I \ f I i I I J ( A! I III i i; i t’ I •i Annals of the Cape Provincial Museums Natural History QH I /?6/J Ann. Cape Prov. Mus. (nat. Hist.) Volume 16 Parti 16th April 1984 Published jointly by the Cape Provincial Museums at the Albany Museum, Grahamstown, South Africa ANNALS OF THE CAPE PROVINCIAL MUSEUMS These Annals are published in two series, Natural History and Human Sciences, the latter series covering cultural history, ethnology, anthropology and archaeology. They are issued in parts at irregular intervals as material becomes available. The primary objective of these Annals is to disseminate the results of research work undertaken by staff members of the Cape Provincial Museums. However, a manuscript by any author is eligible for publication provided it is based wholly or partly on the collection/s of one or more of these museums or if the material on which the work is based is wholly or partly deposited in one of these museums. Parts are obtainable from the Librarian of the Albany Museum. Correspondence in con- nection with these Annals should be addressed to the Editor, Albany Museum, Grahams- town 6140. Editor Dr F.W.GESS: 1978 Contribution to the taxonomy of the southern African species of Ampulex Jurine (Hymenoptera: Sphecidae: Ampulicinae) by F. W. GESS (Albany Museum, Grahamstown) CONTENTS Abstract Introduction Taxonomic descriptions and discussions Ampulex cyaniira Kohl Ampulex bantuae sp. nov Ampulex lesothoensis sp. nov Ampulex montivaga sp. nov Ampulex nigrisetosa sp. nov Ampulex mutilloides Kohl Ampulex timulloides sp. nov Key to females of southern African species of Ampulex in which the pronotum is not posteriorly raised into a conical tubercle Acknowledgements References Page 1 1 2 2 5 9 11 14 16 18 20 22 22 ABSTRACT Descriptions are given of four new species of southern African Ampulex, namely bantuae, lesothoensis, montivaga and nigrisetosa. A. cyanura Kohl is redescribed. The female of A. mutilloides (sensu Arnold) is shown to be specifically distinct from that of A. mutilloides Kohl and is renamed A. timulloides. A key is provided to the females of those southern African species of Ampulex in which the pronotum is not posteriorly raised into a conical tubercle. INTRODUCTION The writing of the present paper was occasioned in the first place by the need to provide a name for an undescribed species of Ampulex, aspects of the ethology of which are the subject of a paper shortly to be published. Initially the species in question was believed to be A. cyanura Kohl or a species closely allied to it and consequently was cited in print either as A. cyanura Kohl or as A. sp. near cyanura Kohl. When the author was subsequently enabled to compare his specimens with the type of A. cyanura Kohl it was immediately apparent that 1 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 though superficially similar they were specifically distinct. Characters of importance in distin- guishing between the two species received no or at best scant attention in both Kohl’s original description of A. cyanura and Arnold’s subsequent rediscription. It has been found necessary therefore not only to describe the new species but also to redescribe A. cyanura. At the same time it has been considered relevant to describe three further species represented in the Albany Museum collection, all of which are seemingly allied to A. cyanura and all of which are apparently new. A. cyanura and the four new species are all medium-sized (length of females 10-16 mm) and are predominantly blue or green lustred (sometimes purple lustred in death). All are further characterized as belonging to that group of species in which the pronotal collar is not posteriorly raised into a conical tubercle. In order to facilitate the identification of the new species a revision was undertaken of that part of Arnold’s 1928 key dealing with those species belonging to the above indicated group. It was therefore pertinent to deal also with a confusion in the literature concerning two species known under the name A. mutilloides . Characters found useful in distinguishing species have largely been concerned with the head — its overall shape and more particularly details of the frontal lobes, frontal carinae, mandibles, clypeus and labrum. These features have all been figured. TAXONOMIC DESCRIPTIONS AND DISCUSSIONS Ampulex cyanura Kohl Ampulex cyanura Kohl, 1893: 471, 493, fig. 48, 9; Arnold, 1928: 206-207, figs 7,7a-d and f, PI. VIII, fig. 5, 9 , (? ; Callan, 1976: 232 [partim, Callan’s material only]. Ampulex capensis Cameron, 1905:254, 9(?) Ampulex africana Cameron, 1905: 256, 6 . [non] Ampulex cyanura Kohl, Callan, 1976: 232, Gess’ prey record [applies to Ampulex bantuae sp. nov.]. Female (Figs 1, 8, 13 and 18) Length 14-16 mm (Holotype 15,5 mm) Head, thorax, abdominal tergites 1-3 and corresponding sternites, coxae partially (espec- ially outer lateral aspect of fore-coxae and dorsal aspects of meso- and metacoxae, not ventral aspects), fore and hind femora and under surface of middle femora, dorsal aspect of hind tibiae black with metallic blue or green lustre (purplish in long dead specimens). Anterior edge of frontal lobes, antennae, median part of clypeus distally, undersides of all coxae, dorsal and lateral aspects of middle femora, fore and middle tibiae, hind tibiae (other than for dorsal aspect), tarsomeres of all legs (other than for ferruginous parts) shining black without metallic blue or green lustre. Mandibles, labrum, apices of first three tarsomeres, most of fourth tarsomere, apex of fifth tarsomere, hind margins of second and third tergites and sternites laterally, normally hidden parts of rest of abdominal segments ferruginous. Wings almost hyaline or somewhat browned (in which case the forewings are darker than the hindwings), beset with semi-erect brown setae the density of which is proportional to the amount of brown pigment in the darker areas of the the wing. Dark areas include: the distal end of the medial cell at RS and the distal end of the submedial cell and the surrounds of the junction of M and Cu-A; in ill-defined cloud extending over the marginal (= radial) cell, the distal half of the first submarginal (= cubital) cell, the entire second submarginal cell, the prox- imal half or less of the third submarginal cell, and the middle of the second discoidal cell. The vestiture consists of short decumbent and semi-decumbent white hairs and of longer upright hairs of two distinct types: stiff coarse black hairs and soft white hairs. 2 GESS: TAXONOMY OF AMPULEX JURINE (HYMENOPTERA; SPHECIDAE: AMPULICINAE) The black pilosity occurs on the anterior margin of the clypeus (where forwardly point- ing), the antennal scape, the frontal lobes, the frons, vertex and occiput, the pronotal collar, the scutum, scutellum, disc of the metanotum, the pleura and ventral surface of the mesotho- rax, the coxa, trochanter, femur, tibia and tarsomeres of all the legs, and sternite 2 (where very sparse). The white upright pilosity occurs on the pleura and ventral surface of the mesothorax (in which areas it is intermixed with black hairs), on the metapleura, and on the sides and declivity of the propodeum. Short decumbent white hairs densely cover the basalar lobe (where the hairs form a tuft) and the postero-ventral extremities of the meso- and metapleura. Puncturation of frons strong, coarse and close with diameter of individual punctures about half that of ocelli; puncturation of head behind ocelli and on vertex with individual punctures larger and with shining interstices between them; pronotum, scutum, scutellum and meso- pleura with punctures as large as those on vertex, on the pronotum as dense as on the vertex but more sparse on the scutum and scutellum; tergite 1 and 2 shining, sparsely, shallowly and finely punctured; lateral areas of tergite 2 more closely punctured and basal half of tergite 3 very closely and finely punctured. Head (Fig. 1). Frontal lobes well developed, moderately raised laterally, not expanded laterally and in dorsal view not covering antennal sockets; frontal carinae diverging posteriorly, evanescent far below level of anterior ocellus. Clypeus (Figs 1 and 8) with elevated and pro- duced median part markedly tectiform, strongly carinate in mid-line and strongly and smoothly curved in profile; anterior median tooth strongly developed, its apex smoothly rounded; lateral teeth large, antero-laterally directed, acutely rounded; lateral wings of clypeus well developed, antero-laterally directed and with their free edges smooth, separated from lateral teeth by a deep emargination and terminating abruptly posteriorly and separated from mandibular socket by a deep emargination. Labrum (Figs 8 and 13) robust, with median part greatly produced and by far exceeding strongly downcurved lateral wings which flank it basally; median part in dorsal view outcurved laterally and wider in distal half than in proximal half, its anterior mar- gin widely bilobed and slightly downcurved, its upper surface generally flat and its ventral sur- face noticeably concave due to lateral thickening; lateral wings each produced into a down- wardly directed lamellate lobe inferior in size to lobes of median part; emargination between lateral wing and median part of labrum with three downwardly directed, short, stout and blunt- ly rounded spatulate setae. Mandibles (Figs 1 and 18) robust, not compressed nor wide in lat- eral view, distinctly downturned near apex, with a well developed emargination or notch dor- sally at the base, with a bluntly triangular lamelliform projection of lower edge just beyond mid-length, with a small tooth immediately below and behind bluntly rounded apex and with a small subapical lamellate cusp on inner surface near upper edge, with a row of short setae be- low cusp and at base of lamelliform projection and few longer setae on inner surface near base. Pronotal collar broader than long, less than twice (1,5-1 ,7) as wide behind as long in the mid-line, shorter than scutum; dorsum of collar plain, without a conical elevation but with a moderate and anteriorly deepened and widened longitudinal sulcus in mid-line; scutum with notauli extending to posterior margin; scutellar disc less than twice as wide basally as long in the mid-line; metanotal disc moderately raised, not expanded posteriorly; propodeum less than twice (1,8) as wide across base as long in the mid-line of its dorsal face, narrowed posteriorly (0,8 times as wide across postero-lateral teeth as across base); propodeal dorsum with a median longitudinal carina (only evident proximally) and four pairs of posteriorly converging lateral carinae; first (innermost) pair moderately lamelliform, inwardly sloping, shagreened, not attaining hind end of dorsum; second pair not joining nor attaining hind end of dorsum; third pair joining posteriorly across propodeal angle and marking end of dorsum; fourth pair not lamelliformly produced below spiracles. 3 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 Figs 1-5. Frontal view of head of; 1, 9 A. cyanura Kohl; 2, 9 A. bantuae sp. nov.; 3, 9 A. lesothoensis sp. nov.; 4, 9 A. montivaga sp. nov.; 5, 9 A. nigrisetosa sp. nov.. Tergite 1 widest in posterior half, less than twice (1,8-1, 9) as wide there as long in the mid-line; tergite 2 about as wide in the anterior half (where widest) as long in the mid-line, less than twice (1,8) as long as tergite 1 and only marginally wider. Forewing with three submarginal (= cubital) cells, with second submarginal cell clearly longer (1,4-1, 5) on the media than on the radius. Male (Fig. 6) Length 9,8-10,6 mm Easily associable with the female by the following shared characters: the form of the frontal lobes and frontal carinae; the form of the pronotal collar; the absence on the metanotal disc of any upright white pilosity and of a dense covering of decumbent white hairs; the simple (non-lamelliform) nature of the fourth pair of lateral carinae of the propodeal dorsum; the pro- portions of the second submarginal cell of the forewings. 4 GESS: TAXONOMY OF AMPULEX JURINE (HYMENOPTERA: SPHECIDAE; AMPULICINAE) Figs 6-7. Frontal view of head of: 6, 6 A. cyanura Kohl; 7, d A. bantuae sp. nov.. Material examined: Cape Province: K.b.sp. [= Cap bonae spei], no date (no collector), Holotype 9 (bearing label: "cyanura det. Kohl Type’) (Zool. Mus. Berlin); Addo, 19. iv. 1896 (Dr Brauns) 9 (bearing label in Brauns’ handwriting: "Ampulex cyanura Kohl 9 confirmed by Kohl himself’) (South African Museum ex National Museum Bulawayo 1981); Addo, 19. iv. 1896 (Dr Brauns) 9 (South African Museum ex National Museum Bulawayo 1981); Addo, 19. iv. 1896 (Dr Brauns) 9 (bearing label in Brauns’ handwriting: "A. cyanura Kohl 9’) (Albany Museum); Grahamstown, Howison’s Poort, 7-14. ii. 1972 (F. W. Gess, Malaise trap) 9 (Albany Museum); Grahamstown, Hilton, 5-9. xi. 1970, 6, 12-30. xi. 1970, 2d d (all F. \V. Gess, Malaise trap) (Albany Museum); Grahamstown, Kudu Reserve (33° 08'S, 26° 42'E), 28.xii.1981 (P. G. Hawkes & P. M. C. Croeser) 9 (In dry river bed bordered by flowering Acacia karroo) (Albany Museum). Zululand: Mfongosi, viii.1911 (W. E. Jones) 9 (bearing label in Arnold’s handwriting: "Ampulex cyanura') (South African Museum); Mfongosi, i.l917 (W. E. Jones) 9 (bearing label in Arnold’s handwriting: "Ampulex cyanura Kohl (= capensis Cam.) (= africana Cam.) 9 A. Tr. Mus. XII III 1928, 206’) (South African Museum). Ampulex bantuae sp. nov. Ampulex cyanura {non Kohl) Callan, 1976: 232 [partini, Gess’ prey record only]. Ampulex sp. near cyanura Kohl, Gess, 1981: 29, 12-11, fig. 36; Gess & Gess, 1981: 27-30, fig. 2. Female (Figs 2, 9, 14 and 19) Length 11,5-14,5 mm (commonly 13,0-13,3 mm; Holotype 14,5 mm) Head, thorax, abdominal tergites 1-4 and corresponding sternites, coxae, femora and tibiae of all legs, first tarsomere of metathoracic leg, usually first tarsomere of mesothoracic leg and occasionally first tarsomere of prothoracic leg, black with metallic blue lustre (sometimes purplish in dead specimens). Anterior edge of frontal lobes, antennae, nasiform median part of clypeus basally, tarso- meres other than those listed above, basal half of claws, exposed parts of terminal abdominal segments, shining black without metallic blue lustre. Mandibles bright ferruginous, sometimes darker at edges, contrasting strongly with metal- lic blue cheeks. Apical half of nasiform portion of clypeus (usually), clypeal teeth and narrow lateral parts of clypeus, labrum and apical half of tarsal claws dark ferruginous. Maxillary and 5 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 labial palps dark brown. (It should be noted that the extreme hind margins of the abdominal tergites and sternites from the second segment onwards may occasionally show ferruginous colour. In an extended abdomen the normally hidden basal parts of the segments are likewise ferruginous.) Wings hyaline or slightly browned, beset with semi-erect brown setae the density of which is proportional to the amount of brown pigment in the darker areas of the wing. Dark areas in- clude; a longitudinal streak in the medial cell, most of the submedial cell, the lower proximal corner of the subdiscoidal cell and part of the wing membrane posterior to these two cells; the marginal (= radial) cell, the distal half of the first submarginal (= cubital) cell, the entire second submarginal cell, the proximal half of the third submarginal cell and a diffuse band across the middle of the second discoidal cell. The vestiture consists of short decumbent and semi-decumbent white hairs and of a mix- ture of longer upright hairs of two distinct types; relatively longer stiff black hairs and rela- tively shorter soft white hairs. The black pilosity occurs on the frons and vertex, the dorsal aspect of the pronotum, the scutum, scutellum, posterior margin of the metanotal disc and parts of the legs (notably the dorsal and outer aspects of the tibiae and first tarsomeres). The white pilosity occurs on the occiput, sides and underside of the head, the anterior declivity, sides and ventral surface of the prothorax, the metanotal disc, the pleura of the meso- and metathorax and the ventral surface of these segments, the sides of the propodeum, the lower half of the propodeal declivity (particularly towards the sides), tergites 1-3 (where sparse and mostly on the sides and posterolateral corners) and sternite 2. It is also present on the coxa, trochanter, femur, tibia and first tarsomere of each leg. Long stiff forwardly-pointing ferruginous hairs are present on the clypeus; shorter ferrugi- nous hairs occur on the labrum and mandibles. Black hairs shorter than those described above are present on the antennal scapes and first five or six flagellar segments; common on the scape and second flagellar segment, the hairs become progressively shorter and sparser with each succeeding segment. Short decumbent white hairs densely cover the disc of the metanotum, the basalar lobe (where the hairs form a tuft), the posteroventral extremities of the meso- and metapleura, the dorsal aspect of the metacoxae, the posterior margin of tergite 2 laterally and the adjacent lateral part of the hind margin of sternite 2. Similar hairs but a lot less dense occur laterally on the hind margins of tergite 3 and sternite 3 and may occasionally form a small patch situated on the posterior third of the pronotum in the midline. Puncturation of frons and vertex reticulate, very coarse and close with diameter of indi- vidual punctures about f that of ocelli and without interstices between punctures; gena with smaller oval punctures and smooth interstices; pronotum, scutum, scutellum and mesopleuron with oval punctures similar in size to those of gena and separated by shining interstices equal- ling or exceeding the punctures in size; tergite 1 and most of 2 highly polished and shiny with small punctures separated by 2-4 times their diameter and with midline of the segments almost inpunctate; postero-lateral areas of tergite 2 and whole of tergite 3 very closely and finely punctured and with a scattering of larger punctures; dorsal aspect of tergites 4 and 5 with minute punctures, tergite 6 virtually impunctate. Head (Fig. 2). Frontal lobes well developed, only weakly raised laterally, arcuately expanded laterally and in dorsal view completely covering antennal sockets; frontal carinae parenthesis-like, evanescent well below level of anterior ocellus. Clypeus (Figs 2 and 9) with elevated and produced median part markedly nasiform, carinate in mid-line and only weakly curved in profile; median part sub-parallel-sided distally, apically not tooth-like but wide and somewhat flattened, subtruncate and very weakly trilobed; lateral teeth large, downwardly directed, acutely rounded; lateral wings of clypeus poorly developed, downwardly rather than 6 GESS; TAXONOMY OF AMPULEX JURINE (HYMENOPTERA: SPHECIDAE: AMPULICINAE) Figs 8-12. Lateral view of anterior part of head of; 8, 9 A. cyanura Kohl; 9, 9 A. hantuae sp, nov.; 10, 9 A. leso- thoensis sp. nov.; 11, 9 A. montivaga sp. nov.; 12, 9 A. nigrisetosa sp. nov.. antero-laterally directed and thus appearing to be strongly receding when seen from above, with a slight angle in region below antennal socket and with free edge thinly lamelliform. Labrum (Figs 9 and 14) robust, triangular in overall shape, greatly but regularly widened from base to apex and with extensive lateral wings exceeding median part, almost three times as wide across acute antero-lateral angles (i.e. apices of lateral wings) as long in the middle; median part raised above level of lateral wings over its entire length and strongly conically pro- duced in basal half where furnished with a high pointed tubercle, its apical margin weakly bilobed with a small backwardly pointing median tooth; emargination between lateral wing and median part of labrum with a pair of forwardly directed, short, stout and bluntly rounded spatulate setae. Mandibles (Figs 2 and 19) robust, compressed, wide in lateral view, widest a little beyond midlength but lacking any lamelliform projection of lower edge and neither downturned near apex nor having a subapical tooth, with a well developed emargination dorsally at the base but 7 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 with the part basad to emargination not produced over the latter to form an acute notch, with the interior upper edge in dorsal view not a smooth curve but in three steps, with two small inwardly directed cusps on apex (at least in newly emerged individuals). Pronotal collar broader than long, more than twice (2, 1-2,3) as wide behind as long in the midline, shorter than scutum: dorsum of collar plain, without a conical elevation but with a narrow, shallow and even longitudinal sulcus in the midline; scutum with notauli extending to posterior margin; scutellar disc slightly more than twice as wide as long in the midline; meta- notal disc moderately raised, expanded posteriorly and therefore exceeding lateral wings in breadth; propodium more than twice (2,3) as wide across base as long in the midline of its dorsal face, narrowed posteriorly (0,7 times as wide across postero-lateral teeth as across base); propodeal dorsum with a median longitudinal carina (only evident proximally) and with four pairs of posteriorly converging lateral carinae; first (innermost) pair markedly lamelli- form, inwardly sloping, shagreened and contrasting with very shiny areas on either side, not attaining hind end of dorsum; second pair joining posteriorly across propodeal angle and mark- ing end of dorsum; fourth pair lamelliformly outwardly and upwardly produced at their origins below spiracles. Tergite 1 widest in posterior half, more than twice (2, 1-2,4) as wide there as long in the midline; tergite 2 evenly curved laterally in dorsal view, about one and a quarter times as wide in the middle (where widest) as long in the midline, twice as long as tergite 1 and about one and one-tenth as wide. Forewing with three submarginal (= cubital) cells, with second submarginal cell sub- quadrate and only one and one-fifth as long on media as on radius. Male (Fig. 7) Length 7,8-12,0 mm (commonly 10-11 mm; Allotype 11,3 mm) Easily associable with the female by the following shared characters: the form of the frontal lobes and frontal carinae; the form of the pronotal collar; the possession of a mixture of longer upright hairs of two distinct types; the presence of a dense covering of short decumbent white hairs on the metanotum; the form of the first pair of lateral carinae of the propodeal dorsum; the lamelliform nature of the fourth pair of lateral carinae of the propodeal dorsum; the subquadrate second submarginal cell of the forewings. Material examined: Cape Province: Grahamstown, 29. iv. 1966 (C. Jacot-Guillarmod) ?,(?; Grahamstown, Hilton, 19-23. xi. 1975, 9, 16. xi. 1977, S (both F. W. Gess, Malaise trap), 27.xii.1973, 9, 26.ii.1974, 9, 15.iii.l974, 9, 19. xi. 1976, 9, 9.xii.l976, 9 (all F. W. & S. K. Gess) (all captured in the field, in association either with natural nests or with trap-nests), 10.xi.l977 (F. W. Gess) S (on flowers of Lasiospermum bipinnatum, Compositae). In addition to the above specimens, all of which were collected as adults, there are in the Albany Museum collection a further 329 9 and 49 o d which were reared from eggs, larvae or pupae obtained from natural nests at Hilton, from trap-nests at Hilton, and from laboratory-based trap-nests which had been utilized by caged wasps of the Hilton population. Both the Holotype 9 and the Allotype 6 belong to the reared material. Holotype 9 : Reared from the cockroach Bantua dispar (Burmeister) collected (dead) during April, 1975, at Hilton from an old gallery of Ceroplesis hottentota (F.) (Ceram- bycidae) in Acacia karroo. Wasp emerged 18.xii.l975. Allotype (?: Reared from the cockroach Bantua dispar (Burmeister) collected (live) on 25. xi. 1973 at Hilton from a trap-nest in Acacia karroo. Wasp emerged 21. i. 1974. Paratypes: 389 9 , 51d d (the bulk of the above listed material). GESS: TAXONOMY OF AMPULEX JURINE (HYMENOPTERA: SPHECIDAE: AMPULICINAE) Etymology; The name is derived from a southern African cockroach genus, Bantua Shelford (Blattariae; Derocalyminidae), a species of which, B. dispar (Burmeister), is the only known prey of the wasp. In the South African Museum collection were found two specimens determined as A. cyanura which upon examination proved to be A. bantuae. Label data and comments are given below: (a) Cape Province: Mossel Bay, i.l899 (T. W. Overbeek) 6 (bearing label in Brauns’ handwriting: "Ampulex cyanura Kohl S'). The specimen is without any doubt an A. bantua 6. (b) Cape Province: Vryburg, 190? (E. Jons) 9 (bearing label in Arnold’s handwriting; 'Ampulex cyanura Kohl r. rhodesianus Arn. 9’). The specimen is definitely not what it is labelled. Though in the ferruginous colouration of its clypeus, of the anterior parts of its frontal lobes and of its first antennal segment it is similar to the subspecies rhodesianus , it differs from it and from A. cyanura sensu stricto in the shape of the clypeus and labrum, in the form of the frontal lobes and the mandibles, in the propor- tions of the pronotum etc. In all these characters it agrees closely with A. bantuae of which it would appear to be a slight geographical variety. Ampulex lesothoensis sp. nov. Female (Figs 3, 10, 15 and 20) Length 13 mm Head, thorax, abdomen, all legs to end of femur, hind tibiae dorsally black with metallic blue, green or purple lustre. (As with other species also, it appears that not only does the met- allic lustre turn from blue or green to purple but the underlying basic black fades to a rich mahogany or ferruginous. It is possible that in a live specimen all the tibiae may be lustred.) Antennae (including scapes), median projection of clypeus, labrum, palps, mandibles, tibiae and tarsi and normally hidden parts of abdominal segments dark ferruginous. Wings strongly and fairly uniformly browned, without noticeable fasciae, beset with short semi-erect brown setae. The vestiture consists in the main of sparse very short decumbent white hairs and of longer upright stiff but fine black hairs, the latter occurring on the head, thorax, propodeum, abdomen (where extremely sparse), antennal scapes and first two flagellar segments. Some- what longer coarser forwardly directed ferruginous hairs are present on the clypeus; shorter ferruginous hairs are present sparsely upon the mandibles. Short decumbent white hairs densely cover the posterovental extremities of the meso- and metapleura. Puncturation of frons and vertex reticulate; reticulum on frons fine and without interstices between punctures, becoming less fine at level of ocelli and course with moderately shining interstices between elongate punctures on vertex; punctures on gena large and elongate with shining interstices; pronotum, scutum, scutellum and mesopleuron with large coarse punctures similar in size to those of gena; tergites and sternites highly polished and shiny all over and fairly uniformly covered with small shallow and widely separated punctures. Head (Fig. 3). Frontal lobes well developed, very strongly raised laterally, not expanded laterally and in dorsal view not covering antennal sockets; frontal carinae strongly divergent posteriorly, strongly raised over their entire length and ending abruptly far below level of anterior ocellus. Area between frontal carinae somewhat elevated medially. Clypeus (Figs 3 and 10) with elevated and produced median part tectiform, weakly carinate in mid-line and moderately curved in profile; anterior median tooth moderately developed, its apex acutely rounded; lateral teeth moderately developed, antero-laterally directed, rounded, well separated 9 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 Figs 13-17. Frontal view of entire labrum (Figs 13-14) or of anterior part of labrum (Figs 15-17) of: 13, 9 A. cyanura Kohl; 14, 9 A. hanluae sp. nov.; 15, 9 A. lesothoensis sp. nov.; 16, 9 A. montivaga sp. nov.; 17, 9 A. nigrisetosa sp. nov.. from median tooth; lateral wings of clypeus well developed, antero-laterally directed and with their free edges smooth and slightly upturned, weakly separated from lateral teeth by narrow and shallow emargination and terminating abruptly posteriorly and separated from mandibular socket by a deep emargination. Labrum (Figs 10 and 15) robust, with median part greatly for- wardly produced and by far exceeding strongly downcurved lateral wings which flank it basally; median part more or less square in crosssection, in dorsal view with sides subparallel (converg- ing very slightly anteriorly) and with upper apical margin pointedly bilobed, with apical face concave and ventral face flat; downcurved lateral wings lamelliform distally, widely rounded and superior in size to apical lobes of median part; emargination between lateral wing and median part of labrum with a pair of downwardly directed, short, stout and bluntly rounded spatulate setae. Mandibles (Figs 3 and 20) robust, not compressed, narrowed towards apex and slightly downturned, with a poorly developed emargination dorsally at the base and with part basad to emargination not produced over the latter and not forming an acute notch, with a 10 GESS; TAXONOMY OF AMPULEX JURINE (HYMENOPTERA: SPHECIDAE: AMPULICINAE) very low longitudinal lamelliform ridge projecting from near mid-length to shortly before apex, without a small tooth immediately below and behind apex but with two low cusps on outer sur- face apically and with apex itself incurved to form a sharp cusp on inner surface, with a large pointed subapical cusp on inner surface near upper edge. Pronotal collar broader than long, more than twice (2,2) as wide behind as long in the mid- line, shorter than scutum; dorsum of collar plain, without a conical elevation but with a wide and deep longitudinal sulcus in the mid-line; scutum with notauli deep and wide and extending to posterior margin; scutellar disc less than twice as wide basally as long in the mid-line; meta- nota! disc very strongly raised throughout and especially laterally where subtuberculate, not expanded posteriorly; propodeum less than twice (1,9) as wide across base as long in the mid- line of its dorsal face, narrowed posteriorly (0,8 times as wide across postero-lateral teeth as across base); propodeal dorsum with a median longitudinal carina and four pairs of posteriorly converging lateral carinae; first pair moderately developed, sublamelliform, inwardly sloping, not shagreened, not attaining hind end of dorsum; second, third and fourth pairs somewhat indistinct due to coarse reticulate surface sculpturing; second pair not joining nor attaining hind end of dorsum but third pair doing so and continuing around and margining hind edge; fourth pair not lamelliformly produced below spiracles. Tergite 1 widest in posterior half, less than twice (1,8) as wide there as long in the mid- line; tergite 2 about one and one-tenth as wide in the anterior half (where widest) as tong in the mid-line, less than twice (1,6) as long as tergite 1 and of the same width. Forewing with three submarginal (= cubital) cells, with the second submarginal cell clear- ly longer (1,6) on the media than on the radius. Male Unknown. Material examined: Lesotho (formerly Basutoland): Mahlatsa, 30.xii.l951 (A. Jacot- Guillarmod), Holotype 9 (Albany Museum). Mahlatsa (29° ll'S, 27° 58'E) is situated in the Berea district of Lesotho in the foothills at an altitude somewhat in excess of 1 800 m. The vegetation is sparse and consists mostly of grass. Etymology: The name, an adjective, is derived from the geographical name, Lesotho, and re- fers to the provenance of the described specimen. Ampulex montivaga sp.nov. Female (Figs 4, 11, 16 and 21) Length 10-13 mm (Flolotype 10 mm) Head, upper surface of scapes, thorax, abdomen, all legs to end of tibia, first tarsomere of metathoracic leg, black with metallic blue, green or purple lustre. Antennae (except for upper surface of scapes), mandibles to a variable extent, tarsomeres (other than blue first tarsomere of metathoracic leg and ferruginous apices of all tarsomeres) black. Mandibles at base and apex, parts of tarsomeres and normally hidden parts of abdominal segments dark ferruginous . Wings subhyaline, very lightly browned, without noticeable fasciae, beset with short semi- erect brown setae. The vestiture consists in the main of sparse very short decumbent white hairs and of longer upright stiff but fine black hairs, the latter occurring on the head, thorax, propodeum, abdomen (where extremely sparse), legs, antennal scapes and first two flagellar segments. Somewhat longer coarser forwardly directed black hairs are present on the clypeus; shorter 11 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 Figs 18-22. Lateral view of right mandible of; 18, 9 A. cyanura Kohl; 19, 9 A. bantuae sp. nov.; 20, 9 A. lesothoensis sp. nov.; 21, 9 A. montivaga sp. nov.; 22, 9 A. nigrisetosa sp. nov.. coarse black hairs are common upon the mandibles. Short decumbent white hairs densely cover the basalar lobe and the postero-ventral extremities of the meso- and metapleura. Puncturation of frons and vertex reticulate; reticulum on frons close with very narrow shining interstices between punctures, becoming less close at level of ocelli and on vertex coarse with wider shining interstices; pronotum, scutum, scutellum and mesopleura with large coarse punctures similar in size to those of vertex; tergites and sternites of first two abdominal segments highly polished and uniformly covered with small shallow widely separated punc- tures; tergite 3 more closely punctured expecially basally. 12 GESS: TAXONOMY OF AMPULEX JURINE (HYMENOPTERA: SPHECIDAE: AMPULICINAE) Head (Fig. 4). Frontal lobes well developed, strongly raised laterally, not expanded lat- erally and in dorsal view not covering antennal sockets; frontal carinae diverging posteriorly, evanescent far below level of anterior ocellus. Clypeus (Figs 4 and 11) with elevated and pro- duced median part markedly tectiform, strongly carinate in the mid-line and weakly but smoothly curved in profile; anterior median tooth well developed, its apex rounded; lateral teeth moderately developed, antero-laterally directed, acutely rounded, well separated from median tooth; lateral wings of clypeus well developed, antero-laterally directed and with their free edges smooth, well separated from lateral teeth by a wide but shallow emargination and terminating abruptly posteriorly and separated from mandibular socket by a narrow but deep emargination. Labrum (Figs 11 and 16) robust, with median part greatly produced and by far exceeding strongly downcurved lateral wings which flank it basally; median part in dorsal view with sides diverging anteriorly and with anterior margin widely bilobed (lobes antero-laterally pointing and noticeably bent down anteriorly), its upper surface more or less flat and its ventral surface concave due to lateral thickening; lateral wings lobate but not lamellately downwardly produced; emargination between lateral wing and median part of labrum with a pair of downwardly directed short, stout and bluntly spatulate setae. Mandibles (Figs 4 and 21) robust, neither compressed nor wide in lateral view and lacking a lamelliform projection of the lower edge, narrowed at apex and distinctly downturned, with a poorly developed emargina- tion or notch dorsally at the base, with a small tooth immediately below and behind apex and with a large pointed subapical lamellate cusp on inner surface near upper edge. Pronotal collar broader than long, twice as wide behind as long in the mid-line, shorter than scutum; dorsum of collar plain, without a conical elevation but with a narrow and shallow longitudinal sulcus in mid-line; scutum with notauli extending to posterior margin; scutellar disc less than twice as wide basally as long in the mid-line; metanotal disc moderately raised, not expanded posteriorly; propodeum less than twice (1,6) as wide across base as long in the mid-line of its dorsal face, narrowed posteriorly (0,8 times as wide across postero-lateral teeth as across base); propodeal dorsum with a median longitudinal carina and four pairs of poste- riorly converging lateral carinae; first pair moderately developed, sublamelliform, shagreened, not attaining hind end of dorsum; second, third and fourth pairs indistinct (particularly poste- riorly) due to weak development and coarse surface sculturing; second pair not joining nor attaining hind end of dorsum but third pair doing so and continuing around and margining hind edge; fourth pair not lamelliformly produced below spiracles. Tergite 1 widest in posterior half, less than twice (1,75) as wide there as long in the mid- line; tergite 2 about one and one-tenth as wide in the anterior half (where widest) as long in the mid-line, less than twice (1,75) as long as tergite 1 and about one and one-tenth as wide. Forewing with three submarginal (= cubital) cells, with the second submarginal cell clear- ly longer (1,5-1, 8) on the media than on the radius. Male Unknown. Material examined: Lesotho (formerly Basutoland): Haha-la-Sekhonyana, 30.xii.l946 (N. C. Mokhehle), Holotype 9, same locality and date (A. Jacot-Guillarmod), Paratype 9 (both Albany Museum). Haha-la-Sekhonyana (29° 22'S, 28° 19'E) is situated in the Maseru district of Lesotho in the mountains at an altitude of 2 450-2 750 m. The vegetation, on basalt, is sparse and consists mostly of grass. Etymology: The name, derived from the word montivagus -a -um: wandering over the moun- tains, refers to the species’ occurrence in a mountainous situation. 13 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 Ampulex nigrisetosa sp. nov. Female (Figs 5, 12, 17 and 22) Length 10-13 mm (Holotype 10,7 mm) Head, thorax, abdominal tergites 1 — 3 and corresponding sternites, coxae and femora of all legs and dorsal aspect of tibiae of hind legs black with metallic blue lustre (sometimes purplish in dead specimens). (The metallic lustre is poorly developed on the abdomen and legs and is not apparent in all the specimens; sometimes the entire abdomen is lustreless black.) Tibiae of all legs (except for dorsal aspect of hind tibiae) black, apparently without metallic lustre; tarso- meres varying from black to dark ferruginous as is the case also with the antennal flagellum. 2 mm Figs 23-24. Frontal view of head of: 23, $ A. mutilloides Kohl; 24, ? A. timidloides sp. nov.. Mandibles, antennal scapes (at least on underside) and normally hidden parts of abdominal segments ferruginous . Wings lightly browned, beset with semi-erect brown setae the density of which is propor- tional to the amount of brown pigment in darker areas of the wing. The dark areas include: a longitudinal streak in the medial cell, most of the submedial cell, the proximal edge of the sub- discoidal cell and part of the wing membrane posterior to these two cells; the marginal (= radial) cell, the distal half of the first submarginal (= cubital) cell, the entire second submarginal cell, the proximal half of the third submarginal cell and a diffuse band across the middle of the sec- ond discoidal cell. The vestiture consists in the main of very short sparse decumbent white hairs and of black upright hairs. Long, very coarse and stiff black pilosity occurs on the frons and vertex, on the dorsal surfaces of the pronotal collar, the scutum, scutellum and metanotal disc, and on the hind femora and tibiae; shorter and finer black pilosity occurs on the mesopleura, underside of the pro- and mesothorax, pro- and mesothoracic legs and sternite 2 (where very sparse). Short decumbent white hairs densely cover the basalar lobe and the postero-ventral extremities of the meso- and metapleura. 14 GESS: TAXONOMY OF AMPULEX JURINE (HYMENOPTERA: SPHECIDAE: AMPULICINAE) Puncturation of frons, vertex, genae, pronotum, scutum, scutellum and mesopleura coarse with shining interstices; interstices on dorsum and sides of pronotum strongly raised and fused to form pronounced longitudinal rugae; abdominal tergites moderately covered with small shallow punctures. Head (Fig. 5). Frontal lobes well developed, moderately raised laterally, not expanded laterally and in dorsal view not covering antennal sockets; frontal carinae low, diverging poster- iorly, evanescent far below level of anterior ocellus. Clypeus (Figs 5 and 12) with elevated and produced median part markedly tectiform, strongly carinate in mid-line and strongly curved in profile; anterior median and lateral teeth of equal size, strong and acutely rounded, separated by a deep but narrow emargination; lateral wings of clypeus well developed, antero-laterally directed and with their free edges serrate, separated from lateral teeth by a small emargination and extending posteriorly without interruption and without emargination the entire distance to the mandibular socket. Labrum (Figs 12 and 17) robust, with median part greatly produced and by far exceeding downcurved lateral wings which flank it basally; median part in dorsal view with sides converging anteriorly, with distal end widely rounded (not bilobed) and slightly bent down, with its upper surface more or less flat but its ventral surface progressively raised in the mid-line from apex to base; lateral wings small, not separated from median part as situated underneath base of latter rather than to each side, without lamellate projections but each with three downwardly and slightly inwardly directed, short, stout and bluntly rounded spatulate setae. Mandibles (Figs 5 and 22) robust, not compressed nor wide in lateral view and lacking any lamelliform projection of lower edge, very slightly downturned near apex, with a well de- veloped emargination dorsally at the base but with part basad to emargination though raised hardly produced over the latter and not forming a deep acute notch, with at most a small, low and rounded cusp immediately below and behind apex, with an elongate subapical lamelliform cusp on inner surface near upper edge. Pronotal collar broader than long, less than twice (1,3-1 ,4) as wide behind as long in the mid-line, of same length as scutum; dorsum of collar plain without a conical elevation but with a narrow and mostly shallow longitudinal sulcus in the mid-line; scutum with notauli extending to posterior margin; scutellar disc less than twice (1,75) as wide basally as long in the mid-line; metanotal disc moderately raised, not expanded posteriorly; propodeum less than twice (1,6) as wide across base as long in the mid-line of its dorsal face, narrowed posteriorly (0,8 times as wide across postero-lateral teeth as across base); propodeal dorsum with a median longitudinal Carina and four pairs of posteriorly converging lateral carinae; first pair well developed but not lamelliform nor shagreened, not attaining hind end of dorsum; second, third and fourth pairs somewhat indistinct due to coarse reticulate surface puncturing; second pair not joining nor at- taining hind end of dorsum but third pair doing so and continuing around and margining hind edge; fourth pair not lamelliformly produced below spiracles. Tergite 1 widest in posterior half, less than twice (1,7) as wide there as long in the mid- line, tergite 2 as wide in anterior half (where widest) as long in the mid-line, almost twice (1,9) as long as tergite 1 and about one and one-tenth as wide. Forewing with three submarginal (= cubital) cells but loss of the first intersubmarginal veinlet (= first transverse cubital vein) in some wings results in only two submarginal cells be- ing present. (In the four specimens examined, one has the veinlet present but weak in both wings, two have the veinlet absent in both wings, and one specimen, the Holotype, has it pres- ent in the left-hand wing and absent in the right-hand wing.) Second submarginal cell when present clearly longer (1,5) on the media than on the radius. Male Unknown. Material examined; Cape Province: Grahamstown, Avenue Park, 23. iv. 1967 (D. Brothers), Holotype $ (with prey) (Albany Museum); Grahamstown, Hilton, 21-31. x. 1970, 1-16.1.1971 15 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 Figs 25-26. Lateral view of anterior part of head of: 25, 9 A. mutilloides Kohl; 26, 5 A. timuUoides sp. nov.. 27 I- I mm 28 Figs 27-28. Frontal view of entire labrum of: 27, 9 A. mulilloides Kohl; 28, 9 A. timuUoides sp. nov.. and 17-31. i. 1971 (all F. W. Gess, Malaise trap) 3 Paratype 9 $ (all Albany Museum). The prey taken with the Holotype is a 9,2 mm long apterous female nymph, possibly a species of Perisphaeria. Etymology: The name, a combination of two words, is derived from niger -ra -rum: black and setosus -a -urn: bristled, and draws attention to the characteristic setae. Ampulex mutilloides Kohl Ampulex mutilloides Kohl, 1893: 456, 468-469, fig. 61, 9; Arnold, 1928: 205-206, fig. 6 \par- tim, 6 only]; Jacot-Guillarmod, 1951: 236 (plant visiting); Gess, 1981: 17, 53-54 (specu- lation re. nesting and prey). Ampulex sanguinicollis Brauns, 1899: 394-395, 6 . [non] Ampulex mutilloides Kohl, Arnold, 1928: 205, fig. 6a, 9 [= A. timuUoides sp. nov.]. Ampulex mutilloides was described by Kohl (1893: 468-469 and fig. 61) from a single female collected by Drege in South Africa (Afr. austr.). A. sanguinicollis, described by Brauns 16 GESS: TAXONOMY OF AMPULEX JURINE (HYMENOPTERA: SPHECIDAE: AMPULICINAE) (1899: 394-395) from a single male collected by himself at Uitenhage not far from Port Eliza- beth, was sunk into synonomy with mutilloides by Arnold (1928: 205). Arnold’s belief that A. sanguinicollis represents the male of A. mutilloides is supported by the study of material of both males and females collected by Jacot-Guillarmod at Mamathes in Lesotho where the species appears to be not uncommon. In his revision of the Sphecidae of South Africa Arnold (1928: 205-206, figs 6, 6a) in dealing with A. mutilloides gave new descriptions in English of both sexes, the original descriptions of both Kohl and Brauns having been in German. It is clear from Arnold’s statement (1922: 104) that the material which he had at his disposal in- cluded Brauns’ but not Kohl’s types. Arnold’s description of the female was therefore based not upon Kohl’s type but upon new material believed by him to be A. mutilloides and consist- ing of two females from the Transvaal, one from Carolina and the other from Pretoria. Comparison of Kohl’s and Arnold’s descriptions and figures pertaining to the female show considerable descrepancies of which the most immediately obvious involve the puncturation of the head and the form of the clypeus. Concerning the puncturation Kohl stated: “Kopf gross und wie der Thorax dicht und ungemein grob . . . punktirt; zwischen den Punkten verlaufen Runzeln” (head large and like the thorax closely and unusually coarsely punctured; between the punctures run wrinkles). Arnold, on the other hand, stated: “Head strongly and rather irregularly punctured; the punctures smallest on the middle of the face and on the temples, on the face fairly far apart; the sides of the face are closely punctured and somewhat rugose. The vertex and occiput have widely spaced, large and deep punctures, the spaces between them smooth and shining”. With respect to the clypeus, that figured by Kohl (fig. 61) appears to be very different from that figured by Arnold (fig. 6a). The suspicion is therefore aroused that the females from Carolina and Pretoria purported to be A. mutilloides are not in fact conspecific with Kohl’s female but rather represent a closely related but distinct species. Examination of the specimens included under A. ‘mutilloides’ in the Albany Museum and in the South African Museum collections (the latter holding Arnold’s material formerly belonging to the National Museum in Bulawayo) showed that, of the total of eighteen females, fourteen could be assigned to A. mutilloides sensu Kohl and four to A. mutilloides sensu Ar- nold and that the differences between the two groups of specimens were sufficient to warrent specific separation. Fig. 29. Lateral view of right mandible of 9 A. mutilloides Kohl. 17 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 The four females answering to A. mutilloides sensu Arnold are assigned to A. timulloides sp. nov.. Major differences between A. mutilloides and A. timulloides are discussed below under the latter species. Figure citations, measurements and list of examined material given immediately below pertain to A. mutilloides Kohl. Female (Figs 23, 25, 27 and 29) Length 10,5-16,5 mm (average of 15 specimens 13,4 mm) Adequately described by Kohl {loc. cit.). Male Length 7,1-11,1 mm (average of 15 specimens 9,1 mm) Adequately described by Brauns and by Arnold {loc. cit.). Material examined: Cape Province: Algoa Bay, no date (Dr Brauns) 9 (South African Museum ex National Museum Bulawayo 1981); Aliwal North, x.1945 (N. C. Mokhehle) 9 (Albany Museum); Grahamstown, 20.iii.l969 (J. Pringle) 9 (Albany Museum); Grahams- town, Hilton, 25. i. 1974 (F. W. & S. K. Gess) 9 (Albany Museum). Lesotho (formerly Basuto- land): Mamathes, Sept. -Jan., 1942-1953 (C., A. & F. Jacot-Guillarmod, D. Wickham and N. C. Mokhehle) 89 9, 14c? d (59 9, 126 6 in Albany Museum; 3 9 9 , 2d d in South African Museum ex National Museum Bulawayo 1981). Natal: Frere, iii.1891 (G. A. K. Marshall) 6 (South African Museum). Zimbabwe (formerly S. Rhodesia): Salisbury, iv.l913 (no collector) 9 (South African Museum). No locality, no date (no collector) 29 9 (Albany Museum). Ampulex timulloides sp. nov. Ampulex mutilloides {non Kohl) Arnold, 1928: 205, fig. 6a \partim, 9 only]. Female (Figs 24, 26 and 28) Length 14,8-18 mm (average of 4 specimens 16,2 mm; Holotype 18 mm) Adequately described by Arnold {loc. cit.) (as mutilloides), it is very similar in colouration and general facies to A. mutilloides Kohl from which it may, however, be readily distinguished on the basis of the characters which follow. The head (Fig. 24) is of relatively greater width than that of A. mutilloides (Fig. 23), its width across the eyes being equal to c. 0,29 of the total body length as opposed to c. 0,26. However, as accurate measurement of the total body length is complicated by factors such as varying degrees of body flexion and of abdominal telescoping, head width is better compared to a body measurement not subject to such variations. Thus in Fig. 30 the width of the head across the eyes is plotted against the length from the anterior margin of the scutum to the brink of the propodeal declivity measured along the dorsal midline. Regression lines fitted to the sets of points show that whereas the relationship of head width to scutum-propodeal declivity length is consistent within each species, the relative head width of A. timulloides is consistently greater. The vertex and occiput have widely spaced, large and deep punctures, the spaces between them being smooth and shining whereas in A. mutilloides the vertex and occiput have close and coarse punctures separated by rugae. The lateral wings of the clypeus are poorly developed and receding (Figs 24 and 26). An- teriorly they are in the same plane as the lateral teeth from which they are not separated but from which they extend in a smooth, very shallow and extremely wide curve to their weakly or 18 GESS; TAXONOMY OF AMPULEX JURINE (HYMENOPTERA; SPHECIDAE: AMPULICINAE) Fig. 30. Regression lines fitted to width, in mm, of head across eyes (vertical axis) plotted against length, in mm, from anterior margin of scutum to brink of propodeal declivity measured along dorsal midline (horizontal axis) for 4 females of A. timulloides sp. nov. (above) and 14 females of A. mutilloides Kohl (below). only moderately produced tooth-like postero-lateral extremities. In mutilloides, by contrast, the lateral wings (Figs 23 and 25) are well developed and produced. Anteriorly they are in a plane distinct from that of the lateral teeth from which they are also clearly separated. Arising from behind and slightly below the level of the lateral teeth they are relatively short and mod- erately straight and are increasingly strongly produced and flange-like towards their postero- lateral extremities where they are abruptly truncated. The median part of the labrum (Figs 26 and 28) is only weakly raised above the level of the lateral wings and is also only weakly concavely dished whereas in A. mutilloides the median part of the labrum (Figs 25 and 27) is both strongly raised and strongly concavely dished. Male Unkown. Material examined; Transvaal; Carolina, 13. i. 1917 (G. A. H. Bedford) Holotype ? (South African Museum ex National Museum Bulawayo 1981); Pretoria, 24.x. 1914 (no collector) Paratype ? (South African Museum ex National Museum Bulawayo 1981); Rustenb[urg], iii. [18] 85 (L. Schunck) Paratype 2 (South African Museum ex National Museum Bulawayo 1981) (bearing two additional labels: 127; AmpulexI mutilloides! Kohl 9/ determ/ Dr. Brauns); Ermelo, 21. xi. 1948 (N. C. Mokhehle) Paratype 9 (Albany Museum). Etymology: The name timulloides, suggested to the author by his colleague, the late Mr C. F. Jacot-Guillarmod, is an anagram of mutilloides and thereby draws attention to the similarity shown by these two species. 