Dorsal shield as in female; seta px$ present. DL= 404-409^; DW=266^; DL/DW=i-5. Holoventral shield expanded behind coxae IV, bearing 21 setae, excluding the first pair of sternal setae which is situated on the presternal area. Chaetotaxy of legs as in the female. Anterior and posterior setae of coxae I approximately equal in length and thickness. Femur II has one stout ventral spine and tarsus II has two stout ventral spines. HOST AND LOCALITY : Anomaluris derbianus Gray from Kumba, British Cameroons. One female paratype received on loan from Dr. M. M. J. Lavoipierre, and two males from the type host and locality received from the Liverpool School of Tropical Medicine, by courtesy of Miss M. Johnson. Androlaelaps spreo (Zumpt & Till) Haemolaelaps spreo Zumpt & Till, 1956, Z. Parasitenk. 17 : 290, fig. 9. FEMALE (figs. 141-143): Androlaelaps spreo is very closely allied to A. dasymys, differing from it mainly in the structure of the chelicerae. The pilus dentilis is different in shape and the arthrodial filaments are subequal in length. The chelae are 54~56pL long. The dorsal shield bears 39 pairs of setae as well as 4-6 accessory setae between the / series. All the setae are relatively long, the lengths of setae 24 and Z2 being at least 1-5 times the distance between their bases. The marginal setae are slightly longer and thicker than those on the median part of the shield and are barbed. Seta Z5 is 17-1-8 times as long as seta /5. DL=849~936[x; DW=525~6o4fz; DL/DW=i-6. Sternal shield reticulate; SL= 114-140^; SW=i66-i75[A; SL/SW=07~o-8. Genital shield slightly widened posteriorly, with a pronounced reticular pattern extending to the anterior margin. GW^ 158-175^; GW 2 = 114-1 23^; GWJ GW 2 =i-3-i-4. Genital setae relatively long, slightly more than 2/3 GW 2 . Anal shield approximately as long as broad and with a reticular pattern; AL= 140-162^ ; AW= 140-158(4. Anus situated at slightly more than its length from the anterior margin of the shield; paranal setae about 4/5 as long as postanal seta. Integument bears about 33 pairs of ventral and marginal setae, the latter being 74 W. M. TILL longer and thicker than the former and provided with barbs. Metapodal plates oval, 1-5-1-8 times as long as broad. Chaetotaxy of legs normal. Posterior seta on coxa I slightly longer and thicker than anterior seta. Tarsus II has three stout ventral bristles but no blunt spines. Trochanter IV has a stout dorsal bristle. Length of tarsus IV 6|~7| times width at base. 141 142 FIGS. 141-143. Androlaelaps spreo (Zumpt & Till), female. Venter (Fig. 141); dorsum (Fig. 142); chelicera (Fig. 143). MALE (figs. 144-145) : Chelae as illustrated, pilus dentilis resembling that of female. Dorsal shield bears 8 accessory setae, otherwise chaetotaxy as in female. DL= 675^; DW=447[A; DL/DW=i-5. Holoventral shield reticulate, expanded, bearing 23 setae. Chaetotaxy of legs as in female. HOSTS AND LOCALITIES: Spreo bicolor (Gmelin), Springs, Transvaal (type series, S.A.I.M.R.). Riparia paludicola (Vieillot), Johannesburg, Transvaal (S.A.I.M.R.). ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 75 Androlaelaps steyni (Till) Haemolaelaps steyni Till, 1959, /. ent. Soc. S. Afr. 22 : 430, figs. FEMALE (figs. 148-150) : This species is very closely allied to A. mesopicos. Dorsal shield relatively shorter and broader than in A. mesopicos and bearing 38 pairs of setae. DL= 578-648^; DW= 368-420^; DL/DW=i-5-i-6. Sternal shield relatively shorter than in A. mesopicos. SL= 53-79(1 ; SW= 144 147 146 FIGS. 144-147. Androlaelaps spreo (Zumpt & Till), venter (Fig. 144) and chelicera (Fig. 145) of male. Androlaelaps steyni (Till), venter (Fig. 146) and chelicera (Fig. 147) of male. 149-166(1; SL/SW=o-3-o-5. Genital shield similar to that of A. mesopicos. Anal shield triangular, with a straight anterior margin; AL= 96-105(0.; AW 105-114^; AL/ AW = 0-9-1-0. Postanal seta more than twice as long as paranal setae. Coxa III has a greatly thickened anterior spine; tarsus IV has only one stout terminal spine. Length of tarsus IV about 5 times width at base. MALE (figs. 146-147) : Chelicera as figured, fixed digit bearing a relatively short, moderately inflated pilus dentilis. Dorsal shield similar to that of female, but 7 6 W. M. TILL with longer marginal setae. DL= 530-534^; DW=3i5-333[j.; DL/DW= 1-6-1-7. Coxa III has a stout anterior spine as in the female. HOSTS AND LOCALITIES: Lybius torquatus (Dumont), Mabelikwa, Transvaal (type series, S.A.I.M.R. and B.M.N.H.). Trachyphonus vaillanti Ranzani, near Johannesburg, Transvaal (S.A.I.M.R.). Lamprotornis nitens (Linnaeus), Rustenburg, Transvaal (S.A.I.M.R.). FIGS. 148-150. Androlaelaps steyni (Till), female. Venter (Fig. 148) ; dorsum (Fig. 149) ; ventral view of tip of tarsus IV (Fig. 150). Androlaelaps suncus sp. nov. FEMALE (figs. 151-153) : Deutosternum provided with 6 rows of 2-5 teeth. Chelae about 30[j, long ; movable digit with a pronounced terminal hook which arches over the tip of the fixed digit ; pilus dentilis inflated, hour-glass-shaped in some positions ; arthrodial filaments subequal in length. Dorsal shield oval, reticulate, widest about the level of seta Si. DL= 608-694^; ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 77 DW=385-456{x; DL/DW= 1-5-17. The shield bears the usual 39 pairs of setae and one accessory seta. Lengths of setae 14 and 22 about 2/3 the distance between their bases ; seta Z$ nearly twice as long as seta 75. Setae Z$ and 53-55 are barbed, the others appear to be simple. Sternal shield reticulate, except in the postero-median third. SL = 90-105^ ; 151 152 FIGS. 151-153. Androlaelaps suncus sp. nov., female. Venter (Fig. 151); dorsum (Fig. 152); chelicera (Fig. 153). = 128-148^; SL/SW=o-7. Metasternal setae and first pair of sternal setae subequal in length, 2nd and 3rd pairs of sternal setae slightly longer. Genital shield slightly widened behind the genital setae, broadly rounded posteriorly, with a fairly well-marked, overall, reticular pattern. GW^ 105-120^ ; GW 2=84-10311; GWj/GW^i-2. Length of genital setae about 3/5 GW 2 . Anal shield approximately as long as broad; AL 103-120^; AW= 106-115^. Anus slightly more than its length from the anterior margin. Paranal setae level approximately with middle of anus, about 5/6 as long as postanal seta. 78 W. M. TILL Integument of idiosoma bears about 23 pairs of ventral and marginal setae, the latter being barbed. Metapodal plates elongate, length to width about 4:1. Chaetotaxy of legs normal. Posterior seta on coxa I is i| times as long as anterior seta. Length of tarsus IV about 5 times its width at the base. MALE: Not known. HOSTS AND LOCALITIES: Suncus varius (Smuts) from Tibedi, Basutoland, 9 October, 1953 (4 females). Crocidura sp. from Pilgrim's Rest, Transvaal, 14 July 1951 (one female not included in the type series). Holotype and one female paratype in the collection of the South African Institute for Medical Research, two female paratypes (1962.6.12.18-19) in the collection of the British Museum (Natural History). Androlaelaps tachyoryctes (Radford) Haemolaelaps tachyoryctes Radford, 1941, Parasitology 33 : 309, figs. 3-4; Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 236, figs. 19-21. Hypoaspis (Haemolaelaps} tachyoryctes, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 71. FEMALE (figs. 154-157) : Deutosternum provided with 6 rows of 2-4 very small teeth. Chelae 34-36^ long; pilus dentilis moderately inflated, elongate, tapering to a point; arthrodial filaments subequal in length. Dorsal shield granular, with a faint reticulate pattern. DL= 580-627^; DW= SSS-SS ^ i DL/DW= 1-7-1-8. The shield bears 39 pairs of setae in the type specimens, but in the other material examined several of the r series are sometimes off the plate. Most of the setae are relatively short and fine ; setae 24 and Z2 about half as long as the distance between their bases; seta Z$ at least twice as long as seta /5 and considerably thicker. Sternal shield granular, presternal area reticulate; SL=jo-jg\; SW=i45-i48(x; SL/SW=o-5-o-6. First pair of sternal setae situated on presternal area, as long as metasternal setae, 2nd and 3rd pairs of sternal setae longer. Genital shield granular, slender, tapering behind the genital setae; the length of the latter about 3/4 GW 2 . GW^ 76-86^; GW 2 =8o^. Anal shield granular, very strongly convex anteriorly. AL=ii4-i22[z; AW= 72-80^; AL/ AW =1-4-1 -6. Anus situated at 1/3 of its length from the anterior margin. Paranal setae level approximately with middle of anus, about 4/5 as long as postanal seta and thinner. Integument of idiosoma bears 23 pairs of ventral and marginal setae, two pairs of ventral setae bordering the genital shield. Metapodal plates small, nearly triangular in shape, length to width about 2:1. Chaetotaxy of legs normal. Setae on coxa I subequal in length. Tarsus II bears 3 stout ventral setae and 3 stout, blunt, apical setae. Tarsus III has 3, and tarsus IV has 4 stout, rather blunt, apical setae. Length of tarsus IV slightly less than 4 times its width at the base. MALE (figs. 158-159) : Fixed digit of chelicera bears a slightly inflated, tapering pilus dentilis. Chaetotaxy of dorsal shield as in the female, but all the setae are ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 79 156 157 158 FIGS. 154-159. Androlaelaps tachyoryctes (Radford). Venter (Fig. 154), dorsum (Fig. 155) and chelicera (Fig. 156) of female; ventral view of tarsus II of female (Fig. 157). Venter (Fig. 158) and chelicera (Fig. 159) of male. 8o W. M. TILL situated on the shield. DL 475-49450.; DW= 290-304(0.; DL/DW 1-6-1-7. Holoventral shield greatly expanded behind 4th pair of coxae and bears 23 setae, 5 pairs in the region between coxae IV and the anus. Integument of idiosoma bears 13 pairs of setae. Chaetotaxy of legs as in female. HOSTS AND LOCALITIES: Tachyoryctes splendens (Riippell) from several localities in Kenya (types in B.M.N.H., other material presented by Miss J. B. Walker to S.A.I.M.R.). Lemniscomys striatus (Linnaeus), Njoro, Rift Valley Province, Kenya (Keegan, 1956). Androlaelaps taterae (Zumpt & Patterson) Hypoaspis (Haemolaelaps) taterae Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 83, fig. 6. Haemolaelaps taterae, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 244. FEMALE (figs. 160-162): A. taterae is very closely allied to A. villosissimus, differing from it chiefly in being more densely haired. This is particularly noticeable in the region between the genital and anal shields, where there are about 13 setae, and in the area between the genital shield and the meiapodal plates. The chelae of A. taterae are 44-48(0, long; pilus dentilis strongly inflated basally, terminal half slender and sharply bent; arthrodial filaments variable in size, the longest being about twice the average length. Dorsal shield faintly reticulate and has a dense overall covering of setae. DL= 813-900(0,; DW= 480-550(0.; DL/DW= 1-6-1-7. Sternal shield reticulate, sharply demarcated from the reticulate and granular presternal area. SL= 128-148(0,; SW= 166-175(0,; SL/SW=o-8-o-9. Metasternal setae nearly as long as first pair of sternal setae ; the latter are barbed. Genital shield with an overall reticular and granular pattern, slightly widened posteriorly. GW X = 150-162(0.; GW 2 = 135-140(0.; GW 1 /GW 2 =i-i-i-2. Length of genital setae about 3/5 GW 2 . Anal shield reticulate; AL= 152-162(0,; AW= 124-138(0,; AL/AW= 1-1-1-2. Anus situated at approximately its length from the anterior margin. Paranal setae level with middle of anus, about 4/5 as long as postanal seta. Metapodal plates roughly oval, length about if 2\ times the width. Genu III bears 10 setae, tibia III bears 9, otherwise the chaetotaxy of the legs follows the normal pattern for the genus. Posterior seta on coxa I slightly longer than anterior seta; many leg setae barbed. Length of tarsus IV is 9 times its width at the base. MALE (figs. 163-164) : Chelicera as figured, the fixed digit bearing a pilus dentilis similar to that of the female. Dorsal shield and integument as in female; DL= 599(0.; DW=359(o,; DL/DW=i-7. Holoventral shield reticulate, bearing numerous setae on the posterior portion. HOSTS AND LOCALITIES: Widely distributed in Southern Africa on Tatera afra (Gray). Single specimens have been recorded from Rattus natalensis (Smith), Rattus chrysophilus (de Winton) and Otomys irroratus (Brants) (S.A.I.M.R., B.M.N.H., and M.R.A.C.). ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lal. 162 164 163 FIGS. 160164. Androlaelaps taterae (Zumpt & Patterson). Venter (Fig. 160), dorsum (Fig. 161) and chelicera (Fig. 162) of female. Venter (Fig. 163) and chelicera (Fig. 164) of male. ZOOL. 10, i 82 W. M. TILL Androlaelaps tateronis (Radford) I schnolaelaps -tateronis Radford, 1939, Parasitology 31 : 247, fig. 4. Hypoaspis (Haemolaelaps} tateronis, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 71. Haemolaelaps tateronis, Zumpt & Till, 1953, Ann. Inst. Med. trap. Lisboa 10 : 242. Haemolaelaps radfordi Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 241, figs. 24-25 (syn. nov.}. FEMALE (figs. 165-167) : Deutosternum provided with 6 rows of 2-4 teeth. Chelae 165 166 FIGS. 165-167. Androlaelaps tateronis (Radford), female. Venter (Fig. 165); dorsum (Fig. 166); chelicera (Fig. 167). about 36^ long ; pilus dentilis inflated basally, terminal portion slender and curved ; arthrodial filaments subequal in length. Dorsal shield oval, reticulate. DL= 656-664^; DW= 400-409^; DL/DW= 1-6-1-7. The shield bears 39 pairs of setae of moderate length; lengths of setae 44 and Z2 approximately equal to the distance between their bases; seta Z$, the longest on the shield, is about twice as long as seta 75. Sternal shield reticulate, sharply demarcated from granular and reticulate ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. hit. 83 presternal area. SL= 101-116(0.; SW= 136-1 37^; SL/SW^o-y-o-g. Metasternal seta about as long as first sternal seta, second and third setae slightly longer. Genital shield narrow, tapering behind genital setae, with a reticulate pattern which extends to the anterior margin. GW 2 = 106-113^; genital setae extend beyond bases of first pair of flanking setae, their length about 3/4 GW 2 . Anal shield pear-shaped. AL= 124-139(0,; AW= 96-106(0. ; AL/AW=i-3. Anus situated at about its length from the anterior margin; paranal setae level approxi- mately with middle of anus, about as long as postanal seta. Integument of idiosoma bears 21 pairs of setae, of which 2 pairs flank the genital shield, a third pair being placed between the genital and anal shields. Some of the longer setae have very fine barbs. Metapodal plates slender, length to width about 6 : i. Tibia III has 9 setae; the number on tibia II is not clear in all the specimens examined, but is either 9 or 10. Posterior seta on coxa I twice as long as anterior seta and considerably thicker. Anterior seta on coxa II and both setae on coxa III thickened. Three of the apical setae on tarsus II stout and spine-like, one of them blunt. Length of tarsus IV 6 times the width at the base. MALE: Not known. HOSTS AND LOCALITIES: Tatera valida (Bocage), Gulu District, Uganda (syntype of /. tateronis in B.M.N.H.). " Rat ", Yei, Sudan (paratype of H. radfordi in B.M.N.H.). Tatera nigricauda Peters, Lali Hills, Kenya (S.A.I.M.R.). Unidentified rodents from Uganda (B.M.N.H.) and from the Cameroons (I.R.S.C.). Androlaelaps tauffliebi sp. nov. FEMALE (figs. 168-170) : Deutosternum provided with 6 rows of 3-4 teeth; anterior rostral seta twice as long as capitular seta. Chelae 24(0. long ; movable digit divided into two parts, a ventral (medial) part which is edentate, except for the terminal hook, and a dorsal (external) part which is broad and dentate. Fixed digit weakly sclerotized, bearing a flagellar pilus dentilis. About 4 of the arthrodia] filaments conspicuously long and stout. Dorsal shield widest about level of seta ?% with a granular appearance and very faint indications of a reticular pattern. Laterally it has a distinct double outline. DL=532-56o[o,; DW= 400-418(0.; DL/DW=i-3-i-4. The shield bears the usual 39 pairs of setae, *3-*'5, 22-2:3, /i-/5 and px2-p%3 being reduced to microsetae. Seta Z5 five times as long as seta /5. In all the specimens examined the anterior margin of the shield is tucked under, so that setae si become marginal in position and ii and n are ventral. Sternal shield about 3 times as wide as long, deeply concave posteriorly, granular in appearance. SL= 43-48^; SW= 138-143(0.; SL/SW=o-3O-o-35. Sternal seta III nearly i| times as long as sternal seta I ; metasternal seta short and fine, about half as long as sternal seta I. Genital shield large, broadly rounded posteriorly, granular. GW X = 142-157(0.; GW 2 =i20-i33(x; GW 1 /GW 2 =i-2-i-3. Genital setae short, about 1/3 GW 2 . W. M. TILL 170 171 FIGS. 168-172. Androlaelaps taujfiiebi sp. nov. Venter (Fig. 168); dorsum (Fig. 169) and chelicera (Fig. 170) of female. Venter (Fig. 171) and chelicera (Fig. 172) of male. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lal. 85 Anal shield with some granulation and striations laterally. AL=77~79[x; AW=9i-98[j.; AL/AW=o-8-o-9. Anus situated at about half its length from the anterior margin. Paranal setae inserted near anterior third of anus, about the same length as the postanal seta. Integument of idiosoma bears about 20 pairs of setae, some of the marginal ones having fine barbs. Metapodal plates very long and slender, length about 10 times width. Chaetotaxy of legs normal, setae simple, without any marked modifications. Posterior seta on coxa I approximately i| times as long as anterior seta and slightly thicker. Length of tarsus IV about 6 times its width at the base. MALE (figs. 171-172) : Deutosternum and gnathosomal setae as in female. Chelicera provided with an elongate spermadactyl ; minute teeth present at tip of movable digit; fixed digit reduced, membranous. Total length of chela about Dorsal shield as in female. DL=4i8[x; DW=295(i; DL/DW=i-4. Sterni- ventral shield separated from anal shield, granular in appearance, with faint indica- tions of reticulation, and bearing 8 pairs of setae. Anal shield approximately as long as broad, otherwise similar to that of female. Metapodal plates relatively broader than in the female. Chaetotaxy of legs as in female, except that three setae on leg II are strongly inflated, namely, the upper, antero- ventral seta on femur II, upper ant ero- ventral seta on tibia II, and postero-ventral seta on tarsus II. HOST AND LOCALITY: Seven females and one male from Cryptomys mechowi Peters, Lac Calundo, Angola, 6 January, 1955 (from the collection of Dr. R. Taufflieb, I.R.S.C., leg. A. de Barros Machado). Holotype (1962.2.12.20), allotype (1962.6.12.21) and two female paratypes (1962.6.12.22-23) in the collection of the British Museum (Natural History) ; two female paratypes presented to Dr. R. Taufflieb, Institut de Recherches Scientifiques au Congo ; two female paratypes presented to the Dundo Museum, Angola. Androlaelaps theseus Zumpt Hypoaspis (Androlaelaps) theseus Zumpt, 1950, Parasitology 40 : 301, figs. 3-4. Turkiella theseus, Zumpt & Till, 1953, Ann. Inst. Med. trap. Lisboa 10 : 225. Androlaelaps theseus, Cooreman, 1954, Ann. Mus. Congo Beige Tervuren, Zool. i : 163, figs. 1-4. FEMALE (figs. 173-175) : Deutosternum provided with 6 rows each of 4-8 teeth. Chelae io6-io8[j. long; pilus dentilis slender, arthrodial filaments subequal. Dorsal shield reticulate and granular. DL= 1350-1570^; DW=iioo-i235fji; DL/DW=i-2-i-3. The shield bears 40 pairs of setae, an additional pair being present in the ax position. The setal lengths vary in mites from different hosts. Seta r-L distinctly longer than seta ii ; setae 14 and zz slightly longer than the distance between their bases; seta Z$ at least i| times as long as seta 75. Sternal shield granular and reticular, anterior and posterior margins concave. SL=i43-i70|ji ; SW=266-2?6y. ; SL/SW=o-5-o-6. First sternal seta reaches almost to base of third; second and third sternal and metasternal setae at least i| times as long as first sternal seta, 86 W. M. TILL Genital shield granular, slender, very slightly widened behind genital setae, then tapering posteriorly. GW X = 205-247^; GW 2 = 190-220^; GW 1 /GW 2 =i-i-i-2. Length of genital setae exceeds the width of the shield at the level of their bases. Anal shield reticulate and granular. AL= 180-205^; AW= 209-238(1; AL/AW o-8-i-o. Anus situated at approximately its length from the anterior margin. Paranal setae level with posterior half of anus, about f as long as postanal seta. Integument of idiosoma bears about 42 pairs of ventral and marginal setae, many of which appear to be barbed. Metapodal plates broadly oval, length approximately ij times breadth. 174 FIGS. 173-175. Androlaelaps theseus Zumpt, female. Venter (Fig. 173); dorsum (Fig. 174); ventral view of femur, genu and tibia of leg II (Fig. 175). Chaetotaxy of legs normal. Anterior and posterior setae on coxae I subequal in length, the posterior one slightly thicker. Spur on second femur relatively small, its length being about 1/5 the width of the femur. Inner and outer ventral setae on tibia II subequal in length and thickness ; inner and outer ventral setae on genu II both simple. Tarsus II has 3 stout, blunt, terminal spines. Length of tarsus IV about 7 times its width at the base. MALE (fig. 176-177) : Gnathosoma as in female ; fixed digit of chelicera elongate, membranous, bearing a slender pilus dentilis. Dorsal shield resembles that of female; DL= 1283-1302^; DW= 920-940^; DL/DW=i-4- Holoventral shield strongly reticulate and also granular, not ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lot. 87 expanded behind coxae IV, bearing 23 setae of which 10 (6+4) are situated in the preanal region behind coxae IV. Chaetotaxy of legs as in female, but with inner ventral seta on tibia II stouter, and inner ventral seta on genu II shorter and stouter. HOSTS AND LOCALITIES: Widely distributed in Southern Africa on Tatera afra (Gray), and found occasionally on Rattus natalensis (Smith), Rattus chrysophilus FIGS. 176-178. Androlaelaps theseus Zumpt, venter (Fig. 176) and chelicera (Fig. 177) of male. Androlaelaps villosissimus (Berlese), venter female (Fig. 178). (De Winton), Acomys cahirinus (Desmarest), Saccostomus campestris Peters and Steatomys pratensis Peters (Zumpt & Till, 1961). Recorded from Tatera afra (Gray) in the Belgian Congo (Cooreman, 1954). Material in S.A.I.M.R. and B.M.N.H. Androlaelaps villosissimus (Berlese) Hypoaspis (Haemolaelaps) villosissimus Berlese, 1918, Redia 13 : 128; Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 72, fig. 5. Haemolaelaps villosissimus, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 244; Keegan, , Trans, Amer, micr, Soc. 75 : 314, figs. 1-2, 88 W. M. TILL FEMALE (fig. 178) : Deutosternum provided with 6 rows of 3-4 small teeth. Chelae 40^ long; pilus dentilis and arthrodial filaments indistinct. Dorsal shield with only faint indications of a reticular pattern. DL=7i8-744(z; DW=429~455[x; DL/DW= 1-6-1-7. The shield has a dense overall covering of setae which completely obscures any pattern. Most of the setae appear to have fine barbs. Sternal shield reticulate, sharply demarcated from presternal area. SL= no-i2O(ji; SW= 153-162^; SL/SW= 0-7-0-8. Metasternal setae as long as first pair of sternal setae. Genital shield granular, with an overall reticular pattern. GW^ 140-144^; GW 2 =n8-i23(ji; GW 1 /GW 2 =i-i-i-2. Length of genital seta about half GW 2 . Anal shield reticulate anteriorly and antero-laterally and also granular in appear- ance. AL= 105-1 i4[x; AW=96-ioOfj,; AL/ AW == 1-1-1-2. Anus less than its length from the anterior margin of the shield. Paranal setae level with middle of anus, about 4/5 as long as postanal seta. Integument of idiosoma bears numerous setae, 8 or 9 being situated between the genital and anal shields. The ventrally placed setae appear to be simple, whereas the marginal ones are barbed. Metapodal plates oval, length about twice width. Genu III bears 10 setae and tibia III bears 9; otherwise the chaetotaxy follows the normal Androlaelaps pattern. Posterior seta of coxa I slightly longer than anterior seta; many leg setae barbed. Length of tarsus IV 10 times width at base. MALE: Not known. HOSTS AND LOCALITIES : Saccostomus campestris Peters, Zululand (Berlese, 1918) and Bulawayo, Southern Rhodesia (S.A.I.M.R.). Unidentified rodent from the Cameroons (I.R.S.C.). Androlaelaps walker ae sp. nov. FEMALE (figs. 179-181) : The hypostomal processes form two pairs of prominent brushes, one ventral in position and the other dorsal. Deutosternum provided with 6 rows of 2-3 teeth. Chelae about 48^ long; pilus dentilis slender; arthrodial filaments subequal in length. Dorsal shield oval, reticulate, showing pronounced hypertrichy which extends anterior to setae 23 and which is not confined to the median part of the shield. The basic chaetotactic pattern is thus partially obscured. DL= 760-780^; DW=475-494|x; DL/DW=i5-i-6. Sternal shield reticulate, anterior margin not sharply defined. SL=95-iiO(j.; SW= 133-142(0.; SL/SW= 0-7-0-8. Sternal and metasternal setae subequal in length. Genital shield flask-shaped, with a reticular pattern extending to the anterior flap. GWj= 133-142^; GW 2 =86-io5[j.; GW 1 /GW 2 =i-4-i-5. Genital setae as long as first pair of sternal setae and approximately equal to GW 2 . Anal shield arched anteriorly. AL= 142-152^ ; AW= 103-110^; AL/AW=i-4. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 89 Anus situated at approximately its length from the anterior margin of the shield. Paranal setae level with middle of anus, i| times as long as the barbed postanal seta. 179 180 FIGS. 179-181. Androlaelaps walkevae sp. nov., female. Venter (Fig. 170); dorsum (Fig. 1 80); chelicera (Fig. 181). Integument of idiosoma provided with numerous setae, nearly all of which appear to be barbed. Metapodal plates oval or irregular in shape, length to width not more than 3:1. Chaetotaxy of legs normal except that tibia I has 14 and tibia III has 9 setae. Tarsus II has 3 stout terminal and 3 stout ventral bristles, and many of the leg go W. M. TILL setae are barbed. Posterior seta of coxa I slighty longer than anterior seta. Length of tarsus IV 7 times width at base. MALE: Not known. HOST AND LOCALITY: Sixty-three females from Tachyoryctes splendens (Riippell), Muguga North, Kenya, 6 July, 1954, presented by Miss J. B. Walker, East African Veterinary Research Organization, to the South African Institute for Medical Research. Holotype and 38 paratypes in the collection of the S.A. Institute for FIGS. 182-184. Androlaelaps wilkini (Till), female. Venter (Fig. 182); dorsum (Fig. 183); chelicera (Fig. 184). Medical Research, 25 paratypes (1962.6.12.23-33) in the collection of the British Museum (Natural History). Androlaelaps wilkini (Till) Haemolaelaps wilkini Till, 1959, J. ent. Soc. S. Afr. 22 : 432, figs. FEMALE (figs. 182-184) : Deutosternum provided with 6 rows of 3-4 small teeth. Chelae about 40^ long; pilus dentilis small, inflated; arthrodial filaments subequal in length. Dorsal shield oval, reticulate; DL= 709-770^; DW=455-5oo(x; DL/DW=i'5-i-6. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 91 The shield bears 37 pairs of setae, ^4 and r6 being inserted on the adjacent integu- ment. Marginal setae of the r and 5 series barbed, the S setae being at least twice as long and thick as those of the / series. Seta Z$ sinuous, 2-^-3 times as long as seta /5. Setae on the median part of the shield (i series posterior to 13, J series, Z2-Z3, Zi-Z4, p%2-px3) subequal in length, setae 24 and 22 slightly more than half as long as the distance between their bases. Sternal shield and presternal area both reticulate. SL^o-ygyi; SW= 145-158^; SL/SW=o-5- First pair of sternal setae slightly shorter than 2nd and 3rd pairs and situated on presternal area. Metasternal setae as long as first pair of sternal setae. Genital shield relatively slender, very slightly widened behind the genital setae, lateral margins almost parallel, posterior portion with a reticulate pattern. GW 1 =i23-i49[x; GW 2 =96-ii4n; GW 1 /GW 2 =i -1-1-3. Anal shield pear-shaped, with a reticulate pattern anteriorly and antero-laterally. AL=i23-i40[x; AW=ii4-i27[o.; AL/AW=i-i-i-2. Paranal setae on a level between middle and posterior margin of anus, their length not more than 2/3 that of the postanal seta. Integument of idiosoma bears 24 pairs of ventral and marginal setae, the longer ones being provided with fine barbs; posterior terminal pair long and whip-like. Metapodal plates slender, slightly curved. Tibia III has 9 setae; anterior and posterior setae on coxa I uniform in length and thickness. Tarsus II bears two blunt terminal spines; tarsi III and IV also have a pair of terminal spine-like bristles, but these are more slender and pointed than the spines on tarsus II. Length of tarsus IV 5 times width at base. MALE: Not known. HOSTS AND LOCALITIES: Myrmecocichla formicivora (Vieillot), Debeete, Bechuana- land (type series, S.A.I.M.R. and B.M.N.H.). Merops apiaster Linnaeus, Johannesburg, Transvaal (S.A.I.M.R.). Androlaelaps zulu (Berlese) Hypoaspis (Haemolaelaps) inops var. zulu Berlese, 1918, Redia 13 : 125. Haemolaelaps inops var. zulu, Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 234, fig. 18. FEMALE (figs. 185-187) : Deutosternum provided with 6 rows each of 4 narrow pointed teeth. Chelae 40(j. long ; pilus dentilis strongly inflated basally, distal third or fourth slender, with recurved apex; arthrodial filaments very short, subequal in length. Dorsal shield reticulate, bearing 39 pairs of setae and 1-3 unpaired accessory setae between the / series. Setae on median part of shield relatively long, lengths of setae 14 and 22 about i times the distance between their bases. Marginal setae slightly longer and thicker, with fine barbs. Seta Z5 at least i times as long as seta 75. DL=627-76ofji; DW= 400-494^; DL/DW=i-5-i-6. Sternal shield reticulate, sharply demarcated from reticulate presternal area. SL=ioo-io5(x; SW=i28-i52{x; SL/SW=o7-o-9. Metasternal setae about as long as first pair of sternal setae. 92 W. M. TILL Genital shield reticulate, moderately expanded behind genital setae. GW X = 143-170^; GW 2 = 105-1 14^; GW 1 /GW 2 =i-3-i-5. Genital setae relatively short, slightly more than half GW 2 . Anal shield reticulate antero-laterally. AL=g6-i2O[i.; AW= 103-127^; AL/AW =0-9-1-0. Anus situated less than its length from the anterior margin of the shield. Paranal setae between middle and posterior end of anus, reaching beyond base of, and about 4/5 as long as, postanal seta. 185 FIGS. 185-187. Androlaelaps zulu (Berlese), female. Venter (Fig. 1 86); chelicera (Fig. 187). (Fig. 185); dorsum Integument of idiosoma bears about 22 pairs of marginal and ventral setae, of which 3 pairs border the genital shield. Marginal setae coarser than the truly ventral ones, and provided with fine barbs. Outer metapodal plates variable, usually cigar-shaped, length 2^-3-J- times the width, ETHIOPIAN MITES OF THE GENUS A N D R O L A E L . IP S BE RLES E 5. tat. 93 Chaetotaxy of legs normal. Anterior seta on coxa I about 4/5 as long as posterior seta. Length of tarsus IV at least 6 times width at base. MALE: Not known. HOSTS AND LOCALITIES: Lemniscomys griselda (Thomas), Rattus natalensis (Smith) and Otomys irroratus (Brants) from the Transvaal (S.A.I.M.R.). 188 189 FIGS. 188-190. Androlaelaps zuluen&is Zumpt, female. Venter (Fig. 188); dorsum (Fig. 189); ventral view of femur, genu and tibia of leg II (Fig. 190). Tatera afra (Gray) from Zululand (S.A.I. M.R.). (Mites from the above hosts were previously identified by Zumpt & Till (1956) as forms of Haemolaelaps glasgowi.) Keegan (1956) has recorded this species from several rodents in Egypt. 94 W. M. TILL Androlaelaps zuluensis Zumpt Hypoaspis (Androlaelaps} zuluensis Zumpt, 1950, Parasitology 40 : 300, fig. i. Turkiella zuluensis, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 218, figs. 2-5. Androlaelaps zuluensis, Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 229, fig. 14. FEMALE (figs. 188-190) : Deutosternum provided with 6 rows each of 3-4 very small teeth. Chelae 46^ long; pilus dentilis slender; arthrodial filaments short and subequal in length. Dorsal shield reticulate, oval. DL= 656-675^; DW 403-486^; DL/DW=i-5. The shield bears 39 pairs of setae as well as 2 or 3 unpaired accessory setae between the J series. Lengths of setae 24 and Z2 approximately equal to the distance between their bases; seta Z$ about i| times as long as seta /5. The marginal setae appear to be barbed. Sternal shield reticulate, posterior margin almost straight or slightly irregular. SL=io5[j.; SW=n8-i24[A; SL/SW=O'9. Sternal and metasternal setae all relatively short, a little more than half the length of the sternal shield. Genital shield with a pattern of striations on its posterior portion; very long, extending almost to the anal shield. GW 1 =iyo[i; GW 2 = 118-124^; GWJGWg = 1-4. Genital setae very short, less than half GW 2 . Anal shield with a reticulate pattern; AL= 105-109^; AW=io5[j.; AL/AW=i-o. Anus situated at less than its length from the anterior margin of the shield. Paranal setae near posterior margin of anus; postanal seta broken in both the specimens examined. Integument of idiosoma bears 20-22 pairs of marginal and ventral setae. Meta- podal plates slender, length 3-6 times the breadth. Chaetotaxy of legs normal. Anterior seta on coxa I slightly shorter than posterior seta. Spur on femur II rather slender, thumb-like, its length a little less than half the width of the femur at its base. Inner ventral seta on genu II short, spur-like; inner ventral seta on tibia II considerably stouter than outer one; tarsus II terminates in pointed setae only. Length of tarsus IV about 8 times its width at the base. MALE (fig. 191) : Deutosternum with 6 rows of 2-5 small teeth. Spermadactyl 6o(A long, fixed digit bears a slender pilus dentilis. Dorsal shield similar to that of female; DL=52Opt. (approx.) ; DW=36o[u.; DL/ DW=i-4- Holo ventral shield reticulate, expanded behind coxae IV, bearing 23 setae, 5 pairs of setae being situated in the preanal region behind coxae IV. Chaetotaxy of legs similar to that of female, except that inner ventral setae on genu II and tibia II are more slender. HOSTS AND LOCALITIES : Rattus namaquensis (Smith) from Abercorn Pont, Zululand (holotype in S.A.I.M.R.), and from Kapps Siding, S. W. Africa (S.A.I. M.R.) Rattus chrysophilus (De Winton), Southern Africa (Zumpt & Till, 1961). Petromyscus collinus (Thomas & Hinton), Kamanjab-Ohopoho Road, S.W. Africa (S.A.I.M.R.). Galago senegalensis Geoffroy (the figured specimens) from Karamoja, Uganda (S.A.I.M.R.). ETHIOPIAN MITES OF THE GENUS AND1WLAELAPS BERLESE s. hit. 95 Keegan (1956) records this species from Arvicanthis niloticus (Desmarest) and Mus musculus Linnaeus in Egypt. Androlaelaps zutnpti sp. nov. FEMALE (figs. 193-195): Deutosternum bears 6 rows of 3-5 very fine teeth. Chelae about 40^. long; pilus dentilis inflated; one arthrodial filament strikingly longer than the others ; an extension of the arthrodial membrane between the chelae is elongate, reaching the tips of the chelae. 