19 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 KEY TO FEMALES OF SOUTHERN AFRICAN SPECIES OF AMPULEX IN WHICH THE PRONOTUM IS NOT POSTERIORLY RAISED INTO A CONICAL TUBERCLE The genus Ampiilex may for convenience be divided into those species in which the pronot- al collar is posteriorly raised into a conical tubercle and those in which such a conical tubercle is lacking. Of the 42 species of Ampulex listed for the Afrotropical Region by Bohart and Menke (1976: 77-78) at least 12 species belong to the tubercle-less group. To their number may be added the five species here described as new. The present key dealing with the tubercle-less group is limited to those species occurring in southern Africa. It is further limited to the females for, as only one of the newly described species, bantiiae, is known from both sexes, little could be added to the key to males given by Arnold (1928). In accordance with these restrictions the following species have been omitted: honesta Kohl and splendidiila Kohl, both described from females and known respectively from Gabon and Zaire and from central Africa; overlaeti Leclercq and pilipes Kohl, both described from males and known respectively from Zaire and from Guinea and Mozambique. A. chalybea Smith, described from a female from an unspecified part of Africa, is omitted as the description is entirely inadequate. 1. Pronotal collar without a conical tubercle behind 2 — Pronotal collar with a conical tubercle behind See Arnold (1928: 200) 2. Pronotal collar with two transversely placed small, low and blunt teeth at about or a little behind middle; body black, without metallic lustre (Forewing with two cubital cells) 3 — Pronotal collar unarmed; body with at least some metallic lustre (blue, green or purple) 4 3. Legs with coxae and femora fusco-ferruginous and tibiae and tarsi ferruginous (occasionally legs are entirely ferruginous); abdomen dull, microscopically reticulate- punctate (Cape Province, Natal, Swaziland, Transvaal and Zimbabwe) denticollis Cameron — Legs with coxae, femora and tibiae black and last three tarsal segments brownish- black; abdomen smooth and shining, second tergite with a few very small scattered punctures (Malawi, Mozambique and Urundi) lugubris Arnold 4. Head (to a varying degree), pronotal collar, scutum and scutellum bright ferruginous and without metallic lustre 5 — Head, pronotal collar, scutum and scutellum without ferruginous colour 6 5. Head width across eyes equal to or slightly less than length from anterior margin of scutum to brink of propodeal declivity measured along dorsal midline (see Fig. 30); vertex and occiput with close and coarse punctures separated by rugae; lateral wings of clypeus (Figs 23 and 25) well developed and produced and anteriorly separated from lateral teeth; median part of labrum (Figs 25 and 27) strongly raised above lateral wings and strongly concavely dished (Cape Province, Lesotho, Natal and Zimbabwe) mutilloides Kohl — Head width across eyes exceeding by a mean factor of 1,14 length from anterior margin of scutum to brink of propodeal declivity measured along dorsal midline (see Fig. 30); vertex and occiput with widely spaced, large and deep punctures, the spaces between them smooth and shining; lateral wings of clypeus (Figs 24 and 26) poorly developed and receding, anteriorly not separated from lateral teeth; median part of labrum (Figs 26 and 28) only weakly raised above lateral wings and only weakly con- cavely dished (Transvaal) timulloides sp. nov. 6. Clypeus produced into a broad, subquadrangular and transversely convex plate, not carinate longitudinally in middle and without a median tooth on anterior margin; fun- 20 GESS; TAXONOMY OF AMPULEX JURINE (HYMENOPTERA: SPHECIDAE: AMPULICINAE) damental sculpture on head and pro-mesonotum exceedingly fine, consisting of a microscopic and very close puncturation, so that those parts are dull; fourth carina of epinotal dorsum obsolete. (Forewing with two cubital cells.) Small species, 8 mm long (Griqualand West and Zimbabwe) arnoldi Brauns — Clypeus differently formed, tectiform or nasiform, carinate longitudinally in middle, carina ending on anterior margin in a distinct tooth or on a wide, subtruncate lobe {bantuae) 7 7. Clypeus (Figs 2 and 9) nasiform; lateral wings of clypeus poorly developed and like lateral teeth downwardly directed and thus appearing to be strongly receding when seen from above; labrum (Fig. 14) triangular in overall shape, greatly but regularly widened from base to apex and with extensive lateral wings exceeding median part; frontal lobes (Fig. 2) arcuately expanded laterally and completely covering antennal sockets; frontal carinae parenthesis-like (Cape Province) bantua sp. nov. — Clypeus (Figs 1, 3, 4 and 5) tectiform; lateral wings of clypeus well developed and like lateral teeth antero-laterally directed; labrum (Figs 13, 15, 16 and 17) differently formed, with median part greatly produced and by far exceeding strongly down- curved lateral wings which flank it basally; frontal lobes (Figs 1, 3, 4 and 5) not arcuately expanded laterally and not covering antennal sockets; frontal carinae diver- ging posteriorly 8 8. Lateral wings of clypeus extending postero-laterally without interruption and without emargination entire distance to mandibular sockets (Fig. 5); forewings with a ten- dency towards loss of first intersubmarginal veinlet leading to frequent coalescence of first and second submarginal cells; (head, thorax and legs with sparse, very coarse black pilosity) 9 — Lateral wings of clypeus not extending postero-laterally without interruption but ter- minated abruptly and separated from mandibular sockets by an emargination (Figs 1, 3 and 4); forewings not showing this tendency and therefore always with three sub- marginal cells 10 9. Claws of all legs bifid (inner ramus of claws apically pointed, wider than outer ramus and only slightly shorter); body metallic blue and green, last three tergites and apical half of third, last four sternites, clypeus, mandibles, scapes and pedicels ferruginous (Cape Province and Natal) apicalis Smith — Claws of all legs toothed (tooth acutely pointed, situated midway along length of claw and at right angles to it); body metallic blue (sometimes abdomen is lustreless black) with only mandibles and antennal scapes ferruginous (Cape Province) nigrisetosa sp. nov. 10. Small species, 7,5 mm or less in length (mandibles, apex of clypeus, antennae, apical segment of abdomen, legs various shades of brown; wings hyaline, forewing with a transverse fuscous band) (Cape Province) nebulosa Smith — Larger species, 10-16 mm in length 11 11. Labrum (Figs 10 and 15) with median part more or less square in cross-section, in dorsal view with sides subparallel and with apical margin pointedly bilobed, with apical face concave and ventral face flat; wings strongly and fairly uniformly browned, without noticeable fasciae; (frontal lobes very strongly raised laterally; frontal carinae strongly divergent posteriorly; puncturation of head and thorax coarse) (Lesotho) lesothoensis sp. nov. — Labrum (Figs 8 and 13; 11 and 16) with median part flattened, in dorsal view with sides diverging anteriorly and with anterior margin widely bilobed and slightly down- curved, with ventral face concave; wings subhyaline, lightly browned, with fasciae (cyanura) or without fasciae (montivaga) 12 21 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 1, APRIL 1984 12. Mandibles black (except for tips), other than for absence of metallic lustre not contrasting with colouration of genae (mountains of Lesotho) montivaga sp. nov. — Mandibles ferruginous, contrasting markedly with colouration of genae 13 13. Mandibles in lateral view with a pronounced bluntly triangular lamelliform projection of lower edge (Fig. 18); punctures of first two tergites shallow, fine to moderate in size 14 — Mandibles in lateral view with only a very low lamelliform projection of lower edge; punctures of first two tergites deep, very distinct, moderate in size; clypeus ferrugi- nous (mountains of eastern Zimbabwe) cyanura monticola Arnold 14. Clypeus, scapes and frontal lobes black (Cape Province and Zululand) cyanura cyanura Kohl — Clypeus, scapes and anterior portions of frontal lobes ferruginous (Zimbabwe) cyanura rhodesiana Arnold ACKNOWLEDGEMENTS The author wishes to thank Dr F. Koch of the Museum fiir Naturkunde an der Humboldt- Universitat zu Berlin and Mr M. C. Day of the British Museum (Natural History), London, for making it possible for him to examine the holotype of Ampulex cyanura Kohl. Dr V. B. Whitehead of the South African Museum (Natural History), Cape Town, is thanked for the loan of the types of A. cyanura rhodesiana Arnold and A. cyanura monticola Arnold, of material determined as either A. cyanura {sensu stricto) or as one or other of the above subspecies, and of material determined as A. mutilloides Kohl. Sarah Gess is thanked for her interest in and encouragement of the Ampulex study and especially for the pains she took in the drawing of and preparation of the figures. Gratitude is expressed to the C.S.I.R. for running expenses grants which facilitated the study. REFERENCES Arnold, G. 1922. The Sphegidae of South Africa. Part 1. Ann. Transv. Mus. 9 (2): 101-138. Arnold, G. 1928. The Sphegidae of South Africa. Part 9. Ann. Transv. Mus. 12 (3): 191-279. Bohart, R. M. and Menke, A. S. 1976. Sphecid wasps of the world: a generic revision. Berkeley; University of Cali- fornia Press. Brauns, H. 1899. Zur Kenntnis der sudafrikanischen Hymcnopteren. Annin naturh. Mus. Wien 13 (4): 382-423. Callan, E. McC. 1976. Notes on Ampulicinae with special reference to African species and prey (Hymenoptera: Sphecidae). Rev. Zool. afr. 90 (1): 228-234. Cameron, P. 1905. On some new genera and species of Hymenoptera collected by the Revd. J. A. O’Neil, S. J., chiefly at Dunbro/dy, Cape Colony. Rec. Albany Mus. 1 (4): 245-265. Gess, F. W. 1981. Some aspects of an ethological study of the aculeate wasps and the bees of a karroid area in the vicinity of Grahamstown, South Africa. Ann. Cape Prov. Mus. (nat. Hist) 14 (1): 1-80. Gess, F. W. and Gess, S. K. 1981. Three Acacia insects: a chain of dependence. The Naturalist 25 (3): 27-30. Jacot-Guillarmod, C. 1951. A South African leguminous plant attractive to Hymenoptera. Ent. inon. Mag. 87: 235-236. Kohl, F. F. 1893. Ueber Ampulex Jur. (s.L.) und die damit enger verwandten Hymenopteren-Gattungen. Annin naturh. Mus. Wien 8 (3 & 4): 455-516. 22 m ■A' - ->i^:-'vf ■ :AA ■ ' S*'' ■ ■ ' ^ a 'Aa ^ ’uy ’3sJ«8fV*J’:» ■••■•=•;. ....- •■...■“ ' •■ ;- -. . '«■ J*«. , -:' ■ P A .: . :'. •• ■'= • * ;v, .; ;W-‘ - ,:' 'f .."S--,**' ;:• 'W ‘ • '• k.‘ ;.r , .a4A‘... • V -' F- \U-’' •■-' ■ ■" ^ i* i,v. •S£- ;’")-'"*w: '■<' ■'.^^^S$!^^tj:^r'^'#s ' >r-: ■ • ■ -#'■'" '.i?^^ ■- ■' ’ ■■ ' '^' A!:< . i/'. •■T C?‘f' ' V :... ' ’ • ■ / ‘iStlV^ljltiiy fk^'-f!jiij:^Ag}- iS'AT --ji^.^ ai's =■' . •. 'it)^ ' '•' ' ' *»o m -Mr: -- '^'- '■ - •-' ; :<■.-■ -My:- %?:s'5i> ■-_;V Wi ■i'< 4:,:2 .<3 I INSTRUCTIONS TO AUTHORS LAYOUT of recent issues should be examined for details of acceptable style and format. MANUSCRIPTS should be submitted in triplicate, typed on A4 size paper, double- spaced, on one side of the paper only and with wide margins all round. Only generic and specific names should be underlined to indicate italics. All other mark- ing up should be left to the editor. Footnotes should be avoided as far as possible. All measurements should be metric. Systematic papers should follow the rules of the International code of zoological nomen- clature or the International code of botanical nomenclature , as applicable. ILLUSTRATIONS should bear the author’s name, title of paper and number in soft pencil on the reverse side. Diagrams, line drawings and photographs are all regarded as figures, should be termed as such and should be numbered in a single series. In addition to the original illustrations a photocopy of each should be submitted. Diagrams and line drawings should be reducible to a size not exceeding 14 x 16 cm (17,5 cm including caption). Photographs for half-tone blocks should be final size, of high contrast and printed on glossy paper. Colour illustrations are only printed under special circumstances and their financing must be arranged by the author. A scale should appear with all illustrations or the magnification should be given in the caption. A list of captions should be typed on a separate sheet. TABLES should be typed on separate sheets. Foldouts and tables printed sideways should be avoided. REFERENCES should be listed alphabetically at the end of the paper. Abbreviations of periodical titles should conform with those in the World list of scientific periodicals (4th ed.). Layout must be as follows: Alexander, J. 1969. The indirect evidence for domestication. In: Ucko, P. J. and G. W. Dimbleby eds. The domestication and exploitation of plants and animals. London: Duckworth, pp. 123-129. De Winter, B. 1969. Plant taxonomy today. S.Afr.J.Sci. 65 (8): 229-242. JuBB, R. A. 1967. Freshwater fishes of southern Africa. Cape Town: Balkema. GALLEY PROOFS are sent (when practicable) to authors for checking printer’s errors. Adjustments of language, style and punctuation and any other amendments of the original manuscript, if made at this stage, may be charged to authors. AUTHOR’S COPIES (50) are issued free of charge. When more than one author is involved the fifty free copies will be sent to the senior author. Additional copies ordered before printing are supplied at a reduced cost. BD2195— CTP Book Printers, Cape ■' ■' '- V - ■' - ■ •: ■■ 'T'"' .'V ■ ' / '■' ''•'.y'' . :'r^ I '■'■ '■■ jv-.''''^''? V 'S’' ;i yisa pa. *(i. -. - ' ' ft.-f ' - ■ L^-. *jku.-., ,c.*» ; ’■•^■‘^ •• I -s ■ ^ ■-- ■ ■ ^ ::v •'■■' -■ ■ :■ ■ . ... V"? ..■'nift-V’V^ ■: Annals of the Cape Provincial Museums Natural History Ann. Cape Prov. Mus. (nat. Hist.) Volume 16 Part 2 16th April 1984 Published jointly by the Cape Provincial Museums at the Albany Museum, Grahamstown, South Africa ANNALS OF THE CAPE PROVINCIAL MUSEUMS These Annals are published in two series. Natural History and Human Sciences, the latter series covering cultural history, ethnology, anthropology and archaeology. They are issued in parts at irregular intervals as material becomes available. The primary objective of these Annals is to disseminate the results of research work undertaken by staff members of the Cape Provincial Museums. However, a manuscript by any author is eligible for publication provided it is based wholly or partly on the collection/s of one or more of these museums or if the material on which the work is based is wholly or partly deposited in one of these museums. Parts are obtainable from the Librarian of the Albany Museum. Correspondence in con- nection with these Annals should be addressed to the Editor, Albany Museum, Grahams- town 6140. Editor DrF.W.GESS: 1978 Some aspects of the ethology of Ampulex bantuae Gess (Hymenoptera: Sphecidae: Ampulidnae) in the Eastern Cape Province of South Africa by F. W. GESS (Albany Museum, Grahamstown) CONTENTS Page Abstract 23 Introduction 24 Description of the nesting sites 24 Identification, distribution and biology of the prey 25 Description of the nest 25 Nesting Prey selection 26 Hunting, grasping and stinging, malaxation and conveyance of the prey to the nest- ing cavity 27 Positioning of the prey within the nesting cavity, oviposition and nest closure 32 Life history, voltinism, mating, longevity and fecundity 33 Sex of wasp in relation to size of prey upon which it developed 36 The effect of the sting upon the prey 37 Discussion 37 Acknowledgements 39 References 39 ABSTRACT Some aspects of the ethology of Ampulex bantuae Gess (Hymenoptera: Sphecidae: Ampulicinae) in the Eastern Cape Province of South Africa are described. Particular attention is paid to: the situation and nature of the nest; the identification and biology of the prey; prey selection; the hunting, stinging, malaxation and conveyance of the prey to the nesting cavity; the positioning of the prey within the nesting cavity, oviposition and nest closure; life history, voltinism, mating, longevity and fecundity. A definite relationship between the size of the prey and the sex of the wasp which develops upon it is demonstrated. Some aspects of the hunting and nesting behaviour of nine Ampulex species are briefly reviewed. It is found that the nesting behaviour of A. bantuae does not differ to any significant extent from that of these species and it is concluded that the ethology of Ampulex is character- ized by marked interspecific uniformity. Various adaptations which enable Ampulex to lead an arboreal life are discussed. 23 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 2, APRIL 1984 INTRODUCTION A very brief outline of some aspects of the ethology of Ampulex bantuae Gess (as Ampulex sp. near cyanura Kohl) and a discussion of its role within a community of insects has previously been published as part of a general ethological study of the aculeate wasps and the bees of a karroid area in the vicinity of Grahamstown, South Africa (Gess, 1981: 29 and 72-75, Figs 34-36). The present publication deals in greater detail with those aspects of the ethology which were previously merely outlined and furthermore introduces other aspects which were omitted from the earlier publication as not relevant to its subject. Ampulex bantuae is a medium-sized, metallic-blue wasp not often seen in the field and consequently little known and poorly represented in collections. However, it proved that the species’ rarity in the field is apparent rather than real and that the impression of rarity is large- ly due to the wasp’s clandestine life history. At the site of the study A. bantuae is a not un- common insect within its restricted habitat. The present study resulted from the utilization for nesting purposes by A. bantuae of trap- nests of the type described by Krombein (1970) placed in the field during the summers of 1973-74, 1974-75, 1975-76 and 1976-77. In consequence of the discovery that A. bantuae nested only in those trap-nests suspended within Acacia karroo shrubs and trees, a close exam- ination of these was undertaken, which in turn led to the exciting unfolding of the particulars of the nesting of A. bantuae under natural conditions. The hunting and nesting behaviour of A. bantuae is such that the physical requirements necessary for the normal performance of these activities can readily be provided in the labora- tory. This made it possible in the present study to observe the behaviour of captive individuals and to carry out some experiments which would have been impossible with free wasps in the field. The present paper is the fourteenth in a series of publications dealing with the ethology of certain solitary wasps occurring at Hilton, a farm situated 18 kilometres WNW of Grahams- town (33° 19'S, 26° 32'E) in the Albany Division of the Eastern Cape Province of South Africa. A detailed account of various aspects of the ecology of Hilton has previously been given (Gess, 1981: 3-9). DESCRIPTION OF THE NESTING SITES The natural nesting sites of A. bantuae at Hilton are the disused galleries of Ceroplesis hottentota (Fabr.) (Cerambycidae). The larvae of C. hottentota in the course of their feeding bore galleries in finger-thick terminal branches of A. karroo. After they have completed their larval feeding, have pupated and have metamorphosed into adults, the beetles break out and abandon their galleries with the result that the latter become available for the use of other insects which cannot themselves hollow out such cavities. The natural nests of A. bantuae were discovered as a result of making a survey of trap- nesting wasps and bees at Hilton. A. bantuae nested only in those trap-nests hung in A. karroo and then only in a restricted area (Fig. 1). This led to a closer examination of these shrubs or small trees. The trap-nests utilized by A. bantuae were of the type described by Krombein (1970). Borings of three sizes, 6,4 mm, 9,5 mm and 12,7 mm, were offered but only those of the two smaller sizes were utilized. These trap-nests were suspended from horizontal living branches at heights between 10 cm and 200 cm above the ground. 24 * GESS: ETHOLOGY OF AMPULEX BANTUAE GESS (HYMENOPTERA: SPHECIDAE; AMPULICINAE) Fig. 1. Hilton, iv.l974. One of a scattering of small Acacia karroo trees in which were hung trap-nests which were util- ized by Ampulex hanluae. IDENTIFICATION, DISTRIBUTION AND BIOLOGY OF THE PREY At Hilton, A. bantuae was found to provision its cells with a single species of cockroach, Bantua dispar (Burmeister) (Blattariae: Derocalymmidae), one of four species of the genus Bantua Shelford endemic to southern Africa. B. dispar appears to be restricted in its distribu- tion to the south-eastern parts of the Cape Province. Princis (1963; 130) recorded the species from Grahamstown (the type locality). Resolution (18 km NNE of Grahamstown), Great Fish River (48 km E of Grahamstown), Port Elizabeth, Dunbrody and Willowmore. Specimens in the Albany Museum collection are from several localities around Grahamstown. Adult males of B. dispar are fully winged whereas females are totally apterous. Male and female nymphs are similar in appearance but may readily be distinguished in all instars by dif- ferences in the form of the terminal abdominal sternites. At Hilton B. dispar was found to be restricted to Acacia karroo where during the day it hides under loose bark or in old galleries of Ceroplesis hottentota. Both the nymphs and the adults were very commonly found sheltering in trap-nests suspended from branches of A. karroo. Individuals kept in captivity spent the day in hiding in dark places but came out at night when they fed eagerly upon A. karroo gum which had been placed in the cages. It is believed that it is this gum, which exudes copiously from any injuries to the bark including those caused by the activities of C. hottentota, that constitutes the natural food of B. dispar and that the restriction of the cockroach to A. karroo is due to this food preference. DESCRIPTION OF THE NEST The nest of A. bantuae is single-celled. The only building material added to the pre-exist- ing cavity is debris used for constructing the cell closure. No preliminary plug and no closure to the gallery opening are constructed. 25 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 2, APRIL 1984 Thirteen natural nests were found. Thirty-six nests constructed in trap-nests in the field were obtained, twenty-four were in 6,4 mm borings and twelve in 9,5 mm borings. An additional fifty-four nests were constructed in trap-nests by captive females. Due to the fact that the loosely-packed nest closures were displaced on splitting open the hollowed out A. karroo branches, no accurate measurements were obtained from natural nests. In trap-nests from the field, nests consisting of the cell and its closure were 43-145 mm long with a mean length of 97 mm in the 6,4 mm borings, and 54-77 mm long with a mean length of 63 mm in the 9,5 mm borings. The closures were 16-83 mm long with a mean length of 36 mm in the 6,4 mm borings, and 27-41 mm long with a mean length of 35 mm in the 9,5 mm borings. The materials comprising the closures were loosely compacted and consisted of a mixture of various kinds of detritus collected by the wasp on or beneath the Acacia karroo shrub or tree from which the trap-nest was suspended. TTie materials utilized in natural nests and in trap-nests in the field consisted in order of frequency of use of: small pieces of A. karroo bark, sometimes lichen covered; A. karroo leaflets, mostly dry; faecal pellets of caterpillars; frag- ments of insect exo-skeletons, including the pronotal shield of B. dispar and an elytron of C. hottentota\ the exuviae of B. dispar-, small twigs; A. karroo inflorescences, both fresh and dry; A. karroo seeds; a length of dry grass leaf blade; a dried-out tick; rodent droppings and fragments of dung of larger mammals; a live lepidopterous pupa (Limacodidae). It appears that any available detritus is utilized for constructing the closures. Wasps nesting in captivity took small rounded quartzite pebbles if no other materials were provided. NESTING Prey selection The prey of A. bantnae, Bantiia dispar, is a hemimetabolous insect. All stages, from fresh- ly hatched first-instar nymphs to adults of both sexes therefore occur together upon Acacia karroo and are met with by the wasp during the course of her hunting. Like all species of Ampidex, A. bantnae provisions each of her cells with but a single cock- roach. As this has to provide all the nutritional requirements of the wasp larva which will feed upon it, it must clearly be selected by the wasp to satisfy certain physical parameters. These could to a large degree be established from prey found in natural nests and in trap-nests at Hilton and from the study of the response by captive female wasps to cockroaches differing in size, age and sex. Forty-two cockroaches utilized as prey in natural nests and in trap-nests at Hilton were examined. Their lengths varied from 15-22 mm. The majority, thirty-seven, were females, both nymphs and adults, with lengths ranging from 17-22 mm and averaging 20,2 mm. The remainder, five, were male nymphs with lengths of 15, 17, 18 and 19 (two) mm. No adult (winged) males were represented amongst the prey. Given cockroaches of both sexes and a range of sizes, captive A. bantnae females showed prey preferences similar to those established for free females. Fifty-eight cockroaches were taken as prey. The majority, forty-eight, were female nymphs and adults ranging in length from 13,0-23,3 mm (average 19,3 mm). The remainder, ten, were male nymphs ranging in length from 14,7-19,7 mm (average 18 mm). As in the sample from field nests no adult (winged) males were represented amongst the prey. Nymphs at the extreme lower end of the size range were taken only when larger cock- roaches were withheld. The smallest nymph taken, a female of 13,0 mm, was stung and trans- ported to a nesting cavity but was abandoned there prior to oviposition. The smallest prey to be oviposited upon was a slightly larger nymph, a female of 13,5 mm, from which a very small 26 GESS: ETHOLOGY OF AMPULEX BANTUAE GESS (HYMENOPTERA: SPHECIDAE: AMPULICINAE) male wasp was successfully reared. Wasps could not be induced to take adult (winged) males even when confined with them in the absence of wingless cockroaches. It is apparent from the above that there are two criteria which individuals of B. dispar have to fulfil in order to be acceptable as prey: they must be larger than a certain minimum size (c. 13,5 mm long) and they must be apterous. The wasp’s acceptance as prey of larger male nymphs, including those in the last nymphal instar, but its consistent rejection of adult males indicates its reaction to some character exhibi- ted by the adult male cockroach. Whether this character is merely a morphological one (such as the presence of wings), a behavioural one depending upon the presence of wings (such as the male cockroach’s habit of stridulating when it is inspected by a wasp), or a chemical one (such as the possession of a pheromone) was not established. However, whatever the means by which the wasp recognises an adult male, its rejection of the latter as prey is of survival value as it would represent a deficient investment in terms of the amount of food it could provide for a wasp larva developing upon it. It was found that for female nymphs and adults and male nymphs (including those in the last nymphal instar) there is a directly proportional relationship between body length and body weight which is common to all three categories. Adult (winged) males, however, have a weight which is less than their body length would suggest on the basis of the above weight/length relationship. Thus the weight of adult male cockroaches was found to be only about 67% of the weight of male and female nymphs of the same body length. The loss in weight at the time of the final moult results in the adult winged male having a weight be- low that found necessary for successful development of wasp larvae upon wingless male nymphs and female nymphs and adults. Adult males are therefore unsuitable for the same basic reason as are nymphs below a minimum size. As stated in the previous section A. bantuae at Hilton restricts its choice of prey to Bantua dispar (Derocalymmidae) which it also readily takes in captivity. In order to establish whether A. bantuae is indeed species-specific with respect to its prey, two cockroaches of other species were offered to captive females. Both met the requirements with respect to body length, body weight and winglessness established for the acceptance by the wasp of individuals of B. dispar. The first, an adult female Pseudoderopeltis sp. (Illattidae) was confined with a female wasp for a period of sixteen days, during the first four days in company with three B. dispar and during the last two days in company with one B. dispar. All four B. dispar were taken as prey but the Pseudoderopeltis sp. was not in any way molested and during the period laid two oothecae. The second, a female, nymph of ?Perisphaeria sp. (Perisphaeriidae) was confined with three female wasps for a period of eleven days and was similarly left unmolested though five B. dis- par introduced into the cage during the period were all taken. The species specificity seen in the field with regard to the prey taken was therefore confirmed in the laboratory. Hunting, grasping and stinging, malaxation and conveyance of the prey to the nesting cavity Hunting is done on foot, the wasp moving hurriedly but by fits and starts, twitching the wings which are held folded over her back. Very noticeable are the antennae which are held out in front of the head and curve out to the sides and which are kept in constant motion, seemingly ‘drumming’ the surface on which the wasp is walking. Close observation of the rapidly moving antennae leads to the conclusion that each antenna as a whole is actually de- scribing an elliptical path, the direction of movement of one being clockwise, that of the other anticlockwise. Together they are brought down towards the surface, move inwards towards each other, are raised up and move apart. At the same time the abdomen is continually moved up and down and is also frequently moved in and out telescopically. A suitable cockroach having been located and, if attempting to escape, pursued, the wasp advances upon it with widely separated mandibles and having positioned herself favourably lunges forward and seizes it by the anterior, antero-lateral or lateral margin of the pronotal 27 ANN. CAPE PROV, MUS. (NAT. HIST.) VOL. 16, PT. 2, APRIL 1984 Figs 2 and 3. Female Ampulex bantuae holding a female Bantua dispar by pronotal shield subsequent to initial stinging (Fig. 2) and flexing abdomen beneath same to administer additional sting (Fig. 3). (Both x4,25) 28 GESS: ETHOLOGY OF AMPULEX BANTU AE GESS (HYMENOPTERA: SPHECIDAE: AMPULICINAE) shield, one mandible being positioned on the upper surface of the pronotal shield, the other on the lower surface , the wasp’s head being held sideways. The wasp then flexes her abdomen downwards and forwards between its wide-spread legs to bring the tip of the abdomen, which is seen to extend telescopically, into position on the underside of the prey, the wasp all the time retaining her hold on the pronotal shield. Probing with the tip of the abdomen in the neck region and between the legs then takes place. In one observed case several stings appeared to be given to the prey, the last being between the prothoracic legs or in the neck region. In another case it appeared that only one sting in the neck region was given. The probing and stinging may take five or more minutes. After the initial stinging the wasp may once or twice straighten her abdomen, pull the cockroach into a slightly different position and again flex the abdomen into the stinging position (Figs 2 and 3). The cockroach having been stung it immediately stops struggling and lies still, its legs un- able to support its weight. The wasp straightens her abdomen and ceases the antennal ‘drumming’ which had been carried on during the stinging and handling of the cockroach. The wasp then usually relinquishes her hold of the prey and withdraws to a short distance where she commences grooming. In grooming, the antennae are cleaned with the fore-legs and the abdomen on its dorsal surface and its sides by the hind-legs. The fore- and middle-legs of each side are rubbed against each other and the hind-legs are rubbed one against the other and also together against one of the middle-legs. The cockroach appears to recover gradually from the initial collapse caused by the stinging and stands up, feebly supporting its weight on its legs. It may do a small amount of grooming and has been seen to raise the thorax on one side and to try with its mouthparts to reach the sting sites on the underside of the thorax. Whilst the cockroach is recovering, the wasp inspects it from time to time. As before, she holds it by the pronotal shield and in addition sometimes tugs it in the direction of the cavity in which nesting is to take place. When the cockroach is sufficiently recovered to stand, the wasp reaches with her head and mandibles beneath the pronotal shield covering the cockroach’s head and takes hold of and draws out forwards one of the antennae (Figs 4, 5 and 6). Taking hold of this antenna near its base the wasp then proceeds to walk backwards leading the cock- roach towards the nesting cavity. The cockroach at the end of its antennal ‘leash’ walks for- wards following the wasp and makes no attempt to resist or to escape even when the wasp occasionally releases her hold in order to run off to inspect the nesting cavity and the way to it. On returning to the cockroach the wasp once again reaches under the pronotal shield in order to grasp the antenna. Occasionally a wasp does not lead away her prey immediately upon first grasping the antenna but rather passes the antenna through between her jaws until a point near the end is reached when the jaws tighten their grip and the wasp proceeds vigorously to pull and tug at the antenna which can be seen to be pulled taut (Fig. 7). The tugging is accompanied by wing buzzing on the part of the wasp and a short length of antenna, the distal five segments or so, is severed and discarded. The wasp retains the end of the antenna between her mouthparts for some time and it appears as if, at this juncture, blood is being imbibed by the wasp from the now slackly held antenna (Figs 8 and 9). While she is imbibing blood the wasp may groom her abdomen and legs. Occasionally both antennae are severed. Therefore it appears that at least in some cases the wasp imbibes blood oozing from a wound or wounds deliberately inflicted upon the prey for this purpose. It is noteworthy that the reaction of the stung cockroach to an- tenna-pulling varies according to where the antenna is grasped. Whereas the cockroach walks forwards when grasped near the base of the antenna as is the case during its conveyance to the nesting cavity, it stands firm when grasped near the tip of the antenna. This behaviour enables the wasp after some effort to pull off the distal segments. 29 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 2, APRIL 1984 Figs 4-6. Female Ampulex hantuae reaching beneath pronotal shield of a stung female Bantua dispar to take hold of an antenna (Figs 4 and 5) and drawing out the same (Fig. 6). (All x2,5) 30 GESS: ETHOLOGY OF AMPULEX BANTUAE GESS (HYMENOPTERA: SPHECIDAE: AMPULICINAE) Figs 7-9, Female Ampulex bantuae pulling taut an antenna of a stung female Bantua dispar prior to severing its distal portion (Fig. 7) and imbibing blood from the cut end of the now slackly held antenna (Figs 8 and 9). Note the severed distal portion of the antenna at bottom right-hand corner of Fig. 9. (All x2,5) 31 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 2, APRIL 1984 Positioning of the prey within the nesting cavity, oviposition and nest closure On arrival at the nesting cavity the wasp enters it backwards and continues backing in until she reaches the inner end, all the while drawing the cockroach in after her so that it comes to face the blind end of the cavity. Oviposition upon the prey then takes place. For this a very characteristic posture relative to her prey is assumed by the wasp, still positioned at the cockroach’s head end. In the case of a cockroach lying on its left side, the wasp’s right man- dible grips the dorsal surface of the pronotal shield and the left mandible the ventral surface of the pronotal shield and the head. The wasp’s left antenna lies dorsally along the ridge of the pronotal shield on the right side and extends as far as the angle of the ridge at the hind end of the shield. The anterior edge of the pronotal shield is situated between the wasp’s antennal bases and at the base of the clypeus. The wasp’s meso- and metathoracic legs extend back- wards towards the blind end of the nesting cavity, its abdomen is flexed downwards and for- wards and extends under the left side of the cockroach with the tip near the base of a meta- thoracic leg where the egg is attached. In a sample of fifty prey cockroaches in trap-nests for which all the necessary details were noted, it was found that there was a random relationship between the number of cockroaches positioned in the cell to lie on their left sides and those positioned on their right sides; similarly that there was a random relationship between the number of times eggs were laid on one side or the other. However, it was found that eggs were more than twice as likely to be laid on the side on which the cockroach was lying than on the opposite side. The egg is cemented over its entire length to the underside of the cockroach and is orien- tated obliquely at an angle of between 35° and 40° to the long axis of the prey. The egg’s poste- rior half is situated upon the basal half of the flattened antero-ventral surface of the meta- thoracic coxa and its anterior half is on the adjoining metathoracic trochantin to which the coxa is articulated (Fig. 10). In only one of the sixty-eight instances in which its position was Fig. 10. Diagrammatic representation of the ventral surface of the base of the right metathoracic leg of a female Bantua dispar showing the positioning upon it of an egg of Ampiilex bantuae. 32 GESS: ETHOLOGY OF AMPULEX BANTUAE GESS (HYMENOPTERA: SPHECIDAE; AMPULICINAE) noted was an egg found situated elsewhere. In this instance it occupied the same position on the coxa and trochantin but the leg involved was that of the mesothorax as opposed to that of the metathorax. The egg did not develop. Oviposition completed, the wasp squeezes past the cockroach and emerges from the nest. Collection of detritus for the sealing of the nesting cavity is thereafter undertaken. Each piece of detritus, held with the mandibles, is carried to the cavity on foot. In captivity in a glass-sided cage with a floor area of 30 cm x 60 cm and furnished with trap-nests and detritus suitable for nest closure it was found that about 2| hours elapsed be- tween the first contact of the wasp with its prey and the completion of the nesting cavity closure following oviposition. LIFE HISTORY, VOLTINISM, MATING, LONGEVITY AND FECUNDITY Eighty-eight of the 100 cockroaches obtained from A. bantuae nests bore viable eggs. Sixty of the young wasp larvae which hatched from these eggs were successfully reared through to the adult stage. The remaining twenty-eight individuals died at various times during their development. For all individuals the dates on which major developmental events occurred were recorded. The egg is off-white and slightly curved. A total of sixteen was measured. All were very similar in size, the average length and the average diameter at the middle being 2,04 and 0,54 mm respectively. The time from oviposition to egg hatch is three to nine days, most commonly five. Larval feeding commences very close to the position of the anterior end of the egg, namely at the weakly sclerotized antero-lateral margin of the trochantin and continues at this point for a period of four to eleven days. During this time the larva moults but retains a position very similar to that previously occupied by the egg. The larva as it grows covers most of the trochantin and the flattened antero-ventral face of the coxa, its posterior end fitting into the angle formed by the distal part of the coxa and the proximal part of the femur. When the larva has attained a length of about 4,5 mm it migrates from the exterior feeding position through the hole created during feeding to within the body of the still living cockroach, disappearing from view. The egg pellicle and the larval exuviae are left adhering to the coxa and trochantin. The larva continues feeding within the cockroach which dies two to six, most commonly three, days after larval entry. The presence of the A. bantuae larva within the cockroach causes its body to be slightly arched, its sternites to be noticeably curved and its legs frequently to be somewhat raised and held away from its body (Fig. 11). After the death of the cockroach its exo-skeleton dries in this configuration and in the field dead cockroaches containing A. bantuae are fairly easy to recognize on this account. The mature A. bantuae larva, having eaten out all the soft tissues inside the cockroach, remains within the now empty but perfectly intact body shell within which it subsequently spins its cocoon and pupates. The cocoon is brown, smooth, shiny, hard and brittle and is terminated at each end by a pronounced nipple-like projection. This hard cocoon is contained within an envelope of off- white spinnings which are adpressed to the inside of the cockroach’s exo-skeleton over most of the length of the cocoon but are more loosely arranged at the cocoon’s two ends where they are also most plentiful. The hard cocoon almost entirely fills the available space within the cockroach. For example, a cocoon which was 11,3 mm long and had a maximum width at the middle of 3,8 mm was accommodated within the body of a cockroach the external length of which was 15 mm. That the anterior end of the cocoon was directed towards the anal end of the cockroach 33 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 2, APRIL 1984 11 12 Fig. 11. Female Banlua dispar showing typical body configuration associated with the presence within it of a feeding Ampulex hantuae larva. (Xl,9) Fig. 12. Hollowed out exo-skeleton of female Banlua dispar with apical segments of abdomen and exposed end of Ampulex hantuae cocoon removed to show orientation of wasp pupa. ( x 1 ,9) Fig, 13. Hollowed out exo-skeleton of female Banlua dispar with female Ampulex hantuae which emerged from it. The anterior end of the wasp’s cocoon is visible within the apical segments of the cockroach’s abdomen cut off by the emerging wasp. (xl,9) was established by breaking off the last few abdominal segments of two cockroaches and care- fully removing the exposed ends of the A. bantuae cocoons within. Each cocoon contained a pre-pupa facing the opened end — that is the posterior end of the cockroach. Pupation occurred within a few days. In both instances the pupae were so positioned that they were the same way up as the cockroach — that is the dorsal surface of the A. bantuae pupa was against the dorsal surface of the cockroach (Fig. 12). The adult A. bantuae emerges by cutting off first the end of its cocoon and then the end of the cockroach’s abdomen (Fig. 13). As the cockroach is always positioned within the cell facing the inner or blind end of the nesting cavity the A. bantuae within the cockroach is not only facing the correct way for emerging easily from the latter but is also facing the outer end of the cell and has only to pass through the detritus plug before emerging into the open. A. bantuae at Hilton is bivoltine and proterandrous. Developmental times for thirty-eight individuals reared during the period from November, 1973 to January, 1975 are plotted in Fig. 14. Developmental times established during other years were found to be very similar. During 1973-74 the first or summer generation resulting from eggs laid in trap-nests in the field during November and December, 1973 developed rapidly. Males emerged from mid January to the end of February, 1974 and females from late January to the end of the first third of March, 1974. Developmental periods from oviposition to adult emergence were about sixty-three days for males and about seventy days for females. The second or overwintering generation resulting from eggs laid in trap-nests in the field and in the laboratory during February and March, 1974 underwent a pre-pupal diapause as shown by the presence of pre-pupae in cocoons opened during early November, 1974. These pre-pupae changed to pupae within one to a few days following cocoon opening. Males of this generation emerged during the first half of December, 1974 and females from mid December, 34 GESS: ETHOLOGY OF AMPULEX BANTUAE GESS (HYMENOPTERA: SPHECIDAE: AMPULICINAE) Fig. 14, Developmental times for 38 individuals (256 6 and 13 9 9) of Ampitlex bantuae reared during the period from November, 1973 to January, 1975. That A. huntiiae is bivoltine is shown clearly. 1974 up to mid or even late January, 1975. Developmental periods from oviposition to adult emergence for second generation wasps ranged from 251-295 days for males and from 276-340 days for females. These developmental periods for second generation wasps of both sexes are clearly unnaturally long. Instead of emerging during December-January as here recorded this generation under natural conditions probably emerges during late October-early November in order that first generation eggs of the following summer season may be laid later in the latter month. The cause for the protracted developmental periods of the second generation wasps may be ascribed to the fact that the trap-nests containing the diapausing wasps were kept in- doors and were therefore not subjected to the temperature and humidity conditions prevailing in the field. The same phenomenon has been found with respect to developmental periods of diapausing Isodontia pelopoeiformis (Dahlbom) (Gess and Gess, 1982). It has been shown that wasps of the non-diapausing first generation take a fixed number of days to develop from egg to adult. The wasps of the diapausing second generation, however, show no such fixed time period for their development. Irrespective of how early or how late the second generation eggs are laid, the development of all the individuals hatching from them is held up at the pre-pupal stage during the autumn, winter and spring months. Further development of all takes place when diapause is broken in early summer and the long delayed moult from pre-pupa to pupa takes place. Post-diapause developmental times differ between males and females for the second generation of wasps, like the first, is proterandrous. Mating in captivity was observed in two instances. The wasps involved were all newly emerged, the males having emerged three to ten days and the females one to two days pre- vious to their being introduced together into a cage. In both cases the males ran about the floor of the cage in great agitation apparently searching for a female. After five to ten minutes, a fe- male having been located, the male immediately mounted her, there being no preliminary courtship. She extruded her genitalia and fusion of the genitalia took place. The male then dis- mounted and the two wasps assumed a tail-to-tail attitude, facing in opposite directions. One pair separated after fifty seconds in this tail-to-tail attitude. The total time taken for copulation was in one instance one minute and in the other two minutes. After separation the male groomed extensively. 35 ANN. CAPE PROV, MUS. (NAT. HIST.) VOL. 16, PT. 2, APRIL 1984 Little was established concerning the longevity and fecundity of A. bantuae. However, some indication will be gained from a consideration of five females kept in captivity. Two were captured as adults. The first, captured in the field at Hilton on 25. ii. 1974 just after ovipositing on a cockroach in a trap-nest, laid a total of thirteen eggs during the following thirty-six days. She died after forty-three days in captivity. The second, captured at Hilton on 15.iii.l974 after ovipositing on a cockroach in a natural nest, laid a total of four eggs during the following twenty-seven days. She died after forty-one days in captivity. The remaining three females emerged in captivity. The first emerged on 28. ii. 1974 from a dead cockroach found in a natural nest. During her adult life of forty-four days she laid a total of eleven eggs. The second emerged on 10.iii.l974 from a trap-nest. During her adult life of eighteen days she laid two eggs. The third emerged on ll.iii.l974 from a trap-nest. During her adult life of forty-nine days she laid three eggs. With respect to all five females more eggs might have been laid had the supply of cockroaches not sometimes failed. SEX OF WASP IN RELATION TO SIZE OF PREY UPON WHICH IT DEVELOPED In the section dealing with prey selection by A. bantuae an account has been given of the preferred size range of cockroach prey. It was also shown that, as the prey has to be apterous, nymphs of both sexes and also adult females are taken but that adult males which are winged are rejected. Whether the size of the prey has any bearing upon the sex of the wasp which develops upon it will now be considered. Forty-nine A. bantuae, twenty-four males and twenty-five females, were reared on cock- roaches ranging in length from 15-23 mm. The differences in the frequency of utilization of different length categories (grouped to the nearest mm) of the cockroach for rearing males and females are shown graphically in Fig. 15. It can be seen that the majority, twenty-one, of the males were reared on cockroaches up to 20 mm in length and the majority, nineteen, of females were reared on cockroaches of 20 mm or more in length. Prey size does therefore have Fig. 15. Graphs to show differences in frequency of utilization (vertical axis) of different length categories (grouped to nearest mm) of Bantua dispur (horizontal axis) by Ampiilex bantuae for rearing males and females. 