191 192 FIGS. 191-192. Androlaelaps zuluensis Zumpt, venter of male (Fig. 191). Androlaelaps zumpti sp. nov., venter of male (Fig. 192). Dorsal shield reticulate, widest in posterior half at level of setae Si and 52. DL= 656-732 [A; DW= 442-494(0. ; DL/DW= 1-3-1-5. The shield bears 39 pairs of setae ; lengths of setae 24 and 22 a little more than half the distance between their bases ; length of seta /i is f that of seta Si ; seta Z$ is 2-2 1 times as long as seta 75. Sternal shield reticulate, sharply demarcated from reticulate and granular presternal area. SL=86-ii4(j.; SW= 124-133^; SL/SW=o-7-o-9. Metasternal 9 6 W. M. TILL and first pair of sternal setae subequal in length, 2nd and 3rd pairs of sternal setae slightly longer. Genital shield scarcely widened behind genital setae, tapering posteriorly, with a well-defined reticular pattern extending over the anterior as well as posterior parts. 193 194 FIGS. 193-195. Androlaelaps zumpti sp. nov., female. Venter (Fig. 193); dorsum (Fig. 194); chelicera (Fig. 195). GW 1 =95-ii4[ji; GW 2 =95-io5[j.. Genital setae about as long as first pair of sternal setae and about 2/3 GW 2 . Anal shield about 133^ long, 90^ wide; AL/AW=i-5. Anus situated at less than its length from the anterior margin of the shield. Paranal setae near posterior margin of anus, about 2/3 as long as postanal seta. Integument of idiosoma bears about 23 pairs of ventral and marginal setae, those ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 97 near the posterior extremity of the body having fine barbs. Metapodal plates slender, length at least 5 times the width. Chaetotaxy of legs normal, anterior seta on coxa I about 4/5 as long as posterior seta. Trochanter IV bears a stout dorsal seta. Length of tarsus IV about 6 times its width at the base. MALE (fig. 192) : Deutosternum bears 6 rows of 4-7 very fine teeth. Fixed digit of chelicera bears a pilus dentilis similar in shape to that of the female. Dorsal shield 55O[j. long, 352[A wide; chaetotaxy similar to that of female. Holo- ventral shield strongly reticulate, bearing 23 setae, 5 pairs being placed in the region between coxae IV and the anus. Chaetotaxy of legs as in the female. HOSTS AND LOCALITIES: Holotype (female), allotype and one male and three female paratypes from Rattus (Aethomys) sp., Richmond, Cape Province, 31 March, 1959, two female paratypes from Rattus (Aethomys) sp., Middelburg, Cape Province, 28 March, 1959. Holotype, one male and two female paratypes in the collection of the South African Institute for Medical Research; allotype (1962.6.12.34) and two female paratypes (1962.6.12.35-36) in the collection of the British Museum (Natural History). NOTE: Several populations of mites have been examined which are very similar to A . zumpti, except that the setae of the dorsal shield are longer (lengths of setae 24 and Z2 about i^ times the distance between their bases) and 2-4 accessory setae are present. These mites, all from the collection of the S.A. Institute for Medical Research, have been taken from the following hosts and localities: Rattus chrysophilus (De Winton), Matopos, S. Rhodesia. Rattus namaquensis (Smith), Bulawayo, S. Rhodesia; Abercorn Pont, Zululand; several localities in S.W. Africa. Rattus paedulcus (Sundevall), Kaokoveld, S.W. Africa. Otomys unisulcatus Cuvier, Luckhoff, Orange Free State. Rhabdomys pumilio (Sparrman), Cape Province and S.W. Africa. Petromyscus collinus (Thomas & Hinton), Kamanjab-Ohopoho, S.W. Africa. Specimens from Parotomys littledalei Thomas, van Rhynsdorp, Cape Province, are intermediate between these and the short-haired form, the lengths of setae 24 and Z2 being slightly less than the distance between their bases. For the present, these will all be regarded as forms of A. zumpti. HOST-PARASITE LIST.* CLASS AVES Order FALCONIFORMES FAMILY AEGYPIIDAE Gyps coprotheres (Forster) Androlaelaps pater soni (Zumpt & Till) * The host names are the same as those used by Zumpt (1961). ZOOL 10, I 98 W. M. TILL Order CORAGIIFORMES FAMILY MEROPIDAE Merops apiaster Linnaeus Androlaelaps wilkini (Till) FAMILY PHOENIGULIDAE Phoeniculus purpureus (Miller) Androlaelaps phoeniculi (Zumpt & Till) FAMILY CAPITONIDAE Lybius torquatus (Dumont) Androlaelaps steyni (Till) Trachyphonus vaillanti Ranzani Androlaelaps steyni (Till) FAMILY PICIDAE Campethera abingoni (Smith) Androlaelaps haydocki (Till) Thripias namaquus (Lichtenstein) Androlaelaps haydocki (Till) Mesopicos griseocephalus (Boddaert) Androlaelaps mesopicos (Radford) Order PASSERIFORMES FAMILY TURDIDAE Myrmecocichla formicivora (Vieillot) Androlaelaps wilkini (Till) FAMILY HIRUNDINIDAE Riparia paludicola (Vieillot) Androlaelaps spreo (Zumpt & Till) FAMILY STURNIDAE Lamprotornis nitens (Linnaeus) Androlaelaps steyni (Till) Spreo bicolor (Gmelin) Androlaelaps spreo (Zumpt & Till) FAMILY PLOCEIDAE Spermestes cucullatus Swainson Androlaelaps congoensis sp. nov. CLASS MAMMALIA Order INSEGTIVORA FAMILY MACROSCELIDIDAE Elephant Shrew Androlaelaps arvicanthis Radford FAMILY SORICIDAE Suncus varius (Smuts) Androlaelaps suncus sp. nov. ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lal. 99 Crocidura spec. Androlaelaps suncus sp. nov. Crocidura manni Peters (?) Androlaelaps hystrici (Zumpt & Till) Order PRIMATES FAMILY LORISIDAE Galago alleni Waterhouse Androlaelaps galagus (Lavoipierre) Galago senegalensis Geoffrey A ndrolaelaps zuluensis Zumpt Order CARNIVORA FAMILY MUSTELIDAE Ictonyx striatus (Perry) Androlaelaps marshalli Berlese FAMILY VIVERRIDAE Suricata suricatta (Schreber) Androlaelaps marshalli Berlese Order LAGOMORPHA FAMILY LEPORIDAE Oryctolagus cuniculus (Linnaeus) Androlaelaps marshalli Berlese Order RODENTIA FAMILY BATHYERGIDAE Bathyergus suillus (Schreber) Androlaelaps scapularis (Berlese) Georychus capensis (Pallas) Androlaelaps capensis (Hirst) Androlaelaps cryptomius (Radford) Androlaelaps georychi sp. nov. Cryptomys holosericeus (Wagner) Androlaelaps capensis (Hirst) Androlaelaps scapularis (Berlese) Cryptomys hottentotus (Lesson) Androlaelaps capensis (Hirst) Androlaelaps marshalli Berlese Androlaelaps scapularis (Berlese) Cryptomys mechowi Peters Androlaelaps tauffliebi sp. nov, ioo W. M. TILL FAMILY HYSTRICIDAE Hystrix cristata Linnaeus Androlaelaps hy strict (Zumpt & Till) FAMILY SGIURIDAE Sciurus carolinensis Gmelin Androlaelaps casalis (Berlese) Heliosciurus gatnbianus (Ogilby) Androlaelaps heliosciuri sp. nov. Xerus inauris (Zimmermann) Androlaelaps casalis (Berlese) FAMILY ANOMALURIDAE Anomalurus derbianus (Gray) Androlaelaps spatuliformis (Lavoipierre) FAMILY PEDETIDAE Pedetes capensis (Forster) Androlaelaps marshalli Berlese FAMILY MUSCARDINIDAE Claviglis spec. Androlaelaps ghanensis sp. nov. FAMILY RHIZOMYIIDAE Tachyoryctes spec. Androlaelaps callosus (Berlese) Tachyoryctes splendens (Riippell) Androlaelaps marshalli Berlese Androlaelaps tachyoryctes (Radford) Androlaelaps walker ae sp. nov. FAMILY MURIDAE Arvicanthis niloticus (Desmarest) Androlaelaps arvicanthis Radford Androlaelaps marshalli Berlese Androlaelaps murinus (Berlese) Rattus spec. Androlaelaps zumpti sp. nov. Rattus chrysophilus (De Winton) Androlaelaps dasymys (Radford) Androlaelaps marshalli Berlese Androlaelaps later ae (Zumpt & Patterson) Androlaelaps theseus Zumpt Androlaelaps zuluensis Zumpt Androlaelaps zumpti sp. nov. Rattus morio (Trouessart) Androlaelaps rhodesiensis (Zumpt & Patterson) Rattus namaquensis (Smith) Androlaelaps dasymys (Radford) Androlaelaps marshalli Berlese ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESK s. hit. 101 Androlaelaps zuluensis Zumpt Androlaelaps zumpti sp. nov. Rattus natalensis (Smith) Androlaelaps arvicanthis Radford Androlaelaps dasymys (Radford) Androlaelaps marshalli Berlese Androlaelaps murinus (Berlese) Androlaelaps later ae (Zumpt & Patterson) Androlaelaps iheseus Zumpt Androlaelaps zulu (Berlese) Rattus paedulcus (Sundevall) Androlaelaps marshalli Berlese Androlaelaps zumpti sp. nov. Rattus rattus (Linnaeus) Androlaelaps casalis (Berlese) Androlaelaps dasymys (Radford) Androlaelaps rhodesiensis (Zumpt & Patterson) Rhabdomys pumilio (Sparrman) Androlaelaps dasymys (Radford) Androlaelaps glasgowi (Ewing) Androlaelaps marshalli Berlese Androlaelaps murinus (Berlese) Androlaelaps zumpti sp. nov. Lophuromys flavopunctatus Thomas Androlaelaps callosus (Berlese) Androlaelaps murinus (Berlese) Lophuromys sikapusi (Temminck) Androlaelaps dasymys (Radford) Oenomys hypoxanthus (Pucheran) Androlaelaps murinus (Berlese) Dasymys incomtus (Sundevall) Androlaelaps dasymys (Radford) Androlaelaps murinus (Berlese) Lemniscomys griselda (Thomas) Androlaelaps marshalli Berlese Androlaelaps zulu (Berlese) Lemniscomys striatus (Linnaeus) Androlaelaps murinus (Berlese) Androlaelaps tachyorycles (Radford) Acomys cahirinus (Desmarest) Androlaelaps theseus Zumpt Saccostomus campestris Peters Androlaelaps marshalli Berlese Androlaelaps oliffi (Zumpt & Patterson) Androlaelaps rhodesiensis (Zumpt & Patterson) W. M. TILL Androlaelaps theseus Zumpt Androlaelaps villosissimus (Berlese) Cricetomys gambianus Waterhouse Androlaelaps cricetomydis sp. nov. Androlaelaps galagus (Lavoipierre) Petromyscus collinus (Thomas & Hinton) Androlaelaps zuluensis Zumpt Androlaelaps zumpti sp. nov. Steatomys pratensis Peters Androlaelaps marshalli Berlese Androlaelaps rhodesiensis (Zumpt Patterson) Androlaelaps theseus Zumpt Otomys spec. Androlaelaps murinus (Berlese) Otomys irroratus (Brants) Androlaelaps dasymys (Radford) Androlaelaps glasgowi (Ewing) Androlaelaps murinus (Berlese) Androlaelaps later ae (Zumpt & Patterson) Androlaelaps zulu (Berlese) Otomys saundersiae Roberts Androlaelaps dasymys (Radford) Otomys sloggetti (Thomas) Androlaelaps dasymys (Radford) Otomys unisulcatus Cuvier Androlaelaps zumpti sp. nov. Parotomys brantsi (Smith) Androlaelaps dasymys (Radford) Parotomys littledalei Thomas Androlaelaps dasymys (Radford) Androlaelaps zumpti sp. nov. Mystromys albicaudatus (Smith) Androlaelaps capensis (Hirst) Androlaelaps dasymys (Radford) Desmodillus auricularis (Smith Androlaelaps marshalli Berlese Androlaelaps oliffi (Zumpt & Patterson) Gerbillus paeba Smith Androlaelaps marshalli Berlese Androlaelaps oliffi (Zumpt & Patterson) Gerbillus pyramidum Geoffroy Androlaelaps centrocarpus (Berlese) Meriones rex Yerbury & Thomas Androlaelaps longipes (Bregetova) ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 103 Tatera spec. Androlaelaps arvicanthis Radford Androlaelaps marshalli Berlese Tatera afra (Gray) Androlaelaps dasymys (Radford) Androlaelaps marshalli Berlese Androlaelaps oliffi (Zumpt & Patterson) Androlaelaps taterae (Zumpt & Patterson) Androlaelaps theseus Zumpt Androlaelaps zulu (Berlese) Tatera nigricauda Peters Androlaelaps marshalli Berlese Androlaelaps tateronis (Radford) Tatera valida (Bocage) Androlaelaps arvicanthis Radford Androlaelaps marshalli Berlese Androlaelaps tateronis (Radford) Taterillus spec. Androlaelaps centrocarpus (Berlese) Taterillus emini (Thomas) Androlaelaps marshalli Berlese Unidentified rodents Androlaelaps arvicanthis Radford Androlaelaps graingeri Zumpt & Patterson Androlaelaps hirsti (Keegan) Androlaelaps marshalli Berlese Androlaelaps tateronis (Radford) Androlaelaps villosissimus (Berlese) REFERENCES BERLESE, A. 1887. In " Acari, Myriopoda et Scorpiones hucusque in Italia reperta." Fasc. XL, N. 6. - 1911. Acarorum species novae quindecim. Redia 7 : 429-435. - 1916. Centuria seconda di Acari nuovi. Redia 12 : 125-177. - 1918. Centuria quarta di Acari nuovi. Redia 13 : 115-190. BREGETOVA, N. G. 1952. New species of mites of the genus Haemolaelaps (Gamasoidea Laelaptidae) parasitizing rodents. Zoo/. Zh. Moscow 31 : 860-882. - 1956. Gamasid Mites (Gamasoidea). A cad. Sci. U.S.S.R., Moskva 61 : 246pp. COOREMAN, J. 1954. Acariens du Congo Beige. Ire serie. Ann. Mus. Congo Beige Tervuren i : 163-168. COSTA, M. 1961. Mites associated with rodents in Israel. Bull. Brit. Mus. (nat. Hist.) Zool. 8 : 1-70. DELFINADO, M. U. 1961. Haemolaelaps travisi, a new species of mite from the Philippines (Laelaptidae : Acarina). Fieldiana, Zoology 44 : 49-51. FONSECA, F. DA. 1959. Notes d'acarologie. XLI. Haemolaelaps Berlese versus Atricholae- laps Ewing et Ischnolaelaps Fonseca; Ornithonyssus Samboa versus Bdellonyssns Fonseca. Mem. Inst. Butantan S. Paulo (1957/58), 28 : 45-54. 104 W. M. TILL GROKHOVSKAYA, I. M. and NGUEN-HUAN-HOE. 1961. Gamasid mites of North Viet-Nam. Part 2. Zoo/. Zh. Moscow 40 : 1633-1646. HIRST, S. 1916. Notes on parasitic Acari. Description of two new African mites of the family Gamasidae. /. Zool. Res. i : 76-81. KEEGAN, H. L. 1956. Ectoparasitic laelaptid and dermanyssid mites of Egypt, Kenya and the Sudan, primarily based on Namru 3 collections, 1948-1953. /. Egypt, publ. Hlth. Ass. 31 : 199-272. - i956a. Original illustrations of Haemolaelaps marsupialis Berlese, 1910 and of five additional Haemolaelaps species described, but not figured by Berlese. Trans. Amer. micr. Soc. 75 : 314-319. LAVOIPIERRE, M. M. J. 1955. A description of a new genus of sarcoptiform mites and of three new species of Acarina parasitic on primates in the British Cameroons. Ann. trop. Med. Parasit. 49 : 299-307. - 1956. A description of a new genus and of three new species of mites (Acarina, Para- sitiformes) parasitic on West African Mammals. Ann. trop. Med. Parasit. 50 : 291-298. RADFORD, C. D. 1939. Notes on some new species of parasitic mites. Parasitology 31 : 243- 254- - 1941. Notes on some new species of parasitic mites. Pt. 4. Parasitology 33 : 306-315. - 1942. New ectoparasitic mites (Acarina) from Uganda. Parasitology 34 : 185-194. - 1942^. New parasitic mites (Acarina). Parasitology 34 : 295-307. 1944. New parasitic mites from rodents. Parasitology 35 : 161-166. STRANDTMANN, R. W. 1949. The blood-sucking mites of the genus Haemolaelaps (Acarina : Laelaptidae) in the United States. /. Parasit. 35 : 325-352. TAUFFLIEB, R. and MOUCHET, J. 1959. Notes sur les acariens (Acarina; Laelaptidae et Spinturnicidae) du Cameroun. Ann. Parasit. hum. comp. 34 : 350-353. TILL, W. M. 1959. Three new Haemolaelaps species (Acarina : Laelaptidae) from birds in the Ethiopian region, and a redescription of Haemolaelaps mesopicos Radford. /. ent. Soc. S. Afr. 22 : 423-435. VITZTHUM, H. 1943. Acarina. In : Bronn, H. G., Klassen und Ordnungen des Tierreiches, Bd. 5, Abt. IV, Buch 5, Leipzig (Becker & Erler). ZUMPT, F. 1950. Notes on parasitic mites. I. Some remarks on the family Laelaptidae (sensu Vitzthum 1943) with descriptions of three new species from African rodents. Parasitology 40 : 298-303. and PATTERSON, P. M. 1950. The Ethiopian species of Hypoaspis subgen. Androlaelaps Berlese (1903), with description of a new species. S. Afr. J. Med. Sci. 15 : 67-74. and PATTERSON, P. M. 1951. Further notes on laelaptid mites parasitic on vertebrates. A preliminary study to the Ethiopian fauna. /. ent. Soc. S. Afr. 14 : 63-93. - and TILL, W. 1953. The genera Turkiella nov. ( Androlaelaps auct.) and Haemolaelaps in the Ethiopian region, with keys and descriptions of three new species (Acarina : Laelaptidae). Ann. Inst. Med. trop. Lisboa 10 : 215-249. and TILL, W. M. 1956. Notes on Haemolaelaps glasgowi (Ewing) and related forms in the Ethiopian region, with descriptions of four new species (Acarina : Laelaptidae). Z. Parasitenk. 17 : 282-291. - and TILL, W. M. 1958. Notes on the classification and synonymy of gamasid mites parasitic on vertebrates (Acarina : Mesostigmata) . /. ent. Soc. S. Afr. 21 : 261-273. and TILL, W. M. 1961. Suborder Mesostigmata. In Zumpt (1961): The arthropod parasites of vertebrates in Africa south of the Sahara (Ethiopian region). Vol. I (Chelicerata) . Publ. S.A. Inst. Med. Res. 9 : 17-91. THE FRANCOLINS, A STUDY IN SPECIATION B. P. HALL BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 10 No. 2 LONDON : 1963 THE FRANCOLINS, A STUDY IN SPECIATION BY B. P. HALL British Museum (Natural History) Pp. 105-204 ; 2 Text-figures, n Maps BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 10 No. 2 LONDON: 1963 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in jive series, corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 10, No. 2 of the Zoological series. Trustees of the British Museum 1963 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued March 1963 Price Thirty-one Shillings THE FRANCOLINS, A STUDY IN SPECIATION By B. P. HALL Associate, Department of Z oology > British Museum (Natural History] CONTENTS Page INTRODUCTION ........... 107 THE GENUS . .......... 108 THE GROUPS . . . . . . . . . . .no SPECIES AND SUBSPECIES . . . . . . . . .no METHODS . . . . . . . . . . . .112 CHARACTERS . . . . . . . . . . .115 THE SPOTTED GROUP (francolinus, pictus, pintadeanus) .... 123 THE BARE-THROATED GROUP (afer, swainsonii, rufopictus, leucoscepus) . 125 THE MONTANE GROUP (erckelii, ochropectus, castaneicollis , jacksoni, nobilis, camerunensis, swierstrai) . . . . . . . .131 THE SCALY GROUP (ahantensis, squamatus, griseostriatus) . . .136 THE VERMICULATED GROUP (bicalcaratus, icterorhynchus , clappertoni, hildebrandti, natalensis, hartlaubi, harwoodi, adspersus, capensis) . . 140 THE STRIATED GROUP (sephaena, streptophorus) . . . . .148 THE RED-WINGED GROUP (psilolaemus, shelleyi, africanus, levaillantoides, levaillantii, finschi) . . . . . . . . . .152 THE RED-TAILED GROUP (coqui, albogularis, schlegelii] .... 160 SPECIES NOT INCLUDED IN ANY GROUP (lathami, nahani, pondicerianus , gularis) ............ 165 CONCLUSIONS . . . . . . . . . . . 168 SUMMARY . . . . . . . . . . . .172 APPENDIX i . Summary of conclusions reached by R. E. Moreau in his paper " The Vicissitudes of the African Biota in the Late Pleistocene " . . 173 APPENDIX 2. Taxonomic Notes and Synonymy . . . . .175 MAPS ............ 180 REFERENCES ........... 200 INDEX ............ 201 INTRODUCTION Thirty-nine species of francolin (Francolinus and Pternistis] were recognised in Peters' Check List of the Birds of the World 2, 1934, of which thirty-four are African and five Asiatic: one (F. ochropectus Dorst & Jouanin, 1952) has been described since. I have rearranged Peters' species slightly, now recognising thirty-six in Africa and five in Asia. As is inevitable a Check List employing conventional nomenclature gives no indication of the relationship between the species which must be the basis for any discussion on speciation. The objectives of this paper are therefore twofold : firstly to set out in the text and on the maps the present relation- ship between the species. This is not as formidable a task as might be expected ZOOL. IO, 2 8 io8 B. P. HALL from the number of species involved, for thirty-seven of the forty-one fall readily into eight groups, seven of which are African and one Asiatic. (In this paper the term " group " is used for either a superspecies, in which all members are largely allopatric, or for a species group which is a rather looser assemblage of related forms containing some species which are partly sympatric with other members.) These groups are discussed and mapped separately with particular reference to their ecology and the relationship of neighbouring or sympatric forms. The four species that do not fall readily into any group are discussed separately at the end. Secondly, from the pattern presented by the birds of each group and by their distribution I have tried to reconstruct the climatic changes, with the resultant isolating and rejoining of populations, which could have formed this pattern. In doing so I am fully conscious that the picture we see today gives only fragmentary clues to the past and that it is impossible to interpret accurately any but the simplest situations. Nevertheless, even in the most complex groups, I believe there is some usefulness in making the attempt, for if a series of studies of this nature suggests changes that from other branches of science are known to have occurred, it may be possible to give some tentative datings to various steps in speciation. Throughout the preparation of this paper I have been particularly fortunate in being able to discuss the problems with R. E. Moreau, who has been preparing concurrently a paper on " The Vicissitudes of the African Biota in the Late Pleistocene ". He has allowed me to quote as Appendix i a summary of his conclu- sions of the major climatic changes that have taken place in Africa in the past 60,000 years. In acknowledging my gratitude to him for this and for invaluable advice and criticism I must also absolve him from any responsibility for some of my interpretations of the evolutionary stages, or the conclusions which I have drawn from them. I am also deeply indebted to M. P. Stuart Irwin who prepared distribution maps for me of all the francolins in the National Museum of Southern Rhodesia: also to Dr. H. Friedmann, C. W. Benson and I. C. J. Galbraith for criticism of the manuscript. Others to whom I am grateful for loans of specimens, advice, or other help, include Dr. Dean Amadon, Professor J. Berlioz, P. Blasdale, Mgr. F. O. Cave, E. M. Cawkell, Dr. James Chapin, P. A. Clancey, Miss M. Court enay-Latimer, I. H. Dillingham, H. J. de S. Disney, Sir Hugh Elliot, W. V. Harris, M. J. Hollis, Cdr. A. M. Hughes (who drew figure i), Gen. Sir Gerald Lathbury, C. W. Mackworth- Praed, Miles Markus, Col. R. Meinertzhagen, Captain C. R. S. Pitman, O. P. M. Prozesky, Professor H. Schouteden. Rev. Dr. W. Serle, R. H. N. Smithers, B. W. H. Stronach, M. A. Traylor, Dr. C. Vaurie, Col. Jack Vincent, Professor V. Van Straelen, John Williams, Dr. J. M. Winterbottom. THE GENUS The francolins are recognised as a branch of the sub-family Phasianinae (Partridges, Quails, Pheasants) which comprises over fifty genera, most of which are Palaearctic or Asiatic. My interest for this paper is primarily in the African members of the genus, but the five Asiatic species are included since they provide THE FRANCOLINS, A STUDY IN SPECIATION 109 some interesting points of comparison. I am not, however, competent to enter into a discussion on the limitations of the genus in Asia where there are thirty- eight other genera in the same sub-family some of which such as Rhizothera, Alectoris, Perdix, Arboricola seem closely related to Francolinus. They have been dis- tinguished on various structural characters such as the number of tail-feathers, the size and shape of the bill, or the length of tail, and I have accepted these distinctions uncritically. It is sufficient to say that the francolins are characterised by a rather longer, more hooked bill than members of the other genera (except Rhizothera), a short tail of fourteen feathers, and an upright stance: in the majority of species the male, at least, is spurred. The range of colours in the plumage is limited to the shades of ochre, chestnut, brown, black and white produced by melanic pigments, but the pattern of individual feathers is complex and varied. The partridge of Madagascar, Margaroperdix, has also some similarity to the francolins (particularly to the forest francolin F. lathami) in colour and pattern, but has various structural differences in bill, tail and legs, which make it doubtful if the two are very closely related. Apart from this in the Ethiopian region there are no gamebirds that can be regarded as similar to francolins in either structure or habits, the only others being the quails (Coturnix and Excalfactoria] , the Stone Partridge (Ptilopachus) a bird confined to the savanna and steppe belt north of the Equator the Congo Peacock (Afropavo), confined to the Congo forest, and the guinea-fowl family (Numididae). I would suggest therefore that the closer affinities of Francolinus with Palaearctic and Asiatic genera indicate that it originated in Asia, becoming separated from the other game-birds as a species particularly adapted to sub-tropical grasslands, and that later it spread into southern Europe and Africa. Once in Africa lack of efficient competition allowed it to spread widely and to exploit varied habitats which, in Asia, are occupied by other related genera. If this hypothesis of the origin of Francolinus is accepted it is apparent that there are factors present in Africa that have encouraged speciation in the com- paratively recent past (since the emergence of the genus). This is readily acceptable in view of the climatic changes known to have taken place (see Appendix i). Fluctuations in humidity and temperature would cause redistribution of the montane and lowland forests, and the woodlands, savannas and steppes, isolating and rejoining the populations of their indigenous birds. Among the francolins various generic divisions have been proposed, the most commonly accepted being Pternistis for the bare-throated francolins (as in Peters', 1934: 84). However, I agree with White (1952) that the bare throat alone is not a generic character and in other respects these birds are closer to some of the other large francolins than are other members of the genus. If a generic division was to be made I would separate the birds of the first five groups (including the Bare- Throated Group) which have relatively plain or vermiculated backs, from the birds of the last three groups with striated or quail-type patterning on the backs, placing F. gularis with the former and F. pondicerianus and F. lathami with the latter. However, I regard such a division as unnecessary and leading to difficulties in respect of the rather atypical species F, nahani. no B. P. HALL For a study on speciation francolins have the advantage of being remarkably sedentary birds for their size, not disposed to fly any great distances, and closely associated with a particular habitat. Morphologically the bold colours and patterns, particularly on the underparts, have the advantage of indicating where interbreeding has or has not taken place between neighbouring forms. Finally their " sporting " and edible qualities have given them an interest to many people other than ornithologists so that their ranges and habits are probably rather better known than those of any comparable group: at the same time these qualities have led to a persecution of the genus over the last fifty years which, combined with the damage done by intensive agriculture to the natural habitats, has almost exterminated species from some areas in which they were common and renders it difficult, if not impossible, to get further information about the habits of species from some areas which are now " developed ". THE GROUPS Grouping the francolins has presented little difficulty, for the close relationship of most members is evident from their appearance, ecology and distribution and has been recognised by previous authors. In a few cases where the relationship is less obvious this is discussed in the text. I have felt that it facilitates discussion to give descriptive names to the groups rather than numbers, and the names chosen indicate an important character common to all members though it has not been possible in all cases to find a character which is exclusive to the group. The order in which the groups are listed and discussed is not intended to be syste- matic except that the first five and the last three appear to form related assemblages. The groups and the species that comprise them are as follows: 1. Spotted Group (francolinus , pictus, pintadeanus) . 2. Bare-throated Group (afer, swainsonii, rufopictus, leucoscepus) . 3. Montane Group (erckelii, ochropectus, castaneicollis , jacksoni, nobilis, camerun- ensis, swierstrai). 4. Scaly Group (ahantensis, squamatus, griseostriatus) . 5. Vermiculated Group (bicalcaratus, icterorhynchus, clappertoni, hildebrandh natalensis, hartlaubi, harwoodi, adspersus, capensis). 6. Striated Group (sephaena, streptophorus) . 7. Red-winged Group (psilolaemus , shelleyi, africanus, levaillantoides, levaillan- tii,finschi). 