36 GESS: ETHOLOGY OF AMPULEX BANTU AE GESS (HYMENOPTERA: SPHECIDAE: AMPULICINAE) a bearing upon whether an egg laid upon a cockroach will be male or female producing, that is unfertilized (haploid) or fertilized (diploid). This situation is reminiscent of that recorded by Brunson (1938) for Tiphia popilliavora Rohwer developing upon the larvae of the Japanese beetle. Three of the A. bantuae females reared in captivity were provided with trap-nests, nesting materials and cockroaches but were not allowed access to male wasps. These females gave rise to nine progeny all of which, having developed from unfertilized eggs, were males. The cock- roaches utilized ranged from 16-23 mm in length. The size of the males produced was not in proportion to cockroach lengths nor did the length range of 8,3-11,3 mm exceed the normal range for males. That is, outsize males were not produced on prey above the size range normally used for rearing males. THE EFFECT OF THE STING UPON THE PREY The paralysing effect of the sting of A. bantuae on its B. dispar prey appears to pass off rapidly and possibly completely. The recovery of the cockroaches from the initial knock-down, evident immediately after stinging, to the state in which they can be led to the nesting cavity has been described. Frequently a stung cockroach, positioned by a wasp at the inner end of a trap-nest, backs along the length of the cell in which it is sealed until its further progress is stopped by the detritus cell closure. Furthermore a stung cockroach bearing an A. bantuae egg or young larva readily escapes from the cell when the trap-nest lid is lifted. This causes difficul- ty when measuring the eggs and recording their position as cockroaches when removed from cells and laid upon their backs immediately right themselves and walk away. It appears there- fore that the detritus cell closure is as important in keeping the cockroach in as it is in keeping other insects out. Two adult female cockroaches laid oothecae some days after being stung. One, found sealed in a trap-nest at Hilton on 25. xi. 1973 and bearing an egg, laid a 12 mm long ootheca containing twenty eggs on 9.xii.l973. The wasp larva which hatched on 27. xi. 1973 died on 9.xii.l973 after developing very slowly. The cockroach died on 19.xii.l973. The eggs in the ootheca developed but the nymphs failed to hatch. The other, a captive, was stung on 11. iv. 1974 but not oviposited upon. It was observed feeding on 13. iv. 1974 and it subsequently laid an ootheca from which nymphs hatched on 18. iv. 1974. Moulting by a stung cockroach was observed. The cockroach, a female bearing an egg, was found sealed in a trap-nest at Hilton on 29.1.1975. The egg failed to develop and conse- quently the cockroach remained alive. It moulted on 10. hi. 1975 and although confined within a gelatine capsule without any food was still active on l.iv.l975. DISCUSSION Some details of the hunting and nesting behaviour of nine of the 118 species of Ampulex recognized by Bohart and Menke (1976) have been published to date. The species, arranged in alphabetical order, and the publications relevant to each are: A. assimilis Kohl (Hingston, 1925), A. canaliculata Say (Williams, 1929; Krombein, 1967), A. compressa (Fabricius) (Reaumur, 1742; Lucas, 1879; Bingham, 1897; Maxwell-Lefroy, 1909; Bordage, 1912; Williams, 1942), A. compressiventris Guerin-Meneville (Sharp, 1901, as A. sibirica), A. dissector (Thunberg) (Sonan, 1927; Kamo, 1957; Kohriba, 1957; Yoshikawa and lida, 1956; Yoshikawa, 1957; in all cases as A. amoena Stal), A. fasciata Jurijie (Picard, 1911 and 1919), A. nigrocaerulea Saussure (Callan, 1976), A. ruficornis (Cameron) (Sharp, 1901) and A. sonnerati Kohl (Sonnerat, 1776). The earlier published accounts of Ampulex biology were reviewed by Williams (1929 amd 1942). Other second- ary sources are Clausen (1940), Iwata (1976) and to a minor extent Guthrie and Tindall (1968). 37 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 2, APRIL 1984 The nesting ethology recorded during the course of the present study does not differ to any significant extent from that published for the above listed species. Indeed, on the basis of present knowledge, it appears that the ethology of An\pulex is characterized by marked inter- specific uniformity. It is, moreover, primitive in several respects. Notably, there is no prepara- tion of a nest prior to hunting, nor indeed is a nest in the true sense constructed at any time. Nesting is within a pre-existing cavity which is neither cleaned out, expanded nor in any way modified to receive the prey. Thus in the present study it was found that foreign objects within trap-nests were not removed prior to prey introduction nor was a preliminary plug ever con- structed even when the inner end of the trap-nest was already occupied by the cells of leaf- cutting Megachile species. Cavity modification is limited solely to its closure by the construc- tion of a detritus plug following the positioning of the prey and oviposition upon it. Consideration of the sequence of behavioural elements in, the nesting oi Ampulex, prey — niche — egg — closure, makes it clear that each sealed cell must be regarded as constituting a complete nest. Multicellular nests therefore do not exist. Where several cells occur in a single cavity as has sometimes been recorded (e.g. in Krombein, 1967 and in the present study) each cell represents a separate and possibly unrelated instance of nesting, not necessarily even by one and the same female. It would appear that Ampulex species are essentially arboreal wasps, hunting their prey on shrubs and trees and nesting within niches present on these plants. The nesting ethology of A. hantiiae described in the present paper is paralleled to a remarkable extent by that of A. assimilis, studied in Iraq by Kingston (1925). This wasp was found to hunt the wingless females of the cockroach Shelfordella tartara Saussure on the trunks of date palms and to nest within old beetle borings in the trunks. Similarly another species, A. compressa, was recorded by Maxwell-Lefroy (1909) as hunting Periplaneta species on the trunks of old Peepul {Ficus religiosa) trees in India. One hole and chink after another was searched for cockroaches and it may be assumed that these cavities were used by the wasp for nesting. Williams (1942) record- ed the same species frequenting the trunk of a coconut palm in New Caledonia. The fact that Krombein (1967) in North America found A. caualiculata utilizing trap-nests suspended 1 to 2 metres above ground from dead limbs of loblolly pines is indicative that this wasp and the cockroaches hunted by it, Parcoblatta species, are also arboreal. The same conclusion may be applied to A. fasciata found in France to prey upon Ectobia livida Fabricius and to nest inside bramble stems hollowed out, it was believed, by solitary bees (Picard, 1911) and in insect borings in fig trees (Picard, 1919). However, several species preying upon synanthropic cockroaches appear to have followed their prey into human habitations. Thus A. compressa, above recorded as arboreal, is more commonly recorded as hunting its prey, Periplaneta species, in houses. Records exist from Mauritius (Reaumur, 1742), Burma (Bingham, 1897) and Reunion (Bordage, 1912). Similarly A. compressiventris in West Africa (Sharp, 1901) and A. sonnerati (Sonnerat, 1776) in the Philippines were recorded hunting inside buildings. In all cases nesting took place in crevices situated within the buildings. It is probable that A. dissector of Formosa and Japan, which like A. compressa preys on Periplaneta species, similarly hunts and nests in buildings. Bohart and Menke (1976) have remarked that, with the exception of A. sikkimensis (Kriechbaumer), all species of Ampulex seen by them have modified tarsomeres. Thus the penultimate or fourth tarsomere has a ventral hair mat but lacks apical setae and the last tarsomere is inserted dorsally near the base of the fourth. This modification and configuration of the tarsomeres, exhibited also by A. bantuae, is very similar to that found in adult Coleoptera of the superfamilies Cucujoidea, Chrysomeloidea and Curculionoidea which nor- mally live on the aerial parts of plants and is clearly an adaptation to an arboreal life. As Ampulex provisions each nest with but a single prey, it follows that the latter is of necessity both larger and heavier than the wasp itself. A consequence of this disparity in size 38 GESS: ETHOLOGY OF AMPULEX BANTUAE GESS (HYMENOPTERA; SPHECIDAE: AMPULICINAE) and weight is that the wasp is unable to pick up the prey and carry it to the nesting cavity. The dragging of a deeply paralysed, inert prey to the nest as practised by the majority of Pompilidae, which, like Ampiilex, provision each cell with a single large prey, is practicable if the way to the nest is over the ground. However, in an arboreal situation such as that in which Ampulex hunts and nests this method of prey transport is not possible for the prey or the wasp and the prey together could easily fall to the ground. Ampulex has overcome the problem in a unique manner — by enlisting the co-operation of the prey itself. The effect of the Ampulex venom upon the prey is not to paralyse but rather to tranquilize it so that at the time of its transport to the nesting cavity.it is incapable of any initiative or independent action such as would be required for escape. However, when grasped by the wasp by an antenna and pulled, the prey is capable not only of walking but, in walking, of clinging to the substrate and there- fore of preventing itself from falling to the ground. In Ampulex, adaptations for an arboreal life are therefore not only morphological (the modified tarsomeres) but also biochemical (the nature of the venom) and behavioural (the nature of prey transport). ACKNOWLEDGEMENTS The author wishes to thank Mr T. C. White of the farm Hilton for his much appreciated kindness over the years in allowing him free access to his land. Sarah Gess is thanked for her interest in the Ampulex study, for valuable discussions, and for her assistance in the preparation of the present publication. Gratitude is expressed to the C.S.I.R. for running expenses grants for field work during the course of which the present study was made. REFERENCES Bingham, C. T. 1897. Hymenoptera. I. Wasps and Bees. In: Blandford, W. T. ed.. The fauna of British India, including Ceylon and Burma. London: Taylor and Francis. Bohart, R. M. and Menke, A. S. 1976. Sphecid wasps of the world: a generic revision. Berkeley: University of Califor- nia Press. Bordage, E. 19t2. Notes biologique recueillies a 1’ He de la Reunion, Bull, scient. Fr. Belg. 46: 29-92. Brunson, M. H. 1938. Influence of Japanese beetle instar on the sex and population of the parasite Tiphia popilliavora. Journ. Agric. Res. 57: 379-386. Callan, E. McC. 1976. Notes on Ampulicinae with special reference to African species and prey (Hymenoptera: Sphecidae). Rev. Zool.afr. 90 (1): 228-234. Clausen, C. P. 1940. Entomophagous insects. New York: McGraw-Hill. Gess. F. W. 1981, Some aspects of an ethological study of the aculeate wasps and the bees of a karroid area in the vicinity of Grahamstown, South Africa. Awn. Cape Prov. Mas. (nai. Hist.) 14(1): 1-80. Gess, F. W. and Gess, S. K. 1982. Ethological studies of Isodontia simoni (du Buysson), I. pelopoeiforniis (Dahlbom) and /. stanleyi (Kohl) (Hymenoptera: Sphecidae: Sphecinae) in the Eastern Cape Province of South Africa. Ann, Cape Prov. Mas. (nat. Hist.) 14(6); 151-171. Guthrie, D. M. and Tindall, A. R. 1968. The biology of the cockroach. London: Edward Arnold. Kingston, R. W. G. 1925. Nature at the desert's edge. London. IwATA, K. 1976. Evolution of instinct: comparative ethology of Hymenoptera. New Delhi: Amerind Publishing Co. for Smithsonian Institution and National Science Foundation, Washington, D.C. Kamo, T, 1957, On the habits of a cockroach hunting wasp Ampulex amoena Stal in Japan. (In Japanese with English summary.) Kontyu 25: 94-98. Kohriba, O. 1957. Some biological notes on Ampulex amoena Stal. (In Japanese with English summary.) Kontyu 25: 99-101. Krombein, K. V. 1967. Trap-nesting wasps and bees: life histories, nests, and associates. Washington, D.C.: Smithsonian Press. Krombein, K. V. 1970. Behavioral and life-history notes on three Floridian solitary wasps (Hymenoptera: Specidae). Smithson. Contr. Zool. 46: 1-26. Lucas, H. 1879. (Chlorion (Ampulex) compressum, Hym.) Ann. Soc. ent. France, (5) 9, Bull. CLIX ( = Bull. Soc 1879.219-220). Maxwell-Lefroy, H. 1909. Indian insect life. Calcutta: Thacker, Spink and Co. Picard, F. 1911. Sur les moeurs et le genre de proie de V Ampulex fasciatus Jurine (Hym. Sphegidae). Bull. Soc. ent. France 1911: 113-116. 39 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 2, APRIL 1984 Picard, F. 1919. Ann. des Epiphyties 1919: 1-144. Princis, K. 1963. Biattariae. Revision der siidafrikanischen Blattarien-fauna. In Hanstrom, B. et alia, eds, South African animal life 9. Stockholm: Swedish Natural Science Research Council, pp. 9-318. Reaumur, R.A.F.de 1742, Memoires pour servir a I'histoire des insectes 6. Paris: Imprim. Royale. Sharp, D. 1901. Insects. Part II. In Harmer, S. F. and Shipley, A. E., eds. The Cambridge natural history 6. London: Macmillan, pp. 1-626. SoNAN, H. 1927. A taxonomic study together with observations of various families of egg laying wasps in Formosa. (In Japanese.) Trans, nat. Hist. Soc. Formosa 17: 121-138. SoNNERAT, P. 1776. Voyage a la Nouvelle Gurnee. Paris: Ruauit. Williams, F, X. 1929. Notes on the habits of cockroach-hunting wasps of the genus Ampulex, sens, lat., with particular reference to Ampulex (Rhinopsis) caniculatus Say. Proc. Hawaii, ent. Soc. 7: 315-329. Williams, F. X. 1942, Ampulex compressa (Fabr.), a cockroach-hunting wasp introduced from New Caledonia into Hawaii, Proc. Hawaii, ent. Soc. 11: 221-233. Yoshikawa, K. and Iida, K. 1956. Behaviour of Ampulex amoena. (In Japanese.) Seiro-Seitai 7: 54-60. Yoshikawa, K. 1957. Parthenogenesis in Ampulex amoena Stal (Hymenoptera: Ampulicidae). (In Japanese with Eng- lish summary.) Seiro-Seitai 7 (2): 131-133. 40 SI, ,|^“;>'^-i:i:4i^ ..^V' ‘ ',. ’ ’ C; “3 ■■' fp/' -'■'■* ^ -■.- ■ * ■ ft .-: " ;.-. ... ; , ., ■ W^J., V .»• ’ ■ •.,. ■■•■ .;'•■• ‘.fl' .V.'^l'"'' * r? " .. INSTRUCTIONS TO AUTHORS LAYOUT of recent issues should be examined for details of acceptable style and format. MANUSCRIPTS should be submitted in triplicate, typed on A4 size paper, double- spaced, on one side of the paper only and with wide margins all round. Only generic and specific names should be underlined to indicate italics. All other mark- ing up should be left to the editor. Footnotes should be avoided as far as possible. All measurements should be metric. Systematic papers should follow the rules of the International code of zoological nomen- clature or the International code of botanical nomenclature, as applicable. ILLUSTRATIONS should bear the author’s name, title of paper and number in soft pencil on the reverse side. Diagrams, line drawings and photographs are all regarded as figures, should be termed as such and should be numbered in a single series. In addition to the original illustrations a photocopy of each should be submitted. Diagrams and line drawings should be reducible to a size not exceeding 14 x 16 cm (17,5 cm including caption). Photographs for half-tone blocks should be final size, of high contrast and printed on glossy paper. Colour illustrations are only printed under special circumstances and their financing must be arranged by the author. A scale should appear with all illustrations or the magnification should be given in the caption. A list of captions should be typed on a separate sheet. TABLES should be typed on separate sheets. Foldouts and tables printed sideways should be avoided. REFERENCES should be listed alphabetically at the end of the paper. Abbreviations of periodical titles should conform with those in the World list of scientific periodicals (4th ed.). Layout must be as follows: Alexander, J. 1969. The indirect evidence for domestication. In: Ucko, P. J. and G. W. Dimbleby eds, The domestication and exploitation of plants and animals. London: Duckworth, pp. 123-129. De Winter, B. 1969. Plant taxonomy today. S.Afr.J. Sci. 65 (8): 229-242. JuBB, R. A. 1967. Fresh water fishes of southern Africa. Cape Town: Balkema. GALLEY PROOFS are sent (when practicable) to authors for checking printer’s errors. Adjustments of language, style and punctuation and any other amendments of the original manuscript, if made at this stage, may be charged to authors. AUTHOR’S COPIES (50) are issued free of charge. When more than one author is involved the fifty free copies will be sent to the senior author. Additional copies ordered before printing are supplied at a reduced cost. BD2t98— CTP Book Printers. Cape ■ • l Y I’ ' i"': 'it® VKbZ Til . . . i*finu; - ‘■•■■■■'iv- .'■ ' ' „ - ''.w-' '. ? . ' • ;■ >> y.- Jf-T ji ■V : ■i y ; ' y ' ■■ - &}/■ i^;- '^/^>.:*Vi '■ :•’ 4lknW'fi . >'u ‘Xi^>- ■;,■*■ r. , ■'■ :•:■ -!T ' y?'r‘ V. •5 <“. i 1 i;,? f ,ij ".vT(;=l i|ji;f,"y:' Sy i' - ^ i.u’^. - ■■'». — . . •■'■ ' li ., > “' I ,Sia ■ Jk- Annals of the Cape Provincial Museums Natural History Ann. Cape Prov. Mus. (nat. Hist.) Volume 1 6 Part 3 1 6th April 1 984 Published jointly by the Cape Provincial Museums at the Albany Museum, Grahamstown, South Africa ANNALS OF THE CAPE PROVINCIAL MUSEUMS These Annals are published in two series. Natural History and Human Sciences, the latter series covering cultural history, ethnology, anthropology and archaeology. They are issued in parts at irregular intervals as material becomes available. The primary objective of these Annals is to disseminate the results of research work undertaken by staff members of the Cape Provincial Museums. However, a manuscript by any author is eligible for publication provided it is based wholly or partly on the collection/s of one or more of these museums or if the material on which the work is based is wholly or partly deposited in one of these museums. Parts are obtainable from the Librarian of the Albany Museum. Correspondence in con- nection with these Annals should be addressed to the Editor, Albany Museum, Grahams- town 6140. Editor DrF.W.GESS: 1978 ERRATA ANN. CAPE. PROV . MUS . (nat. Hist.) 16 (3)- p. 45 - line 20 - delete after (Pfeffer, 1889) "and later figure (Pfeffer, 1889)” p. 50 - 2nd paragraph, line 4 : P ie tschmann 1 9J7 3 = 1913 5th paragraph, line 4 : 1,1 - 3 = 1 - 1,3 p. 57 - 1st paragraph, line 5 : 5+6 = 5-6 line 6 : ph -1+2 = ph + 1 + P- 58 - table 6a in title : spelling Hypura 1 table 6b - allign N vertically. P- 68 - Ac Ic no w 1 e dgeme n t s , line 5 : spelling Audenaerde 69 reference Matthes 1967 E.A.F.F.R.O. V -ri ■ '■ I't" 1 '.I • . 4 p«> 1 j.ct tffij; : S •., 1 . ti.*e 1 ■' It i. i I, !f ri i,’i ■ ■»13' • i r. ,i it- 1 (i in rr. : t ■ 1. ■.- «■ , '«ji < i ■ . -S.1 »■?<»«»» U ■* f w£;t.-3< ».a • l ? 1 A , 3" ■ IT) ON fn o ^ ON ^ NO ON ON ro NO r- ^ On (N oow^rnr^'^Tf fO o" "Tt- — i ^ NO^ r--" w^ + + On no - + + <3 s». Oo in o cN in in^ cn cn ’-4' NO r^" ON o' in oo r--" -f O O CM ir^r^NOONOOoow^ in ^ 00 ON in r4 no On r- rn — h" ^ in ^ ^ t— ( LO 1-H CN rt m nO^ 00 (N NO NO . "t On .. + On 04 00^ No^ ON^ in ON fn m oo" ^ no" on oo" VTi r<^j o o" 00 rn ■ ro NO ^ . On - + -h OO in NO 00 00 o oc i-^inoooNNor^r^ON I - a a ao^ c *= K -3 S I— I f=" -fe-* DX) ^ 3 3 g sp-2i e « xi^'O'a— CCU-T30 Scc-Oicc^S C 5 — --~,S =ay= U isJ3 -■ u ■O'^'O’O ^ >> >-.'0 ^ S — o.^ t§"Oo-- cScgcS(^^-0“003^^t1>'^^nXi ^K^ffi'afflP3UUQ<^CM ^CEcoO 5 -s "3 X) :9 O ^ ^ cd ^2 (U c ^ S c — ^ C3 6o 2 S c 2 O o sx c? 2 ^ ■■g c -a < E 2 o 2 cd e c 5 S U c C« ij o 5 O Dh < cd c _ c 2 13 « ™ ^ ' t O ^ 2 '5 -g '- o 2: 3 33 (U 3 -“ 3 2 O Z ~G cd >. J- 2 -o «u X o c TO Cd a> A:..,-;; • - •':, ■„■=-■•.■ ■■::■'•/% •' ^ ;v: ■■i. ■■:■■' ■: .': ^ :■■ • ; f f - ^ ': = ' . ' ^' ' “ ' ■ ■ . •" ’ ■ = '-:■ '.’. ■' ' ^ ... ' , ' •■ .r' ; ■•■•'■:' J ^■:.' ■■■ m' ■ ^^:' M:: ‘ .4.:^ 'j- ■ -‘ .. 1- ' .. ■" 'kfy'^-':;hy ':.. ,J5i '-r''L ; A. fc; '''■!^:’^ id«a>... INSTRUCTIONS TO AUTHORS LAYOUT of recent issues should be examined for details of acceptable style and format. MANUSCRIPTS should be submitted in triplicate, typed on A4 size paper, double- spaced, on one side of the paper only and with wide margins all round. Only generic and specific names should be underlined to indicate italics. All other mark- ing up should be left to the editor. Footnotes should be avoided as far as possible. All measurements should be metric. Systematic papers should follow the rules of the International code of zoological nomen- clature or the International code of botanical nomenclature, as applicable. ILLUSTRATIONS should bear the author’s name, title of paper and number in soft pencil on the reverse side. Diagrams, line drawings and photographs are all regarded as figures, should be termed as such and should be numbered in a single series. In addition to the original illustrations a photocopy of each should be submitted. Diagrams and line drawings should be reducible to a size not exceeding 14 x 16 cm (17,5 cm including caption). Photographs for half-tone blocks should be final size, of high contrast and printed on glossy paper. Colour illustrations are only printed under special circumstances and their financing must be arranged by the author. A scale should appear with all illustrations or the magnification should be given in the caption. A list of captions should be typed on a separate sheet. TABLES should be typed on separate sheets. Foldouts and tables printed sideways should be avoided. REFERENCES should be listed alphabetically at the end of the paper. Abbreviations of periodical titles should conform with those in the World list of scientific periodicals (4th ed.). Layout must be as follows: Alexander, J. 1969. The indirect evidence for domestication. In: Ucko, P. J. and G. W. Dimbleby eds. The domestication and exploitation of plants and animals. London: Duckworth, pp. 123-129. De Winter, B. 1969. Plant taxonomy today. S.Afr.J.Sci. 65 (8): 229-242. JuBB, R. A. 1967. Freshwater fishes of southern Africa. Cape Town: Balkema. GALLEY PROOFS are sent (when practicable) to authors for checking printer’s errors. Adjustments of language, style and punctuation and any other amendments of the original manuscript, if made at this stage, may be charged to authors. AUTHOR’S COPIES (50) are issued free of charge. When more than one author is involved the fifty free copies will be sent to the senior author. Additional copies ordered before printing are supplied at a reduced cost. BD2232 — CTP Book Printers, Cape SMFTHSONIAN INSTITUTION LIBRARIES lllil 3 9C 188 0' 429 53< 49 Iff •T S, *'^ i 'is® iteiv V. ;'fb= .m 1. ■.. .ji-.'‘^ ''a' 'l. '"ri ••^^.c4ci•n^l.•li eii nA. '^-r f f,' ■•''■. ,i ■' • ,wi; .. ai i»-.-.‘s*l^^|r»;*'iitt«M*<:‘ i !**(<; ^{it Ut" ■■■'• K- thiiaati^.-,-.-^- >I‘''4:'j|l|l|''';jV^s^f/-; ,. fA'T?i/rSt, fjf.« i/stj5,t. ■' 'V ■••■■- -A'^- ^ At ■ '' - ' ^ *' ■* I ' ' .’ " -'. ' ' ■ ... ■.'• •. • t.» ♦ • 'Si[Vf;" <■■>••:»• .V. i k '■; W' •: V, 4-1^ vpmi yliSfr ■'/•<> 'i- i \-> -A Ji 4^‘ Kt ■*;^ .- v:4v.'r ^ vAv;) J^-.>pr>4!> al I \T. ;‘y, ■'■ " :, •• ,.V: ■■"■ irva?l^‘/.w". , av: ,.i' '*"■■ ' ■' A. "^' ■ . .* r. ;i; k' ( W^- . "■ .'■ ■■ T . tf.iw M .l&t' j Annals of the ■^©kpe Provincial Museums Natural History Ann. Cape Prov. Mus. (nat. Hist.) Volume 16 Part 4 21st January 1985 Published jointly by the Cape Provincial Museums at the Albany Museum, Grahamstown, South Africa ANNALS OF THE CAPE PROVINCIAL MUSEUMS These Annals are published in two series, Natural History and Human Sciences, the latter series covering cultural history, ethnology, anthropology and archaeology. They are issued in parts at irregular intervals as material becomes available. The primary objective of these Annals is to disseminate the results of research work undertaken by staff members of the Cape Provincial Museums. However, a manuscript by any author is eligible for publication provided it is based wholly or partly on the collection/s of one or more of these museums or if the material on which the work is based is wholly or partly deposited in one of these museums. Parts are obtainable from the Librarian of the Albany Museum. Correspondence in con- nection with these Annals should be addressed to the Editor, Albany Museum, Grahams- town 6140. Editor DrF.W.GESS: 1978- The distribution and status of bats (Mammalia: Chiroptera) in the Cape Province by J. C. HERSELMAN and P. M. NORTON (Cape Department of Nature and Environmental Conservation, Private Bag 5014, Stellenbosch, 7600) CONTENTS Page Abstract 74 Introduction 74 Survey methods 75 Distribution and status of individual taxa; biological notes 79 Suborder Megachiroptera Family Pteropodidae Epomophorus wahlbergi Wahlberg’s Epauletted Fruit Bat 79 Epomophorus crypturus Peters’ Epauletted Fruit Bat 79 Eidolon helvum Straw-coloured Fruit Bat 81 Rousettus aegyptiacus Egyptian Fruit Bat 81 Suborder Microchiroptera Family Emballonuridae Taphozous mauritianus Mauritian Tomb Bat 83 Family Nycteridae Nycteris hispida Hairy Slit-faced Bat 83 Nycteris thebaica Egyptian Slit-faced Bat 84 Family Rhinolophidae Rhinolophus fumigatus Rtippell’s Horseshoe Bat 87 Rhinolophus darlingi Darling’s Horseshoe Bat 88 Rhinolophus clivosus Geoffrey’s Horseshoe Bat 88 Rhinolophus capensis Cape Horseshoe Bat 90 Rhinolophus denti Dent’s Horseshoe Bat 92 Rhinolophus swinnyi Swinny’s Horseshoe Bat 92 Hipposideros caffer African Leaf-nosed Bat 92 Family Vespertilionidae Myotis seabrai Angola Wing-gland Bat 94 Myotis lesueup Lesueur’s Wing-gland Bat 96 73 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Myotis tricolor Cape Hairy Bat 96 Pipistrellus nanus Banana Bat 97 Pipistrellus kuhlii Kuhl’s Pipistrelle 99 Eptesicus hottentotus Long-tailed House Bat 100 Eptesicus melckorum Melck’s House Bat 101 Eptesicus capensis Cape Serotine 102 Eptesicus nodus Cape Horn-skinned Bat 103 Laephotis wintoni De Winton’s Long-eared Bat 104 Scotophilus dinganii Yellow House Bat 105 Kerivoula lanosa Lesser Woolly Bat 105 Miniopterus fraterculus Lesser Long-fingered Bat 107 Miniopterus schreibersii Schreibers’ Long-fingered Bat 107 Family Molossidae Sauromys petrophilus Roberts’ Flat-headed Bat 109 Tadarida condylura Angola Free-tailed Bat Ill Tadarida pumila Little Free-tailed Bat Ill Tadarida aegyptiaca Egyptian Free-tailed Bat Ill Discussion 113 Conservation of bats in the Cape Province 118 Acknowledgements 120 References 121 Appendix; Gazetteer of localities mentioned in the text 122 ABSTRACT Distribution maps with notes on taxonomy, distribution and status, habits, food and breeding are presented for 29 species of bats occurring in the Cape Province of the Republic of South Africa. Comparisons are made between data from the survey, which lasted from 1976 to 1979, and specimens examined in South African museums. The main motivation for the survey was the conservation of bats. A high priority was placed on collecting from roosts. Some very scarce and endemic species were collected and many interesting new localities were recorded. UITTREKSEL Verspreidingskaarte met aantekeninge oor die taksonomie, verspreiding en status, ge- woontes, voedsel en aanteel van 29 species vlermuise wat in die Kaapprovinsie van die Repub- liek van Suid-Afrika voorkom, word aangebied. Vergelykings word gemaak tussen gegewens wat versamel is tydens die intensiewe opname, wat vanaf 1976 tot 1979 geduur het, en monsters wat bestudeer is in die groter Suid-Afrikaanse museums. Die motivering vir die op- name was die bewaring van vlermuise, en ’n hoe voorkeur is geplaas op die versameling van data in die skuilplekke van elke spesie. Sommige van die baie skaars en endemiese spesies is versamel, en interessante nuwe voorkoms lokaliteite is gevind. INTRODUCTION In their identification manual for African Chiroptera Hayman & Hill (1971) express the opinion that there is unlimited scope for further work on practically every aspect of African bat systematics and biology. Up to now very little more than incidental collecting has been done in the Cape Province. Although naturalists such as Roberts, Shortridge and Allen collected much material and made substantial contributions to bat taxonomy, little information on their ecol- ogy is available. 74 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. The only attempt at an ecological study in the Cape Province was made between 1960 and 1963 when approximately 500 cave bats of the species Rhinolophus capensis, Rhinolophus clivosus, Miniopterus schreibersii and Nycteris thebaica were banded at De Hoop cave, Koegel- been cave and at Die Oog (Kuruman) by B. Copley, O. Siegrist and C. Gow, according to in- formation from the Transvaal Museum bat banding project. Unfortunately there was no fol- low-up and therefore no information on the migration and longevity of the species was collected. The present project was initiated to provide basic information on the conservation status of this large group of mammals and to draw attention to the important role that bats are likely to play in the natural ecosystems of the Cape Province. Whereas the frugivorous species are es- sential in nature as seed propagators, the insectivorous species are of considerable economic importance in the control of nocturnal insect populations. According to Davis et al. (1962) the free-tailed bats of Texas are capable of destroying 6 600 tons of insects yearly. Similarly, Dwyer (1964) calculated that Miniopterus schreibersii destroy about 200 pounds of insects per night in the Macleay Valley in Australia. At the second International Bat Research Conference held in Amsterdam in 1970 delegates from 20 countries agreed that the numbers of almost all bat species are declining, and that further research on the ecology of most species is essential for their survival (Stebbings 1970). The main reasons for the decline in numbers were listed as the effect of insecticides, lost habitat and unnecessary slaughter. They concluded that several species around the world are already endangered. To draw up an effective conservation plan for bats it is necessary to find out their distribu- tion and numbers, and to have a basic understanding of their roosting behaviour and seasonal movements. It was decided to start by concentrating on those species which congregate in large communal roosts, since these would be the easiest places to apply active conservation manage- ment. This paper reports on data on the status and distribution of all bats that were accumu- lated during field work, and more detailed information on population dynamics and migrations in some of the larger roosts will be reported elsewhere. Most recent studies on the regional distribution of mammals in southern Africa include generalized and somewhat repetitive summaries of what is known of the biology of each bat species. Therefore the biological notes included here are mostly restricted to data collected during the study or else information in the literature that refers particularly to bats in the Cape Province. The survey was carried out over the whole of the Cape Province (Fig. 1), an area of 641 500 km^ All previous checklists on the mammals of South Africa were written with the inclusion of Transkei, Bophuthatswana and East Griqualand (the remainder of the Cape Province to the east of Transkei) in the Cape Province. However, these areas were excluded from the present study because the first two are now independent states and East Griqualand has been transferred to the province of Natal. SURVEY METHODS The first priority was to gather information on bat roosts. For this an intensive press cam- paign was carried out. In addition, letters of inquiry were sent to selected high schools and in- dividuals, yielding information on localities of about 120 roosts. The field survey started during the winter of 1976 and lasted until the end of 1979. During this period several additional localities were found. 75 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 As some species undertake seasonal migrations each roost was visited twice, the visits be- ing in different seasons. Except for observations on nocturnal behaviour all visits were made during daylight. For safety at least two people entered the roosts at a time, and each wore a mine-torch, helmet and boots. To keep disturbance to a minimum the stay inside the roost was limited to not more than one hour. During each visit notes on the numbers of each species in the population, the amount of guano, recaptures of banded bats, depth of the cave or mine, habits, temperature and humidity were recorded on field data sheets. A small sample of up to 20 bats of each species was taken with a hand-net, the size of the sample depending on the number in the population. Where roosts were too dangerous to enter, or where the openings were too small, mist-net samples were taken during the evenings at the time of emergence. On windy evenings, when the effec- tiveness of mist-nets was low, bats were shot on emergence with a .22 rifle loaded with dust- shot. 76 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. It soon became apparent that the scarcer species would be overlooked, if collecting at roosts was the only sampling method used, and mist-nets were therefore erected over reser- voirs or streams in suitable areas. The nets were manned for at least five hours per night, start- ing from sunset, and bats were taken from the net at hourly intervals. Although this method yielded little information on the status of the species caught, it was useful for recording occur- rence. In many areas, especially in the drier parts of the north-western Cape, up to 20 bats of five different species were caught per night. At the De Hoop Cave, where the more intensive population study was carried out, a variation of the Tuttle-trap (Constantine 1956, Tuttle 1973) was used for catching the bats. At this site nearly 8 000 individuals of Miniopterus schreibersii, Myotis tricolor, Rhinolophus clivosus and Rhinolophus capensis were banded for the purpose of studying migration (Hersel- man & Norton in prep.) and population dynamics. Bats to be prepared as museum specimens were killed with chloroform, fixed for 24 hours in 5% formalin, and preserved in 70% ethyl alcohol. Pregnant females were dissected to free the embryo for growth studies. The old preservative was replaced with a new solution every six months. All the specimens collected were catalogued in the laboratory using the first three letters of the genus name and a three-digit number from 001 to 999 followed by the letter “J” (for Jonkershoek). For example, the fiftieth collected specimen of Nycteris thebaica would be cata- logued as NYC 050 J. The identification of most of the material was carried out by the senior author, using descriptions and keys in Roberts (1951), Ellerman et al. (1953) and Hayman & Hill (1971). On a few occasions material was sent to Dr I. L. Rautenbach of the Transvaal Mu- seum and to Prof J. Meester of the University of Natal. The entire collection of more than a thousand specimens of 19 species has been deposited in the Kaffrarian Museum in King William’s Town. During the study the Transvaal Museum in Pretoria, the Kaffrarian Museum in King William’s Town, the Albany Museum in Grahamstown, and the South African Museum in Cape Town were visited to examine their bat material from the Cape Province. The authors are not aware of any other significant collections from this region and hope that these four col- lections contain the majority of bat specimens from the Cape Province available in South Africa. A gazatteer was drawn up for all survey and museum specimens (Appendix). For survey material most spot localities were plotted to the nearest minute, whereas localities of museum specimens could seldom be recorded more accurately than the quarter degree grid square. The distribution of survey sampling sites is shown in Fig 2. Localities of both survey and museum records were plotted on the maps using quarter degree squares and localities of special importance mentioned in the literature are indicated by an “S”. To bring the distri- bution of each species into perspective a small map of Africa showing the approximate distri- bution range was drawn from information in Hayman & Hill (1971), Kingdon (1974) and Smithers (1983). The text for each species includes a summary of the material examined, with the number of specimens and the localities where they were collected. Museum records are identified as follows: (TM) for the Transvaal Museum in Pretoria, (KWT) for the Kaffrarian Museum in King William’s Town, (SAM) for the South African Museum in Cape Town, and (AM) for the Albany Museum in Grahamstown. 77 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Fig. 2. Localities where bats were collected in the Cape Province. Numbers indicate major roosts listed in Table 2. The order and taxonomy of all species follows that of Hayman & Hill (1971), with cor- rections and up-dating according to Swanepoel et al. (1980). At the start of the survey the authors traced approximately 900 bat specimens from the Cape Province that had been deposited in South African museums, representing 23 of the 30 species then recorded in the Province. The collection with the largest number of specimens (440) was in the Kaffrarian Museum at King William’s Town. A large number of specimens is also housed in museums overseas, including many type specimens, but these are not easily ac- cessible for taxonomic and distributional study in this country. The present survey yielded more than a thousand specimens of 21 species, which rep- resents a substantial addition to the museum material available in the Cape Province. Nearly all of these specimens have been sent to the Kaffrarian Museum. The specimens were collected from localities distributed throughout the Cape Province, however, sampling was concentrated 78 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. mostly in the mountainous areas of the province with caves and mines suitable for large roosts (Fig. 2). DISTRIBUTION AND STATUS OF INDIVIDUAL TAXA; BIOLOGICAL NOTES Suborder MEGACHIROPTERA: Family PTEROPODIDAE Epomophorus wahlbergi (Sundevall, 1846) Wahlberg’s Epauletted Fruit Bat The genus Epomophorus is easily recognized by the tuft of white hair at the base of the ear. E. wahlbergi differs from all other species of the genus occurring in Africa in that it has only one well-developed post-dental palatal ridge, whereas the others have two (Flayman & Hill 1971). However, Meester et al. (1964) mention the possibility that E. wahlbergi may be a subspecies of E. crypturus. Roberts (1951) gave a range of 80-86 mm for the forearm length of E. wahlbergi. The range for the seven specimens collected during this survey from three localities is 77,5- 85,5 mm. Distribution and status Epomophorus is the most widespread of all African fruit bat genera (Hayman & Hill 1971). E. wahlbergi can be found in a large part of central, eastern and southern Africa. In the Cape Province its distribution seems to be limited to the south-eastern and eastern Cape (Fig. 3). A series of specimens was collected from Grahamstown by J. C. Greig and C. T. Stuart of this Department. The authors collected one specimen from Walmer and one from as far south as Keurbooms River. From discussion with various people in the eastern Cape it ap- pears that the species is fairly common in the area. Habits This species seems to roost in trees and was never found in caves or mines. In Walmer the authors often saw them resting in pine and cypress trees where they uttered their characteristic pinging sound. They appeared to prefer soft fruits which they carried from the low fruit trees to feeding roosts in higher trees. Material examined Survey — 5 Grahamstown, 1 Keurbooms Reserve, 1 Walmer. Museums — 2 Alice, 1 Balossi, 1 Bedford, 1 Blaney, 1 Gladstone, 51 King William’s Town, 2 Peddie, 3 Pirie (KWT); 2 King William’s Town, 14 Port St Johns, 1 Umtata (SAM); 1 Bedford, 30 Grahamstown, 2 King William’s Town, 1 Pirie (AM). Epomophorus crypturus Peters, 1852 Peters’ Epauletted Fruit Bat The single specimen of this species caught during the survey was separated from E. wahlbergi on the basis of the second post-dental palatal ridge, as well as on more general exter- nal features. However, this could not be confirmed because there are no Cape specimens in the South African museums. Roberts (1951) noted a marked difference in forearm length between the two species, with E. crypturus being considerably shorter, but Hayman & Hill (1971) 79 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 showed that the ranges overlap. The single specimen of E. crypturus collected had a forearm of 77,8 mm, which is just within the range of 77,5-85,5 mm the authors established for E. wahlbergi. Distribution and status According to Hayman & Hill (1971) this species occurs from the equator southwards across most of the continent, except for the dry western parts of South West Africa, the west- ern Cape Province and the Karoo. Meester et al. (1964) state that it occurs in the eastern Cape, and Ellerman et al. (1953) record that it has been collected from Pirie. During the survey one specimen was collected in Grahamstown, but no Cape specimens could be traced in the museums. The distribution in the Cape Province must therefore be very limited (Fig. 3), and the status of the species rare. 80 HERSELMAN, NORTON; DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Material examined Survey — 1 Grahamstown. Eidolon helvum (Kerr, 1792) Straw-coloured Fruit Bat This species is distinguished by its long wings, which are the longest for any bat in the Cape Province. They are also longer in proportion to the body than for any other bat in the area. Distribution and status This species occurs as a migrant over the entire continent, but is rarely recorded in south- ern Africa (Meester et al. 1964). A specimen from the Barkly West High School was shot in an orchard during 1976. Photographs of one shot in Middelburg (Cape) during 1970 were also examined. Museum specimens are from localities as widely separated as Hondeklipbaai in the north- west Cape and Tylden in the eastern Cape. Other published localities include Bedford and Steynsburg (eastern Cape) and Griqualand West (Roberts 1951). Thus E. helvum may be found in most of the Cape Province, except for the southern and eastern coastal regions (Fig. 4). However, it is rare in this area and only visits during the summer fruiting season. Habits Little is recorded of the habits of E. helvum in the Cape Province. Where it occurs it is al- most always found feeding in deciduous fruit trees. Material examined Survey — 1 Barkly West. Museums — 1 Mazelsfontein (TM); 1 Hondeklipbaai, 1 Namaqualand, 1 Tylden, 1 Vryburg (SAM). Rousettus aegyptiacus (E. Geoffroy, 1810) Egyptian Fruit Bat This is the only cave-dwelling fruit bat species in the Cape Province. Hayman & Hill (1971) record that the Egyptian fruit bats occurring in the Cape Province belong to the sub- species R. a. leachii, although Roberts (1951) regarded R. leachii as a valid species. Distribution and status The Egyptian fruit bat occurs over a large part of Africa, but in the Cape Province it is re- stricted to the south-western, southern and eastern coastal regions (Fig. 4). Roberts (1951) states that it is rarely found inland from the coastal belt. It is therefore of interest that it was found in the Swartberg near Oudtshoorn and in the Winterhoek mountains near Porterville. Some evidence of migration was found, although the distances involved were not very great. Bats that had been marked by G. R. Robinson at Storms River Mouth were collected during the fruit season of November to February in the Langkloof, about 50 km away. The species is abundant in its distribution area and gathers in large groups of up to a thousand in suitable caves. Habits These are the only fruit bats that are able to use echo-location, and this allows them to roost in dark caves. They were found in larger caves along the coastal areas, usually in associ- 81 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 ation with up to four insectivorous species. They typically occupy the “twilight” areas just in- side the entrances of caves, whereas the insectivorous species roost further in where it is com- pletely dark. They are very noisy, and are easily alerted when the caves are entered. Food They were seen eating peaches, apricots, apples, loquats and the fruit of yellow-wood (Podocarpus) and hard pear (Olinia) trees. It appears that only ripe fruits are eaten, which means that damage in gardens and orchards only occurs during a short period. Thus the extent of damage caused is often exaggerated. For the greater part of the year they feed on wild fruit, and in this way act as important seed propagators of a wide range of indigenous plants and trees. 82 HERSELMAN, NORTON: DIST, AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Breeding Infants were seen at Helderberg Cave in early January, and some of the females caught then were still pregnant. During January 1978 females netted in a peach orchard near to this cave were carrying infants of about two-thirds their weight. Material examined Survey — 1 Bat’s Cave, 3 Bean-se-bos, 1 Die Hel, 1 Helderberg, 2 Keurbooms Reserve, 1 Krakeel River, 1 Robertson, 3 Skeleton Cave, 1 Storms River Mouth, 1 Twee Riviere, 2 Wynberg. Museums — 7 Knysna, 5 Skeleton Cave (TM); 1 Amabele, 1 East London, 2 King William’s Town, 2 Patensie, 2 Pirie (KWT); 1 Knysna, 1 Table Mountain (SAM); 1 Amabele, 3 Bedford, 2 Grahamstown, 1 King William’s Town, 2 Mqanduli, 1 Salem, 1 Swellendam (AM). Suborder MICROCHIROPTERA: Family EMBALLONURIDAE Taphozous mauritianus E. Geoffroy, 1818 Mauritian Tomb Bat This is the only species of the family occurring in the Cape Province. It can be recognized easily because its tail separates from the interfemoral membrane and merges on the upper side about half-way down its length. Distribution and status According to Hayman & Hill (1971) this species is widely distributed over a large part of the continent, in most savanna regions south of the Sahara. Meester et al. (1964) stated that its Cape distribution is limited to the southern and eastern Cape. No specimens were collected during the survey, but the two Cape specimens in the museums confirmed this southern and eastern distribution (Fig. 5). A specimen recently collected at Hartswater (2724 DD) by Erasmus & Rautenbach (in press) suggests that the species may prove to be sparsely distrib- uted over most of the Cape Province. Habits Erasmus & Rautenbach (in press) state that this species has the habit of roosting head downwards against the trunks of large trees or under the eaves of buildings. Material examined Museums — 1 Heidelberg (SAM); 1 Alicedale (AM). Family NYCTERIDAE Nycteris hispida (Schreber, 1775) Hairy Slit-faced Bat References to the occurrence of this species in the Cape Province (Ellerman et al. 1953, Meester et al. 1964) seem to be based entirely on a single dry skin housed in the South African Museum (ZM 5444) and prepared from a specimen that was collected at Port St Johns by Shortridge in 1902. It was caught at the same time as a series of more than 20 N. thebaica adults, but was separated from these on the grounds of its smaller size and lighter coat colour. 83 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 However, it was caught in January, a time when one can expect to find juveniles of most bat species, and comparison with N. thebaica juveniles from other areas yields no reason why this specimen should be separated from the more common species. In the absence of further evi- dence the authors do not consider N. hispida as part of the Cape bat fauna. Nycteris thebaica E. Geoffroy, 1818 Egyptian Slit-faced Bat This is the only species of the family Nycteridae that definitely occurs in the Cape Prov- ince. It is easily recognized by its very large oval ears. Another outstanding character is the deep groove, between the nose-leaves, which extends from the nostril to a line between the base of the ears. Roberts (1951) regarded N. capensis as the senior synonym of N. thebaica, which he incor- rectly attributed to A. Smith, 1834, and gave the subspecies capensis and damarensis. How- ever, Ellerman et al. (1953) and Meester et al. (1964) recognized Geoffroy’s earlier description 84 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. of N. thebaica, but listed the same subspecies. This was followed by Hayman & Hill (1971), although they did not commit themselves on the validity of the subspecies. Distribution and status N. thebaica is widely distributed in open country over most of Africa (Hayman & Hill 1971). In the Cape Province it occurs throughout the western, south-western and eastern parts of the province, although it has not been found in the forests of the southern coastal belt (Fig. 6). In spite of its wide distribution it is nowhere common. Habits These bats normally use caves, disused mines, undisturbed rooms and cellars as resting places, where they hang free from the roofs by their hind legs. They seem to prefer smaller caves or mines, where they gather in small numbers (up to 200 individuals of both sexes). They 85 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 were nearly always found on their own or with Rhinolophus species in the resting places, but in the De Hoop Cave they occurred with Myotis tricolor and Miniopterus schreibersii. Slit-faced bats have characteristically long ears, suggesting that their hearing, and there- fore echo-location ability, is very well-developed. This, combined with their manoeuvrability, meant that they were very difficult to catch in traps or mist-nets. This ability to manoeuvre around obstructions may explain why they were usually found deeper in the caves or mines than any other bat species. They normally fly low, about one metre above the ground, with a slow erratic flight pattern. Some evidence of migration was found. In Namaqualand several caves were deserted for long periods in late summer and winter, and in the Bredasdorp area the slit-faced bats disap- peared from the De Hoop Cave and Spitskop Mine for several months at a time. Too few bats were marked to allow the establishment of a pattern of migration, but it is unlikely that they migrate over very long distances. Food N. thebaica usually forages on or near the ground where it picks up ground insects and then carries them to suitable resting places before eating them. During the survey several such resting places were found. Some were adjacent to, or even in, the roosts. Food remains found in resting places in the Namaqualand and Richtersveld of the north-western Cape showed large numbers of bladder grasshoppers (Pneumoridae), and some scorpions and ordinary grass- hoppers (Acrididae). Breeding Two females, each with a single infant, were taken at Forest Ranch during November and young infants were seen in the De Hoop Cave in December. Three females from Driefontein were caught in early November. All were pregnant with a single implanted foetus. A female collected in October from Sandberg was also pregnant whereas females collected during Janu- ary and February were not pregnant. These data indicate that parturition occurs in November and December only, which is a markedly shorter period than the September — February period given by Smithers (1971) for the species in Botswana. Material examined Survey — 1 Amalinda, 1 Bakleisdrif, 1 Blouputs, 2 Brandkaros, 1 De Hoop, 1 Doringkraal, 2 Doringpoort River, 4 Driefontein, 2 Forest Ranch, 3 Hester Malan, 6 Kersbos, 1 Kroomie, 1 McGregor, 1 Montagu Cave, 3 Okiep, 1 Sandberg, 1 Sewefontein, 1 Skrik van Rondom, 1 Spitskop, 5 Vrolijkheid, 2 Welbedacht Mine, 1 Wonder- gat. Museums — 1 Goodhouse, 11 Grahamstown, 5 Grootvadersbos, 3 Keikamspoort, 11 Kersbos, 6 Twee Riviere (TM); 1 Alice, 2 Atherstone, 3 Compagnies Drift, 1 Garies, 5 Goodhouse, 1 Graaff-Reinet, 14 Rangerton (KWT); 1 De Hoop, 1 Hawston, 1 Hex River, 21 Port St Johns (SAM); 1 Abbotsburg, 7 Brakkloof, 2 Despatch, 1 Fort Brown, 1 Glenlea, 4 Grahamstown, 6 Kleinpoort, 1 Lessendrum, 1 Manley Flats, 1 Moneysworth, 1 Port Alfred, 1 Port Elizabeth, 1 Salem, 1 Stones Hill (AM). 