8. Red-tailed Group (coqui, albogularis, schlegelii). The four species not assigned to any group are the two African forest francolins lathami and nahani, and the two Asiatic species pondicerianus and gularis. SPECIES AND SUBSPECIES In discussing speciation it must always be borne in mind that this is a continuous divergent progress starting when two populations become isolated from each other. In isolation each will develop divergent characters, either morphological or ecological, or both, but there will be a long period during which the two populations are THE FRANCOLINS, A STUDY IN SPECIATION in sufficiently alike for free interbreeding to take place if the isolating barrier is removed (Phase i). If they remain isolated and divergence continues this stage will gradually merge into one in which the divergent characters become such that interbreeding is unlikely (but not impossible) if the two populations rejoin (Phase 2). Again these characters may be ethological, involving different recognition patterns or voice ; or ecological, so that the two are unlikely to occur on the same ground ; or could possibly be a difference in breeding seasons. It is in this stage that the occasional hybrid may be produced in the wild, and in which hybridisation is possible and frequent under artificial conditions. Finally the two birds diverge so completely that they are incapable of interbreeding (Phase 3). In general most populations in Phase i and in the transition period between Phase i and 2 would be regarded as subspecies and those in Phase 2 and 3 as species, but there is no line between them and the decision must rest on the assessment of the taxonomist. This must be based on the degree of divergence and the degree to which they are known to interbreed. In the francolins members of different groups may be assumed to have reached, or almost reached, Phase 3 in their relationship, but the allopatric members of each group may be only in Phases i and 2. Each group provides a case in which I have had to make an arbitary decision whether or not to regard neighbouring forms as conspecific. For example, in the Spotted Group F. francolinus and F. pictus have diverged appreciably, especially in males, and hybrids are scarce, so I treat them as species. In the Bare-throated Group the afer block and the cranchii block have also diverged but interbreed freely, so are treated as conspecific. In the Montane Group F. ochropectus has diverged morphologically and ecologically from F. erckelii and is isolated, so is treated as a species, while atrifrons stands in a similar morphological relationship to castaneicollis but has the same field habits so is considered conspecific. In the Scaly Group the Niger provides a barrier between F. ahantensis and F. squamatus so it is not certain if they are capable of interbreeding but they seem to have diverged sufficiently to justify specific status. In the Vermiculated Group F. hildebrandti and F. natalensis have diverged, especially in the females, but interbreed to a limited extent at one point of contact but not apparently at another so are treated as species. In the Striated Group " rovuma " type birds (with a streaked abdomen) apparently do not interbreed with " sephaena " types (unstreaked) in the southern half of the range but hybridise in the north; the divergence is slight, however, and they are treated as conspecific. In the Red-winged Group the degree of divergence and the relationship between shelleyi and its three neighbours, africanus, whytei, and uluensis varies only slightly, but is just sufficient for me to treat africanus as a species and the other three as conspecific. In the Red-tailed Group the northern forms of coqui have diverged slightly, both morphologically and ecologically, from the southern but seem capable of interbreeding so must be considered conspecific. The foregoing demonstrates the impossibility of defining a species in such con- ditions. For this reason I have tried to avoid differentiating strongly between species and subspecies in the discussion, though greater clarity might have been achieved by setting out the discussion under specific headings (as has been done in the Red-winged Group where rearrangement of the species made it necessary). H2 B. P. HALL On the other hand I have endeavoured to make a clear distinction between those subspecies which I term " potential species " namely those which are either (a) isolated at present, or (6) must be assumed to have been isolated at some period, since they show a degree of divergence from their neighbours which is too great to be attributable to local ecological factors and those subspecies whose divergence can be attributed to ecological factors and which, as a rule, form part of a cline. The characters and ranges of " potential species " are denned equally with those of full species, but other subspecies are noted summarily in smaller print in a section in each group headed " Other variation ", together with any significant local or individual variation. On the maps as much information as possible has been shown on the ranges and relationships of species and subspecies, but this has necessarily been dependent on the complexity of each map. However, I do not consider it practical or desirable to recognise numerous steps in a cline, or slight local variations, and have relegated to the synonymy many names given to such intermediates and micro-populations. There is more justification in naming isolated populations, however small, such as some of the montane ones, and in these cases I have been guided by the degree of divergence shown. For instance, I regard names as desirable for such highly divergent small populations as the montane francolins found on the Plateau du Day in (French) Somaliland (ochropectus) and at Mega in southern Abyssinia (atrifrons), but not for others in the same group in other mountains of southern Abyssinia which differ only slightly from each other. To avoid en- cumbering the main discussion with irrelevant detail, notes on taxonomy and synonymy are confined to an appendix. In the parts of the paper dealing with evolution the ancestral stock of present species or of more than one species is referred to with the prefix " proto " (e.g. proto-coqui and pToto-albogularisjschlegelii) : if clarity demands, the name is followed by " subsp." or " sp.". The name of the author and date of publication are insert- ed only for those subspecies not listed by Peters (1934). METHODS The extensive collection of francolins in the British Museum has formed the basis for this study, supplemented by loans of critical specimens from other museums. All the African specimens examined and all the authentic records I have been able to trace in literature have been plotted over the Vegetation Map of Africa (Oxford 1959), as many species as possible being entered on a single sheet. This has enabled me to make a direct comparison of the range of any species with the vegetational belts and with the range of other species, whether in the same group or not. The Asiatic species have been plotted with accuracy only in the few critical areas, and for this the distribution maps of Indian birds prepared by Whistler, but never published, have been of great assistance. The maps of the groups have been compiled from these key maps. In a distribution map it is always difficult to know to what extent the range should be shown as continuous between actual collecting points, and in this respect the maps differ. For the Asiatic species I have shown the THE FRANCOLINS, A STUDY IN SPECIATION 113 limits of the range only: for the African species I have, as a general rule, shaded only areas from which specimens have been obtained, though this may give a rather distorted picture from parts of the country which are least known ornithologically : in a few cases, most notably for the Bare-throated Francolins (Map 2), I have felt that greater clarity in presentation could be achieved by shading the whole area within the limits of the range: this has been done only with species which, from their known habitat preferences, could be expected to have a reasonably continuous distribution within a vegetation belt. The morphological discussion is based on specimens examined personally unless otherwise stated. For the field notes I have used the standard regional works as a basis but these have been immeasurably enriched by the numerous conversations and letters that I have exchanged with the many field workers whose names appear in the acknowledgments. Where these workers have been responsible for specialised information I have tried to give the credit in the text, but they are responsible also for much of the general information which is incorporated without acknowledgment. I have found it difficult when discussing ecology and habitat to be consistent in the descriptive terms used, especially in correlating the parts of southern Africa with which I am familiar with those of the north for which I have to rely on the descriptions of others. As far as possible I have followed the terminology employed by the Vegetation Map of Africa (Map n). In this the vegetational types which mainly concern the francolins are the Woodlands, Savannas and Steppes (Types 16-25). My own experience in southern Africa and the works of such specialists in ecology as Benson and Stuart Irwin show the importance of the distinction between the woodlands in which acacia is rare or absent (the " brachystegia wood- lands " or " myombo " Types 18 and 19) and the acacia country, whether it is termed acacia " woodland ", " savanna " or " steppe ". This distinction applies not only to the birds of the actual woodlands but also to those of the grasslands and savannas (i.e. grasslands with scattered trees) with which the woods are inter- spersed, many birds being strictly confined to either the brachystegia belt or to acacia. (In the southern and eastern francolins there are, in fact, several species indigenous to acacia, such as F. sephaena, F. leucoscepus, F. rufopictus, F. adspersus and F. levaillantoides , but none which is wholly indigenous to the woodland belt, though the potential species F. s. shelleyi and F. s. whytei and the Bare-throated F. afer nearly qualify.) I have therefore made a clear distinction in the discussions between the brachystegia belt and acacia country and throughout the paper the term " woodland " is used exclusively for types 17-19 unless " acacia woodland " is specifically stated. North of the equatorial forest the woodlands are apparently similar in character to the brachystegia woodlands though composed largely of different species of trees and less continuous (Type 17). I am told also that, because of this discontinuity, the transition from the woodland belt to the acacia is less clearly defined. It is to be expected that there are therefore less clear-cut ecological distinctions in the birds, but this is not easy to determine from published works, the term " savanna " being widely used by different authors without clear definition. I have therefore been less precise in discussing the habitat of northern francolins. (From the maps it H 4 B - P- HALL appears that F. clappertoni and F. coqui spinetorum may be indigenous to acacia, and F. schlegelii, F. albogularis and -F. icterorhynchus to the woodland belt.) The distinction between acacia "woodlands", "savannas" and "steppes" is largely one of aridity and the transition between the types is necessarily gradual. It will be appreciated from the foregoing that savannas are found in both the woodland and acacia belts, but they are necessarily different in character. Where the term is used without qualification it should be plain from the context which belt is under discussion. In attempting to interpret in terms of evolution the pattern presented in each group by divergence, ecology and ranges, I have started in each case with the distributional maps. From these I have tried to find in each group the factor that limits the ranges. More often than not the answer lies in the vegetation map, and the range of a group, species or potential species will be found to coincide with the limits of a vegetation belt, this coincidence being particularly common where acacia savanna or steppe changes to woodland. The presence or absence of moun- tains, hills or rivers is obviously significant in other cases, but here it must be remembered that, while the reason for the present discontinuity in the range of a species may be self-evident, the range must have been continuous in the compara- tively recent past. In one case, the Red-winged Group, the limits of ranges of species coincide not so much with the vegetational belts as with the isohyets of total annual rainfall. In another, the Striated Group, the isolated areas from which one species (F. streptophorus) is found seem to have nothing in common except that they are the two ends of a ridge of higher country. In several cases where there seems no ecological reason for the limits of a species' range the bird is replaced by a member of another group. From facts such as these I have tried to assess in what con- ditions each group would be most likely to thrive at the expense of other groups, and, conversely, what conditions would be sufficiently unfavourable to divide it into isolated populations from which the species developed. In postulating in each group successive climatic changes to account for the isolating and rejoining of the various popualtions I have been guided solely by the pattern presented by the birds themselves, and not by any consideration of what climatic changes are known from other evidence to have occurred, except that I have not, I hope, postu- lated any changes which are, on this evidence, outside the bounds of possibility. However, in the final section of the paper I have tentatively tried to correlate some of the climatic eras postulated with those known, as listed in Appendix i. In the discussion on evolution two assumptions are frequently made which are not necessarily correct but which seem sufficiently probable to justify their use as a basis for hypothesis. The first is the premise that the degree of divergence shown by two isolated forms can be correlated with the length of isolation. This, of course, is not necessarily true, but it can, I think, be accepted that in similar circumstance there is a likelihood that speciation will proceed at approximately the same pace; also that a comparatively long interval must elapse for considerable divergence to take place in isolates. The corollary premise that isolates that have not diverged extensively have not been long separated is more open to doubt, and examples (such as the forest owl Phodilus prigoginei] can be quoted of populations that must THE FRANCOLINS, A STUDY IN SPECIATION 115 have been long isolated but show little divergence. But I regard these as exceptional. The second assumption is that the present species and potential species originated in some part or parts of their present range. This is a premise that is the more questionable in the case of species which are not members of superspecies. However, in the francolins these are exceptional, most of the species being part of a super- species in which all members are allopatric and which has a more or less continuous distribution over a large part of Africa. In these cases it is highly improbable that the origin of any member was in the territory of the present neighbouring forms, although all may have been outside the present range of the group (i.e. when vegeta- tion belts were pushed north or south). In such a contingency each proto-species may be presumed to have held approxi- mately the same position in regard to its neighbours as it does at present. CHARACTERS The study of the francolins in species groups has served to demonstrate that many characters which might be considered as guides to relationship show variable degrees of uniformity in each group. For example, in the Bare-throated Group the extent of bare skin on the face and throat is absolutely constant, and can be regarded as a " group character ", whereas in other groups closely related species may have a variable amount of bare skin. It seems useful therefore to discuss some of these characters separately at the start in order to obtain an understanding of their importance. A summary of some of these characters is set out in Table i. SIZE Little emphasis has been put on size or proportions in this paper, for though there is considerable individual variation there is reasonable uniformity in most related species. In the Red-winged Group the long-billed and short-billed populations of F. shelleyi provide the only example of marked proportional variation. On the whole general variation in size follows Bergmann's Law in relation to altitude and latitude, equatorial and lowland birds usually being smaller, but there are several inconsistencies. It is perhaps worth noting especially the incidence (for which I cannot account) of exceptionally small birds on the coasts of Kenya and southern Angola, and, to a lesser extent, Natal, since this feature is found in other genera besides francolins. PLUMAGE (a) Feather patterns. The complexity of the feather patterns has already been referred to. A number of examples are illustrated below which serve to define some of the descriptive terms I have used, and which have been arranged in sequences to indicate some of the transitional stages between one pattern and another. This is a subject which deserves fuller treatment than I can give it here and much of the discussion by Harrison in his paper " The incidence and origin of spotted patterns in the Estrildidae " (awaiting publication in the Ibis) is relevant to the francolins. It is sufficient here to emphasise that comparatively small changes in feather pattern, B. P. HALL 1 a s? O CD IT) VH H O |S rt ,O 3 o 00 P X< .H o ^.22 S I S u > CD On H * fe - II > V ~ > 8. s S en g P .S en CD S * 3 CD 4J w ^ o ^ e i^ T3 CD -(-> ft en .0 CD C > cn -S

*L c3 5r* C^ ?T* C 3 S to ""ii^ ^^ ni 5i 33 e fill -llll (- IH IH jsl^l 2 ! 2 ! t 2 ^ 1 f o *> !l! Of O _r Of 3 o - rt c* HN Q ^Q o 00 CH ' Of Of Qf H H nT M 2r bo Of "o" I g - o oi S- 2t -" T! 3 *" - - H OH o 5T H M H o o o Of Of Of HN H ^-C. M HI M ^ 1 "^ *O OT *O ^O ^O H S & , *o *o o M *o ^ sS- *o * N (M d III C o o *^3 "^ *C i t t rX S -S3 CO W3 2n CD CD * ^ rj CD ~CD~ c3 be to _O_ d rt | | | 1 r^> ^ rH CD CD S bo bo O o3 o3 ^< l-c T3 O O C S (3 tn .be rG rt rt 2 3^ 2 o . & ' 003 ."* "a j^ S M It en c/1 bD CD CD ^g >->>-. 05 O O O O C fl C C $ O O O a a a v> S Q - . tf) ^ g J J ^ J J ^ J U 3 3 ^ J to .2 <0 I francolinus pintadecmus pictus 1 1 : i s .^ s 1 f 1 S *> I > i | (/) *-s IS 1 1 1 -8 1 I O Q *4 ^* ^i ^ ^ "S 2 e S 6 O Q *rv K Q *l S to > ^ 2 5 5? f H 2 s s totnto to^to^tn 3 S 3 3 .3 3 .2 a -^ WWW ptop , 3 2 rQ ,0 1, n8 B. P. HALL particularly at the tip of the feathers, may greatly alter the appearance of a bird. (Compare the great differences in the general appearance of the breasts of a female schlegelii and a male coqui, and in the underparts of a male and female pintadeanus with the small differences in feathers 2 and i, and 5 and 3 of Figure i.) (b) Underparts. The greatest variety of colour and pattern throughout the genus is found on the underparts (and to a lesser degree on the mantle) and considerable variation is found between closely related birds. The most striking example is illustrated by Bowen (1930) and shows the difference between the subspecies of 8 10 11 \2 14 16 17 Fig. i. Feather patterns. i. Barred (c? coqui, breast). 2. Barred with inverted triangle at tip ($ schlegelii, breast). 3. Transition, barred to spotted ($ pintadeanus, breast). 4. Transition, barred to spotted (<$ shelleyi, breast). 5. Transition, barred to spotted (<$ pintadeanus, breast). 6. Spotted ( 138 B. P. HALL is a richer red-brown in the centres of the feathers both above and below, and has very few white streaks on the abdomen; there is some black in the lores and eye- stripe. The colour, but not the extent of the red-brown patterning in the mantle, is matched in a few individuals of schuetti from various parts of its range, but none approaches this colour below. The third species of the Scaly Group, F. griseostriatus, is confined to the strip of rich vegetation with vestigial patches of forest which is associated with the escarp- ment in western Angola. It is the most distinct member of the group having the leathers of the mantle and wing-coverts chestnut broadly edged with grey, and the rest of the upper parts faintly vermiculated, as in F. s. squamatus and F. ahantensis, but paler: below the feathers of the breast and flanks are chestnut edged greyish or creamy buff and the abdomen is plain creamy buff. While the coloration is brighter and more contrasting, the pattern of the feathers of griseostriatus is fundamentally the same as in the other two species. The bill is mainly blackish on the upper mandible with a bright red base, and below is orange-red : the legs are orange-red, and the three males examined have only a single spur. EVOLUTION The chief features of the distributional pattern of the Scaly Group are the two specific divisions which separate ahantensis and griseostriatus from squamatus; the vast areas in which squamatus exhibits no significant variation : and finally the indications of speciation in the eastern populations. The fact that the ranges of F. ahantensis, and F. s. squamatus coincide so closely with the limits of the main blocks of lowland forest suggest strongly that glades and clearings in this type of forest must be regarded as the natural habitat for the group, and that the populations found outside the perimeter of the forest are relicts from a period in which the forest extended to the extreme limits of the range of the Group. The occurrence of these populations in montane forest, up to at least 9,500 ft. on some mountains, and the rapid adaptation of the birds in Kenya to cultivations indicate, however, that their requirements are not rigid as long as sufficient cover is available. A wide spread of montane forest might, therefore, not be wholly disastrous for the Scaly Group, though they could not expect to compete entirely successfully with members of the Montane Group in these conditions. (It is perhaps significant that the highest altitudes at which Scaly Francolins have been found are on Mount Elgon, which has apparently no representative of the Montane Group.) It seems likely, therefore, that prolonged dry eras have been the chief factors in speciation. In some such periods (possibly coincidental) early in the history of the group proto-ahantensis could have been isolated from proto-squamatus by a belt of savanna stretching to the coast between the Niger and the Volta, and proto- griseostriastus isolated along the escarpment of Angola, where moister conditions prevailed, while the interior of the country was dry (Hall, 19606) . With a return of wetter conditions in West Africa proto-ahantensis spread eastwards. The lack of divergence between populations on either side of the present gap in the forest east of the Volta lends support to the conclusions cited by Moreau that the gap between THE FRANCOLINS, A STUDY IN SPECIATION 139 the two blocks of forest was formerly further east in the neighbourhood of the Niger delta. The Niger itself at present lies between the two species, and since there can be little temptation for sedentary birds such as francolins to fly across a broad river it can be regarded as a barrier sufficiently formidable to inhibit interbreeding. It is possible that the Lower Congo could have acted as a similar barrier between proto-squamatus and pitoto-griseostriatus if in a wet era the escarpment forests of Angola were reunited with the forest of Gabon and Cabinda. In F. squamatus the remarkable lack of variation in the whole species argues that the last spread of forest which linked the present outlying populations was of comparatively recent date. If degree of divergence can be accepted as a guide to time then it can be postulated that the earliest populations to be isolated by the retreat of the lowland forest would be those of Nyasaland and Tanganyika: but in this instance especially this precept must be posed very tentatively for the isolated populations concerned have also become adapted to rather different conditions at higher altitudes which might accelerate divergence. Since the most geographically remote of the remaining populations, those of the Jebel Marra (known from one specimen only) and southern Abyssinia, do not differ significantly from typical schuetti in north-eastern Angola and the southern Congo, the forest may have persisted longest in north-eastern Africa. Finally the distinction and the fairly sharp transition between squamatus and schuetti on the perimeter of the Congo forest merits attention. It suggests that at some time western and eastern birds were divided (presumably in a dry era) and on reuniting the eastern birds tended to favour rather more open habitat on the forest fringes, and the western birds clearings in the deep forest. A summary of the evolution postulated is as follows : Stage i. The group developing in lowland forest, and probably extending to East Africa. Stage 2. In a dry era proto-ahantensis in the Upper Guinea forest becomes cut off from proto-squamatus in the Lower Guinea forest somewhere in the region of the Niger. P-roto-griseostriatus in the escarpment forests of Angola also becomes cut off from proto-squamatus. Stage 3. In a wetter, more favourable, era, the group spreads again but no inter- breeding takes place between the incipient species, the barriers formed by the Niger and Lower Congo possibly reinforcing any tendency to genetic isolation that has developed while they have been separated. Stage 4. A dry era divides the Lower Guinea forest into a western block, in which proto-squamatus subsp. develops and an eastern block in which the birds become adapted to slightly less humid conditions. These eastern birds, proto-schuetti, keeping more to the forest fringes when the two are reunited. Stage 5. A wide spread of lowland forest extends the range of proto-ahantensis to the Gambia, and proto-squamatus sp. to embrace all the outlying localities at which it is found at present (Map 4A). Stage 6. With the retreat of the forest, populations become isolated first in Nyasa- land, proto-doni, and Tanganyika, proto-uzungwensis and proto-usambarae, then on the mountains of the Kenya/Tanganyika border, and finally in the Sudan, Uganda I 4 B. P. HALL and Abyssinia. These isolated populations survive in isolated patches of lowland and riverine forest, and also become adapted to montane forest up to high altitudes, especially where there is no competition from members of the Montane Group, and, in some areas, to cultivations. In West Africa populations of ahantensis become isolated in the Gambia and Portuguese Guinea and the main block of the species is divided by a break in the forest in the region of the Volta. THE VERMICULATED GROUP (MAP 5) (F. Ucalcaratus, F. icterorhynchus, F. clappertoni, F. hildebrandti, F. natalensis, F. hartlaubi, F. harwoodi, F. adspersus, F. capensis) RANGE AND CHARACTERS The Vermiculated Group is the most widespread of all the African groups. I regard it as consisting of a superspecies (containing six allopatric species) and three related species. The superspecies has a more or less continuous distribution from Senegal to Eritrea and southwards to Natal with outlying populations in Morocco and in the mountains of South West Africa and southern Angola. Of the three extraneous species one (F. harwoodi} is confined to a small area of Abyssinia almost encircled by the superspecies: the second (F. adspersus) replaces the super- species in northern Bechuanaland but is partly sympatric with the outlying population in South West Africa: the third (F. capensis] is completely isolated in the extreme south. The ecology of the group is variable and will be discussed in detail under the individual species but it can be said in general that from West Africa to Abyssinia and Uganda the members of the superspecies (F. bicalcaratus, icterorhynchus and clappertoni) are birds of grasslands and cultivations in the woodland, acacia savanna and steppe belts. However, in east and southern Africa this habitat is largely occupied by members of other groups, in particular F. afer of the Bare-throated Group (Chapin 1932: 714), and it appears that members of the Vermiculated Group have been driven into more rocky country on hillsides and into thickets in river valleys and the distribution is consequently more broken (compare Maps 2 and 5). In choosing the term " vermiculated " as the descriptive name for this group I have given emphasis to the most striking character which the majority of the component forms have in common although in some subspecies of F. clappertoni the vermiculations are largely or wholly replaced by extensive U-patterning. All members have brown or grey-brown heads, backs, wings and tail, with lighter vermiculations and/or V- and U-patterning, and sometimes some rufous mottling: the lores are black or blackish and most forms have a white eye-stripe : the under- parts are patterned with dark brown (and sometimes chestnut or maroon) on a whitish or cream background, except in the females of the two sexually dimorphic forms which are plain orange-ochre below: the colours of the bill and legs, and the colour and extent of bare facial skin, is variable and so is the number of spurs (see Table i). THE FRANCOLINS, A STUDY IN SPECIATION 141 SPECIES AND POTENTIAL SPECIES The superspecies consists of six clearly differentiated forms which are usually regarded as species, though it can be argued that two of these, F. hildebrandti and F. natalensis, could be treated as conspecific. The three northern members, F. bicalcaratus, F. icterorhynchus and F. clappertoni, form a homogeneous assemblage occupying grasslands and savannas of various types. F. bicalcaratus is the repre- sentative in West Africa, being common in the acacia and woodland belts from Senegal to the Cameroons, extending also into cleared parts of the Upper Guinea forest. There is also an isolated population in Morocco. Above, it is faintly vermiculated with extensive V- or U-patterning on the mantle and wing-coverts: below, each feather is cream with some chestnut on the edges and with a black drop- shaped centre that is broken by small cream " windows " (Fig. i, 15), giving a partly streaked and partly spotted effect: the bill and legs are greenish, and two fairly well-developed spurs are common in the males. The Morocco population (ayesha) is not conspicuously different, though isolated by 1,500 miles, but may be slightly more rufous and with rather smaller " windows " in the patterning below. Other variation in the species is clinal and will be discussed later. In the Cameroons, at about longitude I5E., F. bicalcaratus is replaced abruptly (but not apparently along any ecological boundary) by a plainer bird, F. icterorhynchus, which is more vermiculated and less V-patterned above, and which is irregularly V-patterned, blotched or spotted with dark brown below: the bill and legs differ from bicalcaratus in being mainly orange, and there is a small patch of yellow skin round the eye. Its range extends eastwards through the woodland belt (Map n, types 16 and 17) to Uganda, where a few specimens are found which have some chestnut streaks on the flanks (as in the type of "F. ugandensis) ." In the acacia belt to the north (Map n, type 20) F. clappertoni replaces F. bicalcaratus between northern Nigeria and Lake Chad, and replaces F. icterorhynchus in the Sudan. It is common also in the Nile and Blue Nile valleys and on grassy hillsides of western Abyssinia and Eritrea below 7,000 ft. (above which it is replaced by the montane francolin, F. erckelii). Western birds (F. c. clappertoni) are very similar to F. icterorhynchus above but a more orange brown, while below they resemble F. bicalcaratus except that there are no " windows " in the dark feather centres and the feathers are edged with maroon rather than chestnut. In the Nile valley and southern Sudan birds become darker and greyer, less vermiculated, but heavily U-patterned above and more heavily patterned below on a whiter ground. In the Blue Nile valley and eastwards birds lack vermiculations entirely but are clearly and regularly V-patterned above and on the breast, giving a lanceolate appearance, and they lack the moustachial streak which is present in the western birds. A single specimen collected at " Ngeem ", Lake Chad (possibly Nguigmi on the north-west edge of Lake Chad), the type of "F. tschadensis ", is very like some of the specimens of F. clappertoni from the Nile valley but the underparts are more buffy and lack any maroon markings, and it is rather less heavily patterned on the back. Neumann suggested (c/Bannerman, i : 327) that it is a hybrid between F. clappertoni and F. icterorhynchus. Since F. icterorhynchus does not now apparently occur as i 4 2 B. P. HALT. far north as this, and since this specimen is rather more heavily patterned than the northern populations of icterorhynchus and clappertoni it might perhaps be better regarded as an aberrant specimen that indicates the close relationship of the two species. East and south of Lake Victoria F. icterorhynchus is replaced by F. hildebrandti, which is a bird largely associated with rocky ground and thickets in river valleys and on hillsides, from sea level to about 8,000 ft., and which is found equally in the acacia savanna and the brachystegia woodland belts. It is common from central Kenya south to the Tabora and Dodoma districts of Tanganyika but in the southern part of Tanganyika and northern Mozambique its distribution is sporadic : however, it is again common in Nyasaland at varying altitudes, and in the upper Luangwa valley. The males are very like F. icterorhynchus above, but have more solid, blotchy, brown markings below: the bill is reddish with a brown culmen and yellowish base, and the legs reddish. The female is quite different having the underparts plain orange ochre. The sexual differences are most complete in the most southerly birds from southern Nyasaland, Mozambique and south-west Tanganyika (F. h. johnstoni] in which the female has no patterned plumage on either the upper mantle or the breast, whereas those from most of Kenya (F. h. altumi) have some patterned " male " plumage on both. Females from the centre of the range (F. h. hildebrandti) are intermediate, having the patterned feathers mostly confined to the upper mantle. F. hildebrandti reaches the southern limits of its range in the Luangwa and Shire valleys, and is replaced further south by F. natalensis, a bird with apparently the same ecological preferences. Birds from the Transvaal and Natal and the southern parts of Southern Rhodesia (F. n. natalensis} differ from hildebrandti in having the sexes alike, the backs darker brown and more heavily patterned and vermiculated and having regular double U-patterning below, unlike the brown blotches of the male hildebrandti. The bill and legs of natalensis are similar to those of hildebrandti except that there is no black on the culmen and the male has commonly one spur and the female none, whereas in hildebrandti the male has commonly two (once three) spurs and the female one or two vestigial spurs. However, the intensity and regularity of the patterning, particularly in the females, decreases northwards, showing some approach to hildebrandti. The populations of the northern districts of Southern Rhodesia, the Zambezi valley and the Luangwa valley (neavei) are plainer and more rufous above and the females tend to have the underparts washed with buff, the abdomen plainer and the U-patterning less defined giving a more spotted or streaked effect, but there is much individual variation. As far as is known the ranges of F. hildebrandti and F. natalensis only approach each other in two areas, in the Luangwa valley between latitudes 12 and I3S and in the neighbourhood of the Shire/Zambezi confluence. In the Luangwa valley there is some indication that the two may sometimes interbreed, or have interbred in the recent past, for some of the specimens of natalensis are more blotched, less U-patterned below than those of F. n. neavei, and some of the females are washed with buff below and have plainer abdomens (see Benson and White, 1957 ' 138) : one male from Jumbe (13 i6'S., 32 07'E.) lent by the National Museum of THE FRANCOLINS, STUDY IN SPECIATION 143 Southern Rhodesia is close to F. n. neavei below but has, like hildebrandti, a dusky ridge to the culmen and two spurs. Nevertheless, there are in this area specimens typical of the two forms taken from within 50 miles of each other so it is probable that there is only sporadic interbreeding. In the Zambezi/Shire area the situation seems rather different. F, natalensis has been collected from points along the Zambezi as far down as Tambara (60 miles below Tete) but not as low as the junction with the Shire, and not from the left bank below the junction of the Mushonganende River, 280 miles above Tete. Within 40-60 miles of Tambara F. hildebrandti has been collected in southern Nyasa- land and neighbouring Mozambique at Zobue, " on rocky hillsides " (Vincent 1934: 336), on Mlunganyama hill " at 500 ft. but not at 200 ft." (Benson, C. W. & F. M., 1948: 3), on Mwananbidzi Hill at 2,500 ft., and near Port Herald in the Shire valley (Long, 1960: 100), but not as far down the Shire as the confluence. There is no evidence in this area of interbreeding between F. natalensis neavei and F. hildebrandti and it has been suggested that here the Zambezi itself, which is as much as two miles wide when in flood, may be a barrier (Benson, et alia: in press). The last member of the superspecies, F. hartlaubi, is confined to the mountains of northern South West Africa and southern Angola and is strictly a bird of the rocks. It is the smallest member of the group and has distinctive rufous mottling above. The females are rufous below like those of F. hildebrandti but the male is finely streaked from chin to abdomen in brown and white. The bill is brownish above and yellow below and at the base: the legs are yellow with spurs poorly developed but often with two present vestigially in both sexes. The first of the three members of the group which are not included in the super- species is a francolin about which little is known, F. harwoodi. Only males have been collected and from only three localities, all in the gorges of the Upper Blue Nile or its tributaries. Of members of the superspecies it resembles F. natalensis most closely, being similar above but with more defined U-patterning on the nape, and below having similar double U-patterning, but rather darker and more clearly defined and extending to the chin, and being absent from the centre of the abdomen, which is cream. It is distinctive in lacking the white eye-stripe but having a circle of bare scarlet skin round the eye. The bill is recorded as red but from the appear- ance of skins probably has a black culmen; the legs are red with two fairly well- developed spurs. F. harwoodi appears to be cut off from other members of the group by the presence of the montane francolin F. erckelii on the plateau above the gorges, in the type of open country which, at lower altitudes, is occupied by F. clappertoni both to the north and the south. In southern Africa along the borders of Rhodesia and Bechuanaland the super- species (represented by F. natalensis) is replaced in similar habitat by F. adspersus, which ranges westwards through both acacia savanna and steppe (Map u, types 20 and 25) to South West Africa where it overlaps the range of the isolated member of the superspecies, F. hartlaubi. Here the two forms are found in the same localities but differ in their ecology and habits, adspersus living in coveys along watercourses and hartlaubi usually in pairs on the rocks. F. adspersus is a rather different-looking bird from those members of the group that have been discussed: its appearance is 144 B. P. HALL grey, the pattern being composed of minute vermiculations on the upper parts and with broader black-and-white vermiculations over all the underparts. It has no eyestripe but an area of bare yellow skin round the eye: the bill and legs are orange-red, the male having usually a single long spur. The last member of the Vermiculated Group, F. capensis, is isolated in the extreme south-west. Like many other members of the group it is found in thick cover in rocky river valleys chiefly in the coastal districts of the Cape. It is not known from Little Namaqualand (possibly because of the lack of suitable rivers) but has been recorded from the gorge of the lower Orange River at Assenkjer. It is the largest member of the group and most like F. natalensis and F. harwoodi in the predominance of brown and white double U- or double V-patterning. However, in capensis the patterning is distinctive in being composed of very fine irregular white lines on most of the back (the rest being vermiculated), and similar on the underparts but with distinct white shaft streaks, giving a streaky effect: it has no eye-stripe and no bare skin round the eye : the bill, as in hildebrandti, is red with a dark culmen, and the legs red, the male having one or two spurs and the female one. OTHER VARIATION In F. bicalcaratus variation is largely clinal, the palest birds being found in the drier savanna areas and the darkest and most heavily patterned in the cleared parts of the forest areas in southern Cameroons. Among the dry country birds those from Senegal, Gambia and the hinterland of Ghana (bicalcaratus) are paler with more rufous heads than those of northern Nigeria and northern Cameroons (adamauae) . The birds from within the borders of the Upper Guinea forest from Sierra Leone to Ashanti (thornei) are slightly whiter, less creamy below than those within the borders of the Lower Guinea forest from south-eastern Nigeria and southern Cameroons (ogilviegranti) which become increasingly dark eastwards. Birds from Accra, between the two forest blocks, are closer to bicalcaratus than to thornei: those of most of southern Nigeria are intermediate between adamauae and ogilviegranti. The isolated ayesha in Morocco (already discussed) is closest to the paler birds above but has heavier patterning below. In F. icterorhynchus variation is clinal and ecological, pale, lightly patterned birds being found in the southern Sudan and in the region of the Albert Nile in northern Uganda (icterorhyn- chus) : dark, heavily patterned birds are found in the more humid areas from the Oubangi and Uelle to southern Uganda (dybowskii). In F. clappertoni the clinal variation is greater than in the two previous species, and features some changes in pattern that warranted discussion in the main section. West of the Nile and in the Nile Valley the paler, less patterned clappertoni merges through the intermediate heuglini to the dark U-patterned gedgii of the southern Sudan. Among the more V-patterned, lanceolated populations of the east, those of Eritrea, and northern Abyssinia (sharpii) are pale above (though not as sandy as clappertoni), those of south-western Abyssinia (nigrosquamatus) are darker and more patterned, and those of the Blue Nile valley (konigseggi) are somewhat intermediate, closest to sharpii but with a greater suffusion of black markings below. In the southern populations of F. natalensis there is some variation according to the habitat, darker, more patterned birds being found in the richer vegetation of Natal, Zululand, and the Zoutspansberg, and paler birds in the drier savanna country of the Transvaal (see Clancey, J 953 : 59)- (Since the range of the darker birds is discontinuous it is impracticable to distinguish the two variations by name.) The available series of F. hartlaubi is inadequate for detailed study, but it would seem that the populations of southern Angola (hartlaubi) are rather smaller than those of South West THE FRANCOLINS, A STUDY IN SPECIATION 145 Africa, and that the western birds from the Kaokoveld and Erongo (crypticus Stresemann, 1939) are rather paler than those from the Waterburg and Otavi further east (bradfieldi) . EVOLUTION It has already been suggested that the rather different ecology of the northern members of the Vermiculated Group compared with those from Kenya southwards may be due to competition with the Bare-throated Group, and this possibility has considerable bearing on the interpretation of the pattern presented by the ranges of the various species. I believe, for example, that while the present can be regarded as generally favourable for the Vermiculated Group as a whole, it is even more favourable for the Bare-throated Group and that the distribution of the southern Vermiculated species is heavily restricted because of this, especially in the areas where alternative habitats to the savanna grasslands are limited. This would account for the absence of any member of the group from the grass- lands south of the Congo forest, also for the sporadic distribution in eastern Africa, and possibly also for the limited ranges of F. natalensis and F. capensis in the coastal districts of the south-east, where the gap between them is occupied by the bare-throated F. afer. In the north-east it is apparent from the isolated population in Morocco that the group is not at present as widespread as it has been at some time in the not distant past (to judge by the lack of divergence in the Morocco population). Elsewhere in the north the range appears to be expanding, for there is good reason to suppose that F. clappertoni has only recently extended its range in the acacia savanna south of the Sahara, where it has supplanted F. coqui (see discussion under Red-tailed Group and Maps 5 & 8), and also has recently extended its range eastwards towards the Red Sea coast, cutting off the Erkowit population of the montane francolin (F. e.pentoni] from the main body of the species (compare maps 3 and 5). From the foregoing it can be assumed that grasslands and scrub in savanna and woodland are the natural habitat for the group, but it is adaptable : the presence of several members in areas where the rainfall in under 10 inches a year shows especially that it can be tolerant of dry conditions if there is adequate cover in watercourses. It is likely, therefore, that humid eras in which there has been a wide spread of forest have been less favourable to the group than the dry eras, and that it is chiefly in these wet eras that the group has been split and speciation has developed in the isolated populations. The morphological differences between F. adspersus and other members of the group, combined with the fact that it has diverged sufficiently to be able to live alongside another member, F. hartlaubi, in complete ecological segregation, suggest that probably the group developed in the east and south-east and that pToto-adspersus was isolated earliest from the original stock. This could conceivably have taken place at the time when a spread of montane forest linked the Natal forests with those of Southern Rhodesia and Nyasaland, cutting off proto-adspersus in the Kalahari/South West African region. Pioto-capensis and proto-harwoodi can be presumed also to have been isolated at an earlier period than any of the six members of the superspecies, but several factors could have caused their segregation. Proto-capensis, as suggested, could have been cut off by the ] 4 6 B. P. HALL intrusion of the bare-throated F. afer. F. harwoodi on the other hand appears to be a " relict " population that has survived in the Blue Nile Gorges at a period when the group was overrun on the plateau above. The fact that the high parts of the plateau immediately surrounding the range of harwoodi are at present occupied by F. erckelii of the montane group suggests that it was first split from other members of the group in an era when montane conditions prevailed all over the plateau, as they do now. Proto-hartlaubi in South West Africa may also have been cut off at an early date from the remainder of the superspecies, in which case it must be regarded as a coincidence that it has diverged along the same very distinctive lines (with marked sexual dimorphism) as the Tanganyika population, hildebrandti. It seems more likely that, following the periods in which adspersus, capensis and harwoodi were isolated, there was a period which favoured the superspecies, during which it became widespread through the grasslands of the tropics and south to Natal and the Transvaal. Meanwhile, proto-adspersus had become adapted to rather drier con- ditions in acacia steppe to the south-west. A subsequent drier era in southern Africa which spread acacia steppe through the Limpopo and Zambezi valleys could isolate proto-natalensis from the rest of the superspecies at a time when proto- hartlaubi and proto-hildebrandti were still linked. A further spread of steppe northwards bringing competition with proto-adspersus or a spread of the Bare- throated Group might then eliminate the superspecies from the areas between the Rhodesias and South West Africa, where there is little alternative habitat. The mountains of the west, however, would provide sufficient ecological variety for proto-hartlaubi to find a different habitat and co-exist in South West Africa with proto-adspersus, and in southern Angola with the bare-throated afer. The tendency to approach hildebrandti shown by the Zambezi population of natalensis (neavei) suggests that these birds are either closest to the ancestral stock of natalensis I hildebrandti or else are the result of a period in which the two had become reunited after their distinctive characteristics had developed, but before there was any bar to free interbreeding. However, the lack at present of an inter- mediate population showing all stages of intergradation between hildebrandti and natalensis suggests that the ranges of the two were again separated in the not too dis- tant past : the presence in the Luangwa valley of a few apparent hybrids among other- wise typical birds indicates that there has been only sporadic interbreeding now that the ranges again join. Minor fluctuations in climate would serve to bring about this splitting and reuniting. The differences between the various populations of hildebrandti are not great enough to suggest any lengthy periods of isolation, but, combined with the present discontinuous distribution, indicate a limited amount of splitting and reuniting since the species developed. The greater amount of "male" plumage exhibited by the females of hildebrandti in western Kenya (altumi) indicates some approach to F. icterorhynchus of Uganda, suggesting that the break between them was at a comparatively recent date. The three northern members of the superspecies, icterorhynchus, clappertoni and bicalcaratus form a homogeneous and completely allopatric assemblage from which THE FRANCOLINS, A STUDY IN SPECIATION 147 it can be postulated that the present extensive range has not been occupied for a long period, since otherwise climatic fluctuations would have caused complex splitting and rejoining of populations over such a wide area. To interpret the main lines of the pattern it is only necessary to visualise one period in which the ancestral stock of the three members spread westwards from Abyssinia to Senegal, followed by one unfavourable period in which it was divided into a western (proto-bicalcara- tus), central (pToto-icterorhynchus) , and eastern (pioto-clappertoni] population. This period seems likely to have been a humid one in which Lake Chad was extensive and forest extended from its southern tip southwards, forming a north-south barrier between the western and central parts of the savanna belt. At the same time it can be postulated that there would have been sufficient riverine forest along the Nile to form a second barrier further east, and that forest in Kenya and Uganda cut off proto-icterorhynchus from proto-hildebrandti in the south. Speciation developed in the three pockets and was almost complete when dry conditions returned. Proto- clappertoni became adapted to rather drier conditions than the other two species and spread eastwards over the Nile north of proto-icterorhynchus. The differences in pattern between the eastern and western populations of clappertoni suggest that at some subsequent period the Nile and its tributaries again formed a barrier between east and west, allowing birds with the distinctive V-patterning of F. c, sharpii to develop in the north-east. At present it appears that clappertoni is extending its range westwards, north of the ranges of icterorhynchus and Ucalcaratus, at the expense of the red-tailed coqui. In the west it is apparent that there must have been a period when sufficiently humid conditions prevailed along the coast to allow the francolins to spread to Morocco. The lack of marked divergence in the Morocco population (ayesha) suggests that this period was of a more recent date than the humid period postulated for the isolation of the ancestral stocks of the three northern species. A summary of the evolution postulated is as follows: Stage i. Ancestral stock of the group develops in the east and south-east. Stage 2. Proto-adspersus isolated in the south, possibly by a spread of montane forest. Stage 3. The rest of the group extends northwards to Abyssinia, and widely through grasslands south of the Congo. Pioto-capensis isolated, perhaps by competition with the Bare-throated Group. Stage 4. The group spreads westwards to Senegal but withdraws from the high plateau of Abyssinia, which is occupied by the Montane Group, leaving a relict population in the Blue Nile gorges, proto-harwoodi. Stage 5. (South). Pioto-natalensis isolated, possibly by extension of acacia steppe in Limpopo or Zambezi valleys bringing competition with proto-adspersus. Stage 5. (North). Proto-bicalcaratus in the west, proto-icterorhynchus in the centre and pioto-clappertoni in the east, isolated from each other, probably by a belt of forest extending northwards to Lake Chad (at a time when its area was more exten- sive), and by riverine forest in the Nile valley. Pioto-clappertoni becoming adapted 148 B. P. HALL to drier conditions than the other two species. Forest in Uganda and Kenya separates proto-hildebrandti from proto-icterorhynchus. Stage 6. (South). Proto-hartlaubi isolated in the mountains of South West Africa and Angola, possibly by an invasion of the central areas by proto-adspersus or by the Bare-throated Group. Stage 6. (North). Minor fluctuations, allowing pToto-clappertoni first to spread westwards across the Nile and then splitting into a western and eastern population, possibly during a return of more humid conditions when the Nile valley again became a barrier. Stage 7. (South). Minor fluctuations causing rejoining and re-splitting of proto- natalensis and proto-hildebrandti, and isolating populations of hildebrandti for short periods. Stage 7. (North). A humid period in which proto-bicalcaratus spreads to Morocco, followed by a drier period in which the Morocco population is isolated. Stage 8. (South). The ranges of natalensis and hildebrandti rejoin, at least in the Luangwa valley, and there is sporadic interbreeding. Stage 8. (North). In the Sudan clappertoni spreads westwards through the drier acacia belt. THE STRIATED GROUP (MAP 6) (F. sephaena, F. streptophoms) RANGE AND CHARACTERS The Striated Group is unlike all others in consisting of only two species which are partly sympatric. The range of the group as a whole is mainly in eastern Africa, from Somalia to Natal, but extending across to South West Africa and southern Angola, and with an isolated population in the Cameroons. Both species are mainly brown above with some chestnut, black and white pattern in the collar, and white shaft streaks on the mantle: the eye-stripes are white, edged with black and there is some chestnut on the sides of the face: the throat is white bordered with chestnut: the underparts are creamy buff with contrasting maroon, chestnut, or black-and-white patterning on the breast (different in the two species) and lighter patterning on the flanks and abdomen. The habitat of the two species is different and will be discussed later. SPECIES AND POTENTIAL SPECIES The two species are quite distinct both ecologically and morphologically. F. streptophoms has a puzzling distribution, being found commonly in the grasslands and " orchard bush " of northern Uganda and sporadically in north-western Kenya usually on stony scrub-covered hillsides (Jackson, 1938: 241). In north-western Tanganyika Dillingham has found it also in clearings of brachystegia. These are the only areas from which it is known in the east but 1,500 miles to the west a single specimen was obtained by Bates among grass and rocks between 3,500 and 4,000 ft., THE FRANCOLINS, A STUDY IN SPECIATION 149 in the Cameroon highlands 40 miles west of Fumban (on the borders of former British and French Cameroons). It is a darker bird than F. sephaena the back being grey-brown with dark brown patches and with fewer and narrower white shaft-streaks. It has a solid chestnut collar on the hind neck above a broken black-and-white collar of barred feathers : there is a considerable amount of chestnut in the facial pattern but only a narrow band bordering the base of the throat: the feathers of the breast are barred black and white occasionally with a chestnut tip: the remainder of the underparts is pale cream marked with some dusky vermiculations and broad blackish streaks on the flanks : the bill is black with yellow at the base and the legs yellowish with only a vestigial spur in the males. The sexes are alike. The Cameroon bird is like those of East Africa but smaller (wing i<$ 145 against 6^ 151-160). In northern Uganda, where both species are found, Pitman tells me that sephaena is strictly confined to the driest areas of scrubby thorn country while streptophorus is in savanna grasslands. Both Jackson (1938: 241) and Dillingham (in litt.} stress that the call of streptophorus is quite unlike that of any other francolin known to them, and is a melodious trill or whistle. Dillingham on the other hand found sephaena' s call closer to other francolins though Vincent, on a label of a bird from Swaziland, describes it as "a loud piping trill " and Benson tells me he has found it very variable, Rhodesian birds being hardly recognisable as the same species as those in Abyssinia. Comparison between the eggs would also be useful for Pitman confirms Jackson's statement that the eggs of sephaena are exceptionally hard, and it would be interesting to know if those of streptophorus, at present un- known, have the same character. F. sephaena is widespread in the acacia savanna and steppe from eastern Abyssinia and (British) Somaliland westwards to the Albert Nile, southwards through Kenya and along the Tankanyika coast to Mozambique and Natal, westwards to South West Africa and southern Angola. It is never far from water and in the driest parts of the range is confined to watercourses. Above it is a rich red-brown with broad white shaft-streaks : the chestnut collar is broken with white and the patterned feathers of the upper mantle are rich dark brown or blackish with elliptical white centres : the females are slightly vermiculated. There is no black on the underparts, the feathers of the breast being cream with triangular chestnut or maroon markings on the tips : the bill is wholly black and the legs carmine, the males having a single long sharp spur. The birds here designated as F. sephaena have sometimes been considered to belong to two species, for two distinct types are found, one, mainly coastal, which has fine, drop-shaped maroon streaks on the abdomen, and the other in which the abdomen may have some dusky patterning but which has no maroon streaks. In the Shire valley of southern Nyasaland birds of the two types appear to behave as species being found close to each other without an ecological division, but without any intergrading : specimens from the southern tip of Lake Nyasa and eastwards are streaked, and so are those from the lower Shire valley at Chiromo and Tangadzi, but a specimen from Lengwe, 35 miles upstream, is unstreaked, as are those from all localities westwards in the Zambezi valley (Benson, C. W. and F. M., 1948: 3). 1.50 B. P. HALL In northern Tanganyika a similar clear division is found between the two types, birds from inland, at altitudes above 1,500 ft., being unstreaked, those from the coastal plain being streaked. However, from the Kenya boundary northwards there is no clear division either geographically or morphologically: in western Kenya, southern Abyssinia (Benson, 1945: 391) and throughout Somalia both types are found, as well as some intermediates with sparse, fine streaks. In (British) Somali- land streaked birds predominate at lower altitudes near the coast but both types are found in the highlands and streaked specimens have been found as far inland as Harar in north-eastern Abyssinia. There is no evidence that the two types behave here as species and it is more probable that they represent a hybrid population. (For this population it is convenient to use the name spilogaster, which was given to a streaked bird from Harar see Rand, 1950 : 384 while rovuma should be confined to the unmixed streaked populations from Tanganyika southwards.) It will be seen from the map that inland birds of the unstreaked type are found in two blocks from northern Tanganyika northwards and from southern Nyasaland southwards, the species being absent from inland areas through most of Tanganyika, Nyasaland and Northern Rhodesia. Birds of the northern block (grantii) have the patterning on the collar and on the breast more restricted (especially in the females) than the southern birds. In birds of the southern block the patterning on the collar is brighter and more extensive and the patterning below extends to the lower breast: birds from the Transvaal, Natal and extreme southern Mozambique (Coguno) have the rest of the underparts fairly heavily vermiculated (sephaend), but northwards through Southern Rhodesia the vermiculations become lighter and birds from South West Africa and southern Angola across the continent to southern Nyasaland are appreciably less patterned (zambesiae) . OTHER VARIATION Throughout the range of F. sephaena there is considerable local variation in size with equatorial birds on the whole being smaller. Exceptionally small birds are found on the Kenya coast and exceptionally large ones in the Upper Zambezi and Chobe valleys. Among birds without the chestnut streaks on the abdomen there is considerable variation in the amount of vestigial patterning on the abdomen, some having ill-defined whitish streaks and irregular dusky U-patterning. On the whole eastern birds in, and bordering, the hybrid zone tend to be more patterned than western. In South West Africa some rather greyer birds are found, probably associated with the limestone pans, but the variation is not constant. EVOLUTION The peculiar distribution of F. streptophorus is the major problem set by the Striated Group. The fact that the eastern and western populations of streptophorus are not strongly differentiated and yet the country in between appears totally unsuitable for the species indicates that these are relicts from an era when the group had a wide distribution north of the Congo forest. Being relicts they may not now be in their natural habitat and it can, therefore, be reasonably postulated from the distribution of sephaena, that the natural habitat of the Group is in acacia steppe, and that proto-streptophorus once occupied this belt from the Sudan westwards, THE FRANCOLINS, A STUDY IN SPECIATION 151 north of the forest and woodland belts. One thinks usually of these belts as moving north and south with successive climatic fluctuations, but it is possible to visualise that the ridge of higher ground that stretches at about latitude 6N from the Cameroons eastwards to Uganda may sometimes dislocate regular north-south movement. It seems possible that, following a dry era in which the belt of acacia steppe lay as far south as this ridge, the return of more humid conditions and richer vegetation might come first to the lower country of the Shari basin and Bahr-el- Ghazal to the north of the ridge, leaving temporarily a pocket of dry country on the ridge while the main belts moved north. If proto-streptophorus became isolated in this pocket it could be expected that, as conditions became more humid and less and less suitable, it would be driven to the only parts where a tolerably dry habitat could still be found, and one which would be unattractive to the indigenous francolins of the encroaching woodlands. Such refuges were provided in the higher ground at the western and eastern extremes of the previous range. It is possible that streptophoms may yet be found in the country between the Cameroons and Uganda, but if so could be expected only in some patches of sparse vegetation. Since so little divergence is shown by the Cameroon bird it can be postulated that it has been separated from the Uganda population at a comparatively recent date. The pattern presented by sephaena points to at least one unfavourable era in which the species was divided for long enough to allow the streaked and unstreaked types to develop in isolation, almost to the specific level. The present distribution shows how small an extension of woodland towards the coast in Tanganyika would divide the species. In a considerably wetter era with a wide spread of woodland or forest it is likely that the northern populations would be confined to the coastal plain of Somalia while the southern populations might develop on the fringes of the Kalahari. If this was followed by a dry era with a spread of acacia the unstreaked southern populations might spread northwards up the Luangwa valley, through Tanganyika and Kenya along the valleys of the rivers which run westwards to the lakes, and then up the Abyssinian Rift to the Somali highlands. Meanwhile the streaked northern populations spread southwards, chiefly in the valleys of the rivers flowing to the sea. It is difficult to account for the fact that the two types seem to behave as species in the south but mix and interbreed to a limited extent in the north. It may be that the two came together at a later date in the south when speciation was further advanced, and it is also possible that in northern Tanganyika the sharper change of altitude on the edge of the coastal plain may provide here a barrier that is lacking in Kenya and eastern Somalia (though not in the mountains of (British) Somaliland). The present gap in distribution between the unstreaked birds of Kenya and northern Tanganyika (grantii) and those of the Rhodesias and southern Nyasaland (zambesiae) is sufficient to account for the divergence between them. The scattered distribution of sephaena along watercourses in the drier areas would account for much of the local variation in both size and patterning that is exhibited, for such populations would be isolated and might be expected to diverge in periods when there is little suitable vegetation between one river valley and another. 152 B. P. HALL A summary of the evolution postulated is as follows: Stage i. Ancestral stock occupying the acacia belt north, east and south of the central forest and woodland block. Stage 2. The stock split into two, one, proto-streptophorus , in the acacia belt north of the central forest, the other, proto-sephaena, in acacia in the eastern districts from Somalia to southern Africa. This split probably taking place in a humid era with a spread of woodland or forest between Uganda and Abyssinia. Stage 3. F. sephaena split into a northern group, probably confined to lowland in Somalia, and a southern group, probably on the borders of the Kalahari. The northern group being the ancestral stock of the birds with streaked abdomens, and the southern group of those with unstreaked abdomens. This split taking place in an even wetter era than Stage 2 with a wide spread of forest and woodland over most of central Africa. Proto-streptophorus still isolated south of the Sahara. Stage 4. (East). A return of arid conditions with increased acacia allows the northern streaked birds to spread down the coast, and the southern, unstreaked birds to spread north up the Luangwa valley, through western Tanganyika and Kenya and the Abyssinian Rift. The two types meet first in the north, and in Somalia they mix and interbreed to a limited extent forming a hybrid zone but with streaked birds always predominant near the coast. Subsequently they meet in northern Tanganyika and later still in southern Nyasaland, but do not interbreed. Stage 4. (North). In a dry era the acacia belt and proto-streptophorus move south to the ridge of high ground from the Cameroons to Uganda and north-western Kenya (Map 6A). Stage 5. (East). An increase in humidity with a spread of woodland in central Africa cuts off the northern unstreaked birds, proto-grantn subsp., from the southern unstreaked birds, proto-zambesiaej sephaena subsp. Stage 5. (North). An increase in humidity, coming first to the low ground of the Shari basin and Bahr-el-Ghazal, north of the ridge, isolates proto-streptophorus along the ridge. Increased humidity and competition with indigenous woodland species drive it into refuges at the extremes of its range where more arid conditions persist. THE RED-WINGED GROUP (MAP 7) (F. psilolaemus, F. shelleyi, F. africanus, 1 F. levaillantoides , F. levaillantii, F.finschi] RANGE AND CHARACTERS This is a very homogeneous group which ranges over most of eastern and southern Africa from Eritrea to the Cape and westwards to Angola. The various species have a complex distribution but are largely allopatric though the ranges of three overlap in the Transvaal. Members of the group are found in a wide variety of habitat and at any altitude, and even individual species are found in varying habitats in different parts of their range. 