86 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Fig. 7. The distribiuion of Rhinolophm darlingi and Rhinolophus fumigatus.in the Cape Province. Family RHINOLOPHIDAE Bats of this family have a distinct horseshoe-shaped nose-leaf with a pointed lancet behind the horseshoe. Species distinction rests mainly on the form of this structure, with dentition and forearm length as secondary characteristics. However, as the taxonomy of several of the species occurring in the Cape Province is rather confused the whole genus probably needs re- vision. Rhinolophus fumigatus Riippell, 1842 Riippell’s Horseshoe Bat There is some argument whether R. fumigatus should include the subspecies aethiops or whether aethiops should be a species in its own right (Hayman & Hill 1971). Either way this large Rhinolophus appears to be very rare in the Cape Province if it occurs at all. All mention 87 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 of its occurrence (Shortridge 1942, Roberts 1951, Ellerman et al. 1953, Hayman & Hill 1971) seems to be based on two specimens collected in 1903 by C. H. B. Grant at Klipfontein, Little Namaqualand (Fig. 7). These are housed in the British Museum of Natural History (J. E. Hill pers. comm.). A Rhinolophus specimen, in the Transvaal Museum, collected by Roberts at Soetendals- vlei in the south-western Cape and labelled as R. fumigatus, appears to be R. clivosus. This identification is supported by Rautenbach (pers. comm.). No specimens were collected during the survey and there are no specimens from the Cape Province in the South African museums. Rhinolophus darlingi K. Andersen, 1905 Darling’s Horseshoe Bat Until recently this bat had not been recorded from the Cape Province. During the survey all of the medium-sized Rhinolophus were identified as either R. clivosus or R. capensis. How- ever, a recent reassessment by Erasmus & Rautenbach (in press) of several specimens col- lected in the Cape Province, including four bats collected during the survey, suggests that R. darlingi may also be present. These authors used one external and fifteen cranial measure- ments for a computerized multivariate analysis and chromosome karyotyping. The results are still somewhat confusing, since clear diagnostic characters did not emerge. Erasmus & Rauten- bach were able to differentiate between R. clivosus and the other two species using some of the measurements and also chromosome structure. However, they were unable to separate R. capensis and R. darlingi with ease since the analysis did not attach much significance to the position of the small premolar. This was used as a diagnostic character by Hayman & Hill (1971) and for the present study. Erasmus & Rautenbach (in press) separated R. darlingi from R. capensis mostly on allopatry. There is a clear need for further work on the medium-sized Rhinolophus group and this is at present being undertaken by Rautenbach (pers. comm.). Distribution and status The localities listed for R. darlingi by Erasmus & Rautenbach (in press) as well as the four survey specimens identified as this species are shown in Fig. 7. These suggest that the species is widely distributed in the northern Cape. The most westerly locality in the Richtersveld (2816 BD) is in the same locus as a series of R. capensis caught during the survey (Fig. 9), which throws some doubt on the complete allopatric separation of the two species. Because of the taxonomic confusion a statement regarding the species’ conservation status cannot be made, although Erasmus & Rautenbach (in press) suggest that it may be fairly com- mon and widespread wherever suitable roosts such as deep caves or old mines are available. Material examined Survey — 3 Glass Nevie, 1 Koegelbeen Cave. Rhinolophus clivosus Cretzschmar, 1826 Geoffroy’s Horseshoe Bat In this species the anterior upper premolar is either absent or lies outside the tooth row, so that the canine and second premolar are in contact (Hayman & Hill 1971). The forearms of 133 specimens caught during the survey range in length from 51,3 mm to 57,8 mm. 88 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Fig. 8. The distribution of Rhinolophus clivosus in the Cape Province. Distribution and status According to Hayman & Hill (1971) R. clivosus occurs over much of the African continent except for the West African and the Congo forest belt. In the Cape Province it is abundant throughout the western, southern and eastern parts, but not the central Karoo (Fig. 8). Speci- mens were collected wherever suitable caves or disused mines were found. Although colonies of up to ten thousand exist in caves such as De Hoop and Koegelbeen, numbers appear to have declined in other roosts due to human disturbance. Habits This is a typical cave-bat, and the great majority of specimens were taken from caves or disused mines. However, on two occasions specimens were collected from cellars beneath 89 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 houses. Elsewhere (Herselman & Norton in prep.) evidence is presented of short-distance migration of up to 10 km in the south-western Cape. They are low altitude fliers and often hunt below treetop level between the shrubs, where they catch low-flying insects. Breeding In population studies carried out at the De Hoop Cave it was found that the infants are born from the middle of December onwards (Herselman unpublished data). Single-foetus im- plantation is the rule. Fertility is high, 92% of a sample of 150 females caught in November having been pregnant. Material examined Survey — 1 Alice Mission, 5 Apies River Forestry Station, 4 Bean-se-bos, 1 Boomslang Cave, 1 Cango Caves, 51 De Hoop, 1 Dennebos, 2 Die Oog, 8 Die Hel, 2 Drie- fontein, 2 Droevlakte, 3 Glass Nevie, 1 Gordon’s Bay, 1 Grahamstown, 1 Jonkershoek, 2 Klipfontein, 5 Koegelbeen Cave, 3 Koo Cave, 13 Maitland Mines, 5 Roodebergskloof, 2 Sandile’s Cave, 1 Simon van der Stel Mine, 4 Skeleton Cave, 2 Skrik van Rondom, 1 Skurweberg Cave, 2 Spitskop, 2 Sterk- stroom, 1 Storms River Mouth. Museums — 2 Berg River, 3 Grootvadersbos, 1 Hester Malan, 3 Klawer, 5 Knysna, 1 Lourens Farm (TM); 2 Ezelfontein, 1 Fort Beaufort, 1 Kei Road, 11 King William’s Town, 21 Kuruman, 3 Pirie (KWT); 1 Boomslang Cave, 1 Camps Bay, 1 De Hoop, 1 Plettenbergbaai, 1 Stellenbosch (SAM); 5 Fort Beaufort, 1 George, 3 Gleniffer, 1 Grahamstown, 2 Pirie (AM). Rhinolophus capensis Lichtenstein, 1823 Cape Horseshoe Bat This species is very similar in colour and general appearance to R. clivosus. Species dis- tinction is based on dentition. R. capensis has the small anterior upper premolar lying within the tooth row, so that the canine and second premolar are not in contact, whereas in R. clivosus this tooth is either absent or lies outside the tooth row. However, in some individuals it was difficult to say with certainty whether the tooth was within or outside the tooth row and this, combined with the fact that the two species were usually found together, made them dif- ficult to separate. The forearm lengths of 50-57 mm for R. clivosus and 47-51 mm for R. ca- pensis given in Hayman & Hill (1971) could not be used as a character, since in the present study a greater overlap of 51-58 mm and 46-53 mm respectively was found. Distribution and status Although Hayman & Hill (1971) record that R. capensis occurs from Zambia southwards through Natal to the Cape Province, recent re-examination of the material suggests that all the specimens from outside the Cape Province are actually R. darlingi (Smithers 1983). This means that R. capensis is probably endemic to the province. During the survey roosts were found all along the coastal areas of the southern and east- ern Cape but none in Namaqualand, an area where specimens had previously been collected from several localities (Fig. 9). Within its restricted range in the southern parts of the Cape Province R. capensis appears to be fairly common, several cave systems having colonies of more than a thousand individuals. 90 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Fig. 9. The distribution of Rhinolophus capensis in the Cape Province. Habits All the survey specimens were found roosting in caves or disused mines, both sexes being found together. In the south-western Cape the species migrates short distances of up to 10 km (Herselman & Norton in prep.). It shows a typically erratic flight-pattern at low altitudes, and feeds mainly on low-flying insects. R. capensis was often found in the same situations as R. clivosus, although clusters of the two species seemed to keep separate. Because of their similarity in size, shape and distribution it is not clear how they differ ecologically, and this is a subject that deserves further study. Breeding At the De Hoop Cave it was found that a single infant is born from the middle of Decem- ber onwards. Infants are carried by their mothers during the day, but are left behind when the 91 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4. JANUARY 1985 adults are feeding. Samples taken from De Hoop in November showed a fertility of 86% among the females. Material examined Survey — 3 Abrie, 1 Bean-se-bos, 10 Bloukrans, 3 Breakfast Vlei, 66 De Hoop, 1 Denne- bos, 1 Die Kelders, 10 Droevlakte, 7 Grootplaas Cave, 6 Keurbosfontein, 4 Klip- fontein, 3 Maitland Mines, 5 Marcus Mud Crawl, 6 Millwood Mines, 2 Montagu Cave, 1 Salem, 7 Skrik van Rondom, 6 Spies Cave, 2 Spitskop, 2 Storms River Mouth, 5 Tunnel Cave, 6 Wondergat. Museums — 4 Klawer, 1 Kleinpoort, 3 Saldanha Bay, 1 Stilbaai (TM); 4 Compagnies Drift, 2 Ezelfontein, 2 Goodhouse, 20 Het Kruis, 10 King William’s Town, 35 Orrelgat, 1 Paddagat (KWT); 2 Cape Town, 1 Clanwilliam, 2 De Hoop, 14 Montagu (SAM); 1 Alicedale, 1 Brakkloof, 1 Grahamstown, 1 Het Kruis, 2 Orrelgat, 10 Slaaikraal (AM). Rhinolophus denti Thomas, 1904 Dent’s Horseshoe Bat This small Rhinolophus is very similar in appearance to R. swinnyi and it is possible that they are conspecific. Neither species has been collected in the Cape Province in the last half century. No specimens of R. denti could be traced in the South African museums, although there are two specimens in the British Museum of Natural History from Louisvale near Upington and Kuruman (J. E. Hill pers. comm.); (Fig. 10). Whether R. denti is included with R. swinnyi or not the species can be considered as ex- tremely rare, possibly even extinct, in the Cape Province. Rhinolophus swinnyi Gough, 1908 Swinny’s Horseshoe Bat R. swinnyi is the smallest species in the genus, with a forearm length of about 42,5-43,8 mm, (Roberts 1951). Distribution and status Hayman & Hill (1971) state that R. swinnyi occurs in the eastern Cape Province and ex- tends northwards to Tanzania. However, it appears to be very rare in the Cape Province. Dur- ing the survey no specimens were collected and only 20 museum specimens, all collected be- fore 1940, were examined (Fig. 10). All these specimens were from only two grid squares, which suggests that R. swinnyi was only marginally distributed in the Cape Province. The lack of recent material indicates that it may no longer occur in the province. Material examined Museums — 3 King William’s Town, 4 Pirie, 1 Ross Mission (KWT); 5 Pirie (SAM); 7 Pirie (AM). Hipposideros coffer (Sundevall, 1846) African Leaf-nosed Bat Before the survey two subspecies had been recorded from southern Africa, H. c. caffer in the eastern Cape and northwards up the east coast, and H. c. angolensis in northern South West Africa (Hayman & Hill 1971). A series of 21 specimens of H. c. caffer from Port St Johns 92 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. in the S.A. Museum was examined. The average forearm length of these specimens is 48,2 mm, which led Roberts (1951) to state that this subspecies is smaller than H. c. angolensis, which has a forearm measuring 50,3 mm. However, the authors collected two specimens with a mean forearm length of 47,4 mm at Brandkaros in the Richtersveld. This measurement tends to identify these specimens as H. c. coffer, but their occurrence in the northern Cape suggests that they should be H. c. angolensis. Therefore criteria other than forearm length should be found for subspecific division of this species. Distribution and status Mentions of the presence of this species in the Cape Province prior to the survey (e.g. Roberts 1951, Hayman & Hill 1971) appear to have been based on very old (before 1930) specimens from Port St Johns and Mqanduli in Transkei (Fig. 10). It has not been collected 93 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 since and this, together with the excision of Transkei, means that the subspecies H. c. caffer is unlikely to occur within the present borders of the Cape Province. The two specimens from the Richtersveld, collected during the survey (Fig. 10), showed that the Angolan subspecies H. c. angolensis occurs in the north-western Cape and three specimens collected recently by Erasmus & Rautenbach (in press) near Marydale (2922AB) suggest that it may be even more widely distributed. Hayman & Hill (1971) record that it is widespread in the drier woodland and savanna regions of Africa. Habits The two specimens were found roosting in a small cave with a very narrow entrance, which they shared with individuals of Nyctehs thebaica. Erasmus & Rautenbach (in press) found three H. caffer roosting in an old mine. Material examined Survey — 2 Brandkaros. Museums — 21 Port St Johns (SAM); 3 Mqanduli (AM). Family VESPERTILIONIDAE This is a very large family, several species of which are widespread in the Cape Province. They are strong fliers and many species look very similar in appearance. Since external fea- tures are often insufficiently marked to make generic identification certain, Hayman & Hill (1971) found it necessary to use dental formulae and tooth patterns in their key to the genera. These, together with ear length, forearm length and a few other characteristics, form the basis of species distinction. Myotis seabrai (Thomas, 1912) Angola Wing-gland Bat Bats of the genus Myotis have 38 teeth, which is more than that in all other genera in the family except Kerivoula. However, the latter is easily distinguished by its highly-elevated cranium. Whereas Roberts (1951) treated both Myotis and Cistiigo as full genera, Ellerman et al. (1953) and Hayman & Hill (1971) recognized Myotis as the genus with two subgenera, Selysius and Cistugo. M. tricolor was then placed in the subgenus Selysius and M. lesueiiri and M. seabrai in the subgenus Cistugo. Bats of the subgenus Cistugo are among the rarest in the world and therefore deserve special attention. They are the only bats on the continent that possess wing-glands. The subge- nus has been divided into two species on the basis of forearm length and the position of the wing-glands. Both of these characters show some variation and therefore more taxonomic work will have to be done to separate them convincingly. The wing-glands of M. seabrai are recorded as being larger than in M. lesueuri and lie in various positions on the wing, whereas those of M. lesueuri are always found close to the fore- arm (Shortridge 1942). Shortridge also records that in several instances there are two glands together on either side, and the same was found in one of the survey specimens. M. seabrai is recorded as being smaller in size than M. lesueuri. The five survey specimens have forearm lengths between 30,2 mm and 33,2 mm, with a mean of 31,9 mm. 94 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Fig. 11. The distribution of Myotis seabrai, M. lesueuri and M. tricolor in the Cape Province. Distribution and status M. seabrai has a very limited distribution in the Cape Province, occurring only in the north-western parts (Fig. 11). From there it extends northwards through South West Africa to Angola. It had previously been collected only from Goodhouse which suggests that it is rare in the Cape. However, Shortridge (1942) stated that it was the most plentiful bat around Good- house Citrus Estate. A similar situation was found at Steyerskraal, 100 km east of Goodhouse, where three specimens from a large number of bats flying next to a river were shot. Two speci- mens were also netted over water at Goegab, 100 km south of Goodhouse. Further collecting may show this bat to be fairly common within a very restricted range in the Cape Province. 95 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Habits Because the survey specimens were cither shot in flight or netted over water no informa- tion on their roosts was obtained. They were collected shortly after sundown while flying a few metres above the ground. Roberts (1951) records this species flying low amongst orange trees and snapping small insects from the leaves. Material examined Survey — 2 Goegab, 3 Steyerskraal. Museums — 8 Goodhouse (KWT). Myotis lesueuri (Roberts, 1919) Lesueur’s Wing-gland Bat This species is very similar in size and colour to Eptesicus. However, the latter can easily be distinguished by its dentition, since it has two minute upper premolars lying transverse to the tooth row. There seems to be some confusion in the literature about the presence of wing- glands. Roberts (1951) states that no glands were visible in the dry skin of the wing of the type specimen, a male collected at Franschhoek in 1919. Similarly, the authors could find no wing- glands on two study skins from specimens collected at Citrusdal in 1937 by Shortridge and now housed in the Kaffrarian Museum. Further, Hayman & Hill (1971) question whether wing- glands are present or not. The answer to the controversy may he in Shortridge’s (1942) sugges- tion that “in dry skins the glands become absorbed and lost to sight”. During the survey a male taken from a trap in Jonkershoek in 1978 shows a small gland close to the forearm on the left wing only. A specimen collected in January 1979 near Beaufort West by staff of the Transvaal Museum also has a small gland on the right wing close to the forearm. Distribution and status In total only eight specimens of this bat have been collected in the Cape. The survey shows that the distribution is not as restricted to the south-western Cape as was previously thought (Fig. 11). The specimens collected in the Great Karoo and in the northern Cape sug- gest that it may occur in most of the Cape Province. However, there is no doubt that it is very scarce and therefore deserves special attention. If it is a valid species separate from M. seabrai it is probably endemic to the province. Material examined Survey — 1 Jonkershoek, 1 Mazelsfontein. Museums — 2 Hex River Estate (KWT); 1 Beaufort West, 1 Saasveld (TM). Myotis tricolor (Temminck, 1832) Cape Hairy Bat Distribution and status M. tricolor occurs in the southern part of the Cape Province, from Citrusdal in the south- west to King William’s Town in the east (Fig. 11). From there it is widely distributed up the eastern side of Africa as far as Ethiopia. It is not very abundant in the Cape Province and only a few museum records exist. During the survey 44 specimens were collected from nine locali- ties. 96 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Habits One specimen from Michell’s Pass was netted flying over water. The others were all found roosting in caves and mines. It appears that this species prefers larger caves that are relatively undisturbed, usually ones that contain pools of water. On every occasion the authors found it in the same roosts as Miniopterus schreibersii and it usually mixed freely with this species in daytime clusters on the walls. It has well-developed jaws and teeth and is difficult to handle. Other bats are sometimes bitten to death when they are placed together with this species in collecting bags. Evidence of migration was found when a Myotis tricolor was recaptured in Montagu Cave, at a distance of 90 km from De Hoop Cave where it was banded. Kingdon (1974) states that all-male and all-female colonies have been found in various parts of the world, indicating that a certain amount of migration must take place. In New England Davis & Hitchcock (1965) found winter and summer migration of a related species M. lucifugus over distances of up to 275 km. Breeding Several juveniles were seen in Gant’s Mine near Port Elizabeth during November 1976. Three pregnant females with single embryos and a female with a juvenile were taken in Montagu Cave towards the end of October 1979. At that time none of the other cave inhabi- tants, Miniopterus schreibersii and Rhinolophus capensis, had young. The same was found in the De Hoop Cave (Herselman unpublished data), where the Myotis tricolor gave birth from the end of October to the middle of November. At night, when the adults emerge to feed, the juveniles are left in thickly-packed clusters against the walls of the cave or mine. Material examined Survey — 7 De Hoop, 4 Die Hel, 3 Droevlakte, 12 Gant’s Mine, 8 Marcus Mud Crawl, 1 Michell’s Pass, 7 Montagu Cave, 1 Skrik van Rondom, 1 Skurweberg Cave. Museums — 10 King William’s Town (KWT); 2 De Hoop, 2 Montagu (SAM); 1 King William’s Town, 1 Mqanduli (AM). Pipistrellus nanus (Peters, 1852) Banana Bat There are two members of the genus Pipistrellus recorded from the Cape Province, P. nanus and P. kuhlii, and both appear to be only marginally distributed in the eastern Cape. They are rather difficult to distinguish and during the survey nearly half of the specimens from the Cape Province in the South African museums were found to be incorrectly identified. It was found that the character of a “hatchet-shaped tragus’’ listed for P. nanus in Hayman & Hill (1971) was rather confusing. To differentiate these two species the present authors pre- ferred to ignore this character and base identification on the second upper incisor, which is re- duced in P. kuhlii and not in P. nanus, as well as on the more domed cranium of P. nanus. As the popular name implies P. nanus is known to roost in banana trees. However, during the survey an intensive search in banana plants on several farms in the eastern Cape yielded no specimens. Their occurrence is therefore based on several rather old specimens from the Albany and South African Museums. 97 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Fig. 12. The distribution of Pipistrellus nanus in the Cape Province. Distribution and status The banana bat has been recorded as widespread in the eastern Cape Province from as far south as East London (Ellerman et al. 1953). The museum records (Fig. 12) support their past occurrence there, although the lack of more recent material suggests that their range may have decreased from the south. A further factor is that a large part of the range mentioned by Ellerman et al. (1953) has been excised to form Transkei, and therefore its present range in the Cape Province is probably restricted. However, this tropical species is widespread up the east coast and in central Africa and lives in plantations of cultivated plants. Therefore its overall conservation status is probably secure. Material examined Museums — 1 Bedford, 2 Port St Johns (SAM); 1 Kei Road (AM). 98 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Fig. 13. The distribution of Pipistrellus kuhlii and Eptesicus hottentotus in the Cape Province. Pipistfellus kuhlii (Natterer, 1817) Kuhl’s Pipistrelle Roberts (1951) separated the species kuhlii from the genus Pipistrellus and placed it in the genus Romicia. However, subsequent authors (Ellerman et al. 1953, Hayman & Hill 1971) do not recognize this separation. Distribution and status This marginal species appears to be very rare in the Cape Province, although it is wide- spread up the east coast of Africa and into Europe (Hayman & Hill 1971). Roberts (1951) and Ellerman et al. (1953) state that it occurs from the forested areas of the southern and eastern Cape as far west as Knysna. During the survey no material was collected and the only museum specimens which could be traced were from the eastern Cape and Transkei (Fig. 13). 99 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4. JANUARY 1985 Habits Little is known of its habits, most specimens being caught when flying at night, either in- side forests or not far from them (Roberts 1951). Smithers (1971) and Rautenbach (1982) found them roosting under the loose bark of dead trees. Material examined Museums — 1 Port St Johns (SAM); 2 Pirie (AM). Eptesicus hottentotus (A. Smith, 1833) Long-tailed House Bat This species is noticeably larger than any other Eptesicus in the Cape Province. Two sub- species, E. h. hottentotus and E. h. pallidior, are mentioned in checklists of mammals of South Africa. Roberts (1951) and Meester et al. (1964) list the distribution range of E. h. hottentotus as the eastern, southern and south-western Cape as far north as the mountains of Little Nama- qualand, and that of E. h. pallidior as the north-western Cape Province and South West Africa. According to Roberts (1951) E. h. hottentotus has a slightly shorter wingspan and a darker colour than E. h. pallidior, which is light yellow in colour. However, the specimens obtained during the study do not support such a clearcut division with respect to wing-length and colour. A female caught at Rocherpan on the west coast in April 1978 is dark in colour and has a fore- arm length of 47,5 mm, thus showing characteristics of E. h. hottentotus. Five specimens caught at Taung in Bophuthatswana, which falls within the distribution of E. h. pallidior, can- not be separated from the Rocherpan specimen. Their colour is also dark brown and their forearm lengths range from 47,0 to 51,5 mm, which is not noticeably greater than that of the Rocherpan specimen. Distribution and status The Taung series was collected about 400 km to the east of the South West African border and about 600 km east of Goodhouse, which was the most easterly site where E. h. pallidior had been collected before the survey. Therefore, it probably also occurs in the western Trans- vaal and southern Botswana although it has never been collected in these areas. The new localities show a wide distribution of E. h. pallidior in the western, north-western and northern Cape Province (Fig. 13). However, the species is rare in the Cape Province, since only six specimens from two localities were collected during the survey. There are a further seven specimens in the South African museums. They are from three localities and were all collected before 1938. Habits The Rocherpan specimens were netted flying low over a marsh and the Taung series was taken from a small hollow in rocks situated at the entrance to an old mine tunnel inhabited by Miniopterus schreibersii. Material examined Survey — 1 Rocherpan, 5 Taung. Museums — 4 Goodhouse, 2 Hex River Estate (KWT); 1 Cape Town (SAM). 100 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Eptesicus melckorum Roberts, 1919 Melck’s House Bat Eptesicus melckorum and E. capensis are very similar in outward appearance but there are marked differences in their forearm and skull measurements. At Herbertsdale and Vrolijkheid both species were found together in roofs. Distribution and status Hayman & Hill (1971) mentioned five Zambian specimens as constituting the only material collected outside the Cape Province and suggested that they might belong to a local subspecies. Most other authors stated that E. melckorum occurs only in the south-western Cape. During the survey the majority of specimens came from this area. However, two fe- males collected at Three Sisters and Graaff-Reinet in the Great Karoo indicate a slightly wider distribution (Fig. 14). 101 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Within its distribution area E. melckorum is quite common and is usually found in hiding places in groups of several individuals. Habits In most cases this species was found roosting under the eaves of buildings but a few speci- mens were mist-netted over water in the early evening. The specimen from Three Sisters was found under the bark of a poplar tree. Breeding Three pregnant females, two with twin foetuses, were found at Sandberg during late October. Material examined Survey — 1 Aurora, 1 Bontebok Park, 1 Bot River, 4 Compagnies Drift, 2 De Hoop, 1 Ebb-and-flow, 1 Graaff-Reinet, 1 Goukamma, 2 Herbertsdale, 2 Hester Malan, 1 Jonkershoek, 2 Rocherpan, 3 Sandberg, 1 Strand, 1 Three Sisters, 4 Vrolijk- heid, 2 Ysterfontein. Museums — 2 Berg River, 1 Norap, 1 Stilbaai (TM); 1 Assegaaibosch, 1 Clanwilliam, 1 Rocherpan (SAM); 1 Compagnies Drift, 1 Oudtshoorn (AM). Eptesicus capensis (A. Smith, 1829) Cape Serotine Roberts (1951), Ellerman et al. (1953) and Meester et al. (1964) list three subspecies of this bat, of which only E. c. capensis occurs in the Cape. However, Meester et al. (1964) state that the subspecies need revision, particularly with respect to their geographic ranges and Hayman & Hill (1971) do not recognize any separate subspecies. Distribution and status E. capensis is fairly common throughout the Cape Province (Fig. 15) and a number of in- dividuals are usually found together in each roost. Because of its habit of roosting in houses it is one of the few bat species that may have increased in numbers as a result of man’s acivities. Habits In the majority of cases E. capensis was found hiding in small crevices under the eaves of buildings. On a few occasions specimens were netted over water. At Brandvlei one specimen was taken from wool-bags hanging from the roof of a garage. Breeding Pregnant females were collected in early November at Barkley West and juveniles were found in late November at Barkly East. Material examined Survey — 1 Barkley East, 6 Barkly West, 5 Bontebok Park, 1 Brandvlei, 2 Canaga, 4 Die Oog, 1 Elliot, 1 Herbertsdale, 1 P.K. le Roux Dam, 1 Riebeeck-Oos, 1 Robert- son, 2 Rolfontein, 1 Steyerskraal, 3 Vrolijkheid. Museums — 2 Keiskamspoort, 1 Kuruman, 1 Vryburg (TM); 3 Atherstone, 6 Compagnies Drift, 2 Ezelfontein, 19 Kaggasmoudt, 22 King William’s Town, 2 Lady Grey, 1 Middelburg, 2 Paddagat, 4 Tambookies Drift, 3 Traveller’s Rest, 13 Lambert’s 102 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P, Bay, 1 Simonstown (SAM); 1 Blythswood, 1 Brakkloof, 1 Compagnies Drift, 12 Grahamstown, 2 Jameston, 1 Kaggasmoudt, 1 Kei Road, 1 Kimberley, 2 King William’s Town, 2 Kleinpoort, 1 Pirie, 1 Stones Hill, 3 Victoria West, 1 Waterloo (AM). Eptesicus notius G. Allen, 1908 Cape Horn-skinned Bat It is unlikely that this is a valid species. It is known only from the type specimen which was collected in Cape Town in 1908 and is now housed in the Museum of Comparative Zoology, Harvard, Massachusetts, U.S.A. Koopman (1975) considered this specimen as a slightly aber- rant individual of E. capensis, possibly infested with nematodes, and this view is followed by Swanepoel, Smithers & Rautenbach (1980). Therefore the present authors do not consider E. notius as part of the Cape bat fauna. 103 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Fig. 16. The distribution of Laephotis wintoni and Scotophilus dinganii in the Cape Province. Laephotis wintoni Thomas, 1901 De Winton’s Long-eared Bat Bats of this genus are very rare in collections and only one specimen has been taken in the Cape Province (Rautenbach & Nel 1978). Hayman & Hill (1971) lumped all specimens under the species L. wintoni, but later Hill (1974) divided the genus into four distinct species with L. wintoni in East Africa, L. botswanae from Zambia and Botswana, L. angolensis from Angola and L. namibensis from South West Africa. This division was based on only 25 specimens. In February 1978 Rautenbach & Nel (1978) netted a specimen over a river at the Algeria Forest Station in the Cedarberg (Fig. 16). These authors compared nine measurements of this specimen with those of the four known species using a multivariate analysis and came up with the rather surprising result that their specimen was by far closest to the East African L. wintoni which is geographically the most distant. This led them to say that the taxonomic status of the genus is not satisfactorily resolved and can only be by the acquisition of more material. 104 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C P. On the above evidence Laephotis wintoni is provisionally included in the bat fauna of the Cape Province. It is not clear whether the specimen represents a well-established local popu- lation. Because of the low number of specimens collected its status in Africa must be described as “rare” although the widespread distribution range suggests that it may occur naturally in very low numbers. Material examined Museums — 1 Algeria (TM). Scotophilus dinganii (A. Smith, 1833) Yellow House Bat The genus Scotophilus is characterized by having only one upper incisor, one upper pre- molar, three lower incisors and two lower premolars on each side. There is some confusion in its taxonomy. Robbins (1978) has shown that S. gigas should be called S. nigrita and that all specimens that have been identified as S. nigrita since the late 1800s must now be called S. dinganii. S. dinganii is the only species that has been collected in the Cape Province. Hayman & Hill (1971), Meester et al. (1964) and Roberts (1951) recognized three subspecies of what they called 5. nigrita and stated that the subspecies pondoensis occurs only in the eastern Cape Province. Distribution and status No material was collected during the survey, but 27 museum specimens were examined (Fig. 16). All these specimens labelled as S. nigrita, were collected in the eastern Cape be- tween 1931 and 1941. Recently Erasmus & Rautenbach (in press) collected a specimen in the Kalahari Gemsbok National Park which they separated from the closely related S. viridis using a morphometric identification model even though its colouring was more like this species than S. dinganii. Thus the species appears to have a very restricted distribution in the eastern and northern Cape although it is fairly widespread in Africa. Habits Roberts (1951) stated that this species usually lives under the eaves of buildings or in deserted nests of woodpeckers and barbets. Erasmus & Rautenbach (in press) record that their specimen was flying with several others, probably of the same species, over a drinking trough in savanna woodland. Material examined Museums — 1 Amabele, 3 Atherstone, 1 Kaggasmoudt, 17 King William’s Town (KWT); 1 Port St Johns (SAM); 1 Gleniffer, 2 King William’s Town (AM). Kerivoula lanosa (A. Smith, 1847) Lesser Woolly Bat This species is characterized by its long, woolly hair, 38 teeth in the dentition (the same as Myotis) and high cranium (similar to Miniopterus). Hayman & Hill (1971) question whether K. harrisoni, K. lanosa and perhaps K. muse ilia represent distinct species or merely local forms of one widespread species. 105 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Fig. 17. The distribution of Kerivoula lanosa and Miniopterus fraterculus in the Cape Province. Distribution and status Seven specimens of this rare species were examined (Fig. 17). There are four museum specimens from the eastern Cape and two from Knysna in the southern Cape. During the sur- vey one specimen was collected at Dennebos near Knysna. It appears to be restricted to the forested regions of the southern and eastern Cape. The overall conservation status of this species depends to a large extent on whether it proves to be clearly distinguishable from K. harrisoni. Habits All localities where specimens of K. lanosa have been collected show that the species has an affinity for forests. Roberts (1951) stated that most specimens were taken from the hanging nests of birds such as weavers. 106 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Material examined. Survey — 1 Dennebos. Museums — 1 Knysna (TM); 1 Cullendale, 1 Knysna, 3 Pirie (AM). Miniopterus fraterculm Thomas & Schwann, 1906 Lesser Long-fingered Bat Bats of the genus Miniopterus are easily distinguished by the elongated second phalanx of the third digit. The four species listed by Hayman & Hill (1971) as occurring in Africa are all very similar in appearance and are usually separated on forearm and skull lengths. However, there is some overlap in both size and distribution ranges, and the taxonomy of the genus is therefore rather confused. Hayman and Hill (1971) record two species for the Cape Province, M. fraterculus and M. schreibersii. The former is generally regarded as being smaller than the latter although the forearm length classes of 42-44 mm and 42-47 mm respectively used by Hayman & Hill (1971) cannot be considered as diagnostic characters. These authors mention that M. fraterculus has variously been included in M. minor and M. schreibersii. However they follow Harrison & Clancey (1952) in declaring it a separate species, without giving clear reasons for the distinc- tion. During the present study different forms of Miniopterus were found on several occasions. In large roosts of Miniopterus occasional individuals had a russet or partly russet colour and appeared to be smaller than the typical dark blackish-brown coloured M. schreibersii. These small, russet individuals were assumed to belong to M. fraterculus, and this identification was supported by Prof. J. Meester (pers. comm.) who examined some of the specimens. However, wet specimens were very difficult to separate and this may have led to mis-identifications in the collection. Distribution and status Roberts (1951) recorded M. fraterculus from Knysna and Hayman & Hill (1971) stated that it occurs in the eastern Cape Province, eastern Transvaal, southern Malawi and Zambia. During the survey at least eight specimens were collected at five roosts in the southern and eastern Cape but no Cape specimens were found in the museums (Fig. 17). Due to confusion with M. schreibersii it is difficult to assess its status. However, it does not appear to be rare in the Cape Province. Habits In every instance this species was found in the same roost as M. schreibersii. No evidence could be found of how the two species differ ecologically. Material examined Survey — 2 Bat’s Cave, 1 Gant’s Mine, 2 Millwood Mines, 2 Sterkstroom, 1 Storms River Mouth. Miniopterus schreibersii (Kuhl, 1819) Schreibers’ Long-fingered Bat This species is by far the most abundant cave-bat found in the Cape Province, where it is represented by the subspecies M. s. natalensis (A. Smith, 1834). 107 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Distribution and status M. schreibersii has a very wide distribution, being found in Africa, Asia and Australia. In Africa it occurs throughout the continent south of the Sahara. The subspecies M. s. natalensis is found from the southern Congo and Zambia southwards (Hayman & Hill 1971). It is abun- dant in the Transvaal (Van der Merwe 1972a) and Natal (Harrison & Clancey 1952). During the survey large colonies were found in all regions of the Cape Province (Fig. 18). At De Hoop 157 specimens were collected for population dynamics studies, and a further 300 specimens were collected from localities throughout the province. The species is very common in the Cape Province. However, its marked decline in some areas due to disturbance at the roosts gives cause for concern. This may be of considerable economic importance, since large colonies are likely to have a substantial influence on local in- sect populations. 108 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Habits Bats of this species congregate in very large colonies, especially in the nursery caves such as the De Hoop Cave where about 80 000 females form a nursery colony during the summer months. They migrate seasonally and banding studies have shown that individuals migrate at least 250 km to and from De Hoop every year (Herselman & Norton in prep.). As found by Van der Merwe (1973b, 1973c, 1975) and Norton & Van der Merwe (1978) in the Transvaal, these movements appear to be associated with the selection of suitable sites for winter hiberna- tion. Within roosts these bats gather in large densely-packed clusters. Densities of up to 2 800 individuals/m^ have been estimated by Norton & Van der Merwe (1978). M. schreibersii is of- ten found in association with other bat species especially, Rhinolophus ctivosus, R. capensis and Myotis tricolor. Breeding Of the 170 females examined at De Hoop in October 158 (93%) were pregnant. Parturi- tion took place from the end of October to the middle of November. The infants were left in dense clusters against the walls while the females fed or rested in the cave. There is some evi- dence that females suckle offspring other than their own, since a pregnant female suckling an infant was collected. Twinning is rare and was found in only one out of more than 150 preg- nant females dissected. During the survey only four nursery caves were located, at De Hoop in the south-western Cape, Koegelbeen Cave in the northern Cape, Bloukrans Cave in the Karoo, and the disused Maitland Mines in the eastern Cape. These nursery caves are considered to play an important role in the survival of the species. Material examined Survey — 9 Bat’s Cave, 28 Bean-se-bos, 21 Bloukrans, 157 De Hoop, 3 Die Oog, 18 Die Hel, 1 Droevlakte, 49 Gant’s Mine, 4 Goegab, 1 Grahamstown, 1 Hansies River Mine, 1 Hot Pot, 1 King William’s Town, 22 Koegelbeen Cave, 8 Marcus Mud Crawl, 8 Millwood Mines, 52 Montagu Cave, 3 Nabeeb Mine, 4 Norlim Mine, 2 Orrelgat, 1 Puntjie Cave, 22 Roodebergskloof, 10 Skrik van Rondom, 2 Skurweberg Cave, 17 Sterkstroom, 4 Storms River Mouth. Museums — 3 Knysna, 3 Table Mountain (TM); 69 King William’s Town, 1 Patensie, 1 Schoenmakerskop (KWT); 2 De Hoop, 1 King William’s Town, 1 Knysna, 1 Wynberg (SAM); 4 Buffelskloof, 1 Dunbrody, 1 Gleniffer, 4 King William’s Town, 6 Mqanduli, 1 Redhouse, 11 Slaaikraal (AM). Family MOLOSSIDAE Sauromys petrophilus (Roberts, 1917) Roberts’ Flat-headed Bat As the common name suggests the skull of this bat is conspicuously flattened, a character- istic which clearly separates it from species of Tadarida, the other free-tailed genus. When originally describing it, Roberts (1917) placed it under Platymops with Sauromys as a sub- genus, but Sauromys has now been elevated to generic rank (Hayman & Hill 1971). The genus as a whole is limited to southern Africa. 109 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Fig. 19. The distribution of Sauromys petrophilus and Tadarida pumila in the Cape Province. Distribution and status During the survey no specimens of this species were collected. Only eight museum speci- mens were located bearing out the statement of Hayman & Hill (1971) that flat-headed bats are not numerous in collections. All the specimens are from the south-western Cape and Namaqualand (Fig. 19), and all, except those from Augrabies Falls, were collected before 1940, suggesting that the species is very rare in the Cape. However, the main reason for their scarcity seems to be that they have very specific habitat requirements and Rautenbach (pers. comm.), who collected the two Augrabies Falls specimens in 1970 and 1977, is of the opinion that they will prove to be more widespread if more intensive searches are made in areas with suitable habitats. 110 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Habits The flattened heads of these bats enable them to roost in very narrow crevices between rocks, where they usually occur singly or in pairs (Smithers 1971). Material examined Museums — 2 Augrabies Falls, 1 Michell’s Pass (TM); 1 Clanwilliam, 4 Goodhouse (KWT). Tadarida condylura (A. Smith, 1833) Angola Free-tailed Bat Roberts (1951), Ellerman et al. (1953), Meester et al. (1964) and Hayman & Hill (1971) state that this free-tailed bat occurs from the eastern Cape Province northwards over a large part of Africa. However, during the survey no specimens were collected and no material from the Cape Province was found in the South African museums. Since a large proportion of the eastern Cape has been excised to form Transkei it is quite likely that this marginal species does not occur within the present borders of the Cape Province. Tadarida pumila (Cretzschmar, 1826) Little Free-tailed Bat As suggested by the common name this species is noticeably smaller than any other free- tailed bat in the Cape Province. Distribution and status Before the survey the little free-tailed bat was only recorded from the eastern parts of southern Africa. However, in February 1978 a series of 13 specimens, all females, was mist- netted over water by Rautenbach & Nel (1978) at the Algeria Forest Station in the Cedarberg. During the present survey a single female was collected in August 1978 from a farm dam at Goegab near Springbok. In September 1979 a series of 10 specimens was mist-netted in the same place. During the same month two males and a female were caught at Algeria and an- other female was caught at Michell’s Pass near Ceres. The range of this species therefore seems to be greater than originally described, probably extending southwards along the mountain chains of the south-western Cape Province (Fig. 19). The present study also refutes the idea of sex segregation mentioned by Rautenbach & Nel (1978). The rate of success with mist-netting suggests that the species may be more common in the western Cape than was previously thought. Material examined Survey — 3 Algeria, 11 Goegab, 1 Michell’s Pass. Museums — 13 Algeria (TM). Tadarida aegyptiaca (E. Geoffroy, 1818) Egyptian Free-tailed Bat This is the largest free-tailed bat in the Cape Province. Roberts (1951) recognizes. T. aegyptiaca and T. bocagei as distinct species, but Hayman & Hill (1971) regard T. a. aegyptiaca and T. a. bocagei as two subspecies of T. aegyptiaca, with the slightly smaller and darker T. a. bocagei occurring in the drier western parts of southern Africa. Ill ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16. PT. 4. JANUARY 1985 Distribution and status According to earlier authors T. aegyptiaca occurs all over the Cape Province and over the greater part of the continent. The present survey confirmed its wide distribution in the prov- ince (Fig. 20), showed that it is fairly common throughout, and has several large colonies of a few hundred individuals. Its conservation status appears to be secure. Its habit of roosting in buildings, especially churches, suggests that it is one of the few insectivorous species to have benefitted from man’s activities. Habits Colonies were found in cracks between rocks. T. aegyptiaca also congregates in crevices in buildings. It has a strong musky odour. As suggested by Shortridge (1942) it is seldom asso- ciated with other bats, although at Herbertsdale it was found in the same roof as Eptesicus 112 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. capensis. It appears to congregate to breed, as indicated by the fact that a series of 10 speci- mens from Junction Farm consists only of pregnant females collected from clefts in rocks near waterholes. Breeding Heavily pregnant females were found at several localities during November, suggesting that births take place towards the end of November or the beginning of December. Material examined Survey — 1 Abraham Kriel, 10 Algeria, 1 Aronegas, 2 Barkly East, 3 Cape Infanta, 2 Dor- ingbaai, 3 Elliot, 4 Goegab, 1 Graaff-Reinet, 1 Grahamstown, 2 Herbertsdale, 1 Herbertshope, 10 Junction Farm, 3 Middelburg, 3 Riversdale, 1 Rolfontein, 2 Wellington, 2 Wilgerfontein, 1 Vrolijkheid. Museums — 4 Algeria, 1 Augrabies Falls, 1 Burgersdorp, 2 Kamferboom, 1 Twee Riviere, 2 Van Wyksvlei (TM); 3 King William’s Town, 3 Upington (KWT); 1 Cape Town, 2 King William’s Town, 1 Pirie (SAM); 2 Aliwal North, 1 Blythswood, 1 Fort Beaufort, 7 Grahamstown, 3 King William’s Town, 10 Middleton, 1 Peddie, 1 Victoria West, 7 Waterloo (AM). DISCUSSION During the survey a considerable amount of information was gathered on the occurrence and numbers of bats in the Cape Province, especially of the communal species which congre- gate in large roosts. However, the conservation status of many species is still not clear. There are four main reasons for this. Firstly, the taxonomy of several species or groups of species is rather confused. Secondly, the excision of Transkei from the Cape Province excludes most of the known range of several of the marginal tropical species from the eastern Cape. Thirdly, a far more intensive netting programme will have to be carried out to form an accurate picture of the solitary species, which may naturally occur in rather low densities. Fourthly, although some of the communal species may be abundant, it was difficult to show whether they were main- taining their numbers or declining. To determine the distribution and status of a species it is essential to be able to separate it from other closely related species. The groups of species from the Cape that need to be exam- ined more closely in order to resolve taxonomic problems involving such differentiation are: Myotis seabrai and M. lesueuri; the four species of the genus Laephotis; Rhinolophus denti and R. swinnyi; Kerivoula lanosa and the East African K. harrisoni; Rhinolophus clivosus, R. capensis and R. darlingi; and the two Miniopterus species, M. schreibersii and M. fratercu- lus. The last two groups deserve special attention because they show some overlap in the diag- nostic characters used and furthermore often occur together in large colonies. Therefore a clear ecological separation will have to be documented before they can be convincingly iden- tified as distinct species. A further pair of species that seems to have been confused is Pipistrellus nanus and P. kuhlii. Initially the authors also found it difficult to separate them using the key in Hayman & Hill (1971), but later found they could be clearly separated on skull characters. 113 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 198.*; Other problems caused by mis-identifications are the records of Eptesicus notius and Nycteris hispida in the Cape. E. notius is now generally considered invalid as a species, being merely an aberrent E. capensis. The record of N. hispida is based on a single specimen that was found with several N. thebaica adults and was most probably a juvenile of this species. Whether or not species are endemic to the Cape Province also depends upon taxonomic interpretation. Myotis lesueuri is recorded as endemic to the Cape but, if it proves to be con- specific with M. seabrai, its range is somewhat greater. Although Kerivoula lanosa is listed only from the Cape Province in Hayman & Hill (1971), Rautenbach (1982) and Smithers (1983) re- cord it from the Transvaal and other areas in southern Africa. If it proves to be conspecific with K. harrisoni, the species’ distribution range will include a large part of Africa. There is too much disagreement on the taxonomy of African bats at the subspecific level for it to be possible at present to reach a conclusion about the presence of endemic subspecies in the Cape Province. Table 1 summarizes what is known of the conservation status of bats in the Cape Province. It shows that 16 species can be considered as rare in the province. This number can be divided into two groups, those that are only marginally distributed and those that are nowhere com- mon. The excision of part of the eastern Cape to form Transkei has affected the status of the marginal tropical species in the Cape Province. Of these Hipposideros caffer and Tadarida condylura had only been recorded from what is now Transkei. At present it seems doubtful whether T. condylura occurs within the present borders of the province. During the survey H. caffer was recorded from the north-western Cape which means that the species can still be considered as part of the Cape bat fauna. Because this study concentrated on the roosts of communal species insufficient data were gathered on the solitary species to determine their true conservation status. Of the 16 species listed as rare in Table 1, nine are nowhere common and may naturally occur at low densities. Most of them appear to have specialized habitat requirements and consequently have seldom been collected. This has resulted in the taxonomic confusion of many of the rare species. The authors are fairly confident that an intensive netting progamme, especially in the drier areas of the Karoo and Namaqualand, would show several of the “rare” species to be fairly common within a restricted habitat. It is difficult to say whether there has been any great change in distribution and status of the more solitary species in recent years, mainly because so little netting has been carried out. What is needed is an intensive netting programme in all of the localities where “rare” species have been collected in the past and then selective trapping for particular species in similar habitats where they are likely to occur. Since most of the more solitary species occur in low densities and in scattered roosts there seems to be little that can be done to conserve them at this stage. The major threats are insec- ticides and unnecessary killing by people. Therefore the priorities are an insecticide testing programme to monitor the effects of pesticides and public education to create a greater aware- ness of the value of bats in nature. The more gregarious bat species, which gather in large colonies, are in a very different pos- ition. Not only is it known that in some places, such as the Kalk Bay Caves on the Cape Penin- sula and the Cango Caves near Oudtshoorn, large colonies have almost disappeared but it is possible to take active steps to prevent further damage in other areas where large colonies are still to be found. 