1 F. afer of Praed & Grant see appendix. THE FRANCOLINS, A STUDY IN SPECIATION 153 All birds of the group have " quail-type " plumage on the backs, with a lattice- work pattern of irregular white or buff bars and shaft streaks on a basic colour of mixed black, brown, grey and chestnut: all have the under wing-coverts and part of the wing chestnut (including the South African form F. africanus, commonly known as the Greywing, in which the chestnut is reduced) : all have a marked facial pattern but this, and the pattern on the underparts, varies in the different forms: all have black bills, usually with a yellowish base, and yellowish legs, the males usually with a single spur. The sexes are alike. SPECIES AND POTENTIAL SPECIES The distribution and relationships of the various forms of this group would be easier to discuss if no previous attempt had been made to designate species or subspecies in the East African forms. I cannot agree with any of the classifications made for this group which seem to place too great a reliance on minor morphological characters, particularly the size of the bill, without regard to the ecology of the forms of their likely evolution. The rearrangement I propose I believe to be more realistic in these respects, though I appreciate that conventional nomenclature fails to give more than an inadequate picture, and it is necessary as well to discuss the relationships in some detail. To facilitate the recognition of the species as re- arranged they will be discussed under separate headings and I have proposed new common names for each. F. psilolaemus The Montane Redwing. F. psilolaemus comprises the montane populations of Kenya, Uganda and Abyssinia known by Praed & Grant as F. shelleyi theresae, F. s. elgonensis, F. afer psilolaemus and F. a. ellenbecki. All these birds are found in heath and grasslands above 8,000 ft. All have a high proportion of rich chestnut in the wings and differ from other red-winged francolins in having some barring on the tips of the primaries: the underparts are rich buff mottled with chestnut and with some dark brown or black markings: the facial pattern is broken and ill denned. Birds from the Shoa mountains round Addis Ababa (psilolaemus) are small with small bills (wing ^164-174, bill ^30-34), and have blackish spots on the throat, indistinct black spotting on the upper breast and rather sparse chestnut and blackish patterning on the abdomen. Birds from the Arussi plateau south of the Abyssinian Rift (ellenbecki) are slightly larger than Psilolaemus, darker on the head and back, the lighter markings being greyer: below the spots on the throat and breast are darker and the patterning is stronger and richer in colour. Birds from Mount Kenya and the Aberdares (theresae) are large (wing (174-191, bill ^35-38) : they have only occasional speckles on the throat but are heavily spotted on the upper breast : they are darker above than Abyssinian birds and the chestnut markings on the abdomen are richer. Birds from Mount Elgon (elgonensis) are similar to theresae but darker and richer. F. shelleyi Shelley's Redwing. F. shelleyi comprises the forms known by previous authors as F. s. shelleyi, ZOOL. 10, 2 II 154 B - P. HALL F. s. whytei, F. afer uluensis, and F. a. macarthuri Van Someren, 1938, but excludes theresae and elgonensis which were previously ascribed to shelleyi (see above). All the birds included in this assemblage differ from F. psilolaemus in having the breast maroon, chestnut and grey, unmarked with black below the necklace and gorget : the remainder of the underparts are patterned all over with black and white, the pattern basically of white spots on a black ground, but the spots frequently merging to form irregular bars. In Kenya and northern Tanganyika birds of this type with rather small bills, white throats, well-defined facial pattern and necklace and rather small patterning on the abdomen are found on hillsides, grassy plateaux, and clearings in montane forest from Mount Kenya southwards to the Crater Highlands and Mount Meru. On Mount Kenya and the Aberdares they are found at lower levels than F. psilolaemus keeping below the montane forest. Since they are confined to uplands above 3,000 ft., their distribution is necessarily discontinuous and some divergence is shown by isolated populations, birds of the Loita Plains in south-western Kenya being rather greyer and those of the Chyulu Hills (macarthuri) rather darker than birds from central Kenya (uluensis). The most southern representative recorded of this small-billed form is from Mbulu at 4,900 feet in the highlands south of Lake Manyara in northern Tanganyika. Fifty miles south-east at Salanga on the Bereku Ridge at about 5,500 feet a large- billed bird has been collected which has larger, more open, patterning on the abdomen: between the two localities there is lower, rather open country probably unsuitable to these francolins. Birds of the large-billed type are found on the Uganda/Tanganyika borders and at scattered localities through western and southern Tanganyika, Mozambique, southern Nyasaland, southern Northern Rhodesia, Southern Rhodesia the Transvaal and Natal. The small-billed, small- patterned form of Kenya (uluensis) and the large-billed, large patterned form of Tanganyika (shelleyi) have been regarded as belonging to different species and the proximity of the two in the Mbulu area lends weight to this view. Nevertheless, they seem to have similar ecological requirements and field habits; and, while in the Mbulu area the topography of the country may inhibit interbreeding, there is a bird from Amani, 200 miles to the west, which is intermediate between the two forms in the patterning on the abdomen, though closer to shelleyi in bill size, which suggests there may be no intrinsic barrier. The long-billed shelleyi is found in varying habitats throughout its range. In southern Nyasaland, as in northern Tanganyika, it may be found on montane grass- land up to 7,000 feet, but it is also found in Nyasaland and Northern Rhodesia in grass under brachystegia or even mopane at low altitudes, though it is nowhere common. In Southern Rhodesia it is common in Mashonaland and the Midlands, in open grass of the plateau, but scarcer westwards in Matabeleland (Smithers et al, 1957: 46). It occurs at least as far south as Pretoria and Legogot, in the Barberton district of the Transvaal, in presumably the same type of country as in Southern Rhodesia, but in Natal is found in acacia country below 2,000 ft. along the foothills of the Drakensburg. In northern Nyasaland, the south-eastern Congo and northern Northern Rhodesia the species (as proposed) is represented by another long-billed form (like shelleyi), but THE FRANCOLINS, A STUDY IN SPECIATION 155 with a small patterning (like uluensis) and which differs from both in having a buff throat with the facial pattern and necklace freckled and indistinct. This form (whytei], like shelleyi, is found at varying altitudes and in both brachystegia and montane grassland and, though there is no ecological barrier between them and they have been collected from within 40 miles of each other at Lundazi and Mzimba, no intermediates are known (Benson, 1951: 79). Benson tells me that he believes this is more likely to be due to lack of collecting and to the scarcity of both forms, rather than to the existence of any barrier to interbreeding. The situation may be found to be comparable to that between uluensis and shelleyi in northern Tanganyika but further research is required. F. africanus (F. afer) the Greywing. I regard the South African Greywing as a monotypic species, and do not associate any of the northern short-billed forms with it. It differs from F. shelleyi in having a white throat flecked with black, the facial pattern obscured by black-and-white freckling, the patterning on the underparts very small, and the rufous in the wing very much reduced. The bill is small as in uluensis. It is found in the grasslands of the mountains of the southern Transvaal, Orange Free State and Cape Province, reaching Natal on the high spurs of the Drakensburg. McLachland & Liversidge (1957 : 94) show an overlap in the ranges of F. shelleyi and F. africanus in the southern Transvaal but I cannot substantiate this, the most northerly specimens o'f africanus recorded being from Potchefstrom and Wakkerstrom, 100 miles south of shelleyi at Pretoria, and in rather different country. In Natal Vincent tells me there is a clear altitudinal distinction between them, africanus being found above 6,000 ft. and shelleyi below 2,000 ft., while the inter- mediate levels are occupied by a third member of the group, F. levaillantii. Further- more F. shelleyi has an earlier breeding season than the other two species, which would contribute to segregation. It will be appreciated therefore that africanus has, in many respects, the same relationship to shelleyi in the Transvaal as shelleyi has to whytei in Northern Rhodesia and to uluensis in Tanganyika, in that it appears to be completely allopatric and to exhibit distinct morphological differences. But, whereas there appear to be no ecological barriers between shelleyi /uluensis and shelleyi /whytei and the morpho- logical differences are slight, there is a distinct ecological barrier between shelleyi and africanus and morphological differences are greater, representing a greater degree of divergence. This to me is just sufficient to warrant giving africanus specific rank. F. levaillantoidss (olim F. gariepensis) Acacia Redwing. F. levaillantoides comprises all the subspecies of southern Africa usually ascribed to this species and in addition the northern populations known by Praed and Grant as F. afer stantoni, F. a. friedmanni, F. a. archeri, F. a. lorti and F. a. gutturalis. It will be seen from the foregoing that I regard the red-winged francolins of the northern and southern acacia steppe as conspecific. They differ from F. psilolaemus and from F. shelleyi in being paler and less patterned below, lacking on the abdomen 156 B. P. HALL the rich chestnut markings of psilolaemus or the black-and-white patterning of F. shelleyi, having instead sparse chestnut and blackish flecks or streaks, more common in the northern birds. In southern Africa they are found in the arid zones from coastal Angola through South West Africa, Bechuanaland, Orange Free State and the Transvaal to the east coast in extreme southern Mozambique, associated, at varying altitudes, with open grassland, hillsides and kopjes, and edges to limestone pans. There are sight records (not included on Map 7) from the Lower Orange River (Winterbottom & Courtenay-Latimer, 1961: 11-12 and in litt.}. The range of levaillantoides overlaps that of africanus in the Orange Free State and that of shelleyi in the Transvaal. There is little information on their relationship here (perhaps because their numbers have been heavily reduced since the country was settled) but it can be expected that levaillantoides would be found at lower altitudes than africanus, and in drier areas than shelleyi. In the northern acacia belt the red-winged francolins are found in much the same habitat as southern birds, in sparse grass on rocky hillsides and mountains between about 2,000 and 7,000 feet. In central Abyssinia, where the Acacia Redwing is found in the same areas as the Montane Redwing, it occupies the lower slopes and F. psilolaemus the higher (Friedmann, 1930: 112-116). In northern Abyssinia and Eritrea there is an isolated population (gutturalis) found in scrub above 6,000 feet (Smith, 1957: 20). These birds appear to be a link between the acacia species and the Montane Redwing F. psilolaemus, for, like the montane bird, they have the facial pattern freckled and ill denned. They also differ from other Acacia Redwings in having the abdomen regularly marked with heavy blackish streaks on the feather centres, though the actual shafts are white, while the flanks are broadly streaked with chestnut. However, in colour and in the distribution of patterning below they seem closer to the Acacia Redwings and are tentatively regarded as conspecific with them. Otherwise variation in pattern among the Acacia Redwings is slight. All have the facial pattern and necklace well defined, but in birds from southern Angola (jugularis) south at least to the Cunene, the black-and-white necklace is enlarged on the breast to a broad gorget. The birds from the Benguela area are exceptionally small, and also rather grey, while Cunene birds are larger and rufous. Throughout both the northern block and other members of the southern blocks there is similar, very marked variation in colour between grey and rufous populations, largely associated with ecology, but there is remarkably little consistent difference between northern and southern birds in spite of the great distance separating them; those from the north have commonly, however, fine dark markings on the abdomen that are usually lacking in those from the south and the abdomens are whiter, less buff. The greyest populations are found in the mountains of eastern (British) Somaliland (lorti), in the Kalahari (kalaharica see appendix), and in the Benguela district (as already noted) . The most rufous birds are found in the Transvaal and Orange Free State (levaillantoides}, while in both South West Africa (pallidior} and southern Abyssinia and southern Sudan (archeri) both rufous and greyish micro-populations are found haphazard (see appendix). THE FRANCOLINS, A STUDY IN SPECIATION 157 It will be seen that, as far as the northern populations of the red-winged francolins are concerned, the subspecies of the acacia zone could be considered conspecific with F. shelleyi, since they are aUopatric. However, since in southern Africa F. levaillantoides and F. shelleyi both occur in the Transvaal, two species must be recognised, and as the northern acacia populations are closer both morphologically and ecologically to the southern acacia birds than to F. shelleyi, it seems logical to associate them in the same species. F. levaillantii Levaillant's Redwing. I have made no rearrangement in this species, which is universally recognised as comprising scattered populations from north-western Kenya to the Cape and west- ern Angola and which differs from all other red- winged francolins in having an ochre collar and the sides of face and edges of the throat ochre inside the black-and-white facial pattern. Such populations are found in north-western Kenya, where west of the Rift F. levaillantii replaces F. shelleyi in grasslands and valleys of the plateaux between 6,000 and 8,000 feet. In Uganda and Ruanda it is found at varying altitudes being common in the lower grasslands west of Lake Victoria (south of Bukobo) and on the plateaux westwards to Lakes Edward and Kivu. It is absent from most of Tanganyika but reappears on the Njombe and Iringa mountains in the south-west, and on the Nyika Plateau of northern Nyasaland, where it is common above 7,500 feet alongside the rarer F. shelleyi whytei (Benson, in litt.}. Another isolated population is found on the Kibera Plateau in Upemba, southern Congo, at about 6,000 feet (Verheyen, 1953 : 256), and scattered populations are found in the grasslands of the Balovale and Mongu districts of north-western Northern Rhodesia westwards through Angola to the Mombolo plateau and the Bailundu highlands. In southern Africa it is found only from the Zoutspansberg southwards through the eastern Transvaal, Natal (between 2,000 and 6,000 feet), Basutoland and the coastal districts of eastern Cape Province, keeping to the moister grasslands and even reed-beds in river valleys. It seems, therefore, in the Transvaal, Natal and Orange Free State, where the four species of Redwing occur, that F. levaillantii is associated with the wettest habitat. The distinctive ochre patterning of the face and throat of F. levaillantii has already been noted : it is otherwise most like F. levaillantoides but is long-billed and darker and richer in colour, with the black-and-white patterning of the necklace and gorget extending in a complete collar round the hind neck below the ochre collar, and the black-and-white stripes from above the eye, which in levaillantoides and shelleyi run down the side of the face, in levaillantii run behind the head to join at the back. In spite of the wide range and broken distribution there is little variation between the populations but in the populations of southern Africa and of the Nyika plateau these stripes continue down the back of the neck to join the black-and-white collar, while in the western and northern populations the ochre collar is unbroken. The Nyika birds (crawshayi) are richer in colour than those of southern Africa (levaillantii), with more rufous in the wing, and with more black markings on the abdomen, while the northern birds without the stripe on the hind neck (kikuyuensis) are otherwise similar to crawshayi. I 5 8 B. P. HALL F. finschi Finsch's Redwing. The last of the red-winged species, F. finschi, is monotypic. It has a very limited distribution being recorded only from about six localities in western Angola, in a strip about 200 miles long from near Nova Lisboa in the south to Vila Salazar in the north, and again from Brazzaville, in the Congo, 350 miles to the north. On Mount Moco in the Bailundu highlands it was found in the brachystegia woodlands and on the bare upper slopes between 6,500 and 7,000 feet (Hall, i96oa: 408) but the specimen from Vila SaJazar (American Museum of Natural History) must have been obtained in very different habitat at not more than 2,000 feet, and at Brazza- ville Malbrandt and Maclatchy (1949: 142) found it in grassland with neighbouring woods in which the birds took refuge. It is possible that the different habitats of F. finschi may be partly due to com- petition with F. levaillantii, which occupies grasslands in the southern part of the range, but is not found north of the Mombolo plateau. F. finschi differs from other red-winged f rancolins in lacking any black-and-white patterning on the face and neck and having the breast grey: it is closest to F. levaillantii in having a long bill and the sides of the face and border of the throat ochre, but in F. finschi the ochre is not so extensive on the hind neck : the abdomen is a mixture of buff and pale chestnut. OTHER VARIATION Considerable individual variation is to be expected in any birds as intricately patterned as the red-winged francolins and is found in all species. There is some local variation, usually in an increase or reduction of rufous and grey in the plumage, associated with ecology. This, as noted, is particularly marked in F. levaillantoides, but since the problems it raises are largely nomenclatorial it is discussed fully in Appendix 2. EVOLUTION The complex pattern displayed by the seven species of the Red-winged Group is not easy to interpret but I believe may be partly explained by comparison with a rainfall map (Map 7A) . From this it will be seen that the range of F. levaillantoides falls largely in the parts where the rainfall is below 20 inches in the year, the range of F. shelleyi falls between the 20-40 inch isohyets, the range of F. levaillantii falls between the 40-60 inch isohyets, while the range of F. finschi coincides remarkably with the only two patches south of the Congo rainbelt where there is over 60 inches in the year. Thus a slight increase or decrease of rainfall may be expected to favour the spread of one species at the expense of others. The discontinuous distribution of all species, and the scarcity of red-winged francolins in so many parts of their range, suggests that the present cannot be regarded as a favourable era for any members. This might be due to competition with other groups, but if so it could be expected that F. levaillantoides would be more common in the Kalahari, where there is no competition. From the association of so many members of the group with mountains I believe that a colder era would be more favourable to the group as a whole, and it was in some such era or eras, with fluctuations in rainfall, that THE FRANCOLINS, A STUDY IN SPECIATION 159 continuity in the ranges of the various species was achieved. Conversely warmer eras could be regarded as least favourable and ones most likely to produce isolated populations in which speciation developed. Even with this hypothesis as a basis a good many interpretations of the pattern present themselves which are largely dependent on which species one believes are most closely related. The following sequences are put forward tentatively on the premise that plumage pattern is the best guide to relationship. I can offer no explanation of the variation in bill size in terms of evolution and relationship but it is worth noting that in this group it is the species and, in shelleyi, the subspecies, associated with the higher rainfall areas that have the long bills. The length may perhaps be correlated with the hardness of the ground in which the birds search for food. The distinctive patterning on the abdomens of F. africanus and F. shelleyi suggests that their common ancestor split at an early date from the rest of the group. The present distribution suggests they developed in the south ; later spreading north to Kenya in a favourable era; later splitting into two blocks, proto-africanus and proto-shelleyi ; later proto-shelleyi splitting into three from which the subspecies shelleyi, whytei and uluensis developed. A favourable era must have followed in which their distribution was continuous over the present range. The species with relatively unpatterned abdomens may have originated in the north. A warm era could be expected to drive members of this block to the extreme north of the range, and also to high altitudes: it can be postulated that proto-psilolaemus developed in such an era to specific status and became adapted to montane conditions : later in a favourable era other members of the northern block spread south and west as far as Angola and the lower Congo. A subsequent un- favourable era isolated first proto-finschi in the Angola mountains, then proto- levaillantii in the mountains of central Africa, and proto-levaillantoides in Abyssinia, Somaliland and Eritrea: in isolation each became adapted to different degrees of humidity: subsequent fluctuations in rainfall allow first proto-levaillantoides to spread widely through the continent and then proto-levaillantii. The small degree of divergence shown between the isolated populations of both species (with the exceptions of F. levaillantoides gutturalis in Eritrea which perhaps was isolated earliest) suggests that these dry and wet cold spells were comparatively recent. A summary of the evolution postulated is as follows: Stage i. The group widespread in eastern Africa in a cold era. Stage 2. The group split into a northern and southern block. Stage 3. The southern block spreads north, at least to Tanganyika, and the northern block south to Kenya (Map 76). Stage 4. Warm and unfavourable, both blocks split into two : in the south proto- africanus isolated in the mountains of South Africa, proto-shelleyi possibly in the mountains of Nyasaland: in the north proto-psilolaemus isolated at high altitude, possibly on Mount Kenya or Elgon, and the ancestral stock of the other species isolated in the extreme north. Stage 5. Favourable. Proto-shelleyi spreads, possibly to the limits of its present 160 B. P. HALL range. Members of the northern blocks (with unpatterned abdomens) spread south and west to Angola. Staged. Unfavourable. Proto-finschi isolated first in Angola: proto-levaillantii isolated next in some mountains of central Africa not occupied by proto-shelleyi, while proto-levaillantoides is cut off in the north. Each becomes adapted to different requirements of humidity. Stage 7. In a long cold era favourable to the group as a whole, fluctuations in rainfall and temperature favour each species in turn. In some unusually cold spell proto- psilolaemus spreads to cover a large part of Abyssinia and Kenya but is later split into isolated populations in the highest mountains. Proto-shelleyi following a favourable era is split into three, from which the subspecies shelleyi, whyteiand uluensis develop almost to specific status. An unfavourable spell in the north isolates proto-gutturalis in Eritrea from the main body of proto-levaillantoides: in a subsequent dry spell (during which the acacia spreads to link the northern and southern belts) proto-levaillantoides spreads south to Bechuanaland. In a wet spell pToto-levaillantii spreads at the expense of shelleyi and levaillantoides, and proto-finschi spreads: where the two come together finschi is pushed into the mountains. Stage 8. The present. Too warm to be favourable to any members of the group, the ranges of all species being discontinuous and the birds uncommon in most areas. THE RED-TAILED GROUP (MAP 8) (F. coqui, F. albogularis, F. schlegelii} RANGE AND CHARACTERS This is a group of small francolins (wing under 150 mm.) which ranges from Senegal to the Sudan, and from central Kenya west to the central Congo and Angola, and south to the Transvaal and Natal: there is an isolated population in the Rift of Abyssinia. Members of the group are found in woodland, savanna and steppe and, in parts of the range, often on the same ground as other francolins, which suggests that their smaller size may make them non-competitive. The group is formed of three species which are allopatric except in the southern Congo and eastern Angola where isolated populations of albogularis are found within the range of coqui. All members of the group have a considerable amount of ochre on the sides of the face, spreading as an ill-defined collar on the hind neck : the crown is plain brown or chestnut, contrasting with the back which has a quail-type pattern (except in the females of schlegelii in which the striations and cross bars are reduced to a minimum), varying in colour from grey and black to vinous rufous: in the redder birds the rufous in the tail is conspicuous but in some of the greyer forms is no more than a pinkish wash on the outer feathers: the underparts are variable but all birds have a white or buff throat contrasting with the breast: all have black bills with some yellow at the base and yellow legs : the males have a single spur. THE FRANCOLINS, STUDY IN SPECIATION r6r SPECIES AND POTENTIAL SPECIES F. coqui is the most widespread of the three species which comprise the group. It is found in short grass in brachystegia woodland from southern Uganda and northern Tanganyika, southwards to the Rhodesias and westwards to Brazzaville and the edge of the Angola escarpment, but is absent from mountains and from the drier acacia steppe of Tanganyika. South of the brachystegia belt it is found in the acacia savanna skirting the Kalahari, from northern South West Africa to the Transvaal, and in the coastal districts of Natal. Scattered populations in Kenya, southern Abyssinia and the Harar area are also found in acacia (Benson, 1945 : 390) finding it common around Yavello and Mega in " open short-grassed country with scattered low thorn bushes at 4,000-4,500 feet ". In West Africa there is a population in the acacia belt of Senegal, the Upper Volta and northern Nigeria that is isolated by 1,000 miles (and the Lower Guinea forest) from the Brazzaville birds, and by nearly 2,000 miles from the Abyssinian birds. In this area coqui is found commonly in cultivations, often alongside F. bicalcaratus (Blasdale: unpublished). However, it seems that at Potiskum, 150 miles east of Kano in northern Nigeria, coqui is replaced abruptly by F. clappertoni which occupies the acacia belt eastwards to Abyssinia (Hollis: unpublished). In Abyssinia clappertoni occupies the northern and western areas to the centre of the Rift but is replaced by coqui in the Harar area in the east and in the Mega/Yavello area of the south. In coqui the quail-type patterning on the back is very consistent and well defined, though the basic colour, and the colour of the head, vary from grey to brown or rufous in different populations. The sexes differ, the males having the sides of the face and throat ochre or light buff without a black eye-stripe or necklace, the females having the sides of the face similar in colour to the males but the throat whiter and having a black necklace and a black eye-stripe that continues as a black line down the sides of the face. The males also have the breast barred while in the females the barring is replaced or overlaid by a pink or grey wash. There is little significant geographical variation in birds of the southern part of the range, from Angola, the Congo, central Tanganyika and coastal Kenya south to South Africa (coqui}. All have grey wings and barred abdomens (though the degree of barring is variable) and females have a pink wash on the breast. In the Abyssinian population (maharao) the wings are rufous and both sexes have light narrow barring from breast to abdomen, the females having only a faint wash of pinkish grey over the barring on the breast. The isolated West African population (spinetorum) also has rufous wings, but is pale above, washed with pink, and there is no barring on the abdomen in either sex: the breast of male is broadly barred and that of the female plain and pinkish-grey. In Uganda, Kenya and Tanganyika the pattern of variation is confused. East of the Rift in Kenya, south to the Arusha area of Tanganyika, there is a population (thikae, Grant & Praed, 1934) known only from a few specimens, which shows some approach to the Abyssinian birds in having rufous wings, light barring, and vestigial barring on the breasts of the females, but with no barring on the abdomens in either sex. From the Rift westwards to the eastern shores of Lake Victoria the birds (hubbardi) are similar to both the West 162 B. P. HALL African spinetorum and the eastern thikae in having plain abdomens but the wings are grey like the southern coqui, the breasts of the females are plain and grey, and the whole upperparts are rather dark and grey. Birds from west of Lake Victoria (ruahdae) are close to the southern coqui, but have the black barring broader and more widely spaced. These birds intergrade with hubbardi in the Tabora area. The second species in the group, F. albogularis, has a scattered distribution. There is a population in the Gambia, found mostly on the edges of cultivation, but it is rare (Hopkinson, 1923: 129). Other populations are found on the coastal plain near Accra and in the hinterland of Ghana, eastwards through Nigeria to Garoua in the (French) Cameroons. Field notes on these birds are scarce and somewhat contradictory, Serle (1957:389) finding them common in "the open savannas and rolling hill country " of Eastern Nigeria, whereas Blasdale tells me that further north he has found them only in thick woodland where there is abundant Isoberlinia doka. Most localities from which they have been collected lie within the woodland belt (Map n, Type 17) but Garoua lies within the acacia savanna belt (Map n, 20) and the specimen was obtained in " open, sandy, sparsely- treed country " (Good, 1952: 67). F. albogularis is absent from the woodland and savanna belt north of the Congo forest (where F. schlegelii is found) but reappears again 1,500 miles away in the south-eastern Congo on the plateaux of Marungu and Kibara, in the Upemba National Park, on the plains of the Balovale district of Northern Rhodesia, and on the treeless dambos of the Zambezi/Congo watershed in eastern Angola. Above F. albogularis is very like F. coqui but the quail-type patterning is less well defined in the females, the shaft streaks and barring being narrower: the wings are more rufous (similar in this respect to the east Kenya population, thikae). As in coqui the females have a black facial pattern and necklace, but it is ill defined, and the males have none, though they differ from coqui in having a white throat contrasting with the ochre cheeks, as in the females. In the rest of the underparts the males of albogularis are quite distinct from any form of coqui being chestnut on the breast with ochre shaft streaks and rich ochre on the abdomen, lacking any dark barring. The females on the other hand are barred, in greater or lesser degree according to the population, with a faint wash of pink or rufous on the breast, and resemble closely the females of the Abyssinian and east Kenya populations of coqui. The population of albogularis in which the females are least patterned is in the Gambia (nominate albogularis) the barring being light and restricted to the breast and flanks (though varying individually). In birds from Ghana to the Cameroons (buckleyi) the barring is generally heavier, usually with only the centre of the abdomen unpatterned. Birds of both sexes from the populations south of the Congo forest are slightly larger and more richly coloured than those from West Africa and in the females of the south-east Congo (dewittei) the barring is heavier but no more extensive than in buckleyi: the darkest birds with the most heavily barred females are found in eastern Angola and possibly the Balovale district (meinertzhageni White, 1944 see appendix). The third species of the group, F. schlegelii, has been considered by Chapin and Peters to be a subspecies of F. coqui. I believe it is distinct enough to warrant THE FRANCOLINS, A STUDY IN SPECIATION 163 specific status and furthermore is closer in several respects to F. albogularis. It is a rare bird, found scattered through the woodland and savanna belt from the eastern Cameroons to the western Bahr-el-Ghazal. From what Cave tells me it seems to be more consistently a woodland bird than albogularis and he has the impression that it is found in rather denser woodland than the sympatric F. icterorhynchus of the Vermiculated Group, especially where there is Isoberlinia doka. Above schlegelii is closer to albogularis than coqui, though the quail type patterning is much reduced and the sexual dimorphism is more pronounced, the males having broad white shaft streaks but few transverse bars while the females are almost unpatterned: both sexes have a soft vinous wash. Below the male of schlegelii is like that of southern races of 'coqui, with ochre sides of the face and throat and narrow black and white barring over the breast and abdomen : the female resembles the male on the head and throat but has the breast grey, patterned with cream triangular marks at the base of