114 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Table 1 Summary of specimens and known conservation status of bats in the Cape Province Species Survey Specimens Specimens in S.A. museums Conservation status Epomophorus wahlbergi 1 142 FAIRLY COMMON in the S.E. and E. Cape. Epomophorus crypturus 1 RARE, marginal in the E. Cape coastal belt but common elsewhere in Africa. Eidolon helvum 1 5 RARE migrant in N. Cape but com- mon elsewhere in Africa. Rousettus aegyptiacus 20 35 COMMON in S.W., S. and S.E. coastal regions. Taphozous mauritianus — 2 RARE in S. and E. Cape but wide- spread in Africa. (Nycteris hispida) — — NOT PRESENT — incorrect identi- fication. Nycteris thebaica 43 121 FAIRLY COMMON throughout. Rhinolophus fumigatus — — VERY RARE, marginal in N.W. Cape but widespread in Africa. Rhinolophus darlingi 4 30+ UNCOMMON in N. Cape, but probable confusion with next two species. Rhinolophus clivosus 133 77 COMMON in coastal areas and N. Cape. Rhinolophus capensis 157 120 FAIRLY COMMON in coastal areas, possibly endemic to Cape Province. Rhinolophus denti VERY RARE, old records only for the Cape, restricted distribution in southern Africa. Possibly conspe- cific with next species. Rhinolophus swinnyi 20 RARE, old records only for the Cape, restricted distribution in southern Africa. Hipposideros caffer 2 26 RARE, marginal in N.W. & E. Cape but more common elsewhere in Africa. My Otis seabrai 5 8 VERY RARE in N.W. Cape, re- stricted distribution on W. coast of southern Africa. Possibly conspe- cific with next species. 115 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Species Survey Specimens Specimens in S.A. museums Conservation status Myotis lesueuri 2 3 VERY RARE ENDEMIC, scattered records from Karoo, N. & S.W. Cape. Myotis tricolor 44 16 UNCOMMON in S.W., S. & E. Cape but widespread up E. coast of Africa. Pipistrellus nanus 4 RARE, marginal in E. Cape, old re- cords only but common elsewhere in Africa. Pipistrellus kuhlii 3 RARE, marginal in E. Cape, old re- cords only but widespread in Africa. Eptesicus hottentotus 6 7 RARE in W. & N.E. Cape, restric- ted distribution in southern Africa. Eptesicus melckorum 30 9 FAIRLY COMMON in Karoo, S.W. & W. Cape, otherwise restricted distribution in Angola & Zambia. Eptesicus capensis 30 122 COMMON throughout. (Eptesicus notius) — — INVALID SPECIES. Laephotis wintoni 1 VERY RARE, only one specimen from S.W. Cape. Taxonomy un- clear. Scotophilus dinganii 1 26 RARE, marginal in E. Cape, histori- cal records only, but widespread in Africa. Kerivoula lanosa 1 6 VERY RARE in S. & E. Cape, re- stricted distribution in southern Africa, but taxonomy needs atten- tion. Miniopterus fraterculus 8 UNCOMMON in S. & E. Cape, although confusion with next species; taxonomy needs attention. Miniopterus schreibersii 470 159 VERY COMMON throughout, apart from central Karoo. Sauromys petrophilus — 8 RARE in W. Cape and restricted in southern Africa. (Tadirada condylura) — — Probably NOT PRESENT, marginal in Transkei. Tadarida pumila 15 13 UNCOMMON in W. Cape but only recently discovered and little known, widespread in Africa. Tadarida aegyptiaca 53 61 Common throughout. 116 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Table 2: The most important bat roosts known in the Cape Province, with the estimated maximum number of bats seen (seasonally numbers may be lower due to migration) Roost Max. No. of bats (1) De Hoop Cave 100 000 Miniopterus (Bredasdorp) 12 000 Rhinolophus 2 000 Myotis 100 Nycteris (2) Koegelbeen Cave 60 000 Miniopterus (Griquatown) 5 000 Rhinolophus (3) Maitland Mines 8 000 Miniopterus (Port Elizabeth) 1 500 Myotis 200 Rhinolophus (4) Die Hel Cave 4 000 Miniopterus (Groot Winterhoek) 2 000 Rousettus 1 500 Rhinolophus 100 Myotis (5) Bloukrans Cave 4 000 Miniopterus (Pearston) 1 000 Rhinolophus (6) Montagu Cave 5 000 Miniopterus (Montagu) (7) Bean-se-bos 1 500 Miniopterus (Hankey) 1 000 Rousettus 500 Rhinolophus (8) Welbedacht Mine 3 000 Miniopterus (9) Droevlakte Cave 1 500 Miniopterus 1 000 Rhinolophus (10) Sterkstroom Mine 1 000 Miniopterus 1 000 Rhinolophus (11) Storms River Mouth 2 000 Rousettus (12) Forest Ranch 1 000 Nycteris 500 Rhinolophus 117 ANN. CAPE PROV. MUS, (NAT. HIST.) VOL. 16, PT. 4. JANUARY 1985 Roost Max. No. of bats (13) Grootplaas Cave 1 000 Rhinolophus (14) Helderberg Cave 500 Rousettus (15) Wellington Church 300 Tadarida Table 2 gives the approximate numbers of individuals found in the larger colonies visited during the survey (see also Fig. 2). Some of these are nursery caves which play a very import- ant role in the survival of all species. For example, in summer the De Hoop Cave attracts about 80 000 pregnant Miniopterus schreibersii females from within a radius of up to 250 km. Disturbance at this critical stage can lead to a high mortality of offspring and the effective pro- tection of such roosts is therefore of the utmost importance. It is clear that the large numbers of bats present in these caves must have a marked effect on the insect fauna of the surrounding areas and must therefore play an important role in the ecosystem. Although it is known that certain roosts have disappeared due to disturbance it is difficult to tell at this stage whether the major roosts are maintaining their numbers or not. It is important therefore that efforts to protect the roosts be accompanied by a low intensity, low disturbance monitoring programme to determine whether conservation measures are being ef- fective or not. The conservation of bats is dependent on a clear understanding of the legislation govern- ing their protection. The Nature Conservation Ordinance (No. 19 of 1974) of the Cape Province lists three cat- egories of protection for all indigenous wild animals. These are ‘Endangered Wild Animals’ (Schedule 1 of the Ordinance), ‘Protected Wild Animals’ (Schedule 2), and ‘Wild Animals’ for all other indigenous species. All bats are therefore ‘Wild Animals’ and all insectivorous bats (Microchiroptera) are classified as ‘Protected Wild Anmimals’. The reason for the exclusion of fruit bats (Megachiroptera) from Schedule 2 is that they are considered to be potential prob- lem animals because of the damage that they sometimes cause to fruit crops. There are several restrictions on how all wild animals may be killed or hunted. As ‘Wild Animals’ no bats may be hunted using prohibited methods (Section 29 of the Ordinance) un- less the ‘hunter’ is the holder of a permit authorizing him so to do. This means that he may not use poison, artificial lights or any type of trap (including nets), and he may not hunt at night. Written permission from the owner of the land is also obligatory. Further restrictions (Section 27) apply to the insectivorous bats (as ‘Protected Wild Ani- mals’) in that they may not be hunted, killed or captured without a permit. This includes bats caught or killed for scientific research. In addition to the above regulations on ‘hunting’, no ‘Wild Animals’ may be kept in cap- tivity without a permit (Section 31), and no ‘Wild Animals’ may be exported from, imported into or transported within the province without a permit (Section 44). These regulations are far from adequately adhered to or enforced at present, and many bats are still killed illegally and unnecessarily. CONSERVATION OF BATS IN THE CAPE PROVINCE 118 HERSELMAN. NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C P. Conservation management For most of the solitary bat species little can be done at present for their protection. There is even little hope of enforcing the existing legislation if people are antagonistic towards them. A public awareness campaign is therefore the only long-term solution. Such a campaign should aim to persuade the public to accept the ecological and economic importance of bats and to discourage all forms of disturbance. Initially the most effective way to reduce antagonism to- wards bats in general is to give good advice on how to avoid problems with roof-dwelling bats and to quell rumours that bats are major carriers of rabies. For the colonial species considerably more can be done. The effective protection of all large colonial roosts should be given highest priority in bat conservation. This involves both the enforcement of the regulations on the killing of bats and the reduction of disturbance dur- ing critical periods. Legislation Whereas legislation to control the killing of bats is adequate at this stage, consideration should be given to including the fruit bats (Megachiroptera) in Schedule 2 as ‘Protected Wild Animals’. This would mean that a permit would have to be obtained under Section 47 of the Ordinance before fruit bats could be destroyed or captured. This should not prevent farmers from protecting their crops as permits should be freely issued where damage is being done. However, it would emphasize to farmers that fruit bats are an important part of natural ecosystems and that indiscriminate killing is not supported. The main advantage of the permit system is that it makes it possible to monitor the extent and severity of the problem in the province and thereby to put into a clearer perspective the often exaggerated claims of damage done by fruit bats. A similar system is already in operation for small antelope such as grysbok and duikers which occasionally cause similar localized and seasonal problems. Research There are two main priorities for further research on bats in the Cape Province. The first is that work should be done on the distribution and taxonomy of the more solitary species oc- curring in the province. This would involve an extensive netting programme, especially in the Karoo and northern parts of the province. It should aim at concentrating on particular species, at following up old records of the occurrence of a species at sites at which collecting has not re- cently taken place, and also at looking in habitats known to be suitable for scarce species. The second research priority should be to determine whether bats in various areas are be- ing markedly affected by pesticides. This would involve chemical analysis of pesticide residues in small samples of bats from areas in which they are most likely to be affected. If significant residues are found, this programme could be expanded to form a full scale research and man- agement project. If no significant residues are found, the information would serve as useful base line data for a low intensity long-term monitoring programme to detect if the use of pesti- cides becomes of importance to bats in the future. Monitoring programme A very important aspect of bat conservation that should receive attention is the establish- ment of a long-term monitoring programme for as many of the larger roosts as possible. The information collected should be a total count that is as precise and repeatable as possible to de- tect changes in the numbers of bats in the different roosts. A maximum intensity of two visits to each roost per year, in summer and in winter, is envisaged. 119 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16. PT. 4. JANUARY IQR.'l ACKNOWLEDGEMENTS The authors thank the Director of Nature and Environmental Conservation under whose auspices the survey was carried out; Drs N. Fairall, I. G. Gaigher and C. J. Loedolff for guid- ance during the study; D. J. Hanekom and R. Pool for technical assistance; Prof. J. Meester for identifying specimens, Mr J. E. Hill and, especially. Dr I. L. Rautenbach for help with many queries on taxonomy and identification; J. C. Greig for criticizing the manuscript; Dr R. H. N. Smithers for information on bat distribution; all our colleagues who assisted with the study, whether by writing articles for newspapers, through public relations work, by collecting information and specimens, or by discussing of ideas; many other individuals and institutions for information regarding roosts; the directors and staff of the Kaffrarian Museum, Transvaal Museum, South African Museum and Albany Museum for access to their collections. 120 HERSELMAN. NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. REFERENCES Constantine, D. G. 1958. An automatic bat-collecting device. J. Wildl. Mgmt 22: 17-22. Davis, R. B., Herreid, C. F. and Short, H. L. 1962. Mexican free-tailed bats in Texas. Ecol. Monogr. 32: 311-346. Davis, W. H. and Hitchcock, H. B. 1965. Biology and Migration of the Bat, Myotus lucifugus, in New Eng- land. J. Mammal. 46: 296-313. Dwyer, P, D. 1964. Seasonal changes in activity and weight of Miniopterus schreibersi blepotis (Chiroptera) in north-eastern New South Wales. Austr. J. Zool. 12: 52-69. Ellerman, j. R., Morrison-Scott, T. C. S. and Hayman, R. W. 1953. Southern African mammals 1758 to 1951: A reclassification. London: British Museum (Natural History). Erasmus, B. H. and Rautenbach, I. L. (In press). New records of occurrence for six species of small mam- mals in the northern Cape Province. 5. Afr. J. Wildl. Res. Harrison, D. L. and Clancey, P. A. 1952. Notes on the bats from a cave in the Pietermaritzburg District of Natal. Ann. Natal Mus. 12: 177-182. Hayman, R. W. and Hill, J. E. 1981. Order Chiroptera. In: Meester, J. and Setzer, N. W. eds. The mam- mals of Africa: an identification manual. Washington D.C.: Smithsonian Institution Press, pp. 1-73. Hill, J, E, 1974. A review of Laephotis Thomas, 1901 (Chiroptera: Vespertilionidae) Bull. Br. Mus. Nat. Hist. Zool. 27(2): 73-82. Kingdon, j. 1974. East African mammals, an atlas of evolution in Africa. Vol. II part A. New York: Academ- ic Press. Koopman, K. F. 1975. Bats of the Sudan. Bull. Amer. Mus. nat. Hist. 154: 353-443. Meester, J., Davis, D. H. S. and Coetzee, C. G. 1964. An interim classification of southern African mam- mals. Pretoria: Zoological Society of Southern Africa and C.S.l.R. Norton, P. M. and van der Merwe, M. 1978. Winter activity of bats in a Transvaal highveld cave. S. Afr. J. Sci.74: 216-220. Rautenbach, I. L. 1982. Mammals of the Transvaal. Pretoria: Ecoplan. Rautenbach, I. L. and Nel, J. A. J. 1978. Three species of Microchiropteran bats recorded for the first time from the South-West Cape Biotic Zone. Ann. Transv. Mus. 31: 157-163. Robbins, C. B. 1978. Taxanomic identification and history of Scotophilus Nigrita (Schreber) (Chiroptera: Ves- pertilionidae). J. Mammal. 59: 212-213. Roberts, A. 1951. The Mammals of South Africa. Johannesburg: Trustees of the Mammals of South Africa Book Fund. Shortridge, G. C. 1942. Field notes on the first and second expeditions of the Cape Museum’s mammal sur- vey of the Cape Province; and descriptions of some new subgenera and subspecies. Ann. S. Afr. Mus. 36: 27-100. Smitkers, R. H. N. 1971. The mammals of Botswana. Salisbury; Trustees of the National Museum of Rhodesia. Smithers, R. H. N. 1983. The mammals of the Southern African Subregion . Pretoria; University of Pretoria. Stebbings, R. E. 1970. Bats in danger. Oryx 10(5): 31 1-312. Swanepoel, P., Smithers, R. H. N. and Rautenbach, 1. L. 1980. A checklist and numbering system of the extant mammals of the Southern African Subregion. Ann. Transv. Mus. 32(7): 155-1%. 1 uttle, M. C. 1973. An improved trap for bats. J. Mammal. 55: 475-477. Van der Merwe, M. 1973a. Aspects of social behaviour of the Natal clinging bat Miniopterus schreibersi natalensis (A. Smith, 1834). Mammalia 37: 379-389. Van der Merwe, M. 1973b. Aspects of temperature and humidity in preferred hibernation sites of the Natal clinging bat Miniopterus schreibersi natalensis (A. Smith, 1834). Zool. afr. 8: 121-134. Van der Merwe, M. 1973c. Aspects of hibernation and winter activity of the Natal clinging bat, Miniopterus schreibersi natalensis (A. Smith, 1834), in the Transvaal Highveld. S. Afr. J. Sci. 69: 116-118. 121 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Appendix: Gazetteer of localities mentioned in the text. For survey localities the co-ordinates are given where possible. Locality Abbotsbury Abraham Kriel Abrie Algeria Alice Alicedale Alice Mission Aliwal North Amabele Amalinda Apies River Forestry Station Aronegas Assegaaibosch Atherstone Augrabies Falls Aurora Bakleisdrif Balossi Barkley East Barkly West Bat’s Cave Bean-se-bos Beaufort-Wes Bedford Berg River Blaney Bloukrans Blouputs Blythswood Bontebokpark Boomslang Cave Bot River Brakkloof Brandkaros Brandvlei Breakfast Vlei Buffelskloof Burgersdorp Camps Bay Canaga Cango Caves Locus Co-ordinates 3224 BA 2824 DA 28°31'S 24°31'E 3322 AC 33° 22 'S 22° 08'E 3219 AC 32° 22'S 19° 03 'E 3226 DD 3326 AC 3226 DD 32° 46'S 26° 50'E 3026 DA 3227 DA 3227 DD 32°59'S 27°51'E 3325 CB 33° 43 'S 25° 19'E 3017 BA 30° 05 'S 17° 42 'E 3318 DD 3326 AD 2820 CB 3218 CB 32° 38'S 18° 28'E 3420 BA 34° 05 'S 20° 30' E 3129 BA 3027 DC 30°58'S 27° 35 'E 2824 BC 28° 30'S 24° 33'E 3327 BB 33° 05 'S 27° 56'E 3324 DB 33° 42'S 24° 49 'E 2923 BA 3226 CA 3320 CC 3327 DC 3225 CC 32° 46'S 25° 14'E 2820 CA 28° 30'S 20° 08'E 3227 BB 3420 AB 34° 04'S 20° 27'E 3418 AB 34° 07'S 18° 26'E 3419 AA 34° 13'S 19° 12'E 3326 BA 2816 BC 28° 29'S 16°41'E 3120 BB 31°00'S 20°51'E 3326 BB 33° 05 'S 26° 46'E 3326 AC 3026 CD 3318 CD 3324 CC 33° 49'S 24°01'E 3322 AC 33° 24'S 22° 12'E 122 HERSELMAN, NORTON: DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P. Locality Locus Co-ordinates Cape Infanta 3420 BD 34° 28'S 20°51'E Cape Town 3318 CD Citrusdal 3219 CA Clanwilliam 3218 BB Compagnies Drift 3218 AB 32° 07 'S 18° 27'E Cullendale (Bedford) 3226 CA De Hoop 3420 AD 34° 25 'S 20°21'E Dennebos 3322 DC 33° 59'S 22° 43 'E Despatch 3325 CD Die Hel 3319 AA 33° 05 'S 19° 05 'E Die Kelders 3419 CB 34° 33'S 19°21'E Die Oog (Kuruman) 2723 AD 27° 27'S 23° 26'E Doringbaai 3118 CA 31°46'S 18° 13'E Doringkraal 3321 CA 33° 45 'S 21° ll'E Doringpoort River 2816 DB 28° 32 'S 16° 55'E Driefontein 3118 DC 31°51'S 18° 36'E Droevlakte 3421 AD 34° 24'S 21°25'E Dunbrody 3325 BC East London 3327 BB Ebb-and-flow 3322 DC 33° 59'S 22° 36'E Elliot 3127 BD 31°20'S 27°51'E Elsies River 3318 DC Ezelfontein 3018 AC Forest Ranch 3227 AB 32° 09 'S 27° 18'E Fort Beaufort 3226 DC Fort Brown 3326 BA Gant’s Mine 3325 CD 33°58'S 25° 17'E Garies 3018 CA George 3322 CD Gladstone 3227 CB Glass Nevie 3027 DD 30° 50'S 27° 52'E Gleniffer 3326 BC Glenlea (Highlands) 3326 AD Goegab 2918 CA 29°41'S 18°01'E Goodhouse 2818 CD Gordon’s Bay 3418 BB 34° ll'S 18° 49'E Goukamma 3422 BB 34° 03 'S 22° 56'E Graaff-Reinet 3224 BA 32° 15'S 24° 34' E Grahamstown 3326 BC 33° 38'S 26°31'E Grootkraal Cave 3322 AC 33° 24'S 22° 12'E Grootplaas Cave 3321 DC 33° 53'S 21°44'E Grootvadersbos 3320 DD Hanover 3124 AB Hansies River Mine 3419 BD 34° 23 'S 19° 38'E 123 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 4, JANUARY 1985 Locality Locus Co-ordinates Hawston 3419 AD Heidelberg 3420 BB Helderberg 3418 BB 34° 02'S 18° 52'E Herbertshope 3326 CC Herbertsdale 3421 BB 34°01'S 21°45'E Hester Malan 2917 DB 29° 39'S 17° 59'E Het Kruis 3218 DA Hex River 3319 DA Hex River Estate 3218 BD Hondeklipbaai 3017 AD Hot Pot 3420 AD 34° 27 'S 20° 26'E Jameston (Highlands) 3326 AD Jonkershoek 3318 DD 33° 58'S 18° 55'E Junction Farm 3226 DD 32° 54' S 26° 56'E Kaggasmoudt 3226 CA Kamferboom 2620BA Keikamspoort 2922 DD Kei Road 3227 DA Kersbos 3018 CC 30° 59'S 18° 14'E Keurbooms Reserve 3323 CD 33° 59'S 23° 24'E Keurbosfontein 3219 AD 32° 28'S 19° 17'E Kimberley 2824 DB King William’s Town 3227 CD Klawer 3118 DC Kleinpoort 3326 BB Klipfontein (Albertinia) 3421 BA 34° 04'S 21° 33'E Knysna 3423 AA Koegelbeen Cave 2823 CB 28° 39'S 23° 20'E Koo Cave 3319 DA 33° 39'S 19° 47'E Krakeel River 3323 CD 33° 49'S 23° 44'E Kroomie 3226 CD Kuruman 2723 AD Lady Grey 3027 CA Lambert’s Bay 3218 AB Lessendrum (Peddie) 3327 AA Lourens Farm 3318 DB Maitland Mines 3325 CD 33° 59'S 25° 17'E Manley flats 3326 BC Marcus Mud Crawl 3322 AC 33° 23 'S 22° 14'E Mazelsfontein 2923 BA 29°01'S 23° 34'E McGregor 3319 DD Michell’s Pass 3319 AD Middelburg 3125 AC 31° 28'S 25°01'E Middleton 3326 AB 124 HERSELMAN, NORTON; DIST. AND STATUS OF BATS (MAMMALIA: CHIROPTERA) IN THE C.P Locality Locus Co-ordinates Millwood Mines 3323 CC 33°53'S 23° 00' E Moneysworth 3326 DA Montagu 3320 CC Montagu Cave 3320 CC 33°50'S 20° lO'E Mqanduli district 3128 DD Nabeeb Mine (Nuwerus) 3118 BA 31° 41'S 18° 37'E Norap 2917 DD Norlim Mine (Taung) 2724 DA 27° 37'S 24° 37'E Olifantskrans Tunnel 3320 CC 33° 54'S 20°0TE Okiep 2917 DB 29° 32'S 17° 52'E Orrelgat 3018 AB 30° 14'S 18° 26'E Oudtshoorn 3322 CA Paddagat 2917 BD ^atensie 3324 DD Peddie 3327 AA Pirie 3227 CB P.K. le Roux Dam 3024 BA Plettenbergbaai 3423 AB Port Alfred 3326 DB Port Elizabeth 3325 DC Port St Johns 3129 DA Puntjie Cave 3421 AC 34° 22'S 21° 30'E Rangerton 3227 CD Redhouse 3325 DC Riebeeck-Oos 3326 AA Riversdale 3421 AB 34° 05 'S 21° 15'E Rocherpan 3218 CB Robertson 3319 DD 33° 48'S 19° 53'E Rolfontein 3024 BD Roodebergskloof 3018 CA 30° 30'S 18° 05'E Ross Mission 3227 CB Saasveld 3322 DC Saldanha Bay 3317 BB Salem 3326 AD 33° 15'S 26° 25 'E Sandberg 3218 BC 32° 17'S 18° 35'E Sandberg Cave 3218 BC 32° 20'S 18° 35'E Sandile’s Cave 3227 CB 32° 43 'S 27° 17'E Schoenmakerskop 3325 DC Sewefontein 3119 CA Simonstown 3418 AB Simon van der Stel Mine 2917 DB 29°38'S 17° 55'E Skeleton Cave 3322 AC 33° 22'S 22° lO'E Skrik van Rondom 3218 DA 32° 34'S 18° 40'E Skurweberg Cave 3319 AD 33° 25'S 19° 22'E 125 ANN. CAPE PROV. MUS. (NAT, HIST.) VOL. 16, PT. 4, JANUARY 1985 Locality Locus Co-ordinates Slaaikraal 3326 AD Spies Cave 3322 AD 33° 24'S 22°21'E Spitskop 3420 AD 34° 19'S 20° 23 'E Stellenbosch 3318 DD Sterkstroom 3126 DA 31°33'S 26° 30'E Steyerskraal 2819 CB Stilbaai 3421 AD Stones Hill 3326 BC Storms River Mouth 3423 BB 34°01'S 23° 54'E Strand 3418 BB 34° 06'S 18° 49'E Swellendam 3420 AB Table Mountain 3418 AB Tambookies Drift 3226 DA Taung 2724 DA 27° 37 'S 24° 37'E Three Sisters 3123 CA 31°35'S 23° 05 'E Traveller’s Rest 3219 AA Twee Riviere 3323 DD 33° 50'S 23° 53'E Tunnel Cave 3422 AA 34° 12'S 22° 07'E Tylden 3227AA Umtata 3128 DB Upington 2821 AD Van Wyksvlei 3021 BD Victoria West 3123 AC Vrolijkheid 3319 DD 33° 55'S 19° 52'E Vryburg 2624 DC Walmer 3325 DC 33° 58'S 25° 35'E Waterloo (Grahamstown) 3326 BC Wellington 3319 BA 33° 38'S 19° OO'E Welbedacht Mine 2917 DB 29°41'S 17° 50'E Wilgerfontein (Pearston) 3225 CB 32° 36'S 25° 19'E Windsor Cave 3421 AB 34° 12'S 21° 18'E Wondergat 2816 BD Wynberg 3318 CD 33° 59'S 18° 24'E Ysterfontein 3318 AC 33° 20'S 18° 09'E 126 i i { / INSTRUCTIONS TO AUTHORS LAYOUT of recent issues should be examined for details of acceptable style and format. MANUSCRIPTS should be submitted in triplicate, typed on A4 size paper, double- spaced, on one side of the paper only and with wide margins all round. Only generic and specific names should be underlined to indicate italics. All other mark- ing up should be left to the editor. Footnotes should be avoided as far as possible. All measurements should be metric. Systematic papers should follow the rules of the International code of zoological nomen- clature or the International code of botanical nomenclature , as applicable. ILLUSTRATIONS should bear the author’s name, title of paper and number in soft pencil on the reverse side. Diagrams, line drawings and photographs are all regarded as figures, should be termed as such and should be numbered in a single series. In addition to the original illustrations a photocopy of each should be submitted. Diagrams and line drawings should be reducible to a size not exceeding 14 x 16 cm (17,5 cm including caption). Photographs for half-tone blocks should be final size, of high contrast and printed on glossy paper. Colour illustrations are only printed under special circumstances and their financing must be arranged by the author. A scale should appear with all illustrations or the magnification should be given in the caption. A list of captions should be typed on a separate sheet. TABLES should be typed on separate sheets. Foldouts and tables printed sideways should be avoided. REFERENCES should be listed alphabetically at the end of the paper. Abbreviations of periodical titles should conform with those in the World list of scientific periodicals (4th ed.). Layout must be as follows: Alexander, J. 1969. The indirect evidence for domestication. In: Ucko, P. J. and G. W. Dimbleby eds. The domestication and exploitation of plants and animals. London: Duckworth, pp. 123-129. De Winter, B. 1969. Plant taxonomy today. S.Afr.J.Sci. 65 (8): 229-242. JuBB, R. A. 1967. Freshwater fishes of southern Africa. Cape Town: Balkema. GALLEY PROOFS are sent (when practicable) to authors for checking printer’s errors. Adjustments of language, style and punctuation and any other amendments of the original manuscript, if made at this stage, may be charged to authors. AUTHOR’S COPIES (50) are issued free of charge. When more than one author is involved the fifty free copies will be sent to the senior author. Additional copies ordered before printing are supplied at a reduced cost. BD3000 — CTP Book Printers, Cape Annals of the Cape Provincial Museums Natural History Ann. Cape Prov. Mus. (nat. Hist.) Volume 16 Part 5 28th February 1986 Published jointly by the Cape Provincial Museums at the Albany Museum, Grahamstown, South Africa ANNALS OF THE CAPE PROVINCIAL MUSEUMS These Annals are published in two series, Natural History and Human Sciences, the latter series covering cultural history, ethnology, anthropology and archaeology. They are issued in parts at irregular intervals as material becomes available. The primary objective of these Annals is to disseminate the results of research work undertaken by staff members of the Cape Provincial Museums. However, a manuscript by any author is eligible for publication provided it is based wholly or partly on the collection/s of one or more of these museums or if the material on which the work is based is wholly or partly deposited in one of these museums. Parts are obtainable from the Librarian of the Albany Museum. Correspondence in con- nection with these Annals should be addressed to the Editor, Albany Museum, Grahams- town 6140. Editor DrF.W. GESS: 1978- A new species of the genus Strongylopiis Tschudi from Namaqualand, Cape Province, South Africa (Anura: Ranidae) by ALAN CHANNING {Biochemistry Department, University of the Western Cape, Private Bag X17, Bellville 7530, South Africa) ABSTRACT Strongylopiis springbokensis sp. nov. is described from Namaqualand. Adults resemble Strongylopiis grayii, but can be distinguished on the basis of body proportions and the posses- sion of a unique male call. The advertisement call, male release call and male aggression call are illustrated. Eggs are laid on land where development proceeds to the pretadpole stage without free water. The tadpole is described and the known distribution of this species is pre- sented. INTRODUCTION Six species of the ranid genus Strongylopiis Tschudi, 1838 are presently recognized (Chan- ning, 1981). The genus is believed to have had a southern origin (Channing, 1981) and to have subsequently migrated northwards. The species comprising this genus occupy a range of habi- tats, from natural forest through grasslands to semi-desert. S. grayii (Smith) is found from the Cape to the Limpopo. It is a pioneering species, fre- quently found in man-made habitats like dams and ditches, even breeding successfully in refuse pits. S. fasciatiis (FitzSimons) is a widespread grassland form found from the south eastern Cape into Malawi, Zimbabwe and eastern Zambia. S. bonaespei is the replacement name for S. montamis (FitzSimons), a form found along the mountains of the south western Cape and the Cape Peninsula. S. hymenopiis (Boulenger) is found on the upper slopes of the Drakens- berg escarpment. S. wageri (Wager) is a forest form collected from the uplands of Natal and the Drakensberg escarpment as well as from near Sabie. During a recent trip to Namaqualand as part of a tadpole survey, the author heard a strange frog call at Springbok and later on the farm Bovlei on top of the Kamies mountains near Leliefontein. Calling males were collected and found to superficially resemble Strongylo- piis grayii. However, morphologically they can readily be distinguished from S. grayii. On the basis of a unique advertisement call and differences in morphology this frog is here described as a new species. 127 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 5, FEBRUARY 1986 DESCRIPTION Strongylopus springbokensis sp. nov. Holotype: a male in the Port Elizabeth Museum collection, PEM A963, collected 10 July 1984 by R. C. Drewes and A. Channing at Springbok in Namaqualand (2917DC). The type locality is a small dam on the southern edge of the town. Paratypes: Nine adults (1 female, 8 males) collected at the same time and place as the holo- type, housed in the Port Elizabeth Museum (PEM A964-A968) and the California Academy of Sciences (CAS 157517-157520). Other material examined; One adult male from the farm Bovlei near Leliefontein, the high- est point (alt: 1 530 m) on the Kamies mountains (CAS 157521), and two specimens from 11 km south of Kamieskroon (PEM A 1274-5). For comparative purposes, 41 specimens of 5. grayii in the Cape Department of Nature Conservation collection, and a series of 9 S. grayii from Kuboos, from the Transvaal Museum collection, were examined. (Localities of these S. grayii are shown later). Measurements of 63 S. grayii from Natal (Channing, 1979) were compared to the Cape material examined in this study and found to be indistinguishable. The holotype (Fig. 1) is an adult male with a pale yellowish-brown dorsum in life, with ir- regular brown markings edged with darker brown. Dorsal surface of the legs barred. A narrow pale vertebral stripe is present in some specimens in contrast to the broad vertebral stripe com- mon in the S. grayii material examined from the western Cape. Venter immaculate. In preservative the colours fade but the pattern remains. The head is relatively wider and the leg shorter than S. grayii. The ratios of head width/tibia and foot/head width (Fig. 2) illustrate this diagnostic difference. The snout is sharp in profile compared with that of S. grayii (Fig. 3). Males in breeding condition have small sharp asperities along the posterior surfaces of the legs, and marginated fingers giving especially the first finger a paddle-like appearance. Four phalanges of fourth toe free of web. DISTRIBUTION The localities of specimens examined and identified as either S. grayii or S. springbokensis are presented in Fig. 4. The northernmost record for S. grayii is a series of animals collected at Kuboos (2816BD) in 1933. This may represent a record of a transient population which moved up the drainage lines from the Orange River during a wet period. Further collecting in the Richtersveld during the rainy season is required. Specimens of S. grayii collected at Wuppertal near Clanwilliam (3218BB) are morphologically intermediate between S. grayii and S. spring- bokensis in terms of head width and leg length. The advertisement call of the Wuppertal speci- mens, however, is the typical “click” of S. grayii. S. springbokensis appears to be confined to the mountainous areas of Namaqualand, north of the Knersvlakte and south of the Orange River. VOCALIZATIONS Advertisement call The male advertisement call (Fig. 5A) consists of a series of two to seven or more notes. Each note has a duration of 0,11 s with a pulse rate of 185/s and emphasized frequencies of 0,8; 128 Measurements of the holotype and paratypes of Strongylopus springbokensis . Head length is measured from the angle of the jaw to the tip of the snout; head height is measured at the level of the tympanum. Eye and tympanum are horizontal distances. The foot includes the metatarsal tubercle and the fourth toe, while the hand includes the metacarpal tubercle. CHANNING: A NEW SPECIES OF GENUS STRONGYLOPUS TSCHUDI (ANURA; RANIDAE) o C/D o "S 2 JZ oi) G .£? 'C s: E D C G C o •n TO TO CL c3 •o "S u Sex o c C/D c3 05 K G 05 K G 05 X Eye E H s H Foo Toe £ D c G X 129 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 5, FEBRUARY 1986 Fig. I. Holotype oi SiiongylopHs springhokensis, PEM A963. 2,6 and 3,6 kHz. Variously emphasized harmonics are visible on the sonagram from 0,8 to 8 kHz. Call duration varies from 0,22 s (two notes) to 1,17 s (seven noles). The calls described were recorded in the field at Bovlei and Springbok with a Uher CR210 cassette recorder. Male aggression call This call (Fig. 5B) was recorded in the laboratory. It is produced when one calling male approaches another calling male. The aggression call consists of a single sharp note with an ex- plosive burst of energy from 0,4 kHz to above 8 kHz, with the highest amplitude between 0,45 and 1,8 kHz. Male release call The release call was recorded in the laboratory. The clasping male continues to utter the advertisement call, while the clasped male produces a long series of squeaking calls (Fig. 5C). Duration up to 3 s and probably more. Emphasized frequencies at 0,8 and 1,2 kHz. OVIPOSITION Eggs were first found while locating calling males in a dry stream bed below a dam wall on the farm Bovlei. The initial discovery of a disused rodent burrow serving as a call site was ex- citing, however, this was soon overshadowed by the discovery of the few hundred eggs found to be lining the burrow walls and floor. Other oviposition sites were subsequently located by tracing calling males, all of which were found with eggs. One such site was between grass roots in a damp crack in a rock pile. This site, like the rodent burrow, was placed about 20 cm above the dry stream bed. The eggs were found during the first week in July, at the start of the winter rain. 130 CHANNING: A NEW SPECIES OF GENUS STRONGYLOPUS TSCHUDI (ANURA: RANIDAE) S I- 4 5 6 7 8 9 (head width / tibia)x10 1,5 1,8 2,1 2,4 2,7 3,0 foot /head width Fig. 2. Comparison of the ratios of head width/tibia and foot/head width. S = S. springbokensis (n = 14), G = S. grayii (n = 50). Range, mean and standard deviation. S. springbokensis Fig. 3. Comparison of the snout profiles of Strongylopus springbokensis and S. grayii to illustrate the sharper snout of S. springbokensis. 131 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 5, FEBRUARY 1986 Fig. 4. Localities where S. springbokensis (solid circles) and S. grayii (open circles) are known to occur, based on material examined in the present study. 132 CHANNING: A NEW SPECIES OF GENUS STRONGYLOPUS TSCHUDI (ANURA; RANIDAE) seconds Fig. 5. Vocalizations of S. springbokensis. A — The advertisement call. B — The male aggression call. C — The male re- lease call. Sonagrams produced with the wide hiter. A batch of eggs was found near a pool in Springbok in a grassy tunnel 15 cm above water level. Some of these eggs were placed in water immediately; these hatched within 30 minutes into free swimming tadpoles nearly 10 mm long. Part of the batch was kept for a week without water in the damp soil from the burrow where they were collected. When added to water they also hatched within 30 minutes. From another batch 25 eggs were placed in water and observed under a microscope. They had all reached stage 22 of Gosner (1960) before being placed in water at room temperature and hatched between 20 minutes and 3 hours afterwards. The eggs seem to develop until the eyes and tail are well formed and then to remain at that stage until they are hydrated. The independance of developing eggs from free water and the rapid hatching of well de- veloped tadpoles soon after water is available seems to be a successful adaptation to life in the arid areas of Namaqualand. TADPOLE This description is based on three specimens (length 15 mm; stage 26 of Gosner (I960)) reared from eggs collected at Bovlei. Mouth situated ventrally, not visible dorsally. Oral disc 0,85 width of head at level of disk. A single row of mental papillae. Suprarostrodont pigmented, serrated across entire margin. Infrarostrodont serrated and pigmented almost to its base. Keratodont formula 1,3 + 3/15-1,2 (Fig. 6). Nostrils sub-circular, rimmed. Opening dorsal, nasal passage only slightly visible dorsally. Width of nostril/internarial distance = 0,19. Rostronasal distance/orbitonasal distance = 0,94. Minute pale pineal spot present. Extra-ocular proportion = 1,5. (See Van Dijk, 1966 for definitions of terminology). Spiracle single, sinistral, visible ventrally but not dorsally, situated 0,65 posteriorly along trunk. Opening slightly constricted, subcircular, visible laterally. Vent dextral, reaching margin of ventral fin. Tail height subequal to trunk height. Tip rounded. Maximum height of fins occurs 0,2 posteriorly along tail. Tail length/length of head and trunk = 2,0. Height of caudal muscles at base of tail/height of trunk = 0,37. 133 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 5, FEBRUARY 1986 Dorsal pigmentation uniform brown. Ventrally unpigmented except for slight stippling around the gut coils. Tail muscles slightly pigmented dorsally (Fig. 7). Fins unpigmented, ex- cept for slight markings on the dorsal fin. Fig. 6. Moiithparts of a tadpole of Slrougvlopiis springhokensis. Fig. 7. The tadpole of Strongxlopus springhokensis . 134 CHANNING: A NEW SPECIES OF GENUS STRONGYLOPUS TSCHUDI ( ANURA: RANIDAE) DISCUSSION The genus Strongylopus is believed to have had an origin in the high mountains of south- ern Africa (Channing, 1981). The Drakensberg massif has been suggested as a moist refuge during interpluvials (Brain and Meester, 1965). The mountains of Namaqualand and the south- western Cape could have served as speciation centres further south. Heleophryne purcelli, H. regis, H. rosei, Capensibufo tradouwi, C. rosei, Breviceps montanus, Strongylopus bonae- spei, S. springbokensis, and Artliroleptella lightfooti are all endemic to the southern mountain ranges, supporting the hypothesis of a southern centre of speciation associated with high relief. S. gray a and S. springbokensis are similar in appearance and presumably closely related. They can be easily separated on the basis of the advertisement call, the long call of S. spring- bokensis being very different from the single click of S. grayii (Channing, 1979). Preserved and other silent material can usually be separated morphologically on the sharper snout, wider head and shorter leg of S. springbokensis. These proportions can be combined as ratios of head width/tibia and foot/head width to give reliable diagnostic characters. The morphological simi- larity of the Wuppertal S. grayii to S. springbokensis remains a puzzle which probably will be solved by a phylogenetic analysis of the group. The distribution of S. springbokensis will only be fully known when museum collections of material presently identified as S. grayii (or Rana grayiil) have been re-examined and the ad- vertisement calls of the populations checked in the field. Further fieldwork is required to estab- lish the geographical boundaries of the new species and the ecological interactions between it an other Strongylopus species, if any. ACKNOWLEDGEMENTS Mr Hannes Mostert kindly permitted access to his farm Bovlei where males were first found calling from oviposition sites. E. Baard and A. de Villiers lent material from the Cape Department of Nature Conservation collection. W. Flaacke sent specimens of S. grayii from the Richtersveld for comparison and W. Branch collected material from Eksteenfontein. W. E. Duellman and J. Frost permitted the author to use a Sonagraph under their control while he was visiting the Museum of Natural History at the University of Kansas. All are thanked for their assistance. Bob Drewes is especially acknowledged for providing an opportunity to visit the type locality and for his enthusiastic help in the field. REFERENCES Brain, C. K. and J. Meester. 1963. Past climatic changes as biological isolating mechanisms in southern Africa. Monogr. Biol. 14: 332-340. Channing, A. 1979. Ecological and systematic relationships of Rana and Strongylopus in southern Africa (Amphibia; Anura). Ann. Natal Mas. 23: 797-831. Channing, A. 1981. Southern origin of the African genus Strongylopus Tschudi, 1938 (Amphibia Ranidae). Monit. zool. ital. (N.S.j Suppl. 15: 333-336. Gosner, K. L. 1960. A simplihed table for staging anuran embryos and larvae with notes on identiheation. Herpetolug- ica 16: 183-190. Van Dijk, D. E. 1966. Systematic and field keys to the families, genera and described species of southern African anu- ran tadpoles. Ann. Natal Mus. 18: 231-286. 135 1 f.v X 1 ■I; INSTRUCTIONS TO AUTHORS LAYOUT of recent issues should be examined for details of acceptable style and format. MANUSCRIPTS should be submitted in triplicate, typed on A4 size paper, double- spaced, on one side of the paper only and with wide margins all round. Only generic and specific names should be underlined to indicate italics. All other mark- ing up should be left to the editor. Footnotes should be avoided as far as possible. All measurements should be metric. Systematic papers should follow the rules of the International code of zoological nomen- clature or the International code of botanical nomenclature , as applicable. ILLUSTRATIONS should bear the author’s name, title of paper and number in soft pencil on the reverse side. Diagrams, line drawings and photographs are all regarded as figures, should be termed as such and should be numbered in a single series. In addition to the original illustrations a photocopy of each should be submitted. Diagrams and line drawings should be reducible to a size not exceeding 14 x 16 cm (17,5 cm including caption). Photographs for half-tone blocks should be final size, of high contrast and printed on glossy paper. Colour illustrations are only printed under special circumstances and their financing must be arranged by the author. A scale should appear with all illustrations or the magnification should be given in the caption. A list of captions should be typed on a separate sheet. TABLES should be typed on separate sheets. Foldouts and tables printed sideways should be avoided. REFERENCES should be listed alphabetically at the end of the paper. Abbreviations of periodical titles should conform with those in the World list of scientific periodicals (4th ed.). Layout must be as follows; Alexander, J. 1969. The indirect evidence for domestication. In: Ucko, P. J. and G. W. Dimbleby eds. The domestication and exploitation of plants and animals. London: Duckworth, pp. 123-129. De Winter, B. 1969. Plant taxonomy today. S.Afr.J. Sci. 65 (8): 229-242. JuBB, R. A. 1967. Fresh water fishes of southern Africa. Cape Town: Balkema. GALLEY PROOFS are sent (when practicable) to authors for checking printer’s errors. Adjustments of language, style and punctuation and any other amendments of the original manuscript, if made at this stage, may be charged to authors. AUTHOR’S COPIES (50) are issued free of charge. When more than one author is involved the fifty free copies will be sent to the senior author. Additional copies ordered before printing are supplied at a reduced cost. BD4769 — CTP Book Printers. Cape Annals of the Cape Provincial Museums Natural History Ann. Cape Prov. Mus. (nat. Hist.) Volume 16 Part 6 28th February 1986 Published jointly by the Cape Provincial Museums at the Albany Museum, Grahamstown, South Africa ANNALS OF THE CAPE PROVINCIAL MUSEUMS These Annals are published in two series, Natural History and Human Sciences, the latter se- ries covering cultural history, ethnology, anthropology and archaeology. They are issued in parts at irregular intervals as material becomes available. The primary objective of these Annals is to disseminate the results of research work under- taken by staff members of the Cape Provincial Museums. However, a manuscript by any author is eligible for publication provided it is based wholly or partly on the collection/s of one or more of these museums or if the material on which the work is based is wholly or partly deposited in one of these museums. Parts are obtainable from the Librarian of the Albany Museum. Correspondence in connection with these Annals should be addressed to the Editor, Albany Museum, Grahamstown 6140. Editor DrF.W. GESS: 1978- Three new species of southern African Bembix, a new synonymy, and biological notes on other species of the genus (Hymenoptera: Sphecidae: Nyssoninae) by F.W. GESS (Albany Musuem, Grahamstown) CONTENTS Page Abstract 137 Introduction 137 Descriptions and synonymy 138 Biological notes 148 Review and discussion of the nature of the prey of the genus Bembix with particular reference to Afrotropical species 155 Acknowledgements 159 References 160 ABSTRACT Three new species of southern African Bembix, B. namibensis sp. nov., B. zinni sp. nov. and B. karooensis sp. nov. are described, B. jimodi Arnold is sunk into synonymy with B. alula Arnold, and biological notes are given for ten other species of Bembix. The nature of the prey of the genus Bembix with particular reference to Afrotropical species is reviewed and discussed. INTRODUCTION An investigation was made of the Bembix holdings of the major collections of southern Africa with a view to undertaking a taxonomic revision of the genus in that region. Three un- described species and one synonymy were discovered. Some other taxonomic problems were identified but due to the inadequacy of the holdings of the relevant species and the consequent need for additional material from specific localities a taxonomic revision was considered to be premature. Consequently the new species and the synonymy only are dealt with in the present paper. Over the years various biological observations on Bembix species have been gathered to- 137 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 6, FEBRUARY 1986 gether. These are constituted from scattered published records, which in some instances are obscure and therefore generally overlooked, from specimen labels of Albany Museum ma- terial, and from personal observations. It is hoped that this represents as nearly as possible a complete compilation of the biological knowledge of Bembix in the Afrotropical Region and that it will as such form a foundation on which to build. DESCRIPTIONS AND SYNONYMY Bembix namibensis sp. nov. Male (Figs 1-8) Black; mandibles excepting their apices, labrum, clypeus, face up to level of anterior ocel- lus (excepting for a pair of elongate wedge-shaped spots above antennal sockets), narrow bands widening below margining posterior orbits, scapes and flagella excepting broad dorsal black streaks, posterior margin and sides of pronotum, narrow streaks on mesonotum margin- ing tegulae, anterior halves of tegulae themselves, bands margining scutellum and metanotum behind, anterior aspect of mesopleura and small marks on same behind pronotal lobes, meso- sternum anteriorly and also along midline and adjacent to coxae, streaks on metapleura, mark- ings of variable extent on sides of propodeum and including spiracular lobe, broad transverse bands on tergites 1-5, large central spot on tergite 6, visible part of gonostyles, sternites I and 2 at base and postero-laterial corners of sternites 1-5, legs (except black inner aspect of fore- femora and black streaks on foretibiae, inner aspect of femora and to a lesser extent of tibiae of middle and hind legs, comb on first tarsomere and distal extremities of all five tarsomeres of each fore-leg, claws and pulvilli of all legs), various shades of yellow. Markings on tergites (except on lateral margins) almost white, markings on head pale yel- low, those on thorax, legs, lateral margins of tergites and also the sternites darker yellow. Extreme tip of tergite 7, centres of sternites I and 2 (and on the latter the raised projec- tion), extreme margin of sternite 6 and whole of sternite 7, inner aspect of femora and to a lesser extent of tibiae of middle and hind legs, light ferruginous . Wings absolutely clear; subcosta to junction with basal vein black, other veins brown (light brown proximally, darker brown distally). Length of body 17-20 mm, length of wings 12-13 mm, hamuli 33-35. Clypeus baso-laterally, face, occiput and back of head, thorax, propodeum, tergite 1 on declivous face and sternite I clothed with white pilosity, sparse everywhere except on meso- pleura below and on mesosternum where dense; vestiture otherwise composed of short, de- cumbent, white pubescence, dense on clypeus and inner orbits below where shining, sparse elsewhere and apparently absent on propodeum above. Puncturation generally fine and close; propodeum with somewhat coarser punctures; ter- gite 7 with large sparse punctures separated by wide shining interstices. Mandibles lightly built, gracile, with the exception of a small tooth some way back from the apex tapering evenly to tip. Clypeus narrow, strongly convex; face narrow, at level of an- tennal sockets only very lightly wider than width of one eye, bluntly carinate below between antennal sockets (this carina carried over onto base of clypeus). Eyes slightly divergent below (Fig. I). Scape robust, less than twice as long (excluding radicle) as maximum width. Flagellum with flagellomere I as long as 2, 3 and half of 4 together; llagellomeres 3-6 finely carinate be- 138 GESS: THREE NEW SPECIES OF S. A. BEMBIX (HYMENOPTERA: SPHECIDAE; NYSSONINAE) Figs i-8. Bembix namibensis sp. nov. 6: i, head (frontal view); 2, right fore tarsus; 3, tergite 7 (dorsal view); 4, abdo- minal segment 7 (lateral view); 5, sternite 2 (lateral view, profile, anterior to top of figure); 6, sternites 5, 6 and 7 (ven- tral view); 7, genitalia (dorsal view); 8, genitalia (ventral view). (Figs 1 , 3-8 are x 12; Fig. 2 is X 25). 139 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 6, FEBRUARY 1986 low, 5 and 6 partially and 7-11 completely excavate below, 7-11 becoming progressively and strikingly larger (last flagellomere equal in length to pedicel and flagellomere 1 together). Fore tarsi (Fig. 2) strikingly modihed; first tarsomere greatly and evenly expanded lateral- ly on one side into a thin, transparent lamella partially covered on the upper surface by fine concolourous pilosity and furnished apically beneath with an acutely pointed projection on the inner side and a pair of short flattened setae on the outer. Arising from the basic shaft of the first tarsomere but on its lower surface are two combs of greatly differing character: the first composed of 16 coarse, long, curved, parallel spines lies immediately beneath the lamella and in the same plane as the latter, the apices of the spines just exceeding the lateral margin of the lamella; the second comb composed of about an equal number (that is 16) of long, fine, curved hairs lies in a plane at right angles to the first. Tarsomeres 2, 3 and 4 also expanded but less so than 1; each (like 1) apically beneath with an acutely pointed projection on inner side and a pair of setae on the outer (these setae, however, longer, finer and not flattened). Middle femora with posterior margin sharply angled, almost carinate, slightly undulate but not serrate nor anywhere toothed; middle tarsi modified. Tergite 7 (Fig. 3) markedly narrowed over apical third, narrowly rounded apically, with lateral lobes (Fig. 4) broad (broadest at level where dorsal portion abruptly narrows) and com- ing to a point apically. Sternite 2 (Fig. 5) with a large, raised, longitudinal, lamelliform projection on middle of disc, rising moderately steeply but smoothly anteriorly and falling in a smooth arc that is a quarter circle posteriorly. Sternites 4-7 modified (Fig. 6); 4 with a weak, hardly raised trans- verse line on middle of disc in posterior half; 5 with a posteriorly raised transverse platform over greater part of disc (margins of platform raised and carinate particularly at evenly round- ed postero-lateral corners); 6 with a posteriorly raised very obtusely pointed triangular plat- form on centre of disc; 7 broadly rounded apically and with a single, weak, median carina. Genitalia as figured (Figs 7 and 8). Female unknown. Material examined: South West Africa (Namibia): Namib Plain, 15 m. E Natab, 29. iv. 1969 (H. D. Brown), Holotype 6, Paratype S (Albany Museum); Awasib Mains, Diamond Area No. 2, 14. V. 1969 (H. D. I3rown), Paratype 6 (National Collection of Insects). Etymology: The name, an adjective, is derived from the geographical name, Namib, and re- fers to the provenance of the described specimens. Bembix zinni sp. nov. Male (Figs 9-17) Black; mandibles excepting their apices, labrum, clypeus (excepting for a pair of narrow triangular markings bordering sides and part of base (Fig. 9), face below and between antennal sockets, broad bands margining inner orbits to level of anterior ocellus, narrow bands widening below margining posterior orbits (but not meeting in the middle behind ocellar area as is the case in female), underside of scapes, posterior margin and sides of pronotum, narrow margins on mesonotum bordering tegulae, bands margining scutellum and metanotum behind, anterior aspect of mesopleura and backwardly curved vertical streaks on middle of same, most of meta- pleura, a pair of converging streaks on propodeum above and on declivity, sides of propodeum 140 GESS; THREE NEW SPECIES OF S. A. BEMBIX (HYMENOPTERA: SPHECIDAE: NYSSONINAE) Figs 9-17, Bembix zinni sp. nov. 6 : 9, head (frontal view); 10, right fore tarsus; 11, left hind femur; 12. tergite 7 (dorsal view); 13, end of abdomen (lateral view); 14, sternite 2 (lateral view, profile, anterior to left of figure); 15, sternites 6 and 7 (ventral view); 16, gonostyle (lateral view); 17, genitalia (dorsal view). (Figs 9, 11-17 are x 12; Fig. 10 is x 25). (except posterior to spiracles) and posterior angles, transverse bands on tergites 1-4 or 5, en- tire tergite 6 (and sometimes 5), posterior half of tergite 7 (sometimes the entire exposed por- tion), sternite 1 (except for a pair of lateral spots), distal two-thirds of sternite 2, entire sternite 3, sides and posterior margin of sternite 4, posterior corners of sternite 5 and narrow lateral margins on sternite 6, legs (except for spots on coxae and trochanters, pulvilli), light coloured. 141 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 6, FEBRUARY 1986 Markings on tergites 1-5 (except on lateral margins) almost white-, markings on head and thoracic dorsum pale yellow, rest of thorax and abdomen as well as legs darker yellow, tergites 6 and 7 orange yellow to light ferruginous. Flagellum dark ferruginous above, lighter below, lanceolate spines on first tarsomere of each foreAeg ferruginous . Wings absolutely clear; subcosta to junction with basal vein black, other veins brown (light brown proximally, darker brown distally). Length of body 13-14 mm, length of wings 9 mm, hamuli 24-34. Labrum, clypeus and inner orbits on lower half of face covered with a very fine, dense, de- cumbent, shining, silvery pubescence, most noticeable on clypeus; rest of head with fairly sparse, fine, white pilosity. Thorax and abdomen with fine, sparse, decumbent, white pubes- cence (not shining); thorax (especially pleura) and also propodeum with fine white pilosity (sparser than on head); last tergite margined by short, stout, light ferruginous bristles. Punctu- ration everywhere exceedingly fine and close except on last tergite which is furnished with large, sparse punctures separated by wide shining interstices. Mandibles lightly built, with a small tooth some way back from the apex. Clypeus some- what flattened in the middle in front (best seen in profile) but depressed area not as clearly de- limited as in female. Inner margins of eyes parallel (Fig. 9). Face carinate below between an- tennae. Flagellum with fiagellomere I about one-half longer than 2; flagellomeres 5-11 excavate below and therefore spined or laterally produced, each fiagellomere up to and includ- ing the penultimate progressively more strongly excavate; ultimate fiagellomere (11) markedly curved, almost angularly bent below. Fore tarsi (Fig. 10) with first four tarsomeres dilated; first tarsomere with a row of about 10 short, stiff, flattened spines inserted below between midline of joint and margin. Middle femora finely serrate over most of the posterior margin; middle tarsi modified; hind femora shallowly but distinctly emarginate below (Fig. 11). Tergite 7 and its lateral lobes as figured (Figs 12 and 13). Sternite 2 (Fig. 14) with a raised, arcuate, median longitudinal carina arising some distance from the anterior margin and ending some distance from the hind margin in a rounded slightly overhanging projection. Sternite 6 (Figs 13 and 15) with a narrow, acutely pointed, posteriorly raised triangular platform; sternite 7 narrow and apically pointed with a strong median and two feeble lateral carinae. Genitalia (Figs 16 and 17) with gonostyles bifurcate; upper ramus long and narrow; lower ramus well sclerotised, untwisted, short and broad, rounded apically, situated below upper ramus and except for outside edges not visible from above. Female (Figs 18 and 19) Black; mandibles excepting their apices, labrum, clypeus, face (excepting for a pair of elongate spots above antennal sockets and a transverse band between eyes at level of ocelli), bands margining posterior orbits and meeting in the middle behind ocellar area, entire scape, flagellum (except for upper surface which narrowly brownish), posterior margin and sides of pronotum, wide lateral bands and a very well developed U-shaped mark on mesonotum, wide bands margining scutellum and metanotum, entire pleura and thoracic sternites, propodeum (except for a basal band margining metonatum and a narrow V-shaped band the point of which is on the declivity above the oriface and the arms of which end above the spiracles where they are downwardly produced), broad transverse bands on tergites 1-4 (that of tergite 1 containing a pair of oval black spots sometimes fused together and then either narrowly or broadly con- nected to basal black band, those of tergites 2 and 3 sometimes with small transverse black 142 GESS: THREE NEW SPECIES OF S. A. BEMBIX (HYMENOPTERA: SPHECIDAE: NYSSONINAE) Figs 18-19. Bembix zinni sp. nov. 2 : 18 head (frontal view); 19, head, thorax and abdomen (dorsal view). Fig. 18 is x 12; Fig. 19 is X 6) spots), entire tergite 5, sternites 1-4 and distal two-thirds of 5, whole of legs (except small spot on hind coxae, pulvilli), various shades of yellow. Tergite 6 and sometimes apex of sternite 5 light ferruginous. Vestiture and puncturation very similar to that of male; shining silvery pubescence most noticeable on depressed region of clypeus. Clypeus distinctly flattened in the middle in front with depressed area clearly delimited (Fig. 18). Other than for secondary sexual characters, very similar to the male. Length of body 11,5-14 mm (Allotype 13,5 mm). Material examined: Cape Province: Merweville Distr., Jan. -Feb. 1947 (H. Zinn) Holotype 6 , Allotype 5, 2 Paratype (3(3, 5 Paratype ??; Merweville, Laingsburg Distr., i.l959 (H. Zinn), 2 Paratype ? $ . Holotype, Allotype, 1 Paratype c3, 4 Paratype 2 $ in South African Museum; 1 Paratype (3 , 3 Paratype 2 2 in Albany Museum. Etymology: The name, in the genitive singular, is formed from the name of the collector of the present material. The author has pleasure in naming this species for the late Mr Humphrey Zinn, for 41 years on the technical staff of the South African Museum, whose enthusiastic col- lecting over the years, not only on official expeditions of which he was a member but also dur- ing his vacations, provided much valuable material so important for the study of our fauna. 143 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 6, FEBRUARY 1986 Bembix karooensis sp. nov. Male (Figs 20-24) Black; mandibles excepting their apices, labrum, clypeus, face up to level of anterior ocel- lus (excepting for a pair of spots below ocellus), posterior orbits, scapes, pedicels, first flagello- meres below, posterior margin and lower half of sides of pronotum, pronotal tubercles (except for a central streak), prosternum bordering coxae, anterior portion of mesopleura and vertical bar below each tegula, mesosternum, two spots on metapleura below, spot of variable size on sides of propodeum and another anterior to each spiracle, legs (except for black markings on middle and hind coxae, streaks on upper side of femora and to a lesser extent of tibiae, lamel- lae of fore-tarsi and distal half on fifth tarsomeres and claws of all legs), most of sternites 1-3 and transverse posterior bands widening at sides on sternites 4 and 5, lemon-yellow . Narrow band bordering sides and hind margin of scutellum, narrow apical band on meta- notum, median transverse bands widening at sides on tergites 1-6 (those on 2-5 sometimes narrowly interrupted at midline and that on 6 sometimes not attaining lateral margins), yellow- ish-white. Flagellomeres (with exception of first) light ferruginous below, dark above. Wings hyaline, veins light brown, subcosta as far as its junction with basal vein darker, almost black proxi- mally. Length of body 15-16,5 mm, length of wings 9-10 mm, hamuli 25-30. Labrum basolaterally, clypeus in greater part and inner orbits on lower half of face cov- ered with a very fine, dense, decumbent, shining, silvery pubescence, most noticeable on clyp- eus; rest of head, entire thorax and propodeum and first abdominal segment with fine, long, erect, white pilosity; rest of abdomen with shorter, semi-erect, white pilosity; sixth tergite post- ero-laterally in addition with a number of dark, short, stout, apically-pointed setae. Punctura- tion everywhere very fine and close except on last tergite where somewhat larger and sparse with wide shining interstices. Mandibles lightly built, with a small tooth some way back from the apex. Clypeus only slightly less than twice as wide as long, moderately but evenly convex. Inner margins of eyes parallel; face at level of antennal sockets 1,2 times wider than width of one eye, carinate below between antennal sockets. Scape robust, only twice as long (excluding radicle) as maximum width. Flagellum with fiagellomere 1 as long as 2, 3 and half of 4 together; flagellomeres 5-11 modified, 5-7 each with a transverse ridge below near base (those on 6 and 7 spine-like in side view) and excavate below distal to ridges, 8-11 excavate below along their entire length. Fore tarsi (Fig. 20) modified; first tarsomere weakly expanded laterally at insertion of spines, more strongly so distally; second to fourth tarsomeres progressively more strongly ex- panded laterally; third and fourth narrowly lamellate; fifth somewhat swollen. Middle femora with fine serrations distally on posterior margin variously developed; mid- dle tarsi modified. Tergite 7 and its lateral lobes as figured (Fig. 21). Sternite 2 (Fig. 22) with a raised, ar- cuate and posteriorly pointed median longitudinal carnina; sternite 6 (Fig. 23) with a small, bluntly rounded, posteriorly raised platform about as wide as long; sternite 7 (Fig. 23) wide, la- melliform and truncate apically, swollen transversely subapically, with two parallel longitudinal carinae bounding sides and raised distally into triangular pointed lamellae bordering ends of transverse subapical swelling, with a triangular raised portion between parallel carinae and having its apex at middle of transverse swelling. 144 GESS: THREE NEW SPECIES OF S. A. BEMBIX (HYMENOPTERA: SPHECIDAE: NYSSONINAE) Figs 20-24. Bembix karooensis sp. nov. J: 20, right fore tarsus; 21, abdominal segment 7 (lateral view); 22, sternite 2 (lateral view, profile, anterior to left of figure); 23, sternites 6 and 7 (ventral view); 24, genitalia (dorsal view). Fig. 20 is X 25; Figs 21-24 are x 12). Genitalia (Fig. 24) with gonostyles bifurcate; upper ramus long and more or less straight; the pair of upper rami subparallel; lower ramus weakly sclerotized, lamelliform. 145 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 6, FEBRUARY 1986 Female Black; mandibles excepting their apices, labrum, clypeus, face up to level of anterior ocel- lus (excepting for a pair of spots below ocellus), posterior orbits, scapes, pedicels, first flagello- meres below, posterior margin and whole of sides of pronotum, pronotal tubercles (except for narrow markings centrally), whole of prosternum, broad wedge-shaped markings bordering tegulae on each side of mesonotum and a pair of narrow longitudinal streaks on disc of same, bands on sides and posterior margins of scutellum and metanotum and band bordering disc of propodeum and extending some way down declivity medially, entire mesopleura, mesosternum and metapleura (except sutures), entire sides of propodeum (except small region behind spira- cles), legs except for black streaks on upper sides of femora, the pulvilli and claws), broad median transverse bands widening abruptly laterally on tergites 1-5, large apical spot on ter- gite 6, entire sternites 1 and 2 and all but basal portion of sternites 3-5, lemon yellow. Flagellum and wings coloured as in male. Vestiture and puncturation likewise similar to that of male. Length of body 14-15 mm (Allotype 15 mm), length of wings 8,3 mm, hamuli 28-29. Material examined: Cape Province: Merweville, Laingsburg Distri., i.[19]59 (FI. Zinn) Holo- type 6 , Allotype 9 (South African Museum); Merweville Distr., ii.l948 (FI. Zinn) Paratype 6 (South African Museum); Dikbome, Merweville, Koup, x.1952 (Mus. Exped.) 2 Paratype 6 6 (Albany Museum). South West Africa (Namibia): 18 km from Ariamsvlei on road to Aroab, 14. V. 1973 (C.F. Jacot-Guillarmod) Paratype 9 (Albany Museum); Barby 26, Bethanie, SE2516DC, 2-7. X. 1972 (no collector) (F9244) Paratype S (State Museum, Windhoek). Etymology: The name, an adjective, is derived from the geographical name, Karoo, and refers to the provenance of the described specimens. B. karooensis is allied to B. cameronis Handlirsch to which it bears a close superficial re- semblance with respect to size, coloration and general facies. The following combination of characters is shared: in both sexes a wide face with parallel inner orbits; in the females short wings, and lateral carinae on tergite 6 enclosing a flattened pygideal area; in the males dilated fore-tarsomeres, postero-distally dentate middle femora, modified middle tarsomeres, a raised longitudinal carina on sternite 2, a raised platform on sternite 6, and two longitudinal carinae on a wide sternite 7. The males may be distinguished by the form of the dilations of the fore-tarsomeres (in cameronis tarsomeres 1-4 are all more or less evenly expanded, the lamellate parts of 2, 3 and 4 being subquadrate) (Fig. 25), by the form of the raised platform on sternite 6 (in cameronis much smaller, narrower than long) (Fig. 26), by the details of the structure of sternite 7 (in cameronis the transverse subapical swelling and the lamelliform ends of the lateral carinae are absent) (Fig. 26), and by the form of the upper ramus of the bifurcate gonostyle (in cameronis wider and inwardly curved). Females may be distinguished by means of the relative proportions of the scapes and of flagellomeres 1 and 2. In karooensis the length of flagellomere 1 is twice that of flagellomere 2 and the combined length of these two flagellomeres exceeds the length of the scape (without radicle) by a factor of almost 1,2. In cameronis, on the other hand, the length of flagellomere 1 is less than twice that of flagellomere 2 and the combined length of these two flagellomeres equals that of the scape (without radicle). 146 GESS: THREE NEW SPECIES OF S. A. BEMBIX (HYMENOPTERA: SPHECIDAE: NYSSONINAE) Figs 25-26. Bembix cameronis Handlirsch (2): 3. Clancey, P. a. (Ed.) 1980. S.A.O.S. Checklist of southern African birds. Johannesburg: Southern African Ornithologi- cal Society. Collett, J. 1982. Birds of the Cradock district. Southern Birds 9: 1-15. Connell, J. H. 1983. On the prevalence and relative importance of interspecific competition. Am. Nat. 122: 661-696. Craig, A. (in press) The diversity and distribution of birds in the Eastern Cape. In: Bruton, M. N. and Gess, F. W. (Eds) Towards an environmental plan for the Eastern Cape. Grahamstown: Rhodes University. Crass, R. 1944. Birds of Sulenkama, Qumbu district, C. P. Ostrich 15: 10-20. Every, B. 1973. New distributional data 4. Ostrich 44: 132. Every, B. 1976a. The Greyheaded Sparrow in the eastern Cape Province. Bee-eater 21 (4): 4-5. Every, B. 1976b. Greyheaded Sparrow breeding in Cape Midlands. Bee-eater 21 (4): 5-6. Every, B. 1981a. Greyheaded Sparrow reaches Port Elizabeth. Bee-eater 32: 7-8. Every, B. 1981b. Notes from Bokvlei. Bee-eater 32: 33. Every, B, 1982a. An annotated list of rare and poorly documented birds recently recorded in the eastern Cape Provin- ce. XVlIl. Bee-eater 33: 21 . Every, B. 1982b. An annotated list of rare and poorly documented birds recently recorded in the eastern Cape Provin- ce, XIX. Bee-eater 33: 40. Every, B. 1983. An annotated list of rare and poorly documented birds recently recorded in the eastern Cape Province, XXI. Bee-eater 34: 23. Every, B, 1985a. An annotated list of rare and poorly documented birds recently recorded in the eastern Cape Provin- ce, XXVll. Bee-eater 36: 21-22. Every, B. 1985b. An annotated list of rare and poorly documented birds recently recorded in the eastern Cape Provin- ce, XXVIII. Bee-eater 36: 34-36. Grobler, j. H. and Braack, H. H. 1984. Additions to the checklist of birds of the Addo Elephant Park. Koedoe 21: 39-43. Hall, B. P. and Moreau, R. E. 1970. An atlas of speciation in African passerine birds. London: British Museum (Natu- ral History). Hamilton, A. C. 1982. Environmental history of East Africa. A study of the Quarternary. London: Academic Press. Hare, H. L. 1915. The birds of Philipstown, Cape Province, with notes on their habits. 7. S.Afr. Orn. Union II: 1-18. Hewitt, J. 1931. A guide to the vertebrate fauna of the Eastern Cape Province. Part 1. Mammals and birds. Grahams- town: Albany Museum. Hill, P. 1985. A swallow, sparrow interlude. Bee-eater 36: 15-16. Irwin, M. P. S. 1981. The birds of Zimbabwe. Salisbury: Quest. JuBB, R. a. 1976. Greyheaded Sparrow nest in fencing post. Diaz Diary 31: 2-4. Leistner, O. a. and Morris, J. W. 1976. South African place names. Ann. Cape Prov. Mus. (nat. Hist.) 12: 1-565. Long, J. L. 1981. Introduced birds of the world. London: David and Charles. MacDonald, I. A. W., Richardson, D. M. and Powrie, F. J. 1986. Range expansion of the hadeda ibis Bostrychia hagedash in southern Africa. S. Afr. J. Zool. 21: 331-342. McLachlan, G. R. and Liversidge, R. 1957. Roberts birds of South Africa. Cape Town: Central News Agency. Markus, M. B. 1967. Distribution of the Greyheaded Sparrow Passer griseus (Vieillot) in the eastern Cape Province. Ostrich 38: 287-288. Martin, R. J. and Martin, E. 1975. New distributional data 5. Ostrich 46: 179-180. Martin, R. J., Neatherway, P. and Neatherway, M. 1981. An extension to the range of the Greyheaded Sparrow. Bee-eater 32: 33. Pike, E. O. 1952. The Greyheaded Sparrow in the Transkei. Ostrich 23: 128-129. Pike, E. O. 1965. The birds of Kobonqaba, Transkei. S. Afr. Avif. Ser. 24: 1-22. Pulliam, H, R. 1985. Foraging efficiency, resource partitioning, and the coexistence of sparrow species. Ecology 66: 1829-1836. ScHOENER, T. W. 1982. The controversy over interspecific competition. Am. Sci. 70: 586-595. Siegfried, W. R. 1973. Breeding success and reproductive potential in the Cape Sparrow Passer melanurus (Muller). In: Kendeigh, S.C. and Pinowski, J. (Eds) Productivity, population dynamics and systematics of granivorous birds. Warsaw: Polish Scientific Publishers. Skead, C. i. 1964. The birds of King William's Town district. S. Afr. Avif. Ser. 15: 1-70. Skead, C. j. 1965. Birds of the Albany (Grahamstown) district. S. Afr. Avif. Ser. 30: 1-46. Skead, C. J. 1967. Ecology of birds in the eastern Cape Province. Ostrich suppl. 1: 1-103. Skead, C. J. 1970. King William’s Town bird list: additions and amendments. Ostrich 41: 264-265. Skead, C. J. 1973. Zoo-historical gazetteer. Ann. Cape Prov. Mus. (nat. Hist.) 10: 1-259. Sulter, D. 1977. Grahamstown news. Diaz Diary 44: 10-11. 199 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 9, DECEMBER 1987 Summers-Smith, D. (in press) Changes in distribution and habitat utilisation by members of the genus Passer. Proc. 19th Ini. Orn. Congr. Ottawa, Canada 1986. Tree, A. J. 1973. New distributional data 4. Ostrich 44: 132. ViERKE.J. 1970. Die Besiedlung Siidafrikas durch den lieiussper\'mg(Passerdomesticus).J. Orn. 111:94-103. Walter, G. H., Hulley, P. E. and Craig, A. J. F. K. 1984. Speciation, adaptation, and interspecihc competition. Oikos 43: 246-248. Ward, P. 1968. Origins of the avifauna of urban and suburban Singapore. Ibis 110: 239-255. 200 /■ 'f- K \ \ v« .£il .4 f i; '7.'E3 I II INSTRUCTIONS TO AUTHORS LAYOUT of recent issues should be examined for details of acceptable style and format. MANUSCRIPTS should be submitted in triplicate, typed on A4 size paper, double- spaced, on one side of the paper only and with wide margins all round. Only generic and specific names should be underlined to indicate italics. All other mark- ing up should be left to the editor. Footnotes should be avoided as far as possible. All measurements should be metric. Systematic papers should follow the rules of the International code of zoological nomen- clature or the International code of botanical nomenclature, as applicable. ILLUSTRATIONS should bear the author’s name, title of paper and number in soft pencil on the reverse side. Diagrams, line drawings and photographs are all regarded as figures, should be termed as such and should be numbered in a single series. In addition to the original illustrations a photocopy of each should be submitted. Diagrams and line drawings should be reducible to a size not exceeding 14 x 16 cm (17,5 cm including caption). Photographs for half-tone blocks should be final size, of high contrast and printed on glossy paper. Colour illustrations are only printed under special circumstances and their financing must be arranged by the author. A scale should appear with all illustrations or the magnification should be given in the caption. A list of captions should be typed on a separate sheet. TABLES should be typed on separate sheets. Foldouts and tables printed sideways should be avoided. REFERENCES should be listed alphabetically at the end of the paper. Abbreviations of periodical titles should conform with those in the World list of scientific periodicals (4th ed.). Layout must be as follows; Alexander, J. 1969. The indirect evidence for domestication. In: Ucko, P. J. and G. W. Dimbleby eds, The domestication and exploitation of plants and animals. London: Duckworth, pp. 123-129. De Winter, B. 1969. Plant taxonomy today. S.Afr.J.Sci. 65 (8): 229-242. JuBB, R. A. 1967. Freshwater fishes of southern Africa. Cape Town: Balkema. GALLEY PROOFS are sent (when practicable) to authors for checking printer’s errors. Adjustments of language, style and punctuation and any other amendments of the original manuscript, if made at this stage, may be charged to authors. AUTHOR’S COPIES (50) are issued free of charge. When more than one author is involved the fifty free copies will be sent to the senior author. Additional copies ordered before printing are supplied at a reduced cost. BD7453-CTP Book Printers, Cape } I ] Annals of th^ Cape Provincial Museums Natural History Ann. Cape Prov. Mus. (nat. Hist.) Volume 16 Part 10 30th May 1988 Published jointly by the Cape Provincial Museums at the Albany Museum, Grahamstown, South Africa ‘M v < -iC . r At " • jL ANNALS OF THE CAPE PROVINCIAL MUSEUMS These Annals are published in two series, Natural History and Human Sciences, the latter series covering cultural history, ethnology, anthropology and archaeology. They are issued in parts at irregular intervals as material becomes available The primary objective of these Annals is to disseminate the results of research work undertaken by staff members of the Cape Provincial Museums. However, a manuscript by any author is eligible for publication provided it is based wholly or partly on the collection/s of one or more of these museums or if the material on which the work is based is wholly or partly deposited in one of these museums. Parts are obtainable from the Librarian of the Albany Museum. Correspondence in connection with these Annals should be addressed to the Editor, Albany Museum, Grahamstown 6140. Editor Dr F. W. GESS: 1978- Editorial Assistant Mrs S. K. GESS: 1980- A taxonomic revision of the redfin minnows (Pisces, Cyprinidae) from southern Africa by P. H. SKELTON* (J. L. B. Smith Institute of Ichthyology, Private Bag 1015, Grahamstown 6140, South Africa.) (*Previous address: Albany Museum, Somerset Street, Grahamstown.) CONTENTS Abstract 201 Introduction 202 Taxonomic history 202 Study area 203 Materials and methods 205 Character analysis 214 Taxonomic conclusions 251 Generic status 262 Species accounts 264 Discussion 304 Acknowledgements 304 References 305 ABSTRACT The redfin minnows of southern Africa form a distinctive group of cyprinid species traditionally placed within the genus Barbus. The species are distributed within the rivers of the Cape Fold Mountains in South Africa and the Drakensberg and Maluti Mountains in Lesotho. The taxonomic history of the species is reviewed with the conclusion that at least two closely similar species pairs require further taxonomic investigation. A broad based character analysis of all the species is presented and taxonomic conclusions made. The character analysis indicates that two natural groups or monophyletic lineages are present, one of which is sufficiently different from all other Barbus species to warrant separate generic status. The genus This paper is dedicated to the memory of Dr R. A. Jubb (1905-1987). 201 ANN. CAPE PROV. MUS. (NAT, HIST.) VOL. 16, PT. 10, MAY 1988 Pseiidobarbus is defined and redescriptions of all the redfin species are provided. A neotype is selected for Pseudobarbus burchelli. Aspects of the conservation of these fishes are discussed. INTRODUCTION This study concerns a group of cyprinid minnows, known as “redfins”, which are found mainly in the streams and rivers draining the Fold Mountains of the Cape Province in South Africa. A hallmark of all the species is the bright red colour of the fins, which has resulted in the species being considered to comprise a natural (or monophyletic) group (Barnard, 1943 and Jubb, 1965 and 1967). Barnard (1943) was the first author to recognise that there is a distinct group of redfin minnows in southern Cape freshwaters. Earlier authors, often working only with preserved specimens, did not always detect the red colour of the fins of certain species. Barnard (1943) and Jubb (1965 and 1967) resolved many of the taxonomical problems concerning redfin species. In spite of the progress by these authors some of the species required further taxonomic clarification. Other matters of a systematic nature concerning the redfins depended on the prior resolution of these basic taxonomical problems. The phylogenetic relationships of the species and their biogeography are particularly interesting in view of the remarkably integrated nature of the species as a group and their circumscribed distribution at the extreme south of the continent. The present study deals with the taxonomy of the redfin minnows. It forms part of a broad systematic study (Skelton, 1980) that includes biogeography (Skelton, 1986), phytogeny (Skelton, in prep, a) and osteology (Skelton, in prep. b). Within the context of the phylogenetic philosophy, the taxonomy depends to a large extent on the interrelationships of the species (Wiley, 1981) and the two should be considered together. In the present case where both the in-group and out-group data baselines are comparatively weak there was a need for a detailed character based taxonomic analysis to precede the phylogenetic analysis. The results of the phylogenetic analysis are nevertheless anticipated and incorporated into the taxonomy. TAXONOMIC HISTORY Smith (1841) gave the first description of a redfin species Barbus (Pseudobarbus) burchelli noting only that it was to be found in “various rivers of the Cape Colony”. An attractive but technically inaccurate illustration accompanied the description and showed clearly the red colour of the fins. Smith (1841) commented that local inhabitants referred to this species as the “rooye vlerk carper”. Castelnau (1861), in describing Gnathendalia vulnerata, also noted that local inhabitants used the common name “redfin”. The name Gnathendalia referred to the Genadendal Mission station on the Riviersonderend (Bree River system) where the specimens were collected. The description was based on three skins, now in the British Museum (Natural History) (BMNH). Shortly thereafter Peters (1864) described Barbus afer from three specimens collected by Krebs in the 1820’s (ffolliott and Liversidge, 1971) but this species was not recognised as a redfin until Jubb’s (1965) study. Gunther (1868) synonomised G. vulnerata with Barbus gobionides Cuvier and Valenci- ennes, 1842. Type specimens of B. gobionides are not known to exist and Barnard (1943) considered this species to be a nomen dubium as the original description was not adequate to 202 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) define any redfin or other species known at the time. Steindachner (1870) described Barbus rnultimaculatus which Boulenger (1905) placed in synonymy with B. vulneratus. Boulenger (1911) used Steindachner’s figure of B. rnultimaculatus to illustrate B. vulneratus. Boulenger (1911) also described two species only subsequently recognised as redfins viz., Barbus burgi and Barbus asper. A number of identification errors made by Boulenger (1911) and repeated by Gilchrist and Thompson (1913-1917) were corrected by Barnard (1943). These included the following: four specimens of B. burchelli from Deelfontein are actually Barbus anoplus Weber, 1897; B. vulneratus from the Baakens River are B. afer\ the illustration of B. anoplus (not a redfin species) is of a B. afer specimen; the illustration of B. afer is of a juvenile B. burgi. In consequence of Boulenger’s mistakes Gilchrist and Thompson (1913-1917) misidentified certain material and also confused specimens of B. asper with B. anoplus and thereby considered the latter to be a redfin species. Smith (1936) described Barbus senticeps, a relatively large-scaled species, from the Kromme River to the west of the Gamtoos. Barnard (1938a and b) described a further four redfin species although one, Oreodaimon quathlambae (Barnard, 1938a), was not recognised as a redfin until the present study (Skelton, 1974a and 1976). The redfins described by Barnard (1938b) were Barbus calidus, Barbus tenuis and Barbus phlegethon . Barnard’s (1943) revision included seven redfin species all from the rivers of the Cape Fold Mountain Belt: B. calidus and B. phlegethon (Olifants River system); B. burchelli (Berg and Eerste River systems); B. vulneratus (Bree and adjacent rivers);B. asper (Gourits and Gamtoos River systems with a variant form in the coastal rivers between the two systems); B. tenuis (Gourits River system); and B. senticeps (the Kromme, Baakens and Swartkops Rivers of the eastern Cape). Jubb (1965) made several nomenclatural changes to the redfins. Barbus afer was substituted for B. senticeps, B. burgi replaced B. burchelli and B. burchelli replaced B. vulneratus which became a junior synonym. The reasons for these changes are given later in the species accounts. A few problems nevertheless remained because certain species were difficult to identify and in some cases identification depended more on the geographical origin of the specimens than anything else. Jubb (1967) also reported an unusual population of B. calidus which required taxonomic attention (Skelton, 1974b). Skelton (1974a) noted that the species Oreodaimon quathlambae (Barnard, 1938a) has bright red patches over the basal fin areas and suggested that the species may be related to the redfins of the southern Cape. Further study supported this suggestion (Skelton, 1976) and O. quathlambae was therefore included in this taxonomic revision. STUDY AREA The redfin minnows are found in the rivers draining the Cape Fold Mountain ranges in the southern and south-western Cape Province of South Africa, as well as streams of the high Drakensberg and Maluti Mountains in Lesotho (Fig. 1). The Cape Fold Mountains are rugged, prominent features of the landscape (Wellington, 1955 and King, 1963), rising to about 2 000 metres. They date from an early Permian orogeny (Halbich et al., 1983). There are two series of ranges, a north-south belt in the west and a double arc of east-west ranges in the south and south-east. These mountains occur between the Great Escarpment and the coast, an important factor in connection with the drainage patterns. The rivers run along the strike of the valleys but in several places turn abruptly to penetrate the mountains through deep, spectacular gorges. 203 ANN. CAPE PROV. MUS. (NAT, HIST.) VOL. 16, PT, 10, MAY 1988 The mountain ranges consist to a great extent of the resistant sandstones of the Table Mountain Group (Cape Supergroup) (Lambrechts, 1979 and Theron, 1983), These are well leached, reworked quartzitic sandstones that impart few soluble salts to the runoff water which is consequently mineral dehcient, acidic and often stained brown to a greater or lesser extent (King et al., 1979). The valley formations are mostly of the Bokkeveld Group consisting largely of shales of marine origin that yield waters of high salinity (Bond, 1946). The drainage of the Fold Mountain region may be conveniently grouped into eight systems or catchment areas. There are four relatively large river systems that drain from the Great Escarpment and penetrate or pass the Fold Mountains to reach the sea. These are, from the west, the (Clanwilliam) Olifants (not to be confused with other rivers of this name in southern Africa), the Gourits, the Gamtoos and the Sundays River systems. Certain rivers adjacent to the Gamtoos and Sundays systems are included with these systems. Other major drainage areas are the west coast which includes the Berg River and adjacent streams, the west Agulhas area in which all the rivers are small and from which no redfin minnows have been recorded, the east Agulhas area including the large Bree River and adjacent streams, and the south coastal area that includes a number of smaller rivers draining the southern flanks of the Outeniqua and Tsitsikama mountain ranges. 204 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) The Orange River system arises in the Drakensberg and Maluti Mountains and, as the Vaal sub-system, on the Transvaal and Orange Free State Highveld, to drain westward to the Atlantic Ocean. The Drakensberg Mountains rise to 3 480 m and form an escarpment of extensive sub-horizontal strata of Karoo sediments capped by thick basalts of the “Drakensberg formation” (Truswell, 1977). The source tributaries often arise from seepage bogs that provide clear, slightly acidic waters. Water quality changes rapidly once the streams reach the underlying Beaufort sediments which contribute large quantities of dissolved and suspended matter (Cambray et al., 1986). MATERIALS AND METHODS Linear and meristic measurements Measurements were taken with vernier calipers and recorded to the nearest 0,1 mm. The pharyngeal bones were measured using a calibrated graticule on a binocular stereo microscope. Samples included specimens of both sexes and initially the measurements were analysed separately for each sex and collectively for the sample, if no difference between the sexes was evident. As far as possible only specimens of adult proportions (i.e. > 50 mm SL) were used for the intra- and interspecific analysis. Wherever possible the samples included 30 specimens and at least one sample of each species included a broad size range of specimens to assess allometry. Linear measurements, as shown in Fig. 2 , were taken according to Hubbs and Lagler (1958) except as follows; (i) measurements from the anterior end of the head were taken from the anterior symphysis of the retracted premaxillae. (ii) measurements on the head were taken from the bony margins of the reference points. (iii) predorsal length was measured along the horizontal to the intersection of the vertical line through the anterior base of the fin. (iv) pectoral to pelvic length is the median distance between the posterior margins of the bases of the fins. (v) pelvic to anal length is the median distance between the posterior margin of the pelvic bases and the anterior margin of the anal fin base. (vi) pharyngeal bones were measured according to Chu (1935) (Fig. 3). The pharyngeal bones were dissected from 10 specimens from each sample, macerated in trypsin solution for a few days and then defleshed with forceps and dried. Pharyngeal teeth were counted in rows and presented in a formula (Eastman and Underhill, 1973) giving the number on the left bone from the minor (outer) row to the major (inner) row followed by the number on the right bone, from the major (inner) to minor (outer) row. (viii) the length of the gut (intestine) was measured according to the method of Snelson (1971). The tract was severed immediately behind the transverse septum of the body cavity and at the anus. Viscera were detached and the tract straightened and pinned to a board without stretching for measurement to the nearest 1,0 mm. Measurements are expressed as percentage Standard length (SL) except for the following which are given as percentage head length (HL): head depth, snout length, orbit diameter, postorbit length and interorbit length. Barbels are given as percentage orbit diameter (OD). Meristic characters are given with the number of individual specimens counted in 205 ANN. CAPE PROV. MUS. (NAT, HIST.) VOL. 16, PT. 10, MAY 1988 Fig. 2. Linear measurements in this study. Abbreviations : Ab — anal fin base; Ah — anal fin height; B — barbel length; Bd — body depth; Bw — body width; CPd — caudal peduncle depth ; CPL — caudal peduncle length; Db — dorsal fin base; Dh — dorsal fin height; HL — head length; Hd — head depth; O — orbit diameter; PI — pectoral fin length; P2 — pelvic fin length; PO — postorbit length; P-A — pelvic to anal fin length; P-P — pectoral to pelvic fin length; PreD — predorsal length; S — snout length; SL — standard length. parenthesis. Scale and fin ray counts were made as in Hubbs and Lagler (1958) (Fig. 4a) except for the predorsal scale count which was taken as in Snelson (1972). This count records the number of oblique scale rows crossing an imaginary line between the anterior base of the dorsal 206 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) Fig. 3. Measurements and orientation of pharyngeal bones, (a) postero-medial view of the left pharyngeal bone: I — major or inner row, II — middle row. Ill — minor or outer row. (b) antero-lateral view of left pharyngeal bone : L — length, W — width, (c) orientation : A — anterior, D — dorsal, V — ventral, M — medial, PS — pitted surface, DS — dorsal surface. 207 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 10, MAY 1988 fin and the posterior margin of the head at the nape. Single scales interposed between two otherwise regular rows are omitted. Fig. 4. Meristic measurements as taken in this study, (a) scale and fin ray counts, (b) post-cranial skeletal meristics. Aptl — first anal pterygiophore, Br — branched fin rays, CP — caudal peduncle scale rows, Dptl — first dorsal fin pterygiophore, L-A — lateral line to anal fin scale rows, L-L — lateral line scales, L-P lateral line to pelvic fin scale rows, PCR — principal caudal fin rays, PUl+Ul — compound ural centrum, Sn — supraneural bones, U — unbranched fin rays, W — Weberian vertebrae, 4-4th vertebra, 10 — last predorsal vertebra, 16 — first caudal vertebra, 17 — last preanal vertebra. Branched and unbranched rays are considered separately in the case of the dorsal and anal fins, but all rays are included in a single count in the case of the pectoral and pelvic fins. Principal 208 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) caudal rays were counted as branched rays plus a single outer dorsal and ventral unbranched or simple ray. Postcranial skeletal meristics were taken from radiographs as shown in Fig. 4b. Vertebral counts include the four Weberian vertebrae and a single ural centrum (PUl + Ul). The first caudal vertebra was taken as the first unit with a closed haemal arch as indicated by a bright exposure point on the radiograph (an indication of increased bone density). Vertebrae before the dorsal (predorsal vertebrae) or anal fins (preanal vertebrae) include all vertebrae before or opposite the respective leading pterygiophore. Supraneural bones include all independent bones between the head and the first dorsal pterygiophore. The number of scale radii is the mean number of radii counted from five scales, each scale taken from one of the following different body regions on the right side: above and below the lateral line in the area before the dorsal fin, and behind the dorsal fin above and below the lateral line. Scales were removed, stained in alizarin red, defieshed and examined under a stereo microscope. Primary radii were considered as radii extending from the focus to or close to the scale margin. Scales with a diffuse focus were excluded from the count. Osteology Specimens were cleared with trypsin and stained for bone with alizarin red (Taylor, 1967). The study was completed before the double cartilage-bone staining methods using alcian blue for cartilage were available. Cartilage components were therefore stained with Victoria blue after dissection of a particular skeletal complex. Phylogenetic analysis The full results of the phylogenetic analysis are being prepared for separate publication (Skelton, in prep. a). At a generic level the phylogenetic relationships have been incorporated into this paper and the methods employed in the analysis are therefore presented here. Species interrelationships were investigated in accordance with the philosophy of cladistics or phylogenetic systematics (Wiley, 1981). The original phylogenetic methods of Hennig (1950 and 1966) included a number of principles and practices of which all but one have been discarded as a means of determining the phylogenetic relationships between taxa. The valid criterion is that related taxa share at least one relatively derived (homologous) character state (synapomorphy) which is not shared with other taxa. A monophyletic group is considered to be one which includes a common ancestor and all of its descendants (Wiley, 1981). Character state was determined according to the method of out-group comparison (Wiley, 1981). The operating principle behind the method is given by Wiley (1981) as the ‘out-group rule’ as follows: given two characters that are homologues and found within a single monophyletic group the character that is also found in the sister group is the plesiomorph (less derived or primitive state) whereas the character found only within the monophyletic group is the apomorphic character. A complementary method of character state evaluation is one using the ontogenetic criterion which is stated by Wiley (1981) as : “given an ontogenetic character sequence which goes from a character found in the outgroup to a character found only within the group considered, the character found only in the group considered is the derived character and the character found in the out-group is primitive.” Abbreviations Institutional abbreviations follow Leviton et al. (1985) and are: AMGP — Albany Museum, Grahamstown (fish collection); BMNH — British Museum (Natural History), London; MRAC 209 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 10, MAY 1988 — Musee de I’Afrique Centrale, Tervuren; MNHN — Museum National d’Histoire Naturelle, Paris; NMP — Natal Museum, Pietermaritzburg; RUSI— J. L. B. Smith Institute of Ichthyo- logy, Grahamstown; SAM — South African Museum, Cape Town; TMP — Transvaal Museum, Pretoria. Character and osteological abbreviations are: Ab — anal fin base; Ah — anal fin length; Apt — anal pterygiophore; B — barbel length; Bd — body depth; Br — branched rays; Bw — body width; CP — caudal peduncle scale rows; CPd-caudal peduncle depth; CPL — caudal peduncle length; Db — dorsal fin base; Dh — dorsal fin height (length); Dpt — dorsal pterygiophore; H — head length; Hd — head depth; L-A — lateral line to anal fin scale rows; L-D — lateral line to dorsal fin scale rows; LL — lateral line scales; L-P — lateral line to pelvic scales; O — orbit diameter; PCR — principal caudal rays; PL — pectoral fin length; P2 — pelvic fin length; PO — postorbit length; P-A — pelvic to anal fin length; P-P — pectoral to pelvic fin length; Pre D— predorsal length; Pul-I-Ul — compound ural vertebral centrum; S — snout length; SL — Standard length; Sn — supraneural bone(s); U — unbranched fin rays; W — Weberian vertebrae (vertebrae 1-4). Materials All redfin material used in this study is recorded under the individual species accounts. The samples used for morphometric and meristic measurements are given in Table 1 and their respective localities are shown in Fig. 5. Table 1. Samples of redfin minnows measured. Species Collection No. Locality No. (Fig. 5) B. burchelli AMG/P 1411, 3463 23 AMG/P 2079 19 AMG/P 2077 20 AMG/P 3472 18 AMG/P 1368 24 AMG/P 1566 21 SAM 18731 22 B. burgi AMG/P 2076, 1578 27 AMG/P 1874, 1875 28 SAM 18747 26 SAM 4695, 5090 25 BMNH 1901.2.11:14-16 27 SAM 4696 27 B. phlegethon SAM 22484 29 SAM 22749 29 SAM 22483 29 AMG/P 722 30 AMG/P 1394 29 AMG/P 1852 32 210 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) Species Collection No. Locality No. (Fig. 5) AMG/P 1863 30 AMG/P 2054 35 B. tenuis AMG/P 2081, 1935, 608 (not plotted) AMG/P 2666 15 AMG/P 2667 16 AMG/P 3186 11 B. afer AMG/P 609 1 AMG/P 745, 2524 2 AMG/P 766 3 AMG/P 2651 8 B. afer (Gamtoos) AMG/P 1415 4 AMG/P 1374, 1375 5 AMG/P 1921 6 B. asper (variant) AMG/P 2652, 2654 9 AMG/P 2656 9 AMG/P 2659 10 AMG/P 1790 11 AMG/P no, 584 11 B. asper AMG/P 1744 7 AMG/P 2663 13 AMG/P 1699 17 AMG/P 607 14 B. calidus AMG/P 1871 29 AMG/P 1862 30 AMG/P 1797, 1371 31 AMG/P 1850 32 AMG/P 1844 33 AMG/P 1855 34 AMG/P 1857 34 B. erubescens AMG/P 1867, 2045, 2049 36 AMG/P 1866, 2075 36 AMG/P 2074 36 O. quathlambae AMG/P 1540, 1877, 1823 37 AMG/P 3473-3478 37 AMG/P 3479 38 AMG/P 3480 39 SAM 19018 40 Delimitation of taxa for analysis A preliminary analysis of the results of the morphometric and meristic characters was made according to the individual samples (populations) measured as in Table 1. The species to which these samples were assigned are those established by Jubb (1965 and 1967), Barnard (1938a) and Skelton (1974b). In the case of Barbus afer and B. asper there was some doubt as to which of the species certain of the samples should be referred. These samples were therefore analysed separately and labelled as B. asper (variant) and B. afer (Gamtoos). B. asper (variant) refers to the populations of redfins found in the coastal streams between the mouth of the Gourits and the 211 ANN. CAPE PROV, MUS. (NAT. HIST.) VOL. 16, PT. 10, MAY 1988 Fig. 5. Localities of samples used for morphometric and meristic analysis (a) south coastal drainages (b) upper Orange and adjacent river drainages. Localities as given in Table 1. 212 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) mouth of the Gamtoos except for the Kromme River population which was included in B. afer (Jubb, 1965). Populations of redfins from the Fold Mountain tributaries of the Gamtoos River system are those referred to B. afer (Gamtoos). Typical B. asper also occurs in the Gamtoos River system but this species was found to be restricted to the large Groot River tributary of this system. As this investigation progressed it became evident that the major character differences lay between B. calidus and B. erubescens on the one hand, and the remaining species on the other. To facilitate repetitive reference to these groups they are named the “serrated” and “flexible” rayed species respectively (in reference to the nature of their last unbranched dorsal fin ray). Fig. 6. The relationship between orbit diameter and length of the snout in three redfin minnows B. asper, B. calidus and B. erubescens. 213 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 10, MAY 1988 CHARACTER ANALYSIS Linear measurements The morphometric measurements for each species are given under the individual descriptive accounts (Tables 14-22). The species are generally similar in form so that many measurements differ only slightly between them and are therefore mainly useful for the purpose of description. Only measurements or proportions which serve to distinguish a species or exhibit noteworthy intraspecific variation will be elaborated on in this section. The largest redfin specimen examined was a B. burchelli of 134 mm SL. Gephard (1978) reports a specimen of O. quathlambae of 143 mm total length (TL) which is also about 130 mm SL. The largest specimen of B. phlegethon measured 71 mm SL making it the smallest redfin species. The maximum size for the other redfin species lies between 85 mm SL {B. tenuis) and 120 mm SL {B. burgi). Of the head proportions the ratio between the snout length and the orbit diameter establishes an interesting difference between the two “serrated” species and the “flexible” species (Fig. 6). In B. calidus and B. erubescens the plot of orbit diameter against snout length is linear but in the flexible rayed species (e.g. B. asper in Fig. 6) a curvilinear function is described and the orbit is relatively smaller in larger specimens. The depth of the head is least in O. quathlambae and B. tenuis and both species have a more slender body profile (depth) than other redfins (Fig. 7). HEAD DEPTH(%HL) Fig. 7. Head depth (%HL) of redfin minnows. 