the feathers, the flanks sparsely barred, and the centre of the abdomen plain cream. OTHER VARIATION In the southern populations of F. coqui there is local and ecological variation in size and colour; exceptionally small birds are found in coastal Kenya and Natal; grey, pale birds in Bechuanaland (vernayi) and pale, rather pinkish birds in South West Africa (hoeschianus Stresemann, 1937) : other exceptionally grey birds occur in southern Mozambique (Coguno) and may be expected locally in dry conditions, while pockets of more richly coloured birds are associated with richer vegetation. There is otherwise little variation between birds from Natal and the Transvaal north to coastal Kenya and west to Northern Rhodesia, except in the extreme west: here and in Angola the birds are consistently less rufous and the bars on the underparts are narrower and rather closer together (angolensis) . Birds from Kasai and western Katanga (kasaicus White, 1945) are similar to coqui above and angolensis below: they intergrade with coqui in the eastern Katanga. In F. albogularis of West Africa individual variation in patterning has already been noted. Serle (1957: 388) also found the general colour individually variable between rufous and grey, though there seems a tendency for the coastal population of Accra to be greyer than inland birds. Individual variation is also found in F. schlegelii especially in the females (see Tray lor, ig6ob: 86). EVOLUTION The most striking features of the pattern presented by the Red-tailed Group are the enormous distances separating the north-western and south-eastern sub- species of both F. albogularis and F. coqui: also, in F. coqui, the complexity of variation in East Africa in contrast to the uniformity of southern birds. I feel confident that the discontinuous range of F. albogularis is largely attributable to the presence of F. schlegelii in the intervening segment of the same vegetational belt, and suggest that the ancestral stock of the two species once had a continuous distribution in the savanna and woodland encircling the Congo forest. To account for the split into two species I suggest this era was followed by a very dry period in which the forest was divided into a western and eastern block each encircled by woodland and separated by a corridor of acacia linking the northern and southern 164 B. P. HALL acacia belts. It can reasonably be postulated that proto-albogularis developed in the belt round the western forest and proto-schlegelii in the belt round the eastern forest (Map SB). When the two forest blocks rejoined with a return of humid conditions and the surrounding belts became one again, the central segment of the belt was occupied by proto-schlegelii and the western and southern segments by proto-albogularis. It is more difficult to interpret the pattern of F. coqui but a logical sequence can be found which will fit into that postulated for the other species. Proto-coqui may have been separated from the ancestral stock of the other two species at a time of spread of montane forest and developed in the east while proto-albogularis / schlegelii developed in the north. With the retreat of the forest proto-coqui spread south and west chiefly in acacia savanna. In an unfavourable spell, either too wet or too dry, proto-maharao in Abyssinia was isolated. In the very dry era postulated above, in which there was a corridor of acacia in the central Congo, proto-coqui could have spread into the northern acacia savanna belt. Subsequently with a return of humid conditions the northern population would be cut off from the southern by woodland and forest. The southern populations nearest the Equator prove adaptable so that, with the return of woodland and savanna and the in- digenous albogularis to the southern Congo and Angola, coqui competes successfully, occupying the woodlands and albogularis the more open grasslands and plateaux. Meanwhile the isolated northern population diverges from southern birds, develop- ing the distinctive unpatterned abdomen in both sexes, and spreads through the northern acacia belt west to Senegal and east through the Sudan to Kenya and Uganda. Here it encounters both the Abyssinian maharao and the southern coqui. The characters of the present populations of Kenya (hubbardi and thikae) and of southern Uganda (ruahdae) suggest that some interbreeding took place at some period. The present appears to be favourable to the Vermiculated Group in the north, at the expense of the Red-tailed Group. In the immediate past it seems that F. clappertoni has spread west through the acacia belt in the Sudan, eliminating coqui between the Nile Valley and northern Nigeria. In the woodland and savanna belt F. bicalcaratus flourishes in the west and F. icterorhynchus in the east, and competition with these two may account for the scarcity and discontinuous dis- tribution of albogularis and schlegelii. However, in West Africa coqui, though apparently unable to compete with clappertoni, is yet able to live on the same ground as bicalcaratus on the southern fringe of the acacia belt. The populations of coqui in eastern and southern Africa also seem able to co-exist on the same ground as other francolins. A summary of the evolution postulated is as follows : Stage i. The group develops in savanna, perhaps in north-eastern Africa. Stage 2. An unfavourable era, possibly a spread of montane forest. The group splits into two with proto-coqui isolated in the east and proto-albogularis I schlegelii to the north of the forest. Stage 3. More favourable. Proto-coqui spreads south and west in rather dry THE FRANCOLINS, A STUDY IN SPECIATION 165 savanna, mostly acacia. PToto-albogularis/schlegelii spreads in the wetter savanna and woodland encircling the lowland forest of the Congo Basin (Map 8 A). Stage 4. A very dry era. The Congo forest is divided into a western and eastern block, each surrounded by its own woodland and savanna belt, with a corridor of acacia between. Proto-albogularis develops in the belt surrounding the western block of forest, proto-schlegelii in the belt surrounding the eastern block. Proto- coqui spreads through the acacia corridor into the northern acacia belt. In an unfavourable period in the north-east proto-maharao becomes isolated (Map SB). Stage 5. More humid. The two blocks of forest rejoin and the two savanna belts become one again, the eastern segment being occupied by proto-schlegelii and the western and southern segments by proto-albogularis. In the southern segment, as woodland and savanna return to districts previously acacia, proto-coqui is not pushed out but becomes adapted to woodland and competes successfully with proto-albogularis which takes to the more open grasslands. In the north proto- spinetorum subsp. spreads west to Senegal and east to Abyssinia and Kenya and diverges from the southern birds. There is limited interbreeding where the northern birds re-encounter other members of the species in the east. Stage 6. Favourable for the Vermiculated Group, especially in the north. F. clappertoni spreads, eliminating F. coqui from the acacia belt in the Nile Valley and Sudan westwards to northern Nigeria. In the woodland and savanna belt F. bicalcaratus and F. icterorhynchus flourish at the expense of F. albogularis and F. schlegelii. In the south F. coqui co-exists more successfully with francolins of other groups. SPECIES NOT INCLUDED IN ANY GROUP (MAPS 9 & 10) Francolinus lathami F. lathami is a forest francolin ranging from Sierra Leone to western Uganda and the southern Sudan. It appears at first sight quite unlike any other African francolin (closer, except in size, to F.francolinus of Asia), being, in the male, largely black below with regular white spots, in the female brown with irregular white spots, the throat black and face patterned, the upper parts mottled rufous and brown with some white streaks on the mantle, the bill black, finer than in other francolins. However, comparison of a pair of lathami with a pair of schlegelii suggests that lathami may be more closely related to the Red-tailed Group than to any other francolin. It will be seen that the pattern and the colouring on the upper parts are basically the same, though lathami, as would be expected in a forest bird, is darker and more mottled. Below the patterning of the females has some similarity in the irregular, rather triangular shaped spots: while the spotted male of lathami looks very different from the black-and-white barred male of schlegelii, this may not be a very significant difference for in odd abdomen feathers of several lathami the spots are less regular and merge as vestigial bars. The chief differences in fact lie in the shape of the bill and the pattern of the face and throat. Furthermore the first plumage of lathami chicks is very similar to the plain vinous mantle of the female schlegelii. 166 B. P. HALL While being confident that there are affinities between lathami and schlegelii I am not prepared to guess at what period in history proto-lathami became isolated and adapted to forest. The likeness to one particular species in the Red-tailed Group suggests that its isolation could have been subsequent to the emergence of species within the group, the adaptation to such different habitat accelerating diver- gence. On the other hand F. schlegelii is the species of the Red-tailed Group that lives in the thickest woodlands and, if proto-lathami had broken from the ancestral stock of the Red-tailed Group before the three species had diverged, the similarities between lathami and schlegelii might be due to convergence. For this reason, and with regard to both its morphological and ecological distinctions, I feel it best to treat lathami independently. It will be seen that I am not prepared to regard the black-and-white spotting in the males of lathami as indicative of relationship with F.francolinus. It has been amply demonstrated in the discussion on groups that the pattern of the underparts is of less significance than size and the pattern above, and while it is interesting that such a striking pattern is repeated in the genus in two such different birds, it cannot be relevant to the grouping of species. From the distribution plotted on Map 9 it would appear that there is a wide gap between the populations of the eastern and western Congo, but this is more likely to be due to lack of collecting in the central area than to a genuine break. There is little geographical variation, in the males of eastern birds (schubotzi) the black- and-white pattern extends further down the abdomen, and the females have the cheeks more rufous than grey, but there is considerable individual variation. Francolinus nahani This little francolin is confined to the heaviest forests of the north-eastern Congo and Uganda. It is a distinctive bird, predominantly dark brown or blackish, with white streaks on the underparts and some white spots on the wing-coverts and flanks. The bill is black at the tip and crimson at the base; there is a patch of crimson bare skin round the eye; the legs are crimson, and without spurs in either sex. It is difficult to assess its affinities with the other francolins. Superficially it appears closest to F. lathami in its small size, general dark coloration and suggestion of spotting, and they occupy the same type of habitat. On the other hand the similarities are in those characters that are to be expected in forest birds, and could well be due to convergence, and the fact that F. nahani in the whole of its range is sympatric with F. lathami, without apparently being ecologically segregated, does not suggest close relationship. I would suggest very tentatively that F. nahani may have broken off at a remote period from the ancestral stock of the Scaly Group (in the same way as I have suggested that lathami may have broken off from the Red-tailed Group), and become adapted to heavy forest instead of the forest clearings. I suggest this because nahani, although much smaller, seems to have some characters in common with the Scaly Group as a whole and with each of the three individual species. In both nahani and all species of the group the sexes are alike, and the legs are red: nahani is like F. squamatus above, having an unpatterned head and THE FRANCOLINS, A STUDY IN SPECIATION 167 vermiculated back; it is like F. ahantensis in the white streaking below; and like F. griseostriatus in the crimson base to the bill. Furthermore in all species of the Scaly Group there is a small area of bare skin below and behind the eye, though this is not brightly coloured and usually overlooked. Chapin considered the absence of spurs to be a generic character and proposed the monotypic genus Acentrortyx for F. nahani, but the fact that F. pictus in India is also without spurs, but yet is so closely related to F. francolinus as to hybridise, demonstrates that this character is not necessarily significant. I see no other reason for regarding F. nahani as generically distinct. F. pondicerianus F. pondicerianus is the smallest of the Asiatic Francolins and is widespread from the shores of the Gulf of Oman eastwards through the plains of India and the dry north-western districts of Ceylon, living in grassland or scrub. It appears to have some affinities with both the Red-tailed and the Striated Groups of Africa and in colour and pattern is closest to F. coqui, in particular to the geographically nearest form F. c. maharao of Abyssinia. It has " quail-type " plumage on the back, but is duller than coqui, with broader transverse white bars and narrower shaft streaks, and with chestnut markings on a grey or huffish ground: the underparts are whitish, irregularly barred with very fine blackish bars: the throat is buff surrounded by a blackish necklace, which is smaller than that found in the females of coqui and starts from the chin and not the sides of the bill. In other respects it is closer to the Striated Group (F. sephaena and F. streptophorus) than to the Red-tailed Group, having a longer and more graduated tail (the tail more than half the length of the wing), and stronger (though not necessarily longer) legs and larger feet, the legs being reddish as in F. sephaena. The bill is distinct from either group being brownish in colour and heavier in front of the nostrils. The tail in colour combines characters of both groups being reddish at the base but otherwise largely blackish. The affinities with two of the African Groups suggest that the ancestral stock of F. pondicerianus was isolated from the African stock before the Red-tailed and Striated Groups had diverged and for this reason I have discussed it independently. It seems that, while the African stock underwent many vicissitudes of climate with consequent isolating and regrouping of populations resulting in speciation, the Indian stock had a comparatively tranquil history for in pondicerianus there is little variation, and that mostly clinal and ecological: birds from the dry districts of Persia and Baluchistan, east to the western Punjab (mecranensis) , are the greyest and palest, and those from southern India (pondicerianus) the darkest and least grey with the greatest amount of chestnut markings. Those from northern and central India.(interpositus) are intermediate. Ceylon birds (ceylonensis Whistler, 1941) are greyer than either pondicerianus or interpositus but darker than mecranensis. F. gularis F. gularis is the swamp francolin of India, being confined to reeds and marshes in the plains of the Ganges, Brahmaputra and their tributaries from central Uttar 168 B. P. HALL Pradesh eastwards into Assam. It is a large francolin, comparable in size and general proportions to the largest member of F. francolinus, but with rather a shorter bill and larger feet (as would be expected in a marsh bird). It is, however, quite different in colour and pattern being predominantly a brown and white bird, with a barred back, streaked underparts and a rufous throat, wings and tail. The sexes are alike. Proto-gularis seems likely to have become isolated from other francolins in a dry period and taken refuge in river valleys, becoming adapted to these conditions. The dry era postulated may even have occurred before the link between the Asian and African francolins was broken since the differences from other francolins shown by F. gularis, both in ecology and morphology, suggests divergence over a long period. CONCLUSIONS THE GROUPS The chief lesson to be learnt from comparison of the relationships between the members of each group re-emphasises the difficulty of assessing from morphological characters the level of speciation which any form has attained. This is brought home most forcibly by finding that such dissimilar birds as cranchii and afer in the Bare-throated Group will interbreed freely, while the three Redwings in the Trans- vaal shelleyi, levaillantoides, and levaillantii, which look more alike, are able, on account of ecological divergence, to live in a small area without interbreeding. It seems that a relatively slight ecological divergence among the francolins may be effective against interbreeding because their sedentary habits and specialised choice of habitat make them reluctant to move out of their selected type of vegeta- tion. These same habits allow also for comparatively small barriers to > act as isolating factors between two populations that are not necessarily ecologically divergent. Two instances have been given where rivers appear to form such a barrier, the Niger between ahantensis and squamatus of the Scaly Group and the lower Zambezi between natalensis and hildebrandti of the Vermiculated Group. In the Red-winged Group fifty miles of unsuitable country in Tanganyika between F. s. shelleyi at Salanga and F. s. uluensis at Mbula serves as barrier to interbreeding between them. The difficulty of designating in these circumstances which members of groups should be called species and which subspecies was discussed at the start of the paper, for the status of most lies somewhere between the two. This should be borne in mind in the subsequent discussion on evolution, for these " species ", which are allopatric members of superspecies, for the most part appear to have attained a level of speciation far below that of those species which are able to live with each other in full sympatry. EVOLUTION The stages of evolution postulated for each group have been based on what seem to me to be the simplest and most likely interpretations of the patterns shown, and they represent the minimum number of stages which can be envisaged to produce THE FRANCOLINS, A STUDY IN SPECIATION 169 such a pattern. They represent a sequence rather than a time-table of change and as such could, in theory, be spread over any number of years. However, there is evidence from the distribution of many montane plants and trees, as well as birds, which shows that there have been periods in which montane forest has spread over vast areas of Africa, linking up the present isolated patches in the Cameroons, Angola, East Africa, Abyssinia, Nyasaland and South Africa, descending as low as 2,000 ft. (see Map 3 A). At the same time (by inference a wet era) there would be extensive lowland forest below that level. This would have the effect of reducing the refuges for grassland and woodland species to comparatively small areas isolated from each other by forests. Such conditions would not only be catastrophic for many species but would obliterate any traces of the pattern of distribution and evolution prior to that period. Thus, although such a spread of forest may have occurred during any extensive cold wet era, it follows that, if the stages of evolution I have suggested are accepted, the conclusion cannot be evaded that they have been accomplished since the last spread. This conclusion carries with it implications on the rate of speciation which need to be examined closely. In the course of these stages I have suggested that twelve species of francolins (the ancestral stock of the eight groups and the four independent species) have increased to forty-one, all the new species being members of superspecies. The accompanying dendrogram (Fig. 2) summarises the relationship postulated of the squamatus ahantensis griseostriatus adspersus capensis natalensis hartlaubi hildebrandti icterorhynchus clappertoni bicalcaratus bar wood i swainsonii rufopictus leucoscepus francolinus pictus pintadeanus gularis- africanus shelley i .psilolaemus levaillantoides levaillanti finschi streptophorus sephaena .pondicerianus Francolinus FIG. 2. Dendrogram showing the relationship postulated of the various groups and species, and their date of origin in respect to the last major spread of montane forest. ZOOL. 10, 2 12 iyo B. P. HALL various groups and species to each other and their date of origin in respect of this last spread. Unfortunately no date can be given yet to this last spread of forest, and there is indeed still uncertainty on the date of onset of the Last Glaciation, opinions varying between 100,000 and 50,000 years ago. Authorities, however, agree that it ended about 18,000 years ago and that there have been at least four glaciations in the course of the Pleistocene. I am suggesting therefore that these changes have taken place in the comparatively short period of the Late Pleistocene (i.e. about the last 100,000 years). If we accept that Francolinus may (among other modern genera of birds) have originated in the Oligocene (approximately forty to thirty million years ago) (Van Tyne & Berger 1959 : 6-17) and has taken all the time prior to the last glaciation to divide into twelve species, it seems that I am postulating an incredible acceleration in the rate of increase of species. But this is not so, for, as stated, the concept of periodic spreads of forest over vast areas of Africa is inseparable from the acceptance of a high casualty rate among the non-forest species at that period. If, for instance, the present was followed by a cold, wet era that brought with it a wide spread of forest, it is highly unlikely that any of the less successful non-forest species of tropical Africa would survive (e.g. rufopictus, streptophorus , finschii, albogularis, schlegelii) . Competition between members of different groups in the limited refuges left would eliminate several of the apparently successful species as well (just as F. dappertoni appears to have eliminated F. coqui from the Sudan) , while members of the same group, forced into unprecedented proximity, would probably interbreed freely. Thus the total number of African species could again be reduced to few (if any) more than the nine which survived the previous forest spread. These would be the ancestors of future groups. Thus while I believe the number of members in any group may fluctuate widely with successive glacial and inter-glacial eras it seems that any increase in the number of groups and independent species would be very slow and not incompatible with the rate of evolution since the inception of the genus. It could also be suggested that any further increase is unlikely in this figure, which represents the number of birds which have reached a sufficiently high level of speciation to be able to live sympatrically without interbreeding, for it is evident that only a limited number of such birds can be supported in any given area. In birds like the francolins, in which two species are rarely found actually on the same ground, this limit is governed largely by the number of alternative types of habitat, and the unequal distribution of the groups in tropical Africa gives some reason to suppose that this limit may have been reached. Disregarding the Montane Group, with its extremely specialised ecology, it will be seen that, apart from a relict population of the Striated Group, only three groups, the Scaly, Vermiculated and Red-tailed, are fully represented in the north-western section (from West Africa to the Sudan and Uganda) . There is little overlap between the ranges of these groups and where two do occur together one is thriving at the expense of the other. Furthermore over a vast area it appears that members of the most successful group, the Vermiculated, have totally driven out representatives of THE FRANCOLINS, A STUDY IN SPECIATION 171 other groups, namely the red-tailed F. coqui from Lake Chad to the Nile, and the striated F. streptophorus between the Cameroons and Uganda. This north-western section is relatively flat with monotonous vegetation, and no very clear-cut divisions between the vegetation belts : the east and south on the other hand, from Abyssinia to South Africa and Angola, is more mountainous in many parts with more varied vegetation, and here no less than six groups are sympatric or partly sympatric. In this section of Africa I have suggested certain instances in which the range of one group has been restricted by competition with another, but it is also evident that several special may thrive in a small area. I am not competent to discuss the situation in Asia where the development and status of the francolins must be dependent on those of the numerous other genera of gamebirds which I have not studied, but it seems unlikely that there is much room for an increased number of full species. In arguing that the members of groups may have developed to the level of members of superspecies within the Late Pleistocene I am conscious that this is postulating a more rapid rate of speciation than is usually accepted, though Rand (1951: 567) has suggested that some of the forest subspecies of West Africa may have evolved in the last 12,000 years, while Moreau (1930) gives examples of Egyptian subspecies which have probably evolved in 5,000 years. However, some weight is added to my argument by the evidence collected by Moreau on the climatic fluctuations that have taken place in this period (Appendix i), which shows that there has been sufficient variation in both temperature and humidity to bring about the wide redistributions of the montane and lowland forest, woodland, savanna and steppe that would be necessary to the different stages proposed. Furthermore, it shows that all the vegetational changes postulated in the evolutionary sequences in the groups (which were founded entirely on the evidence of the birds alone) did, or could, take place, though it is not always easy to satisfy the sequence. To support some of the major changes which I have envisaged, he has shown, for instance, that there have been very dry periods prior to 10,000 years ago in Angola and Katanga (at least) which, if they prevailed elsewhere, could have been sufficient to connect the northern and southern acacia belts in the way postulated in the evolutionary sequences of the Bare-throated, Striated and Red-winged Groups. He gives corroboratory evidence for the existence of such a dry corridor at some period in Tanganyika, and Benson and White (in press) suggest this corridor may have been through the Luangwa and Rukwa valleys, and they list many species of birds and mammals whose present distribution has led them to this conclusion. Moreau shows also that there was a period (undated) at which Lake Chad was 600 miles long, probably with forest at the southern end joining the main Congo forest, so forming a barrier between eastern and western woodland or grassland species, such as was postulated in the Vermiculated Group to split proto-bicalcaratus from proto-icterorhynchus. He quotes too a record that suggests there was probably lowland forest at Oldowai in northern Tanganyika in the early Pleistocene where there is now acacia steppe. Though this dating can hardly be correlated with the spread of lowland forest envisaged in the later stages of the evolutionary sequence of the Scaly Group, it suggests that such a spread has taken place once and could 172 B. P. HALL therefore be repeated in similar climatic conditions: perhaps achieved in even one of the recent minor pluvials in East Africa (less than 35,000 years ago). Similarly he says that late in the Middle Pleistocene there is believed to have been a dry period when the Congo forest was reduced to strips in the east and west (and possibly also on the northern rim) and that there was acacia in the country between. This makes the sequence of evolution postulated for the Red-tailed Group less improbable than it seemed when I first proposed it (before I had seen Moreau's paper), though this sequence remains the most difficult to correlate with known climatic sequences. Finally, the undated dry period in the Late Pleistocene when the Sahara advanced 300 miles southward may perhaps be correlated with the stage postulated in the Striated Group when dry conditions prevailed on the ridge of higher country between the Cameroons and Uganda on which proto-streptophorus may eventually have been isolated. All the other climatic changes I have postulated in Africa are relatively small and the occurrence of similar changes is unquestionable, though not necessarily in correct sequence. However, some of the dating for these in West Africa is interesting for the most recent humid period in the north-west (5,000-2,000 years ago) may be the time at which F. Ucalcaratus spread into Morocco up the north-west coast, and if so the Morocco population, which shows only slight divergence, may have been isolated for only 2,000 years. The lack of divergence between the populations of F. ahantensis on either side of the present gap in the Upper Guinea forest also gives support to Moreau's evidence that the gap was once further east " probably as late as the Late Pleistocene ". In the Spotted Group of Asia, as in the African groups, it was postulated that the members developed to specific status after the last glaciation. While it is easier here to argue that the same sequence of isolating and rejoining of species was likely to take place in any glaciation it seems unlikely that the simple pattern which remains now could be the product of more than one such sequence. Finally, lest any of the foregoing should sound dogmatic, it is necessary to end on the same note of caution that was sounded in the introduction. I am fully aware that all the conclusions are based on circumstantial evidence drawn from my interpretations of the picture as we see it today, which are necessarily fallible since so much of the picture is missing. That it has proved possible to construe the known facts into a logical and coherent story gives some verisimilitude to it and so does the evidence that many of the postulated vegetational changes have actually taken place. However, the story remains hypothetical, but further studies of this nature and researches in other branches of science will corroborate or disprove others of the possibilities suggested so that eventually a picture may emerge that approximates to reality. SUMMARY i. It is found that the genus Francolinus (including Pternistis) is comprised of eight groups, the members of which are largely allopatric, and four independent species. THE FRANCOLINS, A STUDY IN SPECIATION 173 2. Some of the difficulties are discussed of denning the relationship between the members of the groups. 3. The value of different characters as a guide to relationship is illustrated by comparison between the groups. 4. Each group is studied in turn and the characters and ranges of its component members defined. In doing this it was found necessary to review the systematics of the Redwing and Grey wing Francolins of the species africanus (olim. afer), levaillantoides (olim. gariepensis] and shelleyi. 5. An attempt is made to postulate the sequence of evolutionary stages through which each group has developed, and the climatic and vegetational changes which have accompanied each stage. 6. The four independent species are discussed. 7. Comparison between groups suggests that slight ecological divergence may be an especially effective barrier to interbreeding. 8. It is concluded that most members of groups reflect in their characters, relationship, and present ranges, the climatic changes that took place in the Late Pleistocene, subsequent to the last glaciation, or, in Africa, the last extensive spread of montane forest. Consequently it is suggested that the number of species of this status is liable to fluctuate considerably in glacial and inter-glacial periods, though the number of groups and independent species may remain more or less constant. 