214 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) PREDORSAL LENGTH(%SL) Fig. 8. Predorsal length (%SL) of redfin minnows. The results of the analysis of the predorsal length include two pertinent taxonomic features viz., B. burgi has a relatively short predorsal length and B. calidus and B. erubescens have relatively longer predorsal lengths than the other redfins (Fig. 8). In the case of B. burgi there is, however, a considerable amount of intraspecific variation in this character and it is only in the Berg River samples that the predorsal is shorter than in other redfin species. The shorter predorsal length in B. burgi correlates with a shorter head length in Berg River specimens (Fig. 9). It appears that this may account for the phenomenon but it is also noted that B. burgi has the lowest modal predorsal vertebral count (Table 6) which would also tend to reduce the predorsal length. A relatively high predorsal vertebral count in B. calidus may likewise account for the long predorsal length of the species (Fig. 8 and Table 6D). The origin of the dorsal fin in both B. calidus and B. erubescens is placed behind the vertical through the origin of the pelvic fins whereas in the other redfin species the dorsal fin origin lies over or only slightly behind the origin of the pelvics. Barnard (1943) distinguished B. tenuis from B. asper and B. afer on the basis of body depth. Body depth is generally an unreliable cyprinid taxonomic character because it is so easily influenced by the condition of the specimen both in the living and preserved states. In spite of these limitations the redfin species do have fairly characteristic profiles and it is useful to note 215 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 10, MAY 1988 J I I I I I I I I I PREDORSAL LENGTH (%SL) Fig. 9. Intraspecific variation in the predorsal length of B. burgi. that the body depth in these minnows correlates generally with the length of the intestine. The redfins form two fairly distinct groups with regard to body depth, those with a slender profile — B. tenuis, O. quathlambae, B. calidus and B. erubescens, and the rest with relatively deep body profiles (Fig. 10). The former species with shallow profiles all have relatively short guts (see below, p. 243) and the deeper bodied forms have longer more involuted intestines. Some interesting trends are evident in the proportions of the caudal peduncle of redfin species. The peduncle of B. calidus and B. erubescens is shorter than in the flexible rayed species (Fig. 11). Intraspecific variation in this character is not usually marked but an exception is in B. asper where the specimens from the Gamtoos River system have a significantly longer caudal peduncle than those from the Gourits River system (Fig. 12). There is a tendency for the males of flexible rayed redfins to have slightly deeper caudal peduncles than do females (Fig. 13). B. calidus has a relatively narrow caudal peduncle, and this is taxonomically useful in comparison with B. erubescens. Barbels are frequently employed as taxonomic characters in cyprinids and although their value is often overestimated (Gilbert and Bailey, 1972, and Schmidt, 1983), when used in the correct context and perspective, they may be extremely useful features. The redfins have either one or two pairs of simple barbels. The posterior pair is characteristic of all the species but an anterior pair is consistently present only in B. burchelli, B. burgi, B. calidus and B. erubescens (Fig. 14). Occasionally individual specimens of the other species have one or two anterior barbels. Certain populations of B. tenuis have a fairly high incidence of anterior barbels, e.g. 20% of the specimens from the Waterkloof River (Gourits system) had anterior barbels, but 50% of these had only a single anterior barbel. The anterior barbels of B. burgi develop relatively late and are visible in specimens greater than about 45 mm SL (Barnard, 1943). In B. burchelli these barbels develop at an earlier stage and are therefore longer than those of similar sized B. burgi. Short barbels are characteristic of 216 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) I I I I I I 1 I BODY DEPTH(%SL) Fig. 10. Body depth (%SL) of redfin minnows. 217 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 10, MAY 1988 218 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) CAUDAL PEDUNCLE LENGTH (%SL) Fig. 12. Caudal peduncle length (%SL) of B. asper. B. phlegethon and O. qiiathlambae (Fig. 14). The actual length of the barbels is subject to considerable intraspecific variation especially in the widespread species such as B. afer and B. asper (variant) populations. Fins The fins of cyprinid fishes are well known to be labile structures with their shape and proportions subject to the environmental and functional demands placed on them (Hubbs, 1940 and Alexander, 1967). There are no outstanding taxonomic differences between the length of the fins of redfin species. Intraspecific differences are in certain cases at least as large as interspecific differences. In B. tenuis for example the population from the Keurbooms River system has relatively longer fins than conspecifics in the Gourits River system. The Keurbooms catchment receives a higher mean average rainfall compared to Gourits River catchments (Midgeley and Pitman, 1969). This suggests that flow is likely to be higher on average in the Keurbooms, a factor which favours the development of longer fins (ITubbs, 1940) and may be responsible for the observed differences in this case. Sexual dimorphism of the fins is clearly evident in most of the flexible-rayed species (Barnard, 1943). Boulenger (1911) indicated that the pectoral fins of B. asper males were longer than those of the females. Barnard (1943) qualified this statement of dimorphism by referring to the length of the pectoral relative to the base of the pelvic fins. In males of the flexible-rayed species the pectorals reach the base of the pelvics but in the females the pectorals are short of the pelvics. In the case of juveniles the pectoral fins reach the pelvics (except in B. phlegethon). The length of the pectoral and the pectoral to pelvic distance is shown in Fig. 15. Apart from B. calidus and B. erubescens the males have relatively longer pectoral fins but shorter distances between the fins than the females. Interspecific differences and differences between the sexes in the shape of the pectoral fins are evident from Fig. 16. The males have generally broader more rounded fins than the females. 219 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 10, MAY 1988 CAUDAL PEDUNCLE DEPTH(%SL) Fig. 13. Caudal peduncle depth (%SL) of redfin minnows. 220 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) BARBEL LENGTH (%ORBIT DIAMETER)' Fig. 14. The length of the anterior barbel and the posterior barbel (% Orbit diameter) of redfin minnows. Meristics (A) Fin rays The number of fin rays in the dorsal, anal, pectoral and pelvic fins of redfin species are given in Table 2. In the dorsal and anal fins the number of unbranched rays is always difficult to determine accurately because the first one or two rays are very small and hidden from view. The form of the last unbranched ray in the dorsal fin is a useful character in Barbus (e.g. Boulenger, 1911) and is a prominent feature separating B. calidus and to a lesser extent B. erubescens from the other redfin species (Fig. 17). In B. calidus this ray is bony and serrated behind with only the tip being segmented and flexible. This ray is more flexible with a few reduced serrations in B. erubescens, and slender and flexible without serrations in all other redfin species. 221 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 10, MAY 1988 Fig. 15. The length of the pectoral fin and the pectoral to pelvic fin distance (%SL) of male and female redfin minnows. Apart from B. erubescens, for which the modal number is eight, the redfins usually have seven branched dorsal fin rays (Table 2B). Similarly the majority of species have only fivebranched rays in the anal fin but this serves to emphasise the taxonomic significance of the modal six for B. calidus and seven for B. erubescens (Table 2D). Pectoral and pelvic fin rays are seldom reported for African cyprinid species. The intraspecific range of variation in pectoral fin rays is broad which tends to diminish the value of any modal differences between the species (Table 2E). The same is true of the number of pelvic fin rays where all but B. tenuis have a mode of eight rays (Table 2F). The redfins have invariably 10 + 9 principal caudal fin rays. 222 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) Fig. 16. Pectoral fin profiles of redfin minnows, dorsal view, right side. Scale bar = 5 mm. A. B. burchelli, B. B. burgi, C. B. phlegethon, D. B. tenuis, E. B. afer, F. B. asper, G. B. asper (variant), H. O. quathlambae, I. B. calidus, J. B. erubescens. 223 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 10, MAY 1988 Table 2. Fin ray counts for redfin minnows. 4 L. B. C D. Unbranched Branched Unbranched anal Branched anal dorsal fin dorsal fin rays fin rays fin rays rays Species/ (Group) N 3 4 6 7 8 2 3 4 4 5 6 7 8 B. burchelli 179 74 105 8 166 5 130 49 1 171 7 B. burgi 107 93 14 3 102 2 1 103 2 104 2 B. phlegethon 90 63 27 4 85 1 85 5 89 1 B. tenuis 96 69 27 6 88 2 90 6 93 3 B. afer 105 37 68 2 103 105 104 1 B. afer (Gam- toos) 50 14 36 50 48 2 49 1 B. asper 110 11 99 5 104 1 95 15 109 1 B. asper (vari- ant) 142 54 88 5 133 4 137 5 142 O. quathlam- bae 39 39 1 37 1 39 35 4 B. calidus 300 10 290 1 287 12 280 20 2 278 20 B. erubescens 110 33 77 27 83 1 107 2 7 96 7 E. Pectoral fin rays F. Pelvic fin rays Species (Group) N 11 12 13 14 15 16 17 18 1 8 9 B. burchelli 179 1 40 90 44 4 1 160 13 B. burgi 107 4 46 42 12 2 1 1 80 26 B. phlegethon 90 1 4 36 42 7 11 78 1 B. tenuis 96 4 30 48 13 1 70 26 B. afer 105 5 29 41 25 5 98 7 B. afer (Gamtoos) 50 4 29 12 5 2 46 2 B. asper no 7 48 43 12 4 86 20 B. asper (variant) 142 9 41 68 22 2 8 123 11 O. quathlambae 38 8 21 9 4 31 3 B. calidus 280 1 10 141 116 12 28 237 15 B. erubescens no 1 5 62 41 1 14 96 224 SKELTON: TAXONOMY OF REDFIN MINNOWS (PISCES, CYPRINIDAE) Fig. 17. The form of the last unbranched ray of the dorsal fin of redfin minnows. A. B. calidus, B. B. erubescem, C. B. afer. (B) Scale counts. Distribution frequencies of the various scale counts for redfin species are given in Table 3. There is a wide range of variation in the counts and only two species are really distinctive in the group on account of the size of their scales. These two species are O. quathlambae , which has exceptionally small scales, and B. asper, which also has relatively small scales. Scale size, as assessed by scale counts, is an important factor in the problem of delimiting the two species B. asper and B. afer. The distribution frequencies of the lateral line counts of the different populations of the two species and their “variant” groups are given in Table 4 with graphic summaries in Fig. 18. A wide range of counts is characteristic of most of the samples. B. afer samples indicate a low mode of about 32 with the exception of the Kromme River sample where the mode is 29. In the Gamtoos system B. afer samples have a mode of 36 and those of B. asper 37. B. asper (variant) samples show a mode of 34 or 35 lateral line scales. The summaries of lateral line counts for these samples depicted in Fig. 18 provide four taxonomic options. Firstly all the populations can be considered to be of a single polytypic species (Fig. 18A). Although the distribution curve is normal for this option the range of 225 The frequency distribution of the number of scales or scale rows in redfin species. ANN, CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 10, MAY 1988 B. Lateral line to dorsal fin scale rows s m CM 2 vO o 0\ ■Tt CO 37 C" 18 1 65 3 so 133 35 7 19 43 4 101 I22O 72 28 66 56 86 70 6 41 60 38 6 33 3 16 A. Lateral line scales 45 3 r-t 43 -- 40 !!2 1 17 8 38 6 24 1 69 45 ir»^{Ns0^ir)\D<-i tJ-o r-H «-4 (N (N 72 ro 33 4 14 30 1 22 10 20 67 6 51 9 40 36 5 11 1 38 13 2 70 32 14 19 14 16 2 34 2 1 69 18 19 7 8 11 37 00 SO pt 10 23 7 2 24 12 67 V, 7 17 11 6 99 so o 4 14 14 1 65 Os n Nor^-o oo-^ rn <— 1 I— 1 1 35 - 2 ornm.-H ont-1 Tl- <— ( ,-H 34 NO 0\in’^>n*7}"i— 1 On r— iOn Tt ro >-H <-H o m 33 rOfN’-HinrO'— 1 no ■^o (Ni-< iv > ii > i), unwebbed and with well developed subarticular tubercles. Toes long (iv > iii > v > ii > i), webbed and with well developed subarticular tubercles. 311 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 11, MAY 1988 Fig. 2. The type locality of Heleophryne hewitti sp. nov. — the upper reaches of the Geelhoutboom River in the Elandsberg Mountains (3325CC Loerie). In life the dorsal pattern consists of numerous rounded and irregularly-shaped, dark-brown patches superimposed on a uniform light-brown to olive-brown background. The patches are marginated with a thin white line and are not interlinked. Numerous small spots and flecks of a similar colour are present between the larger patches. The hindlimbs are marked with dark, irregularly-shaped transverse bands that lack the white borders of the dorsal patches but remain distinct. Similar bands are present on the forearm, tarsus and outer fingers and toes. The nostrils are hidden in small, dark-brown spots; and there is no interorbital bar. The forearms are slightly swollen and bear elongated nuptial pads. Dorsal skin folds are present but poorly developed. Asperities on the inner surfaces of the forearms are indistinct. Asperities are absent on the outer edges of the upper and lower jaws, and in the chest region. Measurements of holotype SV 47,0 mm; HL 44,1 mm; HW 16,5 mm; DS 4,4 mm; DE 4,9 mm; IN 4,7 mm; lO 8,6 mm; Tib 24,0 mm; Tar 32,7 mm; WH 73,1 mm. 312 BOYCOTT: NEW SPECIES OF HELEOPHRYNE SCLATER (ANURA: HELEOPHRYNIDAE) Description and variation of paratypes Range in snout-vent length 38,0-46,7 mm in males (x 41,6), 42,2-50,3 mm in females (x 44,6). Range in hindlimb length 38,5-46,8 mm in males (x 41,4), 41,3-47,1 mm in females (x 42,8). Morphometric data (ratios) obtained from paratypes: SV/HL 0,94-1,05; SV/Tib 1,74-1,96; SV/HW 2,55-2,79; Tib/HW 1,37-1,54; HW/DE 3,27-3,89; HW/IO 1,79-2,08; lO/DE 1,59-2,04; IN/IO 0,49-0,56; DE/IN 0,93-1,14; DE/DS 0,88-1,18. In 80,9% of males the snout-vent length was equal to or greater than hindlimb length (SV/HL ratio 0,94-1,07), and in all females it was greater (SV/HL ratio 1,02-1,07). In all paratypes the internarial distance was equal to or greater than half the interorbital distance (IN/IO ratio 0,49-0,56). In males, 2 or 2^/2 phalanges of the fourth toe are usually free of webbing. However, in the holotype and one paratype (AMG A 699), the webbing extends to a point between 2^/2 and 3 phalanges from the tip. Females have 3 to 3fi phalanges of the fourth toe free of webbing. The webbing in males extends to the tip of the fifth toe or to a point within half a phalanx of the tip (SAM 45187); in H. regis it frequently falls short of the tip. In females 1 phalanx of the fifth toe is usually free of webbing, but in the largest specimen (PEM A 459) it extends to the tip; female H. regis usually have 2 phalanges of the fifth toe free of webbing. The colouration of the paratypes is similar to that of the holotype. Colour-pattern polymorphism is absent and there is no colour-pattern difference between males and females. Variation in dorsal colour-pattern is minimal. It is mostly evident in the tone and contrast of the markings described for the holotype. A continuous interorbital bar, often formed by a series of interlinked blotches, is present in thirteen of the paratypes including all the females. The background colour of live specimens varies from yellow-brown to olive-brown; during the day a few specimens were seen and they had a dull green background colour. Lighter shades are more evident after dark. As in the holotype, all paratypes are very spotted with numerous smaller spots and flecks scattered between the larger dorsal patches, a condition not found in any of the H. regis specimens. The transverse bands on the limbs are distinct, these are usually indistinct or absent in H. regis. Mating call Recordings were made at the type locality at night and during the day on 3 and 4 October 1980. The call consists of a sequence of soft repetitive whistles. Each call sequence consists of 8 or 9 whistles. The interval between successive notes varies from 1,0 to 2,0 seconds, 1,5 seconds being the most common. Two second intervals were recorded during the day (llhOO, air temperature 26,0° C; water temperature 16,8° C), but after dark (21h00, air temperature 10,0° C; water temperature 12,5° C) 1,0 and 1,5 second intervals were more common. Pauses between each calling sequence varied from 6,0 to 9,0 seconds. Often each calling sequence, particularly those separated by 8,0 or 9,0 second pauses, commenced with a long, loud note produced as a drawn out whistle (Figs 3a and 3c). This may have a duration of almost 0,25 seconds (Fig. 3c), although it is usually between 0,15 and 0,2 seconds long. Successive whistles in the sequence are of considerably shorter duration, varying from 0,06 to 0,11 seconds. They are distributed over a narrow frequency range 1,7-2, 2 kHz (Figs. 3a and 3b). The call of H. regis is distributed over a much wider frequency range, between 1 and 2,6 kHz (Boycott, 1982). The initial drawn-out part of the call, as indicated in Figs 3a and 3c, commences at a higher frequency (2, 0-2, 2 kHz) and sweeps slightly downward. Etymology This species is named after Dr John Hewitt (1880-1961) in honour of his remarkable contributions to South African herpetology. 313 kHz kHz 6 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 11, MAY 1988 4 - 2 - O “T ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1— 0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 seconds a b c Fig. 3. Sonograms of the mating call of Heleophryne hewitti sp. nov. showing: (a) the initial drawn out note and first whistle of a calling sequence; (b) a whistle from a calling sequence; (c) an initial drawn out note. ADDITIONAL MATERIAL Tadpoles and eggs. The tadpoles and eggs of Heleophryne are quite different from those of other southern African amphibians. The tadpole’s most distinctive characteristic is its unique suctorial mouth, which extends right across the ventral surface of the head. Numerous transverse tooth-rows are present, four rows on the upper jaw and more than a dozen on the lower jaw. The 314 BOYCOTT: NEW SPECIES OF HELEOPHRYNE SCLATER (ANURA: HELEOPHRYNIDAE) number of tooth rows depends on the age of the tadpoles {pers. obs.). The eggs of Heleophryne are as distinctive as the tadpoles, being bright yellow and large yolked. Tadpoles and eggs of H. hewitti sp. nov. were collected from various localities. Tadpoles were collected from the type-locality (three series — PEM A 463; SAM 45181, 45182); Martins River (two series — AMG A 719; PEM A 464); Klein River (eight series — AMG A 628, 629; NM 6772, 6773; PEM A 462; SAM 45183, 45184; TM 55202); and Diepkloof River (three series — AMG A 630; TM 55200, 55201). Eight egg batches have been found, six at the type locality and two in the Klein River. The early development of four batches was observed and samples were preserved at different stages (three from the type locality — AMG A 614; SAM 45190, 45191; and one from the Klein River — NM 6775). The total number of eggs in four batches was counted, these do not necessarily refer to the sampled batches as some were counted in situ and left. Assuming that none was overlooked, these totalled 93, 103, 120 and 150. Six egg batches were found in October 1979 and two in October 1980. Most oviposition sites were located beneath submerged or partially submerged rocks of variable size. These were usually in relatively quiet backwaters where the stream flow was not very rapid and the water depth varied from 20 to 60 cm. However, one batch was found under a semi-submerged rock in the middle of a shallow (20 cm), rapidly flowing section of the stream. Indeed, when the rock was lifted most of the eggs were swept away by the current. The eggs have a surprisingly hard capsule which probably protects them should they be dislodged from the original oviposition site. The eggs and oviposition sites of H. hewitti are comparable to those of H. purcelli as described by Visser (1971) and Boycott (1972). HABITAT, DISTRIBUTION AND CONSERVATION It is generally accepted that the tadpoles of Heleophryne spend at least two seasons in the rivers and streams before metamorphosing and the genus is therefore prevented from colonizing non-perennial streams (Rose, 1926; Wager, 1965; Boycott, 1982; Boycott and de Villiers, 1986). The breeding season of H. purcelli and H. regis (Boycott, 1982) and H. hewitti coincide with that time of the year when river flow is reduced. This strategy enhances the survival chances of eggs and young tadpoles at a critical stage of their development. Furthermore, it ensures the longterm survival of tadpoles as any stream that has water at that time of the year will certainly be perennial. The Van Stadensberg Mountains form a south-easterly extension of the Elandsberg Mountains and as they are continuous they are here referred to collectively as the Elandsberg range. The highest peaks in the Elandsberg range, Stinkhoutberg (1 106 m) and Elandsrivier- berg (987 m), receive 1 000 mm of rain per annum. The Elandsberg range has a higher annual rainfall than the Groot Winterhoek range situated to the north and throughout its length several perennial streams are present. Extensive surveys in the Elandsberg range have revealed that H. hewitti has a very limited distribution in these mountains. The species is restricted to four rivers, the Geelhoutboom River, Martins River, Klein River (all 3325CC Loerie) and Diepkloof River (3324DB Cocks- comb/3324DD Hankey). Two of these, the Martins River and Klein River, have perennial tributaries which the other two lack. The sources of the four rivers extend over a distance of just 10 km, and the greatest distance separating any two of these rivers is 3,5 km. All these localities are between 400 m and 550 m above sea level, and are clear, swift-flowing, perennial mountain streams with rocky beds. The banks along the streams are generally steep and thickly vegetated but not forested (Fig. 2). They resemble more closely the typical habitat type of H. purcelli and not that of H. regis. The indigenous vegetation of the Elandsberg range is classified as False Macchia (Acocks, 1975). However, due to the high rainfall most of the mountain slopes have been afforested and very little of the natural vegetation remains. 315 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 11, MAY 1988 Investigations were carried out in six other river systems throughout the Elandsberg range. These are the Honey Clough River (Otterford Forest Reserve), Berg and Bulk rivers (Longmore Forest Reserve), Van Stadens River (Van Stadens Forest Reserve), Keurkloof River (Foerie Forest Reserve) and the Sand River (Hankey Forest Reserve). No Heleophryne specimens were found at any of these localities which emphasises the geographical isolation of the H. hewitti population. The distribution of the species is shown in Fig. 4. Fig. 4. The distribution of Heleophryne Sclater in South Africa. • Heleophryne rosei Hewitt ■ Heleophryne purcelli Sclater □ Heleophryne regis Hewitt ■ Heleophryne hewitt sp.nov O Heleophryne natalensis Hewitt H. hewitti appears to have one of the most restricted distribution ranges of any southern African amphibian. Over the last six or seven years through a combination of unwise forestry practices, severe fires and floods, the habitat of H. hewitti is now seriously threatened. In the early 1980s fires devastated the Otterford and Foerie plantations and most of the trees were left 316 BOYCOTT: NEW SPECIES OF HELEOPHRYNE SCLATER (ANURA; HELEOPHRYNIDAE) Standing. In June 1986 these were cleared and within months, as a result of heavy rain, flooding led to the extensive silting up of the streams, particularly the Geelhoutboom River — the type locality. As a result of this, large tracts of the type locality have been rendered uninhabitable for//, hewitti and at present the habitat hardly resembles that depicted in Fig. 2 (W. R. Branch pers. comm.). Another threat to the continued survival of this species would be the introduction of exotic fish species into the streams. This has already occurred in the Honey Clough River and in some of the other rivers on the periphery of the species’ distribution range {pers. obs.). It is not known whether the construction of dams and reservoirs in those streams, such as the Bulk River and Sand River, on the northern side of the Elandsberg range has affected the distributon of H. hewitti. However, dams constructed at higher altitudes (i.e. in areas where the species does occur) will certainly reduce the amount of suitable habitat available to the species. COMPARISON WITH OTHER SPECIES In the Cape Province because of the isolated nature of the mountain ranges one can expect considerable taxonomic complexity. Attention was initially drawn to this by Poynton and Broadley (1978). More recently, the findings of the present author appear to substantiate this dictum (Boycott, 1982). Heleophryne natalensis and H. rosei are well separated morphologically from the other species of ghost frog in the Cape Province, possibly due to their geographical separation (Poynton, 1964). The Cape Province populations of Heleophryne (excluding rosei) share certain morphological characteristics including secondary sexual characters, dorsal pattern and eye markings. They form a distinct group — the H. purcelli/H. regis species complex. During the breeding season both sexes develop secondary sexual characters (Boycott, 1982). In males these include the development of loose dorsal skin folds, swollen forearms (nuptial pads), and asperities or spines. The distribution and concentration of asperities or spines can be useful species characters (Boycott, op. cit.). There is a marked difference in the structure, distribution and concentration of the asperities in H. natalensis, H. rosei and in members of the H. purcelli/H. regis complex. In H. natalensis ‘jagged’, conspicuous spines are present, and restricted to the axilla and the dorsal aspect of the first three fingers. This is well illustrated in Passmore and Carruthers (1979). However, in H. rosei, H. purcelli and H. regis these structures can be more aptly termed asperities as they are considerably smaller in size, more numerous and widely distributed over the entire dorsal surface. The asperities on the forelimbs in H. rosei are concentrated on the outer surfaces of the forearms, whereas in H. purcelli, H. regis and H. hewitti they are restricted to the inner and dorsal surfaces of the forearms. Despite the relatively large series of specimens examined (n = 27), it was discovered that the asperities in H. hewitti do not appear to be as well developed nor as conspicuous as they are in H. purcelli and H. regis. Nonetheless, the distribution and concentration of asperities in the new species conform more to the typical pattern found in H. purcelli and H. regis. H. natalensis and H. rosei possess a distinctly mottled dorsal pattern, the actual size and shape of the dorsal patches being somewhat ill defined. Individuals representative of the H. purcelli/H. regis complex possess clearly defined, vivid, dorsal patches which are usually bordered by thin lines. All members of the genus possess vertically elliptic pupils which in some are well concealed due to the presence of dark pigment in the eye. H. natalensis and H. rosei lack the dark, transverse band through the eye that is always present in members of the H. purcelli/ H. regis complex. The possession of clearly defined dorsal patches and a dark transverse band through the eye by H. hewitti clearly indicate that this species is part of the H. purcelli/H. regis species complex in the Cape Folded Mountain Belt. 317 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 11, MAY 1988 Unfortunately most of the preserved material of H. natalensis could not be examined by the author. Five specimens (TM 14927, TM 21066, TM 21067, TM 21413, TM 26204) are old, fragile and preserved in such awkward positions as to be unmeasurable. Fresh H. natalensis material is needed. Morphometric data obtained from samples of H. rosei (n = 10); H. purcelli (n = 25); H. regis (n = 25) and H. hewitti (n = 27) revealed few diagnostic differences between the species (Table 1). It appears that the snout-vent length in H. rosei is equal to or less than the hindlimb length (SV/HL ratio 0,93-1,05; x 0,99), whereas in the other species it is normally equal to or greater than the hindlimb length (Table 1). Furthermore, in H. rosei, unlike the other Cape species, the internarial distance is less than half the interorbital distance (INTO ratio 0,45-0,52; X 0,48); and the tibia is usually one and a half times the width of the head (Tib/HW ratio I, 43-1,56; X 1,50). These ratios may be useful diagnostic characters and should be tested when more H. rosei and H. natalensis material becomes available. H. rosei can be separated from the other Cape species on the basis of the ratios SV/HL; Tib/HW; and IN/IO (Table 1). The three species representatives of the H. purcellilH. regis complex are clearly very similar to one another and cannot be separated on morphometric grounds. However, vast differences exist between the mating calls. Table 1. Morphometric data for species of Heleophryne H. n rosei = 10 H. purcelli n = 25 H. n regis = 25 H. hewitti sp. nov. a = 21 Ratio Range Mean S.D. Range Mean S.D. Range Mean S.D. Range Mean S.D. SV/HL 0,93-1,05 0,99 0,04 0,93-1,11 1,02 0,04 0,95-1,16 1,01 0,05 0,94-1,07 1,02 0,03 Tib/HW 1.43-1,56 1.50 0,05 1,25-1.53 1,40 0,07 1,37-1,54 1,46 0.05 1,37-1,54 1,46 0,05 HW/IO 1,83-2,07 1,93 0,09 1,87-2.19 2,00 0,09 1,75-2,15 1,90 0,09 1,79-2,08 1,92 0,08 IN/IO 0.45-0,52 0,48 0.02 0.44-0.58 0,52 0,03 0,47-0,61 0,54 0,03 0,49-0,56 0,53 0,02 DE/IN 0,91-1,20 1,05 0,10 0,94-1,24 1,07 0,07 0,88-1.13 0,99 0.05 0,93-1,14 1,01 0,06 A detailed study of the tadpoles of Heleophryne from the Cape Province (Channing et al. , in prep.) has revealed that most of the sampled populations are distinguishable on a suite of morphological characters. Although the tadpoles of the southern Cape populations are similar, nostril morphology, pigmentation, tail length and proportions suggest that the taxonomic status of other isolated populations, for example on the Kammanassie Moutains, the Baviaanskloof Mountains and the Kouga Mountains warrants further investigation. The tadpoles of H. hewitti can be distinguished from those of H. regis on nostril morphology, tail length and proportions. Distance between populations has been demonstrated on more than one occasion to mean reproductive isolation, even when morphological difference is not great. Therefore as the population under discussion is separable on mating call and tadpole morphology and no intergrades are known to occur in nature, it is proposed that this population be considered specifically distinct. The most diagnostic feature of any frog species is the voice (Blair, 1958; Pengilley, 1971; Littlejohn, 1971; Passmore and Carruthers, 1975; Passmore and Carruthers, 1979; Passmore, 1981; Telford and Passmore, 1981; Boycott, 1982). Examination of the sono- grams of the calls of H. purcelli and H. regis (Boycott, 1982) and H. hewitti (Fig. 3) shows clearly that call difference alone can be used to separate these species unequivocally from each other. 318 BOYCOTT: NEW SPECIES OF HELEOPHRYNE SCLATER (ANURA: HELEOPHRYNIDAE) ACKNOWLEDGEMENTS The research was funded by the Cape Department of Nature and Environmental Conservation and the Board of Trustees of the Port Elizabeth Museum. Thanks are due to the Department of Forestry especially the Regional Directors of the Tsitsikamma and Southern Cape Forest Regions for permission to conduct the study in the regions falling under their control. Furthermore, the assistance and co-operation of staff members of the Department of Forestry are gratefully acknowledged. W. Branch, W. J. Frey, A. Jones, D. Smith-Belton and H. Swanevelder are thanked for assistance in the field. Much benefit was gained from numerous discussions with W. Branch, J. Brytenbach, C. Burgers, J. Greig, R. Haynes, J. Koen, P. Lloyd and N. Passmore. The author is particularly indebted to J. Greig and N. Passmore for their constant encouragement and interest. N. Passmore and S. Telford ae thanked for making the sonograms. Finally, the author would like to thank the Director of the Cape Department of Nature and Environmental Conservation for permission to publish this paper. REFERENCES Acocks, j. P. H. 1975. Veld types of South Africa. Mem. hot. Surv. S. Afr. 40: 1-128. Blair, W. F. 1958. Mating call and speciation of anuran amphibians. Atti. Nat. 92: 7-51. Boycott, R. C. 1972. Ghost Frog eggs. Afr. Wildlife 26 (1): 43. Boycott, R. C. 1982. On the taxonomic status of Ileleophryne regis Hewitt, 1909 (Anura: Leptodactylidae). Ann. Cape Prov. Mils. (nut. Hist.) 14 (3): 89-108. Boycott, R. C. and De Villiers, A. L. 1986. The status of Heleophryne rosei Hewitt (Anura: Leptodactylidae) on Table Mountain and recommendations for its conservation. S. Afr. J. Wildl. Res. 16 (4): 129-134. Channing, a.. Boycott, R. C. and van Hensbergen, H. J. (in prep.). Descriptions of Heleophryne tadpoles from the Cape Province, South Africa (Anura: Heleophrynidae). Littlejohn, M, J. 1971. Amphibians. In Amphibians and Reptiles of Victoria, pp. 1-11. Reprinted from the Victorian Year Book No. 85. Passmore, N. I. 1981. The relevance of the specific mate recognition concept to anuran reproductive biology. Monitore zool. ital. 15: 93-108. Passmore, N. I. and Carruthers, V. C. 1975. A new species of Tomopterna (Anura: Ranidae) from the Kruger National Park, with notes on related species. Koedoe 18: 31-50. Passmore, N. I. and Carruthers, V. C. 1979. South African Frogs. Johannesburg: Witwatersrand University Press. Pengilley, R. K. 1971. Calling and associated behaviour of some species of Pseiidophrvne (Anura: Leptodactylidae). J. Zool. Land. 163: 73-92, Poynton, j, C. 1964. The Amphibia of southern Africa: a faunal study. Ann. Natal Mus. 17: 1-334. PoYNTON, J. C. and Broadley, D. G. 1978. The Herpetofauna. In: Werger, M. J. A. ed, Biogeography and Ecology of southern Africa. The Hague: Junk, pp. 925-948. Rose, W. 1926. Some field notes on the Batrachia of the Cape Peninsula. Ann. S. Afr. Mus 20: 433-450. Telford, S, R. and Passmore, N. I. 1981. Selective phonotaxis of four sympatric species of African reed frogs (Genus Hyperolius). Herpelologica 37 (1): 29-32. VissER, J. D. 1971. Hunting the eggs of the ghost frog Heleophryne purcelli orientalis FitzSimons. Afr. Wildlife 25 (1): 22-24. Wager, V. A. 1965. The frogs of South Africa. Johannesburg: Purnell. 319 INSTRUCTIONS TO AUTHORS LAYOUT of recent issues should be examined for details of acceptable style and format. MANUSCRIPTS should be submitted on floppy disk accompanied by three print-outs (spacing U/a or 2). In exceptional cases typed manuscripts will be accepted. These should be submitted in triplicate, typed on A4 size paper, double-spaced, on one side of the paper only and with wide margins all round. Only generic and specific names should be underlined to indicate italics. All other marking up should be left to the editor. Footnotes should be avoided as far as possible. All measurements should be metric. Systematic papers should follow the rules of the International code of zoological nomenclature or the International code of botanical nomenclature , as applicable. ILLUSTRATIONS should bear the author’s name, title of paper and number in soft pencil on the reverse side. Diagrams, line drawings and photographs are all regarded as figures, should be termed as such and should be numbered in a single series. In addition to the original illustrations a photocopy of each should be submitted. Diagrams and line drawings should be reducible to a size not exceeding 14 x 16 cm (17,5 cm including captions). Photographs for half-tone blocks should be final size, of high contrast and printed on glossy paper. Colour illustrations are only printed under special circumstances and their financing must be arranged by the author. A scale should appear with all illustrations or the magnification should be given in the caption. A list of captions should be submitted on a separate sheet. TABLES should be submitted on separate sheets. Foldouts and tables printed sideways should be avoided. REFERENCES should be listed alphabetically at the end of the paper. Abbreviations of periodical titles should conform with those in the World list of periodicals (4th ed.). Layouts must be as follows: Alexander, J. 1969. The indirect evidence for domestication. In: Ucko, P. J. and G. W. Dimbleby eds, The domestication and exploitation of plants and animals. London: Duckworth, pp. 123-129. De Winter, B. 1969. Plant taxonomy today. S.Afr.J.Sci. 65 (8): 229-242. JuBB, R. A. 1967. Freshwater fishes of southern Africa. Cape Town: Balkema. GALLEY PROOFS are sent (when practicable) to authors for checking printer’s errors. Adjustments of language, style and punctuation and any other amendments of the original manuscript, if made at this stage, may be charged to authors. AUTHOR’S COPIES (50) are issued free of charge. When more than one author is involved the fifty free copies will be sent to the senior author. Additional copies ordered before printing are supplied at a reduced cost. BD8113 — CTP Book Printers, Cape ISSN 0570-1880 / 0 Annals of the Cape Provincial Museums Natural History Ann. Cape Prov. Mus. (nat. Hist.) Volume 16 Part 12 29th July 1988 Published jointly by the Cape Provincial Museums at the Albany Museum, Grahamstown, South Africa ANNALS OF THE CAPE PROVINCIAL MUSEUMS These Annals are published in two series, Natural History and Human Sciences, the latter series covering cultural history, ethnology, anthropology and archaeology. They are issued in parts at irregular intervals as material becomes available The primary objective of these Annals is to disseminate the results of research work undertaken by staff members of the Cape Provincial Museums. However, a manuscript by any author is eligible for publication provided it is based wholly or partly on the collection/s of one or more of these museums or if the material on which the work is based is wholly or partly deposited in one of these museums. Parts are obtainable from the Librarian of the Albany Museum. Correspondence in connection with these Annals should be addressed to the Editor, Albany Museum, Grahamstown 6140. Editor Dr F. W. GESS: 1978- Editorial Assistant Mrs S. K. GESS: 1980- Three new South African species of Aulacoderus la Ferte, a subgenus of Anthicus Paykull (Coleoptera: Anthicidae) by J. C. VAN HILLE (Department of Zoology and Entomology, Rhodes University, Grahamstown) ABSTRACT Descriptions and figures are given of three new species of Anthicus (Aulacoderus), namely halleyi, tardus and propinquus, and their relationships to other species are discussed. INTRODUCTION The present paper constitutes a third supplement to the author’s monograph of Aulacoderus (van Hille, 1984) adding three new species. The two previous supplements (van Hille, 1985a and 1985b) added thirty-eight and six species respectively. An additional three species of Aulacoderus were described in a paper on Anthicidae collected in Botswana (van Hille, 1986). TAXONOMIC DESCRIPTION AND DISCUSSIONS SECTION 5 Anthicus (Aulacoderus) halleyi spec, nov., Eigs 1-4 Size. Length 2,45 mm (1,70-2,62); width over broadest part of elytra 0,81 mm (0,70-0,93). Head (Fig. 2). Matt; testaceous to dark testaceous; posterior arch broadly round. Punctures close together, each surrounded by a round lighter area; with short procumbent hairs. Eyes rather small and somewhat bulging. Prothorax (Fig. 2). Matt; testaceous; longer than broad, slightly broader than head. Punctures as on head, with short recumbent hairs. Without basal lateral constrictions. Elytra. Matt; testaceous, slightly lighter than prothorax. Punctures minute, not surrounded by a lighter area, with fine recumbent hairs, longer than on prothorax. In male not apically pointed (Fig. 4) but with a somewhat elongated notch. Wings. Absent; metatergum hardly sclerotised. Antennae. Slender; light testaceous. Toward apex hardly broadened and not darker. Apical segment somewhat longer than penultimate segment. Undersurface. Light testaceous. With mesepimerite apophyses with spiral markings and a smaller pair of apophyses with spiral markings on anterior margin of metasternum (Fig. 3) projecting into the mesocoxal cavities. Metepisterna immovably fused onto metasternum. 321 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 12, JULY 1988 Figs 1-4. A. (A.) halleyi spec. nov. 1: aedeagus. 2: head and prothorax. 3: meso- and metathoracic sterna. 4: apex of elytron of male. Figs 3 and 4 at same magnification; other figs as indicated. Male abdomen. Aedeagus (Fig. 1). Tegmen slightly sclerotised, with pointed apex; median lobe membranous; in the preparation it cannot be seen as a structure separate from the connecting membrane. Genital opening invisible. The connecting membrane has a number of sclerotised structures: a pair of proximal groups consisting of 12 to 15 straight spines each; the spines are pointed at the median end and point obliquely inward and proximally; more distally lies a single group of 8-10 similar spines pointing inward and distally; subapically lies a pair of short broad dark claws, surrounded by a field of very small thorns which are not pigmented and probably not sclerotised. Spiculum gastrale with a pair of short asymmetrical arms. Last exposed sternum somewhat long, with rounded apex. Last exposed tergum trapezoidal. 322 VAN HILLE: NEW SPECIES AULACODERUS LA FERTE (COLEOPTERA: ANTHICIDAE) Material Examined. South Africa: near old Storms River Bridge, 33° 59'S, 23° 55'E, 25.ii-3.iii.1986, Holotype 6, Paratypes 1205 (probably all 1. vulcania were stocked exclusively with Noctuidae. Prey of A. ferrugineipes consisted mostly of geometrid caterpillars (81 per cent) and only occasional noctuid caterpillars (1,6 per cent). Lycaenidae were included in the prey of both A. ferrugineipes and A. conifera, the latter also taking Pyralidae. The species of prey taken are given in Table 2 which shows that only six out of a total of 42 species of caterpillars were shared between any of the Ammophila species. Calculated niche overlaps for this resource (Table 3) are consequently very low for all species pairs comparisons. Supplementary data obtained from a site in Mkuze Game Reserve over a ten-day period showed no overlap whatever in prey taken by A. ferrugineipes and A. beniniensis. Nests of the latter species (n = 7) were stocked with caterpillars of the noctuid Achaea lienardi, those of the former species (n = 27) with geometrid caterpillars (7) or with noctuid caterpillars of the subfamily Plusiinae (20). In this area noctuid caterpillars formed the main prey of A. ferrugineipes, in contrast to what was found at Hilton (Table 1). Thus, whereas the type of prey taken varied between sites a lack of interspecific overlap persisted. 329 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 13, JULY 1988 Table 2 Caterpillar species taken as prey by Ammophila at Hilton. Species of Ammophila Prey species ARCTIIDAE ^Maenas vocula LASIOCAMPIDAE ■’Beralade prompta -Pachypasa sp -Bombycopsis nigrovittata dolichodera ^ NOCTUIDAE -indet. catocaline Heliothis armigera Craterestra definiens 'Cucullia minuta ,C. consimilLs iNoct. Z ^Blenina squamifera ^Anua selenaris -Eutelia adulatrix ~Hypoplexia externa -Noct. I -Ulotrichopus primulina ^Audea melaleuca insignis ^-^Noct. 3 -? Achaea sp. (Noct. J) Thria robusta indet. noctuid ? Achaea sp / ^ GEOMETRIC AE 'Axiodes bifasciata A A. dochmoleuca ^Omphalucha ditriba 'Geom. D ferrrugineipes ^Geom. F -Syndromodes invenusta ■Tephrina / Semiothisa spp. 'Zamarada spp 'Ligdia pectinicornis "Lomographa indularia 'Prasinocyma scissaria 'Geom. J/X 'Geom. 19+28+30 LYCAENIDAE ''Anthene spp PIERIDAE ''Pieris helice PYRALIDAE indet. pyralid Broken line linking wasp species with prey indicates record from Clifton. Species of Ammophila braunsi dolichocephala beniniensis vulcania . conifera 330 WEAVING: PREY SELECTION OF AMMOPHILA W. KIRBY (HYMENOPTERA: SPHECIDAE) Table 3 Niche overlaps for three species of Ammophila at Hilton. Niche overlaps* for respective species pairs Resource set dolichodera! dolichodera! insignis! insignis ferrugineipes ferrugineipes Prey length 0,559 0,062 0,547 Prey mass 0,146 0,000 0,204 Prey species 0,018 0,000 0,012 Hunting habitat (plant species) 0,099 0,843 0,118 Hunting habitat (generalised) 0,721 0,757 0,238 *Calculated using symmetric MacArthur-Levins formula (Lawlor, 1980). Size of prey Sizes of prey differed markedly between the various species of Ammophila. The mass and length of caterpillars sampled from nests are given in Table 4, grouped according to whether the wasps provision their nests with one or several prey (Weaving, in press). As would be expected, single-prey species on average selected much larger caterpillars than did the multiple-prey species. Further size differences occurred within each group, showing a positive correlation between prey size and wasp size. However, interspecific overlap in prey size, particularly with regard to prey length, was extensive (Fig. 1 and Table 3) due to the wide range in prey size taken by each species. The potential for interspecific competition for prey on the basis of size therefore exists for most of the species. The niche overlap value of 0,559 between the single-prey species A. dolichodera and multiple-prey species A. insignis can be explained on their size difference, the substantially larger size of the latter species counteracting its requirement for smaller prey dictated by its provisioning strategy. A. dolichodera is similar in size to the multiple-prey species A. ferrugineipes and in this case the overlap value is very low (0,062) due to their differing provisioning strategies. Niche overlap values were not calculated for the remaining species because of the small sample sizes. Nevertheless, the data in Fig. 1 and Table 4 show that prey taken by the two single-prey species A. beniniensis and A. braunsi fall entirely within the size range of that taken by A. dolichodera. A similar relationship occurs between the multiple-prey species A. conifera, A. dolichocephala and A. ferrugineipes. Only A. vulcania in the Eastern Cape selected prey entirely outside the range of sizes for all of the other species. An intraspecific comparison of data for A. ferrugineipes from Hilton and the Natal sites confirms that, apart from provisioning strategy, the size of the wasp governs the size of prey taken. Caterpillars taken in Natal were significantly longer than those at Hilton (mean length = 20,6 and 15,9 mm respectively. Student’s t = 2,866 df = 26 p <0,01). Wasp size was also significantly greater in Natal (mean head width = 3,81 and 3,41 mm respectively. Student’s t = 5,228 df = 29 p <0,001). However, there was no significant difference between sites in the ratios of wasp size to prey length (ratio = 0,460 and 0,403 respectively. Student’s t = 1,417 df = 26 p >0,1). This species was therefore taking larger caterpillars in Natal simply because of its larger size. 331 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 13, JULY 1988 Ab(N) Af Ac ♦ Adc • Ai Ab ♦— Abr # Add ♦ 0 1000 ^ Mass of prey (mg) Av 2000 ♦ Af(N) Ab(N) Af # Ac Adc Ai I ■ I 0 10 20 • Ab » Add 30 40 Length of prey (mm) 50 Av 60 Fig. 1, Mean mass/length (•) of prey and range ( ) taken by various species oiAmmophila at Hilton, and in Natal (N) as indicated. Ab — Ammophila beniniensis; Abr — A. braunsi; Ac — A. conifera; Adc — A. dolichocephala; Add — A. dolichodera; Af — A. ferrugineipes; Ai — A. insignis; Av — A. vulcania. 332 WEAVING: PREY SELECTION OF AMMOPHILA W. KIRBY (HYMENOPTERA: SPHECIDAE) Table 4. Sizes of prey used by eight species of Ammophila at Eastern Cape sites except when otherwise indicated. Species of Ammophila Wasp length* mm Wasp mass# mg mass mg Mean size of prey n length mm n Mass of prey Mass of wasp y Single-prey species: braunsi 20 53 439 7 27 3 9,4(5,7-12,5) 4 dolichodera 23 72 466 26 31 29 8,5(4,4-13,3) 5 beniniensis 24 48 408 2 28 2 9,1 1 beniniensis Natal 31 169 1515 2 52 22 9,4(6,7-12,1) 2 vulcania 31 273 1503 7 48 7 3,8(3,0-7,l) 4 Multiple-prey species: conifera 22 53 40 3 17 6 0,3 1 ferrugineipes 22 52 52 106 17 105 l,3(0,7-2,3) 19 dolichocephala 27 108 85 11 21 11 0,6(0,2-2,l) 8 insignis 30 122 185 107 26 126 1,0(0,7-1,7) 6 *Mean of 10 specimens. # Mean of variable number of specimens of each species, y Number of direct prey mass/wasp mass comparisons. Ranges in parentheses. Source of prey The proportions of prey originating from various plant species are shown in Fig. 2 for three well-sampled species oi Ammophila . Twenty-three per cent of the prey of ^4. insignis originated from Diospyros dichrophylla and Rhus macowanii, trees and large shrubs confined to the banks of the watercourses. Acacia karroo, which accounted for a further six per cent, occurred in various growth forms, large trees along the watercourses and elsewhere, smaller trees and shrubs, and low shrubs less than a metre high. A similar proportion (five per cent) came from Lycium campanulatum, small to medium sized shrubs scattered over the area. Thirty-nine per cent of prey was obtained from Pentzia incana (Compositae), a component of the low-growing dwarf karroid scrub, and the remaining 27 per cent from unknown sources. Disregarding the unknown sources, a significant proportion of the prey of A. insignis (32 per cent) therefore originated from trees and large shrubs and this could be as high as 40 per cent if caterpillars from A. karroo came from trees of this species growing along the watercourses. However, the origin of 60 per cent of prey from L. campanulatum and P. incana shows that A. insignis was flexible in its hunting habitat. Nevertheless, it is the only one of the three well-sampled species, A. ferrugineipes , A. insignis and A. dolichodera, which obtained some of its prey from R. macowanii, D. dicrophylla and L. campanulatum. More than 90 per cent of the prey of A. ferrugineipes was from A. karroo, and the few direct observations of this species catching its prey involved moderate-sized shrubs away from the watercourses. At Hilton less than one per cent originated from low-growing plants. In Natal A. ferrugineipes took prey from herbaceous plants too and it is therefore another species showing versatility in its choice of hunting habitat. 333 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 13, JULY 1988 A. dolichodera A. insionis A. ferruaineioes n=34 n=140 n=126 Key to host plants: — Acacia karroo Fig. 2. Sources of caterpillar prey taken by three species of Ammophila at Hilton. Plants linked by dotted lines in the key are dwarf shrubs. Lycium campanulatum is a shrub. Rhus macowanii, Diospyros dichrophylla and Acacia karroo occur as trees mainly along watercourses, the last two species also as smaller trees and small shrubs elsewhere. 334 WEAVING: PREY SELECTION OF AMMOPHILA W. KIRBY (HYMENOPTERA: SPHECIDAE) A. dolichodera obtained 44 per cent of its prey from dwarf shrubs and 50 per cent from A. karroo. A high niche overlap value of 0,843 (Table 3) shows that much of the prey of this species and oi A. ferrugineipes is obtained from the same plant species; interestingly, the same species pair showed no overlap in species of caterpillars selected. In reality, however, the overlap in hunting habitat is probably much less. A. dolichodera was always observed hunting in the smallest of the three size categories of A. karroo mentioned above and therefore probably obtains its prey from low-growing vegetation only. The origins of the prey of the remaining species of Ammophila, for which sample sizes are very small, are indicated in Table 5. A. braunsi hunts in low-growing vegetation, utilising the same type of vegetation as does A. dolichodera. However, the former species’ preference for areas of clay soil (Weaving, 1986) effectively separates the hunting habitats of both species. The hunting habitat of A. dolichocephala, the only multiple-prey species which hunted in low- growing vegetation, overlapped with that of both the above species since it was recorded from sand and clay areas (Weaving, 1986). Table 5 Sources of prey of five poorly sampled species of Ammophila at various localities. Species and locality A. braunsi Hilton (n = 6) Kommadagga (n = 1) A. beniniensis Hilton (n = 3) A. vulcania Hilton (n = 3) Clifton (n = 5) A. conifera Hilton (n = 3) A. dolichocephala Hilton (n = 7) Clifton (n = 3) Source of prey Atriplex semibaccata“ — low-growing herb Atriplex semibaccata“ — low-growing herb One unknown, others below leaf litter Unknown — suspected Rhus macowanii Pappea capensis‘’ — small trees in dwarf karroid scrub Acacia karroo and, probably, Maytenus spp. Chrysocoma tenuifolia — dwarf shrub <0,5 m high Pentzia incana — dwarf shrub <0,5 m high ^ — Chenopodiaceae; — Sapindaceae. At Hilton there were two instances of A. beniniensis obtaining its prey from below leaf litter, behaviour which has not previously been documented for Ammophila. In Natal this species hunted in large trees. The hunting habitat of A. vulcania in the Eastern Cape was similar to that of A. beniniensis in Natal in spite of the difference in the general habitat (Weaving, 1986). A comparison between ranges in the size of prey taken by A. ferrugineipes, A. insignis and A . dolichodera and the maximum sizes attained by prey species and potential prey species reared in the laboratory (Figs 3 and 4) provides further evidence of the hunting habitats used. Relative abundance of the various caterpillar species, as determined by field collections, is shown in Appendix I. Caterpillars of Geom 6 (abundance rating AR = 1), Xylopteryx arcuata (AR = 7) and X. prasinaria (AR = 8) attained suitable sizes and were more abundant than most of the species utilised as prey. Their absence from nests of A. ferrugineipes confirms that this species does not hunt in D. dichrophylla , R. macowanii or Maytenus spp. 335 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 13, JULY 1988 Potential prey from: — 900-1 600 O) E 0) v> ca 300 Species of caterpillar Fig. 3. Ranges in the mass of caterpillars of actual and potential prey species (vertical bars) at Hilton, and limits of mass of individual prey (horizontal shaded bands) taken by three species of Ammophila. Add — A. dolichodera; Af — A. ferrugineipes; Ai — A. insignis. Ak — Acacia karroo; Dd — Diospyros dichrophylla; Ly — Lycium campanulatum; Mh — Maytenus heterophylla; Ml — M. linearis; Rm — Rhus macowanii. 336 WEAVING: PREY SELECTION OF AMMOPHILA W. KIRBY (HYMENOPTERA: SPHECIDAE) Actual prey of: — 337 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 13, JULY 1988 Potential prey from: — r Ak Dd I Rm 1 Ml/Mh Lc 1 T Species of caterpillar Fig. 4. Ranges in the length of caterpillars of actual and potential prey species (vertical bars) at Hilton, and limits of length of prey (horizontal shaded bands) taken by three species of Ammophila. Abbreviations as in Fig. 3. 338 Anthene spp. Prasinocyma scissaria WEAVING: PREY SELECTION OF AMMOPHILA W. KIRBY (HYMENOPTERA: SPHECIDAE) Actual prey of: — Af 1 Ai Add 1 I 1 339 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 13, JULY 1988 During a period when the noctuid Cucullia minuta was extremely abundant on Chrysocoma tenuifolia, it was not taken by A. ferrugineipes even though earlier instars would have been suitably sized. The geometrid Axiodes dochmoleuca also occurred on this plant and was taken by A. dolichocephala; although of a suitable size, it was never recorded as prey of A. ferrugineipes. These observations confirm what has been stated above concerning the different hunting habitats of these two species. Morasa modesta, a hairy lymantriid caterpillar found on R. macowanii (AR = 17) was suitably sized for A. dolichodera but never taken by it. This supports the observation that A. dolichodera hunts only in low-growing vegetation. A. insignis obtained 16 per cent of its prey from D. dichrophylla (Fig. 2). Flowever, in spite of its abundance on this plant, Geom 6 (AR = 1) was never recorded as prey. Although mature caterpillars reached a suitable length (Fig. 4), their maximum mass only slightly exceeded the minimum recorded for A. insignis. Many specimens would probably be smaller and would thus be rejected, which may explain their absence from nests of A. insignis. The caterpillars of a sawfly, Arge sp. (Hymenoptera, Argidae) (AR = 9) occurred on the same plant. Mature specimens reached 25 mm in length and, as indicated by their diameter of 4 mm, were similar in mass to caterpillars of a noctuid Blenina squamifera which were frequently taken as prey from D. dichrophylla. It is of interest that caterpillars of Arge sp. were never taken by A. insignis since sawfly caterpillars have been recorded elsewhere as prey of Ammophila (Bohart and Menke, 1976). The above observations are summarised in Table 6 in which the various species of Ammophila at Hilton have been allocated to one of four categories of vegetation/hunting habitat. From this it is evident that there is potential competition for prey between several species as a result of hunting in the same habitats, especially as it has already been shown that prey size requirements overlap extensively. Table 6 Distribution of species of Ammophila between four categories of hunting habitat at Hilton with supplementary information from other localities. Occurrence of respective species of Ammophila Hunting habitat Single-prey species Multiple-prey species Ab Abr Add Av Adc Af Ai Shrubs and trees bordering watercourses Shrubs and trees away from watercourses Low-growing vegetation Leaf litter below shrubs and trees — -U -1-^ +f -l-‘‘ — + — — +■= + + + Abbreviations: Ab — A. beniniensis', Abr — A. braunsi; Add — A. dolichodera; Av — A. vulcania; Adc — A. dolichocephala; Af — A. ferrugineipes; Ai — A. insignis. “ — suspected; — observations at Clifton; — confirmed by observations at Verdun; — confirmed by observations at Mkuze Game Reserve; ^ — confirmed by observations at Clifton; * — all Natal records involved hunting in large trees in well-vegetated areas. Another aspect of the source of prey, for which there is very little information, is the position that the prey occupies on the host plant. Observations of caterpillars in the field and of those being reared have indicated which species are to be found amongst foliage and those which spend the day either resting along branches or below leaf litter. The most frequently taken prey of A. ferrugineipes for which observations were available were mostly green, foliage-inhabiting 340 WEAVING: PREY SELECTION OF AMMOPHILA W. KIRBY (HYMENOPTERA: SPHECIDAE) forms. A. insignis included both foliage-inhabiting species and those found on branches. The form of dwarf shrubs makes it difficult to draw this distinction, but the prey of A. doUchodera from A. karroo, the lasiocampids Beralade prompta, Pachypasa sp. and probably the unidentified catocaline noctuid, were branch inhabiting forms during the day. Such differences help to explain the almost complete absence of shared prey species even among Ammophila which hunt in the same plant types, but further data are essential before any firm conclusions can be drawn. Relative abundance and seasonality of prey The seasonal distribution of records of the more abundant caterpillar species, including those obtained from sampling wasps’ nests, are shown in Table 7. All species were present for at least half of the period when Ammophila were active (October to May) and many of them for all or most of it. Since both A. ferrugineipes and A. insignis were recorded nesting in every month, and A. doUchodera from October to February, seasonal differences either in availability of prey species or in nesting activity are unlikely to be important in explaining differences in prey selection. Table 7 Seasonal availability of the more abundant actual and potential caterpillar prey species, and nesting periods of three species of Ammophila at Hilton. Wasp and caterpillar species Occurrence of caterpillar species or nesting activity in Oct Nov Dec Jan Feb Mar respective month Apr May Nesting by: A. ferrugineipes + + + + + + + + A. insignis + + + + + + + + A. doUchodera + + + + + Caterpillar species: Ligdia pectinicornis + + + + + Lomographa indularia + + + + Omphalucha nr. ditriba + + + + SemiothisalTephrina + + + + + + + + Syndromodes invenusta + + + + + + Xylopteryx arcuata + + + X. prasinaria + + + + + Zamarada spp. + + + + + + Geom 6 + + + + + + Blenina squamifera + + + + + + + + Eutelia adulatrix + + + + + + Hypocala rostrata + + + + Hypoplexia externa + + + + Ulotrichopus catocala + + + + Noct Z + + + + Beralade prompta + + + + + + + Morasa modesta + + + Maenas vocula + + + + + + Orna nebulosa + + + + + + Anthene spp. + + + + Arge sp. + + + 341 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 13, JULY 1988 The relative overall abundance of the caterpillar species was determined by combining the results of all samples obtained during beating and comparing them with the frequency with which the species appeared in samples of prey taken by Ammophila. The results are shown in Fig. 5 and have a strong positive correlation (r = 0,830). The species of Ammophila were therefore preying on the most abundant caterpillars occurring within their respective hunting habitats. Prey instar Early instars of several species of caterpillars which grew to a large size, eg. the noctuids Ulotrichopiis catocala, Autophlebia bracteata, Ophiusa tirhaca and Sphingomorpha chlorea, were never recorded as prey. The majority of caterpillars taken were apparently mature specimens in their final instars. This observation is supported by the data in Figs 3 and 4 which show that maximum sizes attained by prey species seldom exceeded the maximum size taken by Ammophila. Some prey caterpillars were in fact almost prepupal, there having been several instances of prey “attempting” to pupate. Prey specialisation by individuals Being a mass provisioner, A. insignis was the only multiple-prey species for which the entire contents of fully provisioned nests could be identified with certainty. Out of 25 such nests, 48 per cent were provisioned with a single species of prey, 32 per cent with two species and 20 per cent with three. There was also a tendency for successive nests of individual wasps to be provisioned with the same species of prey. This is illustrated by observations of marked females nesting simultaneously at the same site (Table 8). Further, different wasps were taking different species Table 8 Prey usage by individual Ammophila insignis at Hilton. Wasp number Nest ref. Axiodes bifasciata Number of respective species of prey in nest Hypoplexia Cucullia Noct Z Ulotrichopus externa consimilis primulina Noct J 1 379 0 0 0 7 0 0 2 362 4 1 0 0 0 0 370 1 0 0 1 0 1 371 4 0 0 0 0 0 4 367 0 1 0 2 0 0 383 1 0 1 0 1 0 6 369 0 0 4 0 0 0 375 0 0 0 1 0 1 381 0 0 5 0 0 0 385 0 0 2 0 0 0 10 384 0 2 0 0 0 0 11 387 2 0 0 0 0 0 342 WEAVING: PREY SELECTION OF AMMOPHILA W. KIRBY (HYMENOPTERA: SPHECIDAE) of caterpillars even though they were hunting on the same plant species, as demonstrated by wasps #2 and #6. Thus, although wasps were preying on the more abundant caterpillar species, these were not being taken with the same frequencies by different individuals. The limited nature of the data available makes it impossible to state whether this reflected individual preferences on the part of wasps or was of a more temporary nature, perhaps reflecting localised concentrations of prey to which individuals were repeatedly returning. DISCUSSION Prey specificity can often be demonstrated in the field. However, the factors which govern prey selection are not always readily apparent. O’Neill and Evans (1982) could not explain the ultimate cause of prey partitioning between four sympatric species of the sphecid genus Philanthus, except for some divergences in prey size due to differences in predator size. However, it often transpires that apparent prey preferences are actually the outcome of prey behaviour, size requirements of the predator, differences in the relative abundance of potential prey and in foraging behaviour, or conditioning of the predator to certain prey or hunting environments due to flexibility in the behaviour of the wasps (Coville, 1976; McQueen, 1979; Laing, 1979; Gwynne, 1981). The present study has revealed a remarkable lack of overlap in prey species taken by several sympatric species of Ammophila. The fact that the size of prey taken, within broad limits, was governed by wasp size and provisioning strategy does not necessarily account for the observed partitioning of prey species. Caterpillars pass through a wide range of sizes during their development and therefore the early instars of large species should theoretically be suitable for the smaller species of Ammophila or those employing multiple-prey provisioning. Observations made in the present study suggest that Ammophila prey on caterpillars approaching maturity. It is not known whether this is the result of their selection of mature caterpillars or whether such caterpillars are easier to locate or catch. Similarly, Jennings and Houseweart (1984) reported that certain Eumenidae select late instar caterpillars for provisioning their nests but did not offer any explanation. Nevertheless, the present study shows that the major factor governing this apparent preference for prey species was hunting-habitat selection. Similar findings have been published for species of the sphecid genera Sceliphron and Chalybion (Muma and Jeffers, 1945), two species of the pompilid genus Anoplius (Evans, 1953), and for Pompilidae in general (Evans and Yoshimoto, 1962). Bowden (1964) showed how two sympatric species of the sphecid genus Dasyproctus differed by hunting over grass or amongst shrubs respectively. However, the present data indicate only broadly the differences in hunting habitats selected by Ammophila. The species of prey taken differ in their location on particular parts of their host plants, suggesting that the wasps themselves are either selecting these microhabitats for hunting or that their prey-size requirements govern whether they must hunt amongst foliage or along stems and branches. Further studies are therefore likely to identify even more discrete interspecific differences in hunting habitats and explain why certain common and apparently suitable prey are not taken. The discovery of A. beniniensis taking prey from below leaf litter at Hilton is of special interest; similar behaviour amongst the Ammophilini is known only from species of Podalonia. A. beniniensis observed in Natal obtained their caterpillars from trees, never on the ground. Whether or not this is evidence for the existence of two sibling species can only be answered after 343 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 13, JULY 1988 Fig. 5. Correlation between field (beating) and prey samples of ten caterpillar species taken by three species of Ammophila at Hilton. 344 WEAVING: PREY SELECTION OF AMMOPHILA W. KIRBY (HYMENOPTERA: SPHECIDAE) additional studies of their behaviour and taxonomy have been undertaken. Hunting on the ground may have developed in more arid areas, such as in the Eastern Cape, where branch-inhabiting caterpillars may have been forced to spend the day below leaf litter because of increased pressure from predation in the more open vegetation, though this is only speculation. The methods used by Ammophila to locate prey are to be discussed in more detail elsewhere, but the fact that wasps initially search on the ground for cues in the form of faecal pellets of caterpillars which indicate the presence of prey in the vegetation overhead (Weaving, 1986) means that such a switch in hunting habitat is perfectly feasible. The findings for yl. insignis demonstrate the phenomenon of specialisation on prey species by individual wasps. This has been reported by Tsuneki (1963) for the sphecid genus Sphex. Evans (1966) refers to the possibility of conditioning with respect to hunting sites as an explanation. Rather than reflecting individual prey preferences, this is almost certainly the result of locality learning by the wasps, behaviour which would optimise utilisation of spatially heterogeneous prey populations (Rosenheim, in press). The coexistence of several species of Ammophila at Hilton, and at many other sites, inevitably raises the subject of interspecific competition. In spite of differing habitat preferences for nesting (Weaving, 1986), the mobility of the adults would allow the different species to hunt in the same places and therefore compete for prey. However, observations have indicated that it is unlikely that populations of Ammophila ever become large enough to cause depletion of their prey resources (unpublished data), though the effects of other caterpillar predators, such as eumenid wasps and birds, and parasitoids would have to be included in any evaluation. Whether competition could have brought about the observed differences in prey selection cannot be answered here. The data collected in this study provide measures of “actual” niches rather than “virtual” niches as defined by Colwell and Futuyma (1971). These authors stress the importance of obtaining both types of niche measurement before the existence or otherwise of competition can be demonstrated. Overlap of “actual” niches, or lack of it, may be evidence for or against competition (Colwell and Futuyma, 1971; Strong, 1983). The measurement of “virtual” niches requires manipulation of the predator or prey populations; in fact, carefully planned experiments are essential if the existence or otherwise of competition is to be established, and such an investigation was impracticable in the present study. Even so, a clear demonstration of interspecific competition is very difficult to obtain (Connell, 1983). The results have at least shown that any role competition may have played will have been an indirect one through partitioning of the hunting habitat rather than direct partitioning of prey resources. Whether competition or some other factor was responsible for the development of hunting habitat preferences during speciation remains an open question. This topic will be discussed elsewhere with respect to other aspects of nesting behaviour included in a broader study. ACKNOWLEDGEMENTS Thanks are due to Dr F. W. Gess for showing the author the sites on the farms Hilton, Clifton and Verdun and for introducing him to the owners of these farms; to Mr T. C. White for free access to his farm Hilton where much of this study was conducted; to Mr F. Norton of Clifton and Mr Truter of Verdun. The Director of the Natal Parks, Game and Fish Preservation Board kindly issued permits to conduct research in areas under his control. 345 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 13, JULY 1988 Dr P. E. Hulley and Dr F. W. Gess are thanked for discussions, assistance and encouragement during the present study and for constructive criticism of this manuscript. Dr L. Vari is thanked for his help in identifying many of the Lepidoptera reared from caterpillars. Mr Alfred Bill rendered valuable assistance in the field for which the author is grateful. Gratitude is expressed to the Council for Scientific and Industrial Research for a running expenses grant from 1982 to 1984 and to the Albany Museum for research funding for 1985 and 1986. REFERENCES Bohart, R. M. and Menke, A. S. 1976. Sphecid wasps of the world: a generic revision. Berkeley: University of California Press. Bowden, T 1964. Notes on the biology of two species of Dasyproctus Lep. and Br. in Uganda (Hymenoptera: Sphegidae). Journal of the Entomological Society of southern Africa 26: 425-437. Colwell, R. K. and Futuyma, D. J. 1971. On the measurement of niche breadth and overlap. Ecology. 52: 567-576. Connell, J. H. 1983. On the prevalence and relative importance of interspecific competition: evidence from field experiments. The American Naturalist 122: 661-696. Coville, R. E. 1976. Predatory behavior of the spider wasp, Chalvbion calif ornicum . Pan-Pacific Entomologist 52: 229-233. Evans, H. E. 1953. Comparative ethology and the systematics of spider wasps. Systematic Zoology 2: 155-172. Evans. H. E. 1959. Observations on the nesting behavior of digger wasps of the genus Ammophila. American Midland Naturalist 62: 449-473. Evans, H. E. 1965. Simultaneous care of more than one nest by Ammophila azteca Cameron (Hymenoptera, Sphecidae). Psyche 72: 8-23. Evans, H. E. 1966, The behavior patterns of solitary wasps. Annual Review of Entomology 11: 123-154. Evans, H. E. and Yoshimoto, C. M. 1962. The ecology and nesting behavior of the Pompilidae (Hymenoptera) of the north eastern United States. Miscellaneous Publications of the Entomological Society of America 3: 67-119. Gess, F. W. 1981. Some aspects of an ethological study of the aculeate wasps and the bees of a karroid area in the vicinity of Grahamstown, South Africa. Annals of the Cape Provincial Museums (Natural History) 14: 1-80 Gwynne, D. T, 1979. Nesting biology of the spider wasps (Hymenoptera: Pompilidae) which prey on burrowing wolf spiders (Araneae: Lycosidae, Geolycosa). Journal of Natural History 13: 681-692. Gwynne, D. T, 1981. Nesting biology of the bumblebee wolf Philanthus bicinctus (Mickel) (Hymenoptera: Sphecidae). American Midland Naturalist 105: 130-138. Jennings, D. T. and Houseweart. M, W, 1984. Predation by eumenid wasps (Hymenoptera: Eumenidae) on spruce budworm (Lepidoptera: Tortricidae) and other lepidopterous larvae in spruce-fir forests of Maine. Annals of the Entomological Society of America 77: 39-45. Kurczewski, F. E. and Kurczewski, E. J. 1968. Host records for some North American Pompilidae (Hymenoptera) with a discussion of factors in prey selection. Journal of the Kansas Entomological Society 41: 1-33. Laing, D. j. 1979. Studies on the populations of the tunnel web spider Porrhothele antipodiana. Part 2. Tuatara 24: 1-21. Lawlor, L. R. 1980. Overlap, similarity, and competition coefficients. Ecology 61: 245-251, Mcqueen, D. j. 1979. Interactions between the Pompilid wasp Anoplius relativus (Pox) and the burrowing wolf spider Geolycosa domifex (Hancock). Canadian Journal of Zoology 57: 542-550. Muma, M. H. and Jeffers, W. F. 1945. Studies of the spider prey of several mud-dauber wasps. Annals of the Entomological Society of America 38: 245-255. O’BRIEN, M. F. 1982. Trypargilum tridentatum (Packard) in trap nests in Oregon (Hymenoptera: Sphecidae: Trypoxyli- nae). Pan-Pacific Entomologist 58: 288-290. O’NEILL, K. M. and Evans, H. E. 1982. Patterns of prey use in four sympatric species of Philanthus (Hymenoptera: Sphecidae) with a review of prey selection in the genus. Journal of Natural History 16: 791-801. Powell, J. A. 1964. Additions to the knowledge of the nesting behavior of North American Ammophila (Hymenoptera: Sphecidae). Journal of the Kansas Entomological Society 37: 240-258. Rau, P. 1935. The spider prey of the mud wasp, Sceliphron caementarium (Araneae, Hymen.: Sphegidae). Entomological News 46: 267-270. Rosenheim, J. A. (in press). Host location and exploitation by the cleptoparasitic wasp Argochrysis armilla: the role of learning (Hymenoptera: Chrysididae). Behavioral Ecology and Sociobiology. Strong, D. R. 1983. Natural variability and the manifold mechanisms of ecological communities. The American Naturalist 122: 636-660. Tsuneki, K. 1963. Comparative studies on the nesting biology of the genus Sphex (s.l.) in east Asia (Hymenoptera: Sphecidae). Memoirs of the Faculty of Liberal Arts, Fukui University 13: 13-78. 346 WEAVING: PREY SELECTION OF AMMOPHILA W. KIRBY (HYMENOPTERA: SPHECIDAE) Weaving, A. J. S. (1986). A comparison of nesting behaviour and prey selection in some southern African species of Ammophila (Hymenoptera: Sphecidae). M.Sc. Thesis, Rhodes University, Grahamstown. Weaving, A. J. S. (in press). Nesting strategies in some southern African species of Ammophila (Hymenoptera: Sphecidae) Journal of Natural History. White, T. C. R. 1975. A quantitative method of beating for sampling larvae of Selidosema suaris (Lepidoptera: Geometridae) in plantations in New Zealand. The Canadian Entomologist 107; 403-412. 347 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 13, JULY 1988 APPENDIX I. Species of lepidopterous and hymenopterous caterpillars collected from dominant plant species at Hilton. Family, species, (food plant) Maximum size prior to pupation mass length mg mm Relative abundance Prey of and per cent GEOMETRIDAE Aphilopota patulata (Mh) 28* 16 Ascotis selenaria (Ak,Dd,Mh) 734 47 3 — Axiodes dochmoleuca (Ct) — 26* 7 Adc 85,7 A. bifasciata (Pi) 208 35 7 Ai 23,2 Comibaena leucospilata (Dd) — 22* 24 — Ligdia pectinicornis (Ak) 52 19 10 Af 22,2 Lomographa indularia (Ak) — 18* 17 Af 0,8 Omphaliicha ditriba (Rm) 423 42 19 Ai 0,7 ? Omphalucha sp. (Rm) 117 26 22 — Prasinocyma scissaria (Ak) 69 19 21 Af 0,8 Semiothisa/Tephrina spp. (Ak) 224 25 2 Af 24,6 Sicyodes spp. (Mh) 145 29 18 — Svndromodes invenusta (Ak) 34 23 13 Af 11,1 Xylopteryx arcuata (Rm) 120 26 7 — X. prasinaria (Ml) 226 25 8 — Zamarada spp. (Ak) 74 19 6 Af 14,3 Geom 6 (Dd) 75 25 1 — Geom 12 (Lc) 138 31 24 — Geom 29 (Ml) 279 30 21 — NOCTUIDAE Anua selenaris (Dd) 674 51 24 Ai 0,7 Audea melaleuca (Ak) 832 43 18 Af 1,6 Autophlebia bracteata (Dd) 2046 49 19 Ai 1,4 Blenina squamifera (Dd) 237 20 4 Ai 15,2 Craterestra definiens (Pi) — — 7 Abr Ciicullia minuta (Ct) 479 34 7 Add 14,7 C. consimilis (Pi) — — ? Ai 17,4 Eutelia adulatrix (Rm) 348 23 15 Ai 6,5 Heliothis armigera (Ct) — 40* ? Abr Hypocala rostrata (Dd) 389 31 17 — Hypoplexia externa (Lc) 399 28 22 Ai 5,1 Ophiusa tirhaca (Rm) 1573 62 23 — Sarimarais bicolor (Rm) 580* 38* 24 — Sphingomorpha chlorea (Ak) 1820 53 21 — Thria robusta (Ak) 396 30 20 Ab/Pd Ulotrichopus catocala (Ak) 1161 60 14 — U. primulina (Ak) 426 42 24 Ai 0,7 LASIOCAMPIDAE Beralade prompta (Ak) 609 36 12 Add 47,1 348 WEAVING: PREY SELECTION OF AMMOPHILA W. KIRBY (HYMENOPTERA: SPHECIDAE) Family, species, (food plant) Maximum size prior to pupation mass length mg mm Relative abundance Prey of and per cent Bombycopsis nigrovittata (Ct) 1240 50 9 Add 2,9 Pachypasa truncata (Ak) NOTODONTIDAE 24 Morasa modesta (Rm) 581 33 17 — Porthesia subalba (Ak) ARCTIIDAE — 17* 18 — Maenas vocula (Ct,Mes) LIMACODIDAE 628 35 9 Add 26,4 Coenobasis amoena (Ak) ZYGAENIDAE — 20* 24 — Orna nebulosa (Mh) LYCAENIDAE — 12* 5 — Anthene spp. — 14* 11 Af 15,9 Azanus ubaldus (Ak) PIERIDAE — 12* 24 — Pieris helice (Ct) ARGIDAE (Hymenoptera: Symphyta) — — ? Af 0,8 Arge sp. (Dd) — 25 9 — ^estimated from photographs. Abbreviations : — Plants: Ak — Acacia karroo; Ct — Chrysocoma tenuifolia; Dd — Diospyros dichrophytla; Ly — Lycium campanulatum; Mes — “mesemb”; Mh — Maytenus heterophylla; Ml — M. linearis; Pi — Pentzia incana; Rm — Rhus macowanii. Wasps: Ab — Ammophila beniniensis; Abr — A. braunsi; Adc — A. dolichocephala; Add — A. dolichodera; Af — A. ferrugineipes; Ai — A. insignis; Pd — Podalonia canescens. 349 ji-;? I , \ / - ■:■ ** iV- ■ '4^^,-V- 1 ' ” . ^ - M ^ '■ ‘ > ■j ./ ■'iA' > , . ' 'JiChrt' #' ' ■o', 'if,' I O;. P' T ' «*•■•■ \ > Mill INSTRUCTIONS TO AUTHORS LAYOUT of recent issues should be examined for details of acceptable style and format. MANUSCRIPTS should be submitted on floppy disk accompanied by three print-outs (spacing U/i or 2). In exceptional cases typed manuscripts will be accepted. These should be submitted in triplicate, typed on A4 size paper, double-spaced, on one side of the paper only and with wide margins all round. Only generic and specific names should be underlined to indicate italics. All other marking up should be left to the editor. Footnotes should be avoided as far as possible. All measurements should be metric. Systematic papers should follow the rules of the International code of zoological nomenclature or the International code of botanical nomenclature, as applicable. ILLUSTRATIONS should bear the author’s name, title of paper and number in soft pencil on the reverse side. Diagrams, line drawings and photographs are all regarded as figures, should be termed as such and should be numbered in a single series. In addition to the original illustrations a photocopy of each should be submitted. Diagrams and line drawings should be reducible to a size not exceeding 14 x 16 cm (17,5 cm including captions). Photographs for half-tone blocks should be final size, of high contrast and printed on glossy paper. Colour illustrations are only printed under special circumstances and their financing must be arranged by the author. A scale should appear with all illustrations or the magnification should be given in the caption. A list of captions should be submitted on a separate sheet. TABLES should be submitted on separate sheets. Foldouts and tables printed sideways should be avoided. REFERENCES should be listed alphabetically at the end of the paper. Abbreviations of periodical titles should conform with those in the World list of periodicals (4th ed.). Layouts must be as follows: Alexander, J. 1969. The indirect evidence for domestication. In; Ucko, P. J. and G. W. Dimbleby eds, The domestication and exploitation of plants and animals. London: Duckworth, pp. 123-129. De Winter, B. 1969. Plant taxonomy today. S.Afr.J.Sci. 65 (8): 229-242. JuBB, R. A. 1967. Freshwater fishes of southern Africa. Cape Town: Balkema. GALLEY PROOFS are sent (when practicable) to authors for checking printer’s errors. Adjustments of language, style and punctuation and any other amendments of the original manuscript, if made at this stage, may be charged to authors. AUTHOR’S COPIES (50) are issued free of charge. When more than one author is involved the fifty free copies will be sent to the senior author. Additional copies ordered before printing are supplied at a reduced cost. BD8234 — CTP Book Printers, Cape ^ J e 1 -•■a. ■ X- I m ^ I 1 s ISSN 0570-1880 tt Annals of th^ Cape Provincial Museums Natural History Ann. Cape Prov. Mus. (nat. Hist.) Volume 16 Part 14 29th July 1988 Published jointly by the Cape Provincial Museums at the Albany Museum, Grahamstown, South Africa ANNALS OF THE CAPE PROVINCIAL MUSEUMS These Annals are published in two series, Natural History and Human Sciences, the latter series covering cultural history, ethnology, anthropology and archaeology. They are issued in parts at irregular intervals as material becomes available The primary objective of these Annals is to disseminate the results of research work undertaken by staff members of the Cape Provincial Museums. However, a manuscript by any author is eligible for publication provided it is based wholly or partly on the collection/s of one or more of these museums or if the material on which the work is based is wholly or partly deposited in one of these museums. Parts are obtainable from the Librarian of the Albany Museum. Correspondence in connection with these Annals should be addressed to the Editor, Albany Museum, Grahamstown 6140. Editor Dr F. W. GESS: 1978- Editorial Assistant Mrs S. K. GESS: 1980- A contribution to the knowledge of the taxonomy and the ethology of the genus Masarina Richards (Hymenoptera: Masaridae) by F. W. GESS and S. K. GESS (Albany Museum, Grahamstown) CONTENTS Abstract 351 Introduction 351 Taxonomic descriptions (by F. W. Gess) Masarina strucki sp. nov 352 Masarina hyalinipennis Richards 352 Masarina mixta Richards 353 Key to species 353 Some aspects of the ethology of Masarina familiaris Richards 354 Discussion 361 Acknowledgements 362 References 362 ABSTRACT Descriptions of the female of Masarina strucki sp. nov., and the previously undescribed female of M. hyalinipennis Richards and male of M. mixta Richards are given. A revised key to species is presented. A first account of the nesting of a species of the genus Masarina Richards is given. M. familiaris Richards nests in vertically presented soil in multicellular nests, attaches the egg to the cell wall and provisions with a wet pollen and nectar mixture. It forages on Aspalathus spp. (Leguminosae). INTRODUCTION The genus Masarina Richards, 1962, endemic to southern Africa, when erected included three species, two of which were known only from one sex. The present taxonomic contribution based on material collected by M. Struck in the Hester Malan Nature Reserve, Springbok and by the Gesses in the Clanwilliam district adds a fourth species, at present known from the female only, and completes the descriptions of the two species previously known from one sex only. 351 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 14, JULY 1988 Nothing has hitherto been published regarding the biology of the genus. The data presented are the results of an investigation made during a six day visit to the Clanwilliam district, western Cape Province, South Africa. TAXONOMIC DESCRIPTIONS (by F. W. Gess) Masarina strucki sp. nov. Female Black; a spot at top of tempora behind eyes, a narrow interrupted band along dorsal hind margin of pronotum, a streak on humeral angles, a spot on prepectus, a median spot on posterior half of scutellar disk and a narrow streak along baso-lateral margins of scutellum, transverse posterior bands on tergites 1-5 (somewhat widened medially and at sides), yellowish-white. Underside of flagellomeres (especially those composing club), tegulae, distal extremities of femora, entire tibiae and tarsi of all legs, orange. Mandibles in part reddish-brown. Length 7,5 mm; length of fore wing 4,9 mm. Clypeus strongly raised from sides, disk widely and deeply longitudinally depressed, ventral margin trilobed (that is, with a rounded median lobe and on each side a slightly upturned lamellate angular lobe), surface of entire clypeus markedly longitudinally aciculate; frons with its lower third raised laterally and widely depressed medially to match form of clypeus, over most of its area longitudinally aciculate like clypeus. Puncturation of thorax moderate but close, that of abdomen fine but close. Scutellum moderately raised above mesoscutum, falling vertically into a narrow, smooth anterior furrow. Propodeum with obtuse tubercles near top of angles. Tergite 6 transversely impressed distally. Shorter hind tibial spur simple. Male unknown. Material Examined: [Cape Province;] Namaqualand, [Springbok] 2917 DB, Hester Malan N[ature] R[eserve], 20.viii.l985 (M. Struck) Holotype female. Etymology: The name, in the genitive singular, is formed from the name of the collector of the present specimen, Mr Michael Struck of Hamburg, West Germany, whose help in locating good study areas in the Hester Malan Nature Reserve is hereby gratefully acknowledged. Masarina hyalinipennis Richards Female (hitherto undescribed) Black; small to minute spot on each side of face above ocular sinus, narrow streak at top of tempora behind eyes, tegulae, entire raised disk of scutellum, tergite 1 and 2 (except for anterior declivity of 1, narrow base of 2 and sides of both), a narrow transverse posterior band medially on tergite 3, parts of tibiae of all legs and entire tarsi of fore legs, various shades of red. Length 8, 7-9, 8 mm (av. of 5: 9,1 mm); length of fore wing 6, 8-7, 3 mm. Clypeus moderately raised from sides, disk a little depressed in midline, ventral margin widely and shallowly emarginate with a small, erect, rounded central tooth, lamellate margin wide; surface of clypeus dull, finely aciculate (especially in depression) with moderate punctures becoming closer at sides. Frons similarly finely aciculate (especially medially). Rest of head, thorax and abdomen moderately punctured. Scutellum falling steeply into a wide coarsely crenulate anterior furrow. Propodeal angles weakly developed, almost rounded. Tergite 6 transversely impressed distally. Generally quite similar to the male. 352 GESS: TAXONOMY AND THE ETHOLOGY OF THE GENUS MASARINA RICHARDS Material Examined: [Cape Province:] Namaqualand, [Springbok] 2917 DB, Hester Malan N[ature] R[eserve], 20.viii.l985 (M. Struck) 4 females, 28.viii,1985 (M. Struck) 2 males, 25. ix. 1986 (M. Struck) 1 female. Masarina mixta Richards Male (hitherto undescribed) Black; gastral tergites 1-2 red, contrasting markedly with rest of tergites. Antennal club beneath, parts of tibiae and tarsi (especially those of fore legs), orange. Disk of clypeus, broad supraclypeal marking on face, whole of labrum, proximal half of mandibles, a spot at top of tempora behind eyes, pronotal dorsum medially and humeral angles, yellowish-white . Length 6,5 mm, length of fore wing 4,7 mm. Like the female characterized by the broad, triangular, black tegulae; other shared characters are the simple shorter hind tibial spur, the scutellum which is anteriorly almost on the same level as the mesoscutum, and the propodeum which has obtuse tubercles near the top of the angles. Tergite 7 only weakly transversely impressed, apically rounded; proximal sternites unmodified. Puncturation similar to that of female. Material Examined: Cape Province: Clanwilliam District, Kransvlei (32° 14' 3" S, 18° 50' 49" E), 7- 13. X. 1987 (F. W. and S. K. Gess) 1 male; Clanwilliam District, 5 km W of Clanwilliam, road to Graafwater, 12.x. 1987 (F. W. and S. K. Gess) 1 female (on flowers oi Aspalathus desertorum Bol., Leguminosae); Clanwilliam District, Witelskloof (32° 20' S, 18° 48' E), 13.x. 1987 (F. W. and S.K. Gess) 1 female; Clanwilliam District, Klein Alexandershoek (32° 20' 20", 18° 46' E), 8- 13. X. 1987 (F. W. and S. K. Gess) 2 females (on ground). KEY TO THE SPECIES OF MASARINA RICHARDS The present key is based upon that of Richards (1962: 268) but is augmented by the inclusion of the female of M. strucki sp. nov., the hitherto unknown male of M. mixta Richards and the hitherto unknown female of M. hyalinipennis Richards. 1. Tegula black, broad, triangular. Thorax entirely black in female, black except for yellowish-white markings on pronotal dorsum and on humeral angles in male; markings on abdomen reddish, largely confined to tergites 1 and 2 (and in female to parts of corresponding sternites). Shorter hind tibial spur simple. Male with clypeal disk, broad supraclypeal marking on face, whole of labrum and proximal half of mandibles yellowish white; with tergite 2 unmodified and tergite 7 rounded mixta Richards — Tegula reddish, long, pyriform. Thorax with at least tip of scutellum reddish or yellow; markings on abdomen reddish, yellow, or both, not confined to first two tergites. Shorter hind tibial spur simple or bifid 2 2. Thorax with yellow markings; abdomen black with narrow yellow posterior bands on tergites 1-5. Shorter hind tibial spur simple. (Male not known.) strucki sp. nov. — Thorax without yellow markings but with at least tip of scutellum reddish and other markings, if present, of that colour; abdomen with extensive reddish markings and with yellow, if present, confined to lateral or medial spots. Shorter hind tibial spur bifid ... .3 3. Face on each side above ocular sinus with a large yellow spot; abdomen with yellow markings in addition to reddish ones. Propodeal angles strongly developed, tuberculate. 353 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 14, JULY 1988 Scutelium with a narrow smooth anterior furrow. Puncturation of head, thorax and abdomen coarse. Male with clypeal disk and narrow supraclypeal marking on face yellow, labrum black and mandibles largely reddish; with sternite 2 unmodified and with tergite 7 emarginate apically familiaris Richards — Face on each side above ocular sinus with a small to minute reddish spot (absent in male); abdomen with reddish markings only. Propodeal angles weakly developed, almost rounded. Scutelium with a wide coarsely crenulate anterior furrow. Puncturation of head, thorax and abdomen moderate. Male with clypeal disk and a narrow supraclypeal line on face pure white, labrum and mandibles black; with sternite 2 bearing a bituberculate prominence and with tergite 7 rounded hyalinipennis Richards SOME ASPECTS OF THE ETHOLOGY OF MASARINA FAMILIARIS RICHARDS Geographic distribution Masarina familiaris Richards seems to be an essentially south western Cape species with a distribution extending to the fringes of Namaqualand in the north and to Willowmore in the east. Richards (1962) gives collecting records for Camps Bay, Stellenbosch, Ceres, Clanwilliam, Nieuwoudtville, Ladismith and Willowmore. Additional records (all Albany Museum) are: Ladismith, 23. ix. 1948, 1 female; 24. ix. 1948, 1 female; and 26. ix. 1948, 1 female (all C.F. Jacot Guillarmod). Citrusdal, 2.xi.l966 (J. G. Rozen) 1 male. Clanwilliam District: Kransvlei (32° 14' 3" S, 18° 50' 49" E), 7-13.X.1987 (F. W. Gess and S. K. Gess) 9 females, 4 males. 6-9 km N of Paleisheuvel on road to Clanwilliam, 13.x. 1987 (F. W. Gess and S. K. Gess) 4 females. Clanwilliam Dam (32° 11' 30" S, 18° 53' 42" E), 14.X.1987 (F. W. Gess and S. K. Gess) 10 females, 1 male. Witelskloof (32° 20' 35" S, 18° 48' E), 13.x. 1987 (F. W. Gess and S. K. Gess) 1 male. Klein Alexandershoek (32° 20' 20" S, 18° 46' E), 28.ix.1985 (F. W. Gess and S. K. Gess) 1 male and 8-13. X. 1987 (F. W. Gess and S. K. Gess) 10 females, 5 males. 5 km W of Clanwilliam on road to Graafwater, 12.x. 1987 (F. W. Gess and S. K. Gess) 3 females. Description of the Clanwilliam district Clanwilliam lies in the Olifants River Valley with to the east the Cederberg Mountains and to the west a hilly area with beyond it the coastal plain. Apart from the sandy coastal plain the whole area is classified geologically as Table Mountain Series. The soils are derived from quartzitic sandstone and shale and are therefore a mixture of sand and clay the proportions of each varying from area to area resulting in their being of variable friability. The vegetation of the Olifants River Valley to the north of, around and some way to the south of Clanwilliam is described by Acocks (1953) as Veld Type 31, Succulent Karoo, and that further to the south extending almost to Citrusdal as Veld Type 26, Karroid Broken Veld. That of the high lying areas is Veld Type 69, Macchia (Fynbos) and that of the coastal plain is Veld 354 GESS: TAXONOMY AND THE ETHOLOGY OF THE GENUS MASARINA RICHARDS Type 34, Strandveld. Moll et al (1984) re-described the major vegetation categories in and adjacent to the Fynbos Biome. They categorize the area in the immediate vicinity of Clanwilliam a “Mosaic of Dry Mountain Fynbos and Karroid Shrublands” and the high lying areas to the west and east “Mesic Mountain Fynbos”, changing to "Dry Mountain Fynbos" further to the west on the fringes of the coastal plain. The six sites at which Masarina familiaris was collected were all characterized by the presence of some Fynbos species and, at the time when the wasps were collected, flowering Aspalathus sp./spp. (Leguminosae). Klein Alexandershoek and Witelskloof are situated in Mesic Mountain Fynbos, and the site 6-9 km N of Paleisheuvel and that 5 km W of Clanwilliam on the road to Graafwater are situated in Dry Mountain Fynbos. That at Clanwilliam Dam is a sparsely vegetated slope above the caravan park and that at Kransvlei is a transition area, the vegetation being a mosaic of Succulent Karoo, Karroid Broken Veld and Fynbos. Plants visited During the period 7-14. x. 1987 a wide range of plants in flower in the Clanwilliam district was sampled for wasp visitors. Masarina familiaris was found to be visiting Aspalathus spp. (Leguminosae) exclusively. Three species of yellow Aspalathus , Aspalathus desertorum Boh, Aspalathus vulnerans Thunb. and Aspalathus sp. (small shrub with ericoid leaves) (Fig. 1) were being visited by both females and males. Aspalathus desertorum — Clanwilliam Dam, 10 females, 1 male. Klein Alexandershoek, 9 females, 5 males. Witelskloof, 1 male. 6-9 km N of Paleisheuvel, road to Clanwilliam, 3 females. 5 km W of Clanwilliam on the road to Graafwater, 3 females. Aspalathus vulnerans — 6-9 km N of Paleisheuvel, road to Clanwilliam, 1 female. Aspalathus sp. — Kransvlei, 6 females, 4 males. Provision Provision was obtained from nine cells from four nests from two nesting sites. In all instances it was extremely wet and sticky, not forming a distinct “pollen loaf”. Pollen in all instances was of one size, 22,5 p. in diameter. When compared with that extracted from flowers of Aspalathus desertorum and Aspalathus sp. it was found to be identical. M. familiaris as a possible pollinator o/ Aspalathus desertorum The flowers of Aspalathus desertorum were observed to be visited by the masarid wasps M. familiaris and Ceramius clypeatus Richards, various bees, beetles and some non-masarid wasps, however, the two masarid wasps seemed to be the only species which penetrated deeply into the flowers and spent any considerable time working in a flower. Whilst gathering pollen but more particularly nectar from these flowers it seems likely that they serve the flowers as pollinators. M. familiaris being considerably smaller than C. clypeatus is probably the better suited having a better fit to the small flowers. Description of the nesting sites Four nesting sites of Masarina familiaris were located. Three were at Kransvlei (Figs 1, 2 and 3) and one at Klein Alexandershoek. At all sites nesting was in vertically presented soil, the 355 ANN. CAPE PROV. MUS. (NAT, HIST.) VOL. 16, PT. 14, JULY 1988 Fig. 1. A nesting site of M. familiaris Richards, a roadside bank, Kransvlei. 12, x. 1987. Positions of nests arrowed. Forage plant, Aspalathus sp., in middle distance. Fig. 2. A nesting site of M, familiaris Richards, Kransvlei, 10.x, 1987. Two turrets arrowed. 356 GESS: TAXONOMY AND THE ETHOLOGY OF THE GENUS MASARINA RICHARDS Fig. 3. A nesting site of M. familiaris Richards, Kransvlei, 10.x. 1987. Two turrets arrowed. height of the banks varying from 15-100 cm and the nests being excavated at heights of a few centimetres to half a metre. The soil of the nesting sites at Kransvlei is a very hard, non-friable red clay-sand mixture with a relatively high proportion of clay whereas that at Klein Alexandershoek is a relatively friable, sand coloured clay-sand mixture with a relatively low proportion of clay. That at Klein Alexandershoek can, however, with the addition of water be formed into durable “bricks”. The nests occurred singly and also grouped in the vicinity of an old nest suggesting that there is a tendency for newly emerged females to initiate their nests in close proximity to the nest from which they themselves emerged. Description of the nest The nest of M. familiaris consists of a multicellular burrow with at its entrance a downwardly curved tubular mud turret (Figs 4 and 5). The turret is constructed of mud pellets smoothed on the inside but left rough on the outside. A large number of the interstices are left open so that the turret has a somewhat lacy appearance. The turret and shaft entrance are both of the same diameter. There are one or more sub-horizontal to upwardly or downwardly sloping shafts each ending in a cell (Fig. 5). All those shafts leading to sealed cells are filled with earth (Fig. 5, Nests 3 and 7). A completed nest is finally sealed with a mud-plate at the burrow entrance (Fig. 5, Nest 3). A cell is, over most of its length, of the same diameter as that of the shaft. There is a distinct neck of smaller diameter than that of the cell and shaft. Distally the cell walls slope inwards abruptly to a truncate end wall. 357 ANN. CAPE PROV. MUS. (NAT, HIST.) VOL, 16, PT, 14, JULY 1988 Fig. 4. Turret of M. familiaris Richards, the nearer of the two turrets arrowed in Fig. 3, (x 2) Method of construction of the nest, oviposition and provisioning Water is required for nest construction. At an early stage in burrow excavation turret construction is initiated using pellets extracted from the excavation. At the commencement of turret construction pellets are laid down around the shaft opening in such a way that the turret will have the same diameter as that of the shaft, that is 3,5 mm (range 3,5-5 mm, sample of 6). The thickness of the turret wall is 0,5 mm (outer diameter of turret = 4,5 mm, range nil, sample of 5). Almost from the start the additional pellets are added in such a way that the turret curves over and downwards (Fig. 4). After turret construction has been completed further pellets extracted from the excavation are dropped so that they accumulate in a pile at the base of the bank beneath the nest. The shafts are short and generally slope downwards although they may less commonly slope upwards (in a sample of 15, 13 sloped downwards and 2 upwards). The average angle of slope for the sample was 26°. A shaft is extended without change of angle to end in a cell. Cell excavation is preceded by a reduction of 1 mm in the diameter of the shaft over a short distance to form a neck. After the neck has been created the diameter returns to that of the shaft until the inner end of the cell is approached so that the cell walls are parallel over most of the length of the cell. Shortly before the end of the cell is reached there is a rapid reduction in diameter so that the sides slope inwards to the end of the cell which is truncate, not curved. The average length of the cells of the sample was 11,5 mm. The excavated cell is very carefully smoothed and shaped so that, although a mud cell is not constructed within it, the walls of the cell are stabilized to such a degree that in nests constructed in relatively friable soil parts at least of the cell walls can be separated from the surrounding soil. A cell having been constructed an egg is laid at the inner end. The egg is yellow and curved, and is attached by one end to the upper side of the cell so that it lies across the end wall. 358 GESS: TAXONOMY AND THE ETHOLOGY OF THE GENUS MASARINA RICHARDS F/G Fig. 5. Vertical plans of the turrets and underground workings of nests of M. familiaris Richards, Nests 3, 8, 9 and 10, Kransvlei and Nest 7, Klein Alexandershoek. For key to lettering see Table 1. 359 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 14, JULY 1988 After oviposition provisioning takes place. The provision which is a mixture of pollen and nectar is very wet and sticky. Being wet it has no discrete shape of its own. It occupies about two thirds of the cell. Provisioning having been completed the cell is sealed with a thin mud plate and the shaft is filled with earth. A second shaft is then excavated diverging from the first immediately inside the nest entrance and terminating in a cell which is completed and sealed in the same manner. Several shafts each terminating in a cell may be similarly excavated and completed. When a nest has been completed the burrow entrance is sealed with a mud-plate. Ten nests were excavated. Of these two were newly initiated nests consisting only of a turret and short shaft, three were single-celled, two two-celled, two three-celled and one four-celled. The nature of the contents of these cells is indicated in Table 1. The cocoon spun by the fully grown larva is firmly attached to the cell walls. There was no indication of nests being used by this wasp for more than one season. Table 1. Details pertaining to the 10 nests of Masarina familiaris excavated in the Clanwilliam district. Nest No. Turret No. of cells Nature of each cell, cell contents Remarks 1 Absent 2 A Y Old nest Hoplitis sp. female in nest 2 Present 4 F/G* B/C Z Z M. familiaris female in nest Hoplitis sp. in attendance 3 Present 3 F/G*F/G"F/G* Nest entrance sealed 4 Present 0 — Newly started nest 5 Present 0 — Newly started nest 6 Present 1 H M. familiaris female in nest 7 Present 3 G G G 8 Present 2 G F 9 Present 1 H M .familiaris female in nest 10 Present 1 1 M. familiaris female in nest Key: * indicates that no egg or larva was found A. Cell open, containing old cocoon from which adult wasp has emerged. B. Cell closed, containing pupa in cocoon. C. Cell closed, containing prepupa in cocoon, F. Cell either open or closed, containing still feeding immature larva. G. Cell either open or closed, containing egg with provision. H. Cell open, containing egg without provision. I. Cell open, empty. Y. Old cell containing bee cell. Z. New cell containing bee cell. Associated insects Two of the nests of