9. These conclusions are supported to some extent by evidence provided by R. E. Moreau of climatic changes that have taken place in the Late Pleistocene (Appendix i), but more evidence is needed from all branches of science to evaluate them fairly. APPENDIX i Summary of some conclusions reached by R. E. Moreau on the climatic fluctuations in Africa as set out in his paper " Vicissitudes of the African Biota in the Late Pleistocene " (in press : to be published in Proceedings Zoological Society, London). Africa generally (in conformity with world temperatures) Colder than the present (with associated reduced humidity) in varying degrees for at least 50,000 years prior to 18,000 years ago. In the last extreme phase, extending from about 24,000-18,000 years ago the cooling amounted to about 5^- Around 7,000-5,000 years ago probably 2C. warmer, and for some centuries prior to 150 years ago 1-2 C. cooler, than the present. Mediterranean Africa and Northern Sahara A major pluvial period affected at least the western half of the area in the late Middle Pleistocene, prior to perhaps 100,000 years ago. Palaearctic plants and land- Footnote : Mr. Moreau has very kindly allowed me to include this summary to illustrate my paper but it must be appreciated that most of the conclusions quoted have been qualified in the full text, where reference is given for the authority on which they are based. This summary must not therefore be used as a base for further premises without consulting the full paper. It must be emphasized that the time- correlations of the events referred to in different parts of Africa remain to be worked out. 1 74 B. P. HALL molluscs advanced some 400 miles southward and persisted in Ahaggar until the Neolithic, some 6,000 years ago. Between 5,000 and 2,000 years ago it was more humid (at least in the north-west) than at present. Southern edge of Sahara and sub-Saharan West Africa The southern edge of the desert has fluctuated in the course of the Late Pleistocene (i.e. about the last 100,000 years) between about 300 miles south and 300 miles north of the present position and these changes probably extended across to the White Nile. Since the Neolithic, i.e. in about the last 5,000 years, a belt of country some 300 miles in width has reverted to desert. These fluctuations affected the vegetation of coastal West Africa, especially the Upper Guinea forest, and when the desert came south of its present location the area of forest was greatly reduced and further dissected. At some stage in the Late Pleistocene, not yet dated, Lake Chad increased to an area 600 miles long, extending nearly to Tibesti. This implies a contemporary north- ward extension (or movement) of the equatorial rain-belt by some 300 miles. Prior to its capture by the Lower Niger, some 10,000 years ago, the Upper Niger expended itself in large lakes northwest of Timbuktu, in what is now extreme desert. Congo basin and Angola Late in mid-Pleistocene very dry, with aeolian re-distribution of Kalahari sand. Congo forest reduced to a strip towards the coast, another along the eastern rim of the basin and perhaps a third along the northern edge. Throughout Late Pleistocene, humid except in Katanga and Eastern Angola. Here arid in the latter part of the Late Pleistocene down to about 10,000 years ago. East Africa The " Gamblian " pluvial, with fluctuations in intensity was contemporary with the Last Glaciation and extended down to about 18,000 years ago, and there have been subsequently at least two short periods more humid than the present. Owing to the broken topography the vegetational effects would have been, in the main, local advances and retractions of vegetation types. In Somaliland the fluctuations were minor and the country always to some degree dry. (Note : There is evidence that at some period in the Early Pleistocene lowland forest extended at least to Oldowai in northern Tanganyika. Also at some period in the Late Pleistocene a corridor of acacia probably extended through Tanganyika to Northern Rhodesia). Rhodesia and southern Africa During the Late Pleistocene the rainfall fluctuated between about half and one and a half times the present in the Rhodesias and the Transvaal. Kalahari sand was blowing east as far as the Victoria Falls some 12,000-9,000 years ago. The Orange Free State was drier than the present some 40,000 years ago. South West Africa was never better than semi-arid, THE FRANCOLINS, A STUDY IN SPECIATION 175 APPENDIX 2 Taxonomic notes and synonymy SPOTTED GROUP F. francolinus Five clinal races have been recognised and further sub-division does not seem useful. There- fore caucasicus and sarudyni are placed in the synonymy of francolimts ; festinus Koelz, 1955 and bogdanovi in the synonymy of henrici ; parkerae Van Tyne & Koelz, 1955 in the synonymy of asiae. Also billypaynei Meinertzhagen, 1933, which is a name given to an exceptionally dark micro-population from the Lake of Antioch, is included in francolinus. F. pintadeanus The characters on which wellsi was based do not seem constant and it is accordingly regarded as a synonym of pintadeanus. I accept Delacour's opinion (1951 : 10) that -F. boineti Bourret, 1944 is an aberrant specimen of F. pintadeanus. BARE-THROATED GROUP F. leucoscepus The amount of variation exhibited by the whole species is so slight that numerous sub- divisions are not practical. Accordingly holtemulleri, muhamedbenabdullah, keniensis,* kili- mensis, tokora, oldowai Van Someren, 1939, are placed in the synonymy of infuscatus. F. swainsonii Since Pternistis has been submerged in Francolinus, P. s. chobiensis Roberts, 1932, is pre- occupied and F. s. lundazi used instead (see Benson & White, 1957: 138). A record of the occurrence of F. swainsonii at Karonga, at the north end of Lake Nyasa (Long, 1961 : 25) was based on a misidentification, the specimens being F. afer (Benson, in lift.). F. afer In the block of vermiculated subspecies there is much individual but little geographical variation. Accordingly punctulata and nyanzae are placed in the synonymy of cranchii. In the black-and-white birds of Angola and South West Africa variation between the northern and southern birds is too slight to warrant recognition, so palliditectus White, 1958 (nom. nov. for cunenensis Roberts) is placed in the synonymy of afer. In South Africa an intermediate population between castaneiv enter and notatus was named krebsi, which I place in the synonymy of castaneiv enter. In eastern Africa the name humboldtii was given to a bird from Tete on the Zambezi, a locality lying in the centre of the zone of hybrids between the black-faced melano- gaster and the white-faced swynnertoni : it is here restricted to the unstable population. If names are required for the hybrids between the vermiculated and black-and-white sub- species it is convenient in Angola to refer those populations which are closest to cranchii to manueli White, 1945 (synonyms camabatelae Meise, 1958, and mackenziei White, 1945). an ^ those which are closest to afer to benguellensis (synonyms angolensis Bocage, angolensis Roberts, and chio Meise, 1958 see Traylor ig6oa: 143-146). Similarly in Tanganyika, Nyasaland and the Luangwa valley those populations closest to intercedens may be known as bohmi (synonym tornowi Meise, 1933), and those closest to melanogaster may be known as itigi (synonym aylwinae White, 1947). With the merging of Pternistis in Francolinus F. coqui angolensis Rothschild, 1902, is ante- dated by Pternistes afer var. angolensis Bocage, 1893, and by the laws of homonymy should be renamed. However Bocage 's name angolensis was given in error to the same bird which he had named benguellensis at the head of his article, and the name was forgotten until listed in the synonymy of Pternistis afer benguellensis by Bowen (1930: 15?) Jt thus has remained unused as a senior synonym in the primary zoological literature for more than fifty years and must be considered a forgotten name (nomen oblitum). (Int. Code of Zool. Nomen. 1961. Article * Also pre-occupied by F. squnmatus keniensis Mearns, 1910. 176 B. P. HALL 23b.) Application is being made to the International Commission for it to be placed on the Official Index of Rejected Names. MONTANE GROUP F. castaneicollis It is probable that most of the micro-populations in isolated patches of evergreen forest are slightly distinct, but it is not practical to recognise by name all those that show minor differences: accordingly vottegi and gofanus are placed in the synonymy of castaneicollis, and patrizii Toschi, 1958, in the synonymy of kaffanus. F. jacksoni It has not been established that variation in the amount of white in the underparts is geographi- cal or associated with altitude. Therefore gurae is placed in the synonymy of jacksoni. F. nobilis F. camerunensis ruandae Van Saceghem, 1942, is placed in the synonymy of F. n. nobihs. It is also pre-occupied by -F. coqui ruandae Jackson, 1938. F. swierstrai F. cvuzi Themido, 1935, is placed in the synonymy of F. swierstrai. SCALY GROUP F. squamatus Although there is considerable individual variation no significant geographical variation is recognised from the greater part of the range. Accordingly tetraoninus, zappeyi and dowashanus are placed in the synonymy of schuetti and whitei Schouteden 1954 (nom. nov. for confusa Schouteden 1954) * n tne synomy of squamatus. It is not thought practical to recognise by name any micro-populations isolated on the mountains of the Kenya/Tanganyika border, since the main trend of variation is clinal. Therefore, kapitensis, keniensis and chyuluensis Van Someren, 1939 are placed in the synonymy of maranensis. VERMICULATED GROUP F. bicalcaratus It has been shown that birds become darker and more heavily patterned towards the south- east of the range, but the differences exhibited in this cline do not seem great enough to warrant separating the birds of the extreme south-east from those of the southern (British) Cameroons. Accordingly molunduensis Grote, 1949, is synonymised with ogilviegranti. F. icterorhynchus The clinal variation does not warrant more than a division into two subspecies. Accordingly grisescens is placed in the synonymy of icterorhynchus and emini and ugandensis in the synonymy of dybowskii. Birds showing a few chestnut feathers on the flanks, the character on which ugandensis was founded, occur sporadically in south-eastern Uganda and must be considered aberrant specimens perhaps indicating past hybridisation with clappertoni. F. clappertoni F. tschadensis is placed in the synonymy of F. c. clappertoni (see text). The clinal variation shown in southern populations does not warrant extensive splitting; accordingly cavei Macdonald, 1940, and testis are placed respectively in the synonymy of gedgii and sharpii. F. hildebrandti Numerous subdivisions are not practical; accordingly helleri is placed in the synonymy of hildebrandti, and grotei and lindi in that of johnstoni. F. natalensis The name thamnobium Clancey, 1953, was proposed to distinguish paler birds inhabiting drier areas from those of wetter habitats in Natal and the Zoutspansberg (natalensis), but since this THE FRANCOLINS, A STUDY IN SPECIATION 177 gives natalensis a discontinuous range it is not practical to recognise the differences by name. F. hartlaubi ovambensis is considered a synonym of bradfieldi. F. adspersus kalahari is considered a synonym of adspersus. STRIATED GROUP F. sephaena The name somaliensis Grant & Praed, 1934, was given to streaked birds from Somaliland and was based on the misapprehension that the type of spilogaster was an unstreaked bird (Rand, 1950: 384). As noted in the text I believe spilogaster must be used for the hybrid populations from Somaliland to southern Kenya. However, if it is established that there still exists an unmixed population of streaked birds on the coastal plains which is distinct from the southern rovuma then the name somaliensis is available. Such a population can be assumed to have existed once since it must be regarded as one of the parents of the hybrid spilogaster but it may have been integrated into the hybrid population. Local variations in size, colour and pattern throughout the species do not seem great enough to distinguish by name. Accordingly zuluensis is placed in the synonymy of sephaena : thompsoni, chobiensis, mababiensis in the synonymy of zambesiae : schoanus, ochrogaster Hartlaub, 1882, delutescens, and jubaeensis in the synonymy of grantii: Peters (1934: 73) shows that rovuma antedates kirkii. RED-WINGED GROUP F. psilolaemus On the author's own showing fricki is regarded as a synonym of F. p. ellenbecki (Friedmann, 1930: 113). F. shelleyi There is not sufficient variation between the populations of South Africa and Tanganyika to justify the recognition of trothae which is regarded as a synonym of shelleyi, as is sequestris Clancey, 1960. F. africanus Praed & Grant have been among the few modern authors who have correctly regarded Perdix afra Latham, 1790, as the senior name of the South African Grey wing, known by others as F. africanus Stephens, 1890. However, since the submersion of Pternistis in Francolinus, Perdix afra Latham is pre-occupied by Tetrao afer Muller 1776, the senior name for the bare- throated francolin formerly known as Pternistis afer. The Greywing must therefore now be known as F. africanus. Within the range of the species there is slight local variation but it is not great enough for me to recognise any subspecies: proximus Clancey, 1957, i- s therefore placed in the synonymy. F. levaillantoides Roberts (1936: 321) shows that F. levaillantoides is the original name for this species and gariepensis is a synonym. In this species the haphazard distribution of rufous and grey micro-populations in the southern districts of the Sudan and Abyssinia and in Angola and South West Africa makes it difficult to know how many subspecies can usefully be recognised. I believe it is impractical to attempt to differentiate between these variations if no other differences are shown, and, furthermore, names should not be used for intermediate populations. Accordingly I recognise in the south (a) F. I. levaillantoides. All districts east and south of the Kalahari, intergrading with kalaharica on the fringes of the Kalahari. Richly coloured (even in the greyest specimens) and heavily patterned. Synonyms, ludwigi and gariepensis. (b) F. I, kalaharica (Roberts) 1932. The Kalahari, Pale and grey, birds from central 178 B. P. HALL districts almost unmarked on the abdomens; birds from the fringes rather darker with more markings on the abdomen. Synonym, langi Roberts, 1932. (Although langi has line priority over kalaharica I use kalaharica, as I am entitled to do as first revisor, since the name was given to a very pale, grey bird from the Damara Pan near the centre of the Kalahari, and thus represents the extreme of variation, while langi was given to a rather darker bird from Nkate near the eastern border.) (c) F. I. pallidior. South West Africa. Either rufous or grey but the rufous birds paler in tone than levaillantoides, and the grey birds darker and more heavily marked than kalaharica. As known at present the population of the Windhuk area is more rufous, and that of the Tsumeb area greyer (approaching kalaharica). Synonym wattii Macdonald, 1953. (See Macdonald, 1953, for detailed discussion.) (d) F. I. jugularis. Southern Angola to the Cunene. Differs from all other subspecies in having a broad black-and-white gorget: typical birds from the Benguela area are small and grey but those from the Cunene are larger and more rufous showing an approach to some of the populations of pallidior. Synonyms, cunenensis Roberts, 1932, stresemanni Hoesch & Niethammer, 1940. In the north (e) F. 1. archeri. Extreme southern Sudan to the Golis range of western (British) Somaliland. Similar to F. I. pallidior in having both rufous and grey populations but in both the abdomens are paler and commonly marked with fine lines. Synonyms friedmanni Grant & Praed, 1934, stantoni Cave, 1940. (/) F. I. lorti. Mountains of eastern (British) Somaliland. Very grey, with a very pale abdomen fairly heavily marked with fine lines. (g) F. I. gutturalis. Northern Abyssinia and Eritrea. Quite distinct (see text). Synonym eritrea Zedlitz. F. levaillantii A number of names have been given to the birds of the northern and western populations which have no black-and-white stripe down the ochre collar on the hind neck. However, while the few specimens available indicate that there is a tendency for Angola birds to be paler, the difference is trivial and the series otherwise is remarkably uniform. Accordingly mulemae, adolffriederici, benguellensis Neumann, 1908, clayi White, 1944, and momboloensis White, 1952 (nom. nov. for benguellensis) are placed in the synonymy of kikuyuensis . THE RED-TAILED GROUP F. coqui In the populations of central and south-eastern African there is considerable individual and local variation in size and colour so that consistent geographical variation is hard to determine. I do not believe it is practical to distinguish from nominate coqui the small coastal birds of Kenya since equally small birds are found in Natal, or to distinguish the populations of Natal, Zululand and southern Mozambique, which on the whole are less rufous, for individuals can be matched with others elsewhere. The names stuhlmanni and campbelli are therefore placed in the synonymy of coqui. I regard lynesi Sclater, 1932, also as a synonym of coqui since the type is intermediate between coqui and kasaicus but comes from an area in which the majority of specimens are closest to coqui. For the Uganda race the original spelling " ruahdae " has been used since, while the author may have intended the name to be " ruandae ", there is no clear evidence of this intention in the original publication. (Int. Code of Zool. Nomen. 1961, Article 32a (ii) .) F. coqui ruandae Jackson, 1938, is accordingly a junior objective synonym. (Article 33a (ii) .) F. albogularis Serle (1957: 388) shows that gambagae must be considered a synonym of buckleyi. The isolated populations of Marungu and Upemba in the south-eastern Congo, the Balovale district of Northern Rhodesia, and of eastern Angola seem to show an increase from east to west in the amount of patterning in the females, and the Angola birds seem also to be darker, THE FRANCOLINS, STUDY IN SPECIATION 179 less rufous, than others (Hall, ig6oa: 409). Too few specimens are available at present on which to judge the extent of variation, and until there are more, especially of topotypical meinertzhageni from Rhodesia, it seems best to use meinertzhageni for Rhodesian and Angola birds and dewittei for the Congo birds, though I believe that the Rhodesian birds may prove to be closer to those from the Congo than to those from Angola. F. schlegelii Traylor (1960: 86) has shown that confusus Neumann, 1933, should be considered a synonym of schlegelii. SPECIES NOT INCLUDED IN ANY GROUP F. pondicerianus It has been shown that variation in this species is slight and largely clinal and ecological. It is not practical to recognise more than one intermediate subspecies. Accordingly paganus Koelz, 1954, ti* ar Koelz, 1954, prepositus Koelz, 1954, are au placed in the synonymy of interpositus. SPOTTED GROUP MAP 1 The shaded areas represent the overall ranges of the species. The thick red line indicates divisions between species. The thin red line indicates divisions between subspecies, the zig-zag line denoting intergradation. The red crosses indicate localities from which hybrid specimens have been obtained. F. francolinus F. pictus 1. francolinus 2. arabistanicus 3. henrici 4. asiae 5. melanotus 6. pallidus 7. pictus 8. watsoni F. pintadeanus 9. phayrei 10. pintadeanus MAP. IA. Hypothetical distribution of the ancestral stock of the three species during a glaciation (Stage 2). BARE- THROATED GROUP MAP 2 Overall ranges are shown rather than precise collecting localities. Thick red lines denote specific boundaries. Thin red lines denote subspecific boundaries, plain lines being boundaries between non-intergrading subspecies, zig-zag lines boundaries between intergrading subspecies, toothed lines boundaries of hybrid zones. The red " X " marks the area of hybrids swainsonii x afer. Numbers have been given only to species and subspecies which are important to the discussion on speciation, but the ranges of other subspecies are indicated. i. F. leucoscepus (leucoscepus, infuscatus) iiillii-iiililiiliil 2 - F. rufopictus V.V.'.'.'.'.J 3. F. swainsonii (swainsonii, lundazi, damarensis, gilli) IB F. afer. Black-and-white, " afer-type " subspecies 4. castaneiv enter & notatus 5. lehmanni 6. swynnertoni 6x7. hybrids (humboldtii) 7. melanogaster & loangwae 8. leucoparaeus 9. afer Vermiculated, " cranchii-type " subspecies 10. cranchii & intercedens 11. harterti Hybrids between " afer-type " and " cranchii-type " subspecies. MAP 2A. Hypothetical distribution of the ancestral stock of the four species during a spread of forest (Stage 4). MAP 2B. Hypothetical distribution of the ancestral stock of the four species and two main blocks of afer subspecies during a later dry period (Stage 6) . MONTANE GROUP MAP 3 The letters indicate precise localities from which specimens have been collected. Plain red lines denote divisions between isolates which are regarded as species, and broken lines between those regarded as subspecies. The thickness of the lines is relative to the degree of divergence between neighbouring forms. e F. erckelii i. pentoni 2. erckelii o F. ochropectus 3. ochropectus c F. castaneicollis 4. ogoensis 5 . castaneicollis 6. kaffanus 7. atrifrons j F. jacksoni 8. jacksoni g. pollenorum n F. nobilis 10. nobilis 1 1 . chapini x F. camerunensis 12. camerunensis s F. swierstrai 13. swierstrai MAP 3A. Hypothetical distribution of montane forest (shaded black) in a cold, wet era (Stage i). ZOOL. 10, 2. SCALY GROUP MAP 4 A continuous distribution has been shown for members of the Scaly Group through the Upper and Lower Guinea forests, along the escarpment of Angola, and in the wetter areas of Uganda and Kenya, for within these areas the birds may be expected in suitable clearings or cultivations. Outside this range only general localities from which specimens have been collected have been shaded. Broad red lines indicate divisions between species : thinner red lines denote divisions between subspecies, those that intergrade being indicated by a zig-zag line. Thicknesses of the subspecific lines are relative to the divergence shown by neighbouring forms. F. ahantensis 1. hopkinsoni 2. ahantensis F. squamatus - griseostriatus 3. squamatus 4. schuetti 5. maranensis 6. usambarae 7. uzungwensis 8. doni MAP 4A. Hypothetical map showing extent of lowland forest in a warm, wet era (Stage 5). Limits of forest shown by broken line. ZOOL. IO, 2 VERMICULATED GROUP MAP 5 The distribution of many members of this group is known to be discontinuous. Accordingly only general areas from which specimens have been collected are shaded, but the birds may be expected to occur in some of the intervening areas in which little collecting has been done. Thick red lines denote divisions between species: where there is evidence of hybridisation these lines are toothed. Thin red lines denote divisions between subspecies, those that intergrade being shown with a zig-zag line. The red " X " denotes a hybrid between hildebrandti and natalensis. The red " T " denotes the aberrant or hybrid specimen F. tschadensis. The red " U " denotes the area in which the aberrant " ugandensis "-type birds are found. F. bicalcaratus i. ayesha 2. bicalcaratus 3. adamauae 4. ogilviegranti 5. thornei F. clappertoni 6. clappertoni 8. heuglini 9. sharpii 10. konigseggi 11. nigrosquamatus F. icterorhynchus 12. dybowskii 13. icterorhynchus F. hildebrandti F. natalensis F. hartlaubi F. capensis F. adspersus F. harwoodi 14. altumi 15. hildebrandti 1 6. johnstoni 17. neavei 1 8. natalensis (hartlaubi, bradfieldi, crypticus} tfll STRIATED GROUP MAP 6 Only general areas from which birds have been collected are shaded. The broken thick red line indicates the outer limits of the range of F. streptophorus , and the unbroken thick red line the outer limits of the range of F. sephaena. Thin red lines indicate subspecific divisions, plain lines denoting divisions without intergrading, zig-zag lines denoting intergrading, and toothed lines denoting the boundary of a hybrid area. F. streptophorus F. sephaena i. sephaena 2. zambesiae 3. grantii 3x4. spilogaster 4. rovuma MAP 6A. Hypothetical distribution of the ancestral stock of the two species at the end of a dry era (Stage 4). Arrows indicate the line of advance of encroaching woodlands. ,vif RED -WINGED GROUP MAP 7 The distribution of all members of this group is discontinuous, accordingly only general areas from which birds have been collected are shaded. Thick red lines denote the extreme limits of the range of each species : these lines are broken where they link remotely isolated populations, except in F. levaillantoides where arrows indicate where the lines linking the northern and southern populations might fall. Lineal divisions between subspecies are not shown but different shading has been used for those subspecies which are strongly differentiated: others are numbered. The red cross indicates the Amani specimen which is intermediate between F. s. shelleyi and F, s. uluensis. F. psilolaemus (psilolaemus, ellenbecki, elgonensis, theresae] F. shelleyi uluensis (including macarthuri) F. shelleyi shelleyi F. shelleyi whytei F. africanus F. levaillantoides i. levaillantoides 2. kalaharica 3. pallidior 4. jugularis 5. archeri 6. lorti F. levaillantoides gutturalis F. levaillantii F. finschi 7. levaillantii 8. crawshayi g. kikuyuensis MAP yA. Mean Annual Rainfall (taken from Fitzgerald 1950: 35) under 4 inches MAP ye. Diagrammatic map showing hypothetical colonisation routes and centres of speciation (Stages 3-6). ad RED-TAILED GROUP MAP 8 The distribution of all members of this group is discontinuous and only general areas from which specimens have been collected are shaded. Thick red lines denote the extreme limits of the ranges of species, those linking remotely isolated populations being broken, except between the north-western and southern populations of F. albogularis where the link is indicated by arrows. Subspecific divisions are shown lineally only in the critical area of Kenya and Tanganyika. Here a toothed line surrounds a population which appears to be a result of hybridisation but which does not now intergrade with neigh- bouring subspecies: zig-zag lines indicate intergrading. Outside this critical area all subspecies, if not isolates, intergrade. F. coqui W///A F. schlegelii F. albogularis 1. coqui 2. vernayi 3. hoeschianus 4. angolensis 5. kasaicus 6. ruahdae 7. hubbardi 8. thikae 9. maharao 10. spinetorum 11. albogularis 12. buckleyi 13. dewittei 14. meinertzhageni MAP SA. Hypothetical distribution of the ancestral stock of F. coqui and proto-albogularis/ schlegelii (shown by mixed diagonal and horizontal shading) before the two latter species diverged (Stage 3). Distribution of lowland forest shown in black. MAP SB. Hypothetical distribution of the ancestral stock of the three species in a sub- sequent dry era (Stage 4). SPECIES NOT INCLUDED IN ANY GROUP MAP 9 F. nahani F. lathami 1. lathami 2. schubotzi D F. pondicerianus F. gularis MAP 10 1. mecranensis 2. interpositus 3. pondicerianus 4. ceylonensis 10 VEGETATION MAP OF AFRICA MAP 11 Adapted from " Vegetation Map of Africa south of the Tropic of Cancer", Oxford University Press, 1959. The numbering and terminology of the original map have been used as far as possible, excluding Madagascar. Explanatory notes indicating geographical areas are given in the key in cases where it is not possible to differentiate on the map between vegetation types. KEY 1-4. EVERGREEN FORESTS & MONTANE COMMUNITIES. 1. Montane Evergreen Forest. 2. Temperate and Subtropical Evergreen Forest. Knysna and other districts of the extreme south only, 3. Montane Communities undifferentiated. These include evergreen forest (type i), grassland (type 5) and woodland. 4. Montane Communities with alfro-alpine communities. High moun- tains of central and east Africa and Ethiopia only. (It has not been possible on this scale to distinguish between types i , 3, and 4.) 5-6. MONTANE & HIGH ALTITUDE GRASSLAND. 5. Montane Grassland above 6,000 ft. North-eastern and central Africa. 6. Temperate and Subtropical Grassland. Pure grassland above 3,5oo//. in South Africa. MOIST FOREST AT LOW AND MEDIUM ALTITUDES. FOREST-SAVANNA MOSAIC. 8. Moist forest surrounded by savanna of tall grasses. Found in belts surrounding type 7. 9. Coastal Forest-Savanna Mosaic. East coast only. 10. DRY & DECIDUOUS FOREST with abundant Baikiaea plurijuga. Found only in parts of N. & S. Rhodesia, Bechuanaland and the Caprivi Strip, and sometimes regarded as woodland rather than forest. 12 & 14. THICKETS. 12. Itigi type. Central Tanganyika only. 14. Ethopian evergreen type. Ethiopia only. 15. CAPE MACCHIA. 16. RELATIVELY MOIST WOODLANDS. Types in which the genera Isoberlinia, Brachystegia, and Julbernadia are absent or rare. 17-19. WOODLANDS (interspersed with Savanna). 17. With abundant Isoberlinia doka and /. dalzielii. Northern areas. 18. With abundant Brachystegia and Julbernadia " myombo ". Tan- ganyika to central Northern Rhodesia, Southern Rhodesia and Mozambique. 19. Similar to type 18 but with a number of distinctive species. Prin- cipally on Kalahari sand. Angola to western N. Rhodesia. 20-22. RELATIVELY DRY WOODLANDS & SAVANNAS. 20. With frequent Acacia but other species as well, and savannas of tall grass with certain species of Acacia. 21. Ethiopian types. A heterogeneous assemblage of low deciduous shrubs on stony ground. Ethiopia only. 22. With abundant Colospermum mopane. Southern Africa only. 24. GRASS STEPPE ON KALAHARI SAND. 25. STEPPES. 25. Wooded steppe with abundant Acacia and Commiphora. 26. Grass steppe with thicket clumps. Western Uganda only. 27. Grass steppe Luanda type. Coastal Angola only. 28-31. KAROO & SUBDESERT STEPPE. 28. Karoo Succulent Steppe. Western Cape Province only. 29. Karoo shrub and grass. Cape Province and southern South West Africa. 30. Transitional and mixed Karoo. East-central Cape Province. 31. Northern areas and northern South West Africa and southern Angola. 32. DESERT. 33- SWAMPS. 200 B. P. HALL REFERENCES BANNERMAN, D. A. 1930. The birds of Tropical West Africa, i. London. BENSON, C. W. 1945. Notes on the birds of southern Abyssinia (part). Ibis 87: 366-400. - 1948. Geographical voice-variation in African birds. Ibis 90: 48-71. 1951. Breeding and other notes from Nyasaland and the Lundazi district of Northern Rhodesia. Bull. Mus. Comp. Zool. Harv. 106: 69114. 1958. Some additions and corrections to a " Check List of the Birds of Northern Rhodesia ". Oc. Pap. Nat. mus. S. Rhod. 22b: 190-197. BENSON, C. W. & F. M. 1948. Notes from southern Nyasaland mainly from the Lower Shire Valley at 200 ft. altitude. Ostrich 29: 1-16. BENSON, C. W., IRWIN, M. P. S., WHITE, C. M. N. in press. The significance of valleys as avian zoogeographical barriers. Symposium on causes and problems of animal distribution with special reference to southern Africa. July, 1961. Zool. Soc. of southern Africa. BENSON, C. W. & WHITE, C. M. N. 1957. Check List of the Birds of Northern Rhodesia. Lusaka. BENSON, C. W. & WHITE, C. M. N. in press. Discontinuous Distributions (Aves). BOWEN, W. W. 1930. The relationships and distribution of the bare-throated francolins (Pternistis). Proc. Acad. Nat. Sci. Phil. 82: 149-164. CHAPIN, J. P. 1932. The birds of the Belgian Congo. I. New York. CLANCEY, P. A. 1953. Miscellaneous taxonomic notes on African birds. Durban Mus. Nov. 4(4): 57-64. COLLIER, F. S. 1935. The Ahanta Francolin in Ibadan Province, Nigeria. Ibis (13) 5: 665-666. DELACOUR, J. 1951. Commentaires, modifications et additions a la liste des oiseaux de 1'Indochine Francaise (2). L'Oiseau (5) 21: 1-32. DORST, J. & JOUANIN, C. 1954. Precisions sur la position systematique et 1'habitat de Francolinus ochropectus. L'Oiseau (5) 24: 161-170. FITZGERALD, W. 1950. Africa (seventh edition). London. FRIEDMANN, H. 1930. Birds collected by the Childs Frick expedition to Ethiopia and Kenya Colony. Pt. i. Non-passeres. U.S. Nat. Mus. Bull. 153. GOOD, A. I. 1952. The birds of French Cameroon (Part i). Mem. Inst. Fran. Afr. Noire Nat. 2: 1-203. GRANT, C. H. B. & PRAED, C. W. M. 1946. Notes on Eastern African Birds. Bull. Brit. Orn. Cl. 66: 74-81. HALL, B. P. 1956. Notes on a small collection of birds from Panda Matenga, N.E. Bechuana- land. Ostrich 27: 96-109. HALL, B. P., ig6oa. The ecology and taxonomy of some Angola birds. Bull. Brit. Mus. (Nat. Hist.) Zool. 6 (7): 370-453. I96ob. The faunistic importance of the scarp of Angola. Ibis. 102: 420-442. HARRISON, C. J. O. In press. The incidence and origin of spotted patterns in the Estrildidae. Ibis. 105 HOPKINSON, E. 1923. The game-birds and pigeons of the Gambia. Avic. Mag. (i) 4: 125-31, 166-72, 183-7, 207-15, 23 1 -?- HORSBURGH, B. 1912. The game-birds and waterfowl of South Africa. London. JACKSON, F. 1938. The birds of Kenya Colony and the Uganda Protectorate. Vol. i. LAVAUDEN, L. 1936. Les Francolins. Alauda 8: 301-315. LILFORD, LORD. 1862. On the extinction in Europe of the Common Francolin (Francolinus vulgaris). Ibis. (i). 4: 352-356. LONG, R. C. 1960. The birds of the Port Herald District (i). Ostrich. 31: 85-104. - 1961. Interesting specimens of birds in the Nyasaland Museum. Nyasa Mus. Ann. Rep. and Bull. 1960-61 : 25-27. MACDONALD, J. D. 1953. The races in South West Africa of the Orange River Francolin. Bull. Brit. Orn. Cl. 73: 34-36. 1957- Contribution to the ornithology of western South Africa. London. THE FRANCOLINS, A STUDY IN SPECIATION 201 MALBRANDT, R. & MACLATCHY, A. R. 1949. Faune de 1'Equateur Africain Fran9ais. (I). Oiseaux. Enc. biol. Paris 35. MARKS, M. B. 1959. Red-necked Francolin Pternistis afer in Bechuanaland. Ostrich 30: 44. MCLACHLAN, C. R. & LIVERSIDGE, R. 1957- Roberts' birds of South Africa. Cape Town. MOREAU, R. E. 1930. On the age of some races of birds. Ibis, (12), 6: 229-239. in press. Vicissitudes of the African biota in the Late Pleistocene. Proc. Zool. Soc. London. PETERS, J. L. 1934. Check list of birds of the world. 2. Cambridge, U.S.A. PRAED, C. W. M. & GRANT, C. H. B. 1952. Birds of eastern and north eastern Africa. London. RAND, A. L. 1950. On the name Francolinus sephaena spilogaster Salvadori. Auk 67 384-385- - 1951. Birds from Liberia. Fieldiana: Zoology 32 (9): 561-653. ROBERTS, A. 1936. Some unpublished field-notes made by Dr. (Sir) Andrew Smith. Ann. Trans. Mus. 18: 271-323. SERLE, W. 1957. A contribution to the ornithology of the eastern region of Nigeria. Ibis 99 : 371-418, 628-685. SMITH, K. D. 1957. An annotated check list of the birds of Eritrea. Ibis 99: 1-337. SMITHERS, R. H. N., IRWIN, M. P. S. & PATTERSON, M. L. 1957. A check list of the birds of Southern Rhodesia. Cambridge. TRAYLOR, M. A. ig6oa. Notes on the birds of Angola, non-passeres. Publ. cult. Co. Diam. Ang. Lisboa 51: 129-186. I96ob. Francolinus schlegelii Heuglin in Cameroon. Bull. Brit. Orn. Cl. 80: 86-88. VAN SOMEREN, V. G. L. 1925. The birds of Kenya and Uganda. Part 2. Phasianidae. Journ. E. Afr. & Ug. Nat. Hist. Soc. 23: 95-104. VAN TYNE, J. & BERGER, A. J. 1959. Fundamentals of Ornithology. New York. VERHEYEN, R. 1953. Exploration du Pare National de 1'Upemba, Mission G. F. de Witte. Fasc. 19. Oiseaux. VINCENT, J. 1934. The birds of northern Portuguese East Africa. Part III. Ibis, (13.) 4: 305-340. WHITE, C. M. N. 1952. On the genus Pternistis. Ibis 94: 306-309. WINTERBOTTOM, J. M. & CouRTENAY-LATiMER, M. 1 961. A list of the birds of Little NAMAQUALAND. S. Afr. Av. Ser. Percy FitzPatrick Inst. Afr. Orn. 1: 1-46. INDEX TO SPECIFIC AND SUBSPECIFIC NAMES OF FRANCOLINUS Specific and subspecific names of francolins are listed under the nomenclature used in this paper, those of species and subspecies here recognized appearing in roman type, others in italics. No attention is paid to other generic names that have been given to francolins except in a few cases where the merging of these names in Francolinus has created homonyms. For any species that is here regarded as a member of a species group all the pages of the section dealing with the whole group are listed, regardless of whether the species is actually mentioned on each page. adamauae, F. bicalcaratus, 144, 188 afra, Perdix (F. afer), 152, 155, 173, 177 adolffriederici, F. levaillantii, 178 africanus, F., no, in, 117, 120, 152-160, adspersus, F., no, 113, 117, 120, 121, 140- 169, 173, 177, 192 148, 169, 176, 188 ahantensis, F., no, in, 116, 136-140, 167, afer, F., no, in, 113, 116, 119, 121, 122, 168, 169, 172, 186 125-131, 132, 140, 145, 146, 168, 169, 175, ahantensis, F. a., [136], 186 177, 182 albogularis, F., no, 114, 117, 122, 160-165, afer, F. a., 127, 175, 182 169, 170, 178, 194 202 B. P. HALL albogularis, F. a., 162, 194 altumi, F. hildebrandti, 142, 146, 188 angolensis, Pternistis afer, 175 angolensis, F. coqui, 163, 175, 194 arabistanicus, F. francolinus, 124, 180 archeri, F. levaillantoid.es, 155, 156, 178, 192 asiae, F. francolinus, 124, 175, 180 atrifrons, F. castaneicollis, 112, 133, 135, 136, 184 ayesha, F. bicalcaratus, 141, 144, 147, 188 aylwinae, F. afer, 175 benguellensis , F. after, 175 benguellensis, F. levaillantii, 178 bicalcaratus, F., no, 117, 118, 120, 121, 123, 140-148, 161, 164, 165, 169, 171, 172, 176, 1 88 bicalcaratus, F. b., 144, 188 billypaynei, F. francolinus, 175 bogdanovi, F. francolinus, 175 bohmi, F. afer, 175 boineti, F., 175 bottegi, F. castaneicollis, 176 bradfieldi, F. hartlaubi, 145, 177, 188 buckleyi, F. albogularis, 162, 178, 194 camabatelae, F. afer, 175 camerunensis, F., no, 116, 118, 120, 131-136, 169, 184 campbelli, F. coqui, 178 capensis, F., no, 117, 140-148, 169, 188 castaneicollis, F., no, in, 116, 131-136, 169, 176, 184 castaneicollis, F. c., 133, 135, 176, 184 castaneiventer, F. afer, 128, 130, 131, 175, 182 caucasicus, F. francolinus, 1 75 cavei, F. clappertoni, 176 ceylonensis, F. pondicerianus, 167, 196 chapini, F. nobilis, 134, 184 chio, F. afer, 175 chobiensis, F. sephaena, 177 chobiensis, Pternistis swainsonii, 175 chyuluensis, F. squamatus, 176 clappertoni, F., no, 114, 117, 118, 119, 120, 121, 140-148, 161, 164, 165, 169, 170, 176, 188 clappertoni, F. c., 141, 144, 176, 188 clayi, F. levaillantii, 178 confusa, F. squamatus, 176 confusus, F. schlegelii, 179 cooperi, Pternistis, 128 coqui, F., no, in, 117, 118, 119, 121, 122, 145, 147, 160-165, 167, 169, 170, 171, 178, 194 coqui, F. c., 161, 162, 163, 164, 178, 194 cranchii, F. afer, 127, 168, 175, 182 crawshayi, F. levaillantii, 157, 192 cruzi, F., 176 crypticus, F. hartlaubi, 145, 188 cunenensis, F. levaillantoides, 178 cunen sis, Pternistis afer, 175 damarensis, F. swainsonii, 129, 182 delutescens, F. sephaena, 177 dewittei, F. albogularis, 162, 179, 194 doni, F. squamatus, 137, 139, 186 dowashanus, F. squamatus, 176 dybowskii, F. icterorhynchus, 144, 176, 188 elgonensis, F. psilolaemus, 153, 154, 192 ellenbecki, F. psilolaemus, 153, 177, 192 emini, F. icterorhynchus, 176 erckelii, F., no, in, 116, 120, 131-136, 141, 143, 145, 146, 169, 184 erckelii, F. e., 135, 184 eritrea, F. levaillantoides, 178 festinus, F . francolinus , 175 finschi, F., no, 117, 119, 122, 152-160, 169, 170, 192 francolinus, F., no, in, 116, 118, 119, 122, 123-125, 165, 166, 167, 168, 169, 175, 180 francolinus, F. 1, 124, 175, 180 fricki, F. psilolaemus, 177 friedmanni, F. levaillantoides, 155, 178 gambagae, F. albogularis, 178 gariepensis, F., 155, 173, 177 gedgii, F. clappertoni, 144, 176, 188 gilli, F. swainsonii, 129, 182 gofanus, F. castaneicollis, 176 grantii, F. sephaena, 150, 151, 152, 177, 190 grisescens, F. icterorhynchus, 176 griseostriatus, F., no, 116, 136-140, 167, 169 grotei, F. hildebrandti, 176 gularis, F., 109, no, 117, 167-168, 169, 196 gurae, F.jacksoni, 176 gutturalis, F. levaillantoides, 155, 156, 159, 160, 178, 192 harterti, F. afer, 127, 182 hartlaubi, F., no, 117, 120, 121, 140-148, 169, 177, 188 hartlaubi, F. h., 144, 188 harwoodi, F., no, 117, 118, 120, 140-148, 169, 1 88 helleri, F. hildebrandti, 176 henrici, F. francolinus, 124, 175, 180 THE FRANCOLINS, A STUDY IN SPECIATION heuglini, F. clappertoni, 144, 188 hildebrandti, F., no, in, 117, 118, 140-148, 1 68, 169, 176, 1 88 hildebrandti, F. h., 142, 176, 188 hoeschianus, F. coqui, 163, 194 holtemulleri, F. leucoscepus, 175 hopkinsoni, F. ahantensis, 136, 186 hubbardi, F. coqui, 161, 162, 164, 194 humboldtii, F. afer, 127, 175, 182 icterorhynchus, F., no, 114, 117, 120, 122, 140-148, 163, 164, 165, 169, 171, 176, 188 icterorhynchus, F. i., 144, 176, 188 infuscatus, F. leucoscepus, 129, 175, 182 intercedens, F. afer, 127, 175, 182 intermedius, F., 124 interpositus, F. pondicerianus, 167, 179, 196 itigi, F. afer, 175 jacksoni, F., no, 116, 118, 120, 131-136, 169, 176, 184 jacksoni, F. j., 133, 176, 184 johnstoni, F. hildebrandti, 142, 176, 188 jubaeensis, F. sephaena, 177 jugularis, F. levaillantoides, 156, 178, 192 kaffanus, F. castaneicollis, 133, 134, 135, 176, 184 kalahari, F. adspersus, 176 kalaharica, F. levaillantoides, 156, 177, 178, 192 kapitensis, F. squamatus, 176 kasaicus, F. coqui, 163, 178, 194 keniensis, F. leucoscepus, 175 keniensis, F. squamatus, 175, 176 kikuyuensis, F. levaillantii, 157, 178, 192 kilimensis, F. leucoscepus, 175 kirkii, F. sephaena, 177 konigseggi, F. clappertoni, 144, 188 krebsi, F. afer, 175 langi, F. levaillantoides, 178 lathami, F., 109, no, 117, 119, 123, 165-166, 169, 196 lathami, F. 1., [166], 196 lehmanni, F. afer, 128, 131, 182 leucoparaeus, F. afer, 127, 131, 182 leucoscepus, F., no, 113, 116, 119, 122, 125- 131, 169, 175, 182 leucoscepus, F. 1., 129, 182 levaillantii, F., no, 117, 121, 122, 152-160, 168, 169, 178, 192 levaillantii, F. 1., 157, 192 203 levaillantoides, F., no, 113, 117, 121, 122, 152-160, 168, 169, 173, 177, 192 levaillantoides, F. 1., 156, 177, 178, 192 lindi, F. hildebrandti, 176 loangwae, F. afer, 127, 129, 182 lorti, F. levaillantoides, 155, 156, 178, 192 ludwigi, F. levaillantoides, 177 lundazi, F. swainsonii, 129, 175, 182 lynesi, F. coqui, 178 mababiensis , F. sephaena, 177 macarthuri, F. shelleyi, 154, 192 mackenziei, F. afer, 175 maharao, F. coqui, 161, 164, 165, 167, 194 manueli, F. afer, 715 maranensis, F. squamatus, 137, 176, 186 mecranensis, F. pondicerianus, 167, 196 meinertzhageni, F. albogularis, 162, 179, 194 melanogaster, F. afer, 127, 129, 131, 175, 182 melanotus, F. francolinus, 124, 180 molunduensis, F. bicalcaratus, 176 momboloensis, F. levaillantii, 178 muhamedbenabdullah, F. leucoscepus, 175 mulemae, F. levaillantii, 178 nahani, F., 109, no, 117, 120, 123, 166-167, 169, 196 natalensis, F., no, 111, 117, 121, 128, 131, 140-148, 168, 169, 176, 188 natalensis, F. n., 117, 142, 176, 177, 188 neavei, F. natalensis, 117, 142, 143, 146, 188 nigrosquamatus, F. clappertoni, 144, 188 nobilis, F., no, 116, 120, 131-136, 169, 176, 184 nobilis, F. n., [134], 176, 184 notatus, F. afer, 128, 131, 175, 182 nyanzae, F. afer, 175 ochrogaster, F. sephaena, 177 ochropectus, F., 107, no, in, 112, 116, 131- 136, 169, 184 ogilviegranti, F. bicalcaratus, 144, 176, 188 ogoensis, F. castaneicollis, 133, 184 oldowai, F. leucoscepus, 175 ovambensis, F. hartlaubi, 177 paganus, F. pondicerianus, 179 pallidior, F. levaillantoides, 156, 178, 192 palliditectus, F. afer, 175 pallidus, F. pictus, 124, 180 parkerae, F. francolinus, 175 patrizii, F. castaneicollis, 176 pentoni, F. erckelii, 132, 135, 145, 184 204 B. P. HALL phayrei, F. pintadeanus, 124, 180 pictus, F., no, in, 116, 119, 120, 123-125, 167, 169, 1 80 pictus, F. p., 124, 1 80 pintadeanus, F., no, 116, 118, 123-125, 169, 175, 180 pintadeanus, F. p., [124], 175, 180 pollenorum, F. jacksoni, 133, 184 pondicerianus, F., 109, no, 117, 122, 167, 169, 179, 196 pondicerianus, F. p., 167, 196 prepositus, F. pondicerianus, 179 proximus, F. africanus, 177 psilolaemus, F., no, 117, 121, 122, 152-160, 169, 177, 192 psilolaemus, F. p., 153, 192 punctulata, F. afer, 175 rovuma, F. sephaena, no, 150, 177, 190 ruahdae, F. coqui, 162, 164, 178, 194 ruandae, F. coqui, 176, 178 ruandae, F. nobilis, 176 rufopictus, F., no, 113, 116, 125-131, 169, 170, 182 sarudnyi, F. francolinus, 175 schlegelii, F., no, 114, 117, 118, 160-165, 1 66, 169, 170, 179, 194 schoanus, F. sephaena, 177 schubotzi, F. lathami, 166, 196 schuetti, F. squamatus, 137, 138, 139, 176, 186 sephaena, F., no, in, 113, 117, 118, 121, 123, 148-152, 167, 169, 177, 190 sephaena, F. s., 150, 152, 177, 190 sequestris, F. shelleyi, 177 sharpii, F. clappertoni, 144, 147, 176, 188 shelleyi, F., no, in, 113, 115, 117, 118, 121, 122, 152-160, 168, 169, 173, 177, 192 shelleyi, F. s., in, 113, 122, 153, 154, 155, 159, 160, 168, 177, 192 somaliensis, F. sephaena, 177 spilogaster, F. sephaena, 150, 177, 190 spinetorum, F. coqui, 114, 161, 162, 165, 194 squamatus, F., no, in, 116, 118, 121, 136- 140, 1 66, 1 68, 169, 176, 1 86 squamatus, F. s., 137, 138, 139, 186 stantoni, F. levaillantoides, 155, 178 streptophorus, F., 110, 114, 117, 120, 148- 152, 167, 169, 170, 171, 172, 190 stresemanni, F. levaillantoides, 178 stuhlmanni, F. coqui, 178 swainsonii, F., no, 116, 121, 122, 125-131, 169, 175, 182 swainsonii, F. s., 127, 129, 130, 182 swierstrai, F., no, 116, 131-136, 169, 176, 184 swynnertoni, F. afer, 127, 128, 131, 175, 182 testis, F. clappertoni, 176 tetraoninus, F. squamatus, 176 thamnobium, F. natalensis, 176 theresae, F. psilolaemus, 153, 154, 192 thikae, F. coqui, 161, 162, 164, 194 thompsoni, F. sephaena, 177 thornei, F. bicalcaratus, 144, 188 titar, F. pondicerianus, 179 tokora, F. leucoscepus. 175 tornowi, F. afer, 175 trothae, F. shelleyi, 177 tschadensis, F., 141, 176, 188 ugandensis, F., 141, 176, 188 uluensis, F. shelleyi, in, 154, 155, 159, 160, 168, 192 usambarae, F. squamatus, 137, 139, 186 uzungwensis, F. squamatus, 137, 139, 186 vernayi, F. coqui, 163, 194 watsoni, F. pictus, 124, 180 wattii, F. levaillantoides, 178 wellsi, F. pintadeanus, 175 whitei, F. squamatus, 176 whytei, F. shelleyi, in, 113, 122, 154, 155, 157, 159, 160, 192 zambesiae, F. sephaena, 150, 151, 152, 177, 190 zappeyi, F. squamatus, 176 zuluensis, F, sephaena, 177 PRINTED IN GREAT BRITAIN BY ADLARD AND SON, LIMITED, BARTHOLOMEW PRESS, DORKING TAXONOMY AND BIOLOGY OF THE GENUS LEBETJ7S (TELEOSTEI-GOBIOIDEA) P. J. MILLER BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 10 No. 3 LONDON: 1963 TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS (TELEOSTEI-GOBIOIDEA) BY P. J. MILLER (Department of Zoology, The University, Glasgow, W.2) Pp. 205-256 ; Plate ; 21 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 10 No. 3 LONDON : 1963 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 10, No. 3, of the Zoological series. Trustees of the British Museum 1963 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued March 1963 Price Twenty One Shillings TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS (TELEOSTEI-GOBIOIDEA) By P. J. MILLER ABSTRACT The marine gobioid genus Lebetus Winther 1877 is redefined and previous division into two species orca Collett 1874 and scorpioides Collett 1874 shown to be based on sexual dimorphism in one species, by page priority L. orca. A redescription of the species is provided including details of sensory papillae and osteology. L. orca is eastern Atlantic boreal in distribution, occurring as a small predator on coarser grounds down to 375 m. Specialization of the male genitalia and sexual dimorphism are described and the breeding season provisionally regarded as from March to August. The skeleton of Lebetus is typically gobiid, and evidence is presented for a closer relationship to the genus Buenia Iljin as represented by B. Jeffrey sii (Giinther) than to the Mediterranean Odondebuenia De Buen and Cabotichthys Whitley with which Lebetus had been previously grouped. INTRODUCTION THE genus Lebetus was founded by Winther (1877) to contain two species of gobies dredged by G. O. Sars on the coast of Norway and described by Collett (1874) as Gobius orca and G. scorpioides. These species have subsequently been regarded as among the rarest of European Atlantic fishes and, until recent years, the numbers of demersal individuals known to science have been relatively few. In the last fifteen years, protracted dredging surveys off the Isle of Man (Jones, 1951 ; Hartnoll, 1961) have yielded many examples. Their small size and occurrence on rough grounds makes collecting difficult ; in Manx waters, most specimens have been obtained by means of a scallop dredge lined with shrimp netting. Relative infrequency of capture is probably due to inadequacies of the fishing gear rather than to actual rarity, and occasional hauls providing up to five individuals suggest the presence of appreciable numbers in the area. Other long term faunistic investigations within the range of the genus have produced more or less sizeable collections of both demersal (Grieg, 1913 ; Le Danois, 1913) and planktonic stages (Petersen, 1919 ; Johansen, 1925 ; Russell, 1926-1940). My interest in these fishes was aroused firstly by the difficulty experienced in separating the two species. After endeavouring to use the diagnoses of several authors, it was found possible to achieve this only on the basis of coloration. Then, while examining the testes of various gobies, it was noticed that no female L. orca were present in a collection of twenty two dissected individuals, and only immature males among twenty eight L. scorpioides. The possibility that the two so-called species were in fact based on sexual dimorphism in a single species indicated the need for taxonomic revision. Although such a view was put forward by Smitt (1900), this was not accompanied by any evidence and appears to have been overlooked by later authors. Dissection of the extensive material from the Irish Sea presented oppor- tunities for settling the position of Lebetus among the gobioid families by osteological ZOOL. 10, 3 i4 208 P. J. MILLER study, and also for investigating various aspects of diet and reproduction. A pre- liminary account of some of this work has already been published (Miller, 19616). MATERIAL AND METHODS The following specimens have been available for examination : Isle of Man : 33 $<$, standard lengths 15 -75-30-0 mm., including BMNH 1961 . 10 . 17 . i. 3 5 an( i one m PEM, and 27 ?, 14-5-29-0 mm., inc. BMNH 1961 . 10 . 17 . 2, 4, and one in PEM. Other British localities : 3 $, 12-0-30-0 mm., BMNH 88 . 3 . 22 . 9-10 part and BMNH 1903.4.14.4-7 part, and 2 $$, 10-5 and 17-0 mm., BMNH 1903.4.14.4-7 part. Scandinavia: 13 <$<$, 13-5-26-0 mm.,ZMO J3999 (type of Gobius orca), J40OO, J4OOI, J4023 ; ZMB 771, 1887, 1893, 1966, 4174, 4175, 5294 ; ZMC 89, 91 ; and 4 ??, 21-75-29-0 mm., ZMO J402O, J4O2I (types of G. scorpioides) ; ZMB 536, 2009. Abbreviations for museums are as follows : BMNH British Museum (Natural History). PEM Port Erin Marine Biological Station Museum. ZMB Zoologisk Museum, Universitetet i Bergen. ZMC Universitetets Zoologisk Museum, Copenhagen. ZMO Zoologisk Museum, Universitetet i Oslo. The specimens were viewed by means of low power binocular and dissecting micro- scopes. With suitable illumination the sensory papillae were readily seen without staining after careful removal of mucus from the head and body surfaces. The skeleton was investigated by dissection of alizarin stained whole-mount preparations. THE GENUS LEBETUS WINTHER 1877 Lebetus Winther, 1877: 49 (orthotype Gobius scorpioides Collett 1874); Smitt, 1900: 554; Jordan, 1919 : 392 ; 1923 : 225 ; Duncker, 1928 : 140 ; Iljin, 1930 : 55 ; De Buen, I93oa : 123 ; 19306 : 5, 21 ; Koumans, 1931 : 43, 46 ; Whitley, 1931 : 155 ; De Buen, 1931 : 50, 54, 61 ; 1940 : 3 ; 1951 : 56, 57 ; Norman, MS : 413. Lebistes Jordan, 1920 : 487 ; 1923 : 226 (non Filippi, 1862 : 69). Butigobius Whitley, 1930 : 123. Body moderately elongate, subcylindrical, with 25-29 ctenoid scales in lateral series. Head, nape, and back naked to origin of second dorsal fin, lacking skin folds or barbels. Postorbital length not more than half length of head. Anterior nostril a short tube. Opercle without scales. Preopercle and lower jaw unarmed. Branchi- ostegous membranes attached to sides of isthmus, but not fused across isthmus. Jaws oblique, subequal, maxillae ensheathed and not elongate. Teeth of both jaws in several rows, simple, erect. Tongue truncate to very weakly emarginate, free from floor of mouth anteriorly. Pelvic fins united along entire length, without anterior membrane (frenum) and with fourth branched ray somewhat produced. Pectoral girdle lacking flaps on anterior edge ; uppermost rays of pectoral fin not free from membrane. Dorsal fins separate, first dorsal with 6 rays, not pungent. Caudal fin rounded, not much longer than head. Cephalic canals absent. Sensory papillae prominent but reduced in number. Two pairs of interorbital papillae (/>). Cheek with only suborbital row a, a short longi- TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 209 tudinal row c, and an intermediate papilla ; row d without posterior longitudinal section. Anterior dorsal (occipital) series (excluding h) in two groups of a few papillae. Opercular transverse row ot in two sections. Preoperculo-mandibular series in several parts. In compiling his work on the genera of fishes, Jordan (1919, 1920, 1923) includes, together with Lebetus Winther 1877, the genus Lebistes Smitt 1899. Since Jordan gives " L. scorpioides Smitt " as the orthotype of this genus, the latter, as noted by Koumans (1931), is clearly an erroneous reference to Lebetus for which Smitt (actually 1900, not 1899) indicated Gobius scorpioides Collett as the type. The generic name Lebistes is preoccupied by Lebistes Filippi 1862 among the cyprinodont teleosts and, to replace this name within Jordan's list of gobiid genera, Whitley (1930) proposed Butigobius without consulting Smitt 's paper. In a later account, Whitley (1931) unravels this confused story. THE SPECIES OF THE GENUS The present section embodies a review of the various characters proposed by earlier authors as being of value in the diagnosis of L. orca and L. scorpioides. Coloration. The two species are reputed to differ in both body and fin coloration. The body of Gobius orca was described by Collett (1874, 18750) as uniformly greyish except for darker areas between the second dorsal and anal fins, and at the root of the caudal fin, while that of G. scorpioides was said to possess four or five dark brown vertical bands across the sides. After examination of a living scorpioides, Winther (1877) remarked on the colourless caudal peduncle, and both Holt & Byrne (1903) and Le Danois (1913) emphasized the importance of this pale band, sharply demar- cated anteriorly, as a means for the identification of L. scorpioides. In the original description, Collett (1874) mentioned that the second dorsal fin of G. orca had an indistinct white band in its outer half whereas the same fin in G. scorpioides showed only ill-defined banding. Alternate dark brown and snow white bands were found across the second dorsal fin of a later specimen of G. orca (Collett, 1885). Holt & Byrne (1903) regarded a superior black band to this fin as characteristic of G. orca and in their text referred to a row of white spots immediately below the dark peri- phery. These authors described and illustrated the second dorsal fin in G. scorpioides as banded with red, the markings in some cases being flanked by dark edging. With the collection of Lebetus amassed for this work, it has been possible to segre- gate the individuals into two groups on the basis of differences in the coloration of the second dorsal fin. In the first of these, termed the orca group, the second dorsal fin has a thick black superior edge, below which occur white areas and wide oblique bands, yellow or ochre in life but grey in preserved material. Within the scorpioides group, the black edge to the second dorsal fin is present but much thinner and less intense than in the orca examples. The rest of the fin bears merely four narrow double lines of melanophores, enclosing in the living fish thin red or ferruginous striae, and separated by hyaline fin membrane. Among living scorpioides, madder brown lateral marks are pronounced on the body which, in orca fishes, is more uniform grey or. yellowish grey although comparable markings may be quite prominent. A pale band on the caudal peduncle is more clearly defined in the scorpioides group, especially 2io P. J. MILLER when preserved, but living orca show this feature to a noticeable degree (cf. PI. I and Text-fig. 3) and it is shown in Collett's illustrations of Gobius orca (i875a, 1885). TABLE I. Relation of Coloration Groups to Sex and Maturity Stage No. of males Mature and No. of females Group developing Immature all stages (a) Dissected orca 21 i scorpioides . 6 22 (b) Not dissected orca . . 16 scorpioides . 5 n The relation of these two colour patterns to sex and to gonad maturation is indicated in Table la, based on fifty Manx specimens whose sex and sexual development were confirmed by dissection. The sex and approximate maturity stage in a further thirty two British and Scandinavian examples have been determined by inspection of the genital papilla (Text-fig. 17), and these results are given in Table Ib. It is apparent that mature or developing male scorpioides do not occur, and that female orca are unknown at any stage of maturation. This distribution strongly implies that the orca coloration is the male livery of a species whose females and immature males bear the markings of scorpioides. Incipient development of the orca pattern in the second dorsal fin can, in fact, be noted among some of the immature males in the scorpioides group, which usually exhibit a pronounced black spot on the distal part of the first two interradial membranes of the second dorsal fin (see Le Danois, 1910, fig. 4). The above hypothesis is further supported by the range in standard length found in the three categories shown in Table II. The examples of orca are from 18-0-30-0 mm. TABLE II. Standard Length Frequency in Coloration Groups, Expressed in 2 mm. Length Groups Standard length ( Coloration group orca . cJ scorpioides 9 scorpioides in length while male scorpioides do not exceed 23-0 mm. Taning (1940) has reported juvenile Lebetus orca of 5-15 mm., but relied for the identification on meristic characters whose validity is doubtful. Female scorpioides attain about the same maximum length as that for orca. The first dorsal fin in the orca individuals is a uniform yellowish grey edged with white, but in scorpioides is usually banded with a dark spot on the interradial mem- brane between the fifth and sixth rays. Confluence of these bands frequently occurs, and the spot may not be very distinct. This scorpioides group pattern is stated by 10 12 14 16 18 20 22 24 26 25 1 30 i i IO 16 7 i 2 2 3 3 2 I I I 2 i 2 5 9 5 7 TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS an Holt & Byrne (1903) to be found only in the breeding male of " Gobius scorpioides ". Fage (1918) regarded the presence of a black spot on the first dorsal fin of a 12-5 mm. example from the Dana collections as a sign of precocious development of supposed male coloration. However, in the material under consideration, the spot is well represented in adult females and among male Lebetus occurs only in the immature where it is not uncommonly absent. The single female G. orca mentioned in the literature was taken by Patience (1906) in the Firth of Clyde. This specimen has not been traced. The diagnosis of G. orca given by Patience was not derived from his own examples but seems to be a translation from Collett (1896) and could apply to individuals of either the orca or scorpioides group. TABLE III. Meristic Characters Cited for Lebetus orca and L. scorpioides No. of fin rays First Second \ Verte- dorsal dorsal Anal Pectoral Caudal brae L. orca 1 VI ii 10 18 3/13/3 VI ii 9 17 3/13/3 VII IO-II 9 17 3/12-13/3 VII I/9-io 1/8 17 12-13 VI-VII I/9-io 1/8 17 x/13/x VII IO-II 9 17 3/12-13/3 VI-VII 9-1 1 9-10 VI-VII I/9-io 1 8 (sic) 17-18 12-13 VI-VII I/9-io 1/8-9 VI-VII 10-12 8-10 18-20 7-8/1 1/6-7 28-29 L. scorpioides 2 ' 3 VI 9 8 20 (18) 6/12/6 VI 9 8 1 8-20 6/12/6 VI 9 8 c. 18 12 VI 1/8 1/7 18-20 12 VI 1/8 1/7 18-20 X/I2/X VI VI VI VI VI V-VI VI 9 9 1/9 1/8-9 8-9 1/7-8 1/8-9 8 7-8 1/6 1/6-9 6-8 1/6-7 1/6-7 18-20 18-20 18-20 X/I2/X 12 X/I2/X No. of Scales in 1.1. 28 28 24 25 c. 24 25-28 25-26 24-28 28 28 30 28-30 c. 28-30 28-30 28 28 27-28 28 25-26 25-26 26-30 28-30 Authority Collett, 1874 Collett, 18750 Collett, 1885 Lilljeborg, 1884 Smitt, 1892 Collett, 1896 Holt & Byrne, 1903 Duncker, 1928 De Buen, 19300, b Taning, 1940 Collett, 1874 Collett, 18750 Winther, 1877 Lilljeborg, 1884 Smitt, 1892 Holt & Byrne, 1903 Le Danois, 1910 Le Danois, 1913 Fage, 1918 Petersen, 1919 Duncker, 1928 De Buen, I93oa, b, 1932 1 Scales in 1.1. c. 26-c. 28 (Grieg, 1913). z Scales in 1.1. 26, T) l V (Collett, 1902). 3 Scales in 1.1. 28 (Grieg, 1913)- Meristic characters. Radial formulae, vertebral and scale counts provided by various authors for the two species are shown in Table III. Although Collett origin- ally ascribed 6 elements to the first dorsal fin of both species, he later changed this to 212 P. J. MILLER 7 for G. orca. Lilljeborg (1884) and Smitt (1892) employed this supposed disparity in keys to Scandinavian gobies. L. orca is also regarded as having rather more articu- lated rays in the second dorsal and anal fins than L. scorpioides, and slight differences are indicated in pectoral and caudal fin ray counts. The number of scales along the lateral midline is reputed to be greater in L. scorpioides than in L. orca. For the orca and scorpioides groups of the present material the meristic values in Table IV have been obtained. Methods of counting are described in the systematic section below. To minimise the effect of regional variation, results for British and Scandinavian collections have been kept separate. Mean values are given for the more numerous British material. TABLE IV. Meristic Characters of Coloration Groups : Number of Observations (Highest in Bold Type) Against Value. M Mean Value for British Material Origin of specimens Coloration group orca scorpioides orca scorpioides orca scorpioides orca scorpioides orca scor orca scorpioides orca scorpioides British Isles 5 6 26 34 First dorsal fin rays Scandinavia 6 7 7 I 6 Second dorsal fin articulated rays 6 7 8 9 10 ii M 8 9 10 ii i 18 7 9-23 531 i 2 4 21 6 8-85 2 4 i Anal fin articulated rays 5 6 7 8 9 M. 56789 i 19 6 7 -19 54 32 22 7 6-97 i 4i Pectoral fin rays 15 16 17 18 19 20 21 M 17 18 19 20 21 2 15 20 15 18-92 742 i 4 7 9 30 10 i 18-56 4 2 Caudal fin branched rays 8 9 10 ii M 8 9 10 ii 24 i 9 -04 5 2 31 i 8-97 2 Scales in lateral series 25 26 27 28 29 M 25 26 27 28 29 2 18 10 6 2 26-68 i 3 2 i i 17 17 6 i 26-74 i i 2 Vertebrae including urostyle (Manx specimens only) 26 27 28 29 M 2 13 i 27-94 i 3 17 27-76 TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 213 Except for pectoral ray and scale counts in the Scandinavian fishes, where the number of observations is small, the most common value for each meristic character is the same in the two coloration groups and mean values are only slightly different. The wider range and somewhat lower means for fin ray counts in scorpioides are due 17- < 15- 13- Original standard Present First Second Pect- Scales Sex ZMO No. . + Caudal fin standard Dorsal dorsal Anal oral in 1.1. 6* J 3999 26-0 + 6-0 25-0 d. I/IG 1/8 d. 26-27 ? J 4020 total 28-0 21-5 VI 1/9 1/7 d. d. ? J 4 21 30-0 + 7-0 29-0 VI 1/9 1/7 20 27-28 TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 219 GENERAL DESCRIPTION. Body subcylindrical, laterally compressed towards caudal fin ; in standard length, depth at origin of pelvic fin 5-1-6-65 (mean 5-65), at origin of anal fin 5-7-7-5 (6-53), of caudal peduncle before origin of caudal fin 8-5-12-3 (10-12). Head rounded ; in standard length, horizontal length (snout to upper origin of opercle) 3-2-4-1 (371), maximum width (between upper origins of opercles) 5'25-5 '95 (5 '68). Horizontal distance from tip of snout to origin of first dorsal fin 2.7-3-05 (2-9), to origin of second dorsal fin 1-7-1-9 (1-81), to anus 1-75-1-95 (1-86), to origin of anal fin 1-6-1-75 (1-69), to origin of pelvic fin 3-2-4-0 (3-47), all in standard length. Caudal peduncle (end of anal fin base to origin of caudal fin) horizontal length 4-15-5-1 (4-71) in standard length, depth 1-8-2-35 ( 2 *7) m own length. In head length, snout 3-8-5-6 (4-7), eye 2-8-3-4 (3-0), postorbital length 2-0-2-7 (2*34), cheek 3-8-4-85 (4-34) ; eye and cheek direct, rest horizontal measurement. Inter- orbital minimum width (bony) 4-3-7-4 (5-5) in eye length. Snout less than diameter of eye, with moderately sloping profile ; nostrils adjacent, anterior on short tube, not projecting over upper lip and with entire distal rim lacking local prolongation ; posterior nostril also tubular but shorter. Eyes large, close together. Postorbital region never exceeding half head length ; profile more or less flat, without dermal ridges. Upper jaw relatively wide, about equal to preorbital area. Mouth oblique, lower jaw at most only slightly in advance of upper ; maxilla ends posteriorly under anterior quarter of eye. Lower jaw unarmed, without barbels. Cheek smooth, without dermal ridges. Preopercle unarmed. Branchiostegous membrane attached to entire lateral margin of isthmus from immediately anterior to lower edge of pectoral lobe and without transverse fusion across isthmus (Text-fig. 5). Teeth simple, caniniform, erect in both jaws. Dentition of jaws an outer row of larger teeth with narrow band of two to four rows of smaller teeth. Pharyngeal teeth relatively large, recurved, in two superior and one inferior median patch (see description of skeleton). Tongue more or less truncate, with anterior edge rounded laterally ; at most pronounced, a very slight median emargination (Text-fig. 6s). Gill rakers small protuberances, not spinous, about 6-8 on first arch. Sagittal otolith rectangular, anterior and posterior edges slightly concave, inferior edge slightly convex, superior edge with angular projection near posterior end ; outer surface convex, more flattened in middle, inner surface convex, with furrow in middle and shallower groove to posterior edge (Text- fig. 6A). FINS. In the following counts, spinous elements are represented by large roman, articulated rays by arabic numerals. The terminal bifid ray in the second dorsal and anal fins is counted as one element. Only in the caudal and pelvic fins are articu- lated rays branched. The range of values is given in parentheses after the most common number, frequency distribution being shown in Table VII. First dorsal fin VI (VI-VII) ; second dorsal fin 1/9 (6-n) ; anal fin 1/7 (5-8) ; caudal fin (branched rays) 9 (8-10) ; pectoral fin 19 (15-21), pelvic fin 1/5 -f- 1/5. In standard length, first dorsal base 7-95-8-95 (8-24), second dorsal base 3-75-4-55 (4-24), anal base 4-85-5-8 (5-4), longest caudal ray 3-2-4-45 (for standard lengths up to 20-0 mm., mean = 3-55 ; more than 20-0 mm., mean 3-92), longest pectoral ray 3-7-4-45 (4-0), pelvic origin to tip of fifth pelvic articulated ray 3-85-4-55 (S.L. up to 25 mm., mean = 4-05 ; more than 25-0 mm., mean = 4-2). First dorsal fin P. J. MILLER TABLE VII. M eristic Characters of Lebetus orca : Number of Observations against Value Locality British Isles Scandinavia British Isles Scandinavia British Isles Scandinavia British Isles Scandinavia British Isles Scandinavia British Isles Scandinavia Isle of Man First dorsal fin rays 5 6 7 60 Second dorsal fin articulated rays 6 7 8 9 10 ii i 2 5 39 13 2941 Anal fin articulated rays 567*9 3 3 4i 13 i 9 5 Pectoral fin rays 75 16 I^ 18 19 20 21 i 4 9 24 50 25 i 784 Caudal fin branched rays 8 9 10 ii 2 55 2 7 Scales in lateral series 25 26 27 28 29 3 35 27 12 3 1433 Vertebrae including urostyle 26 2j 28 29 i 5 30 i TABLE VIII. Dorsal Fin Ray Lengths Expressed in Standard Length for Mature and Immature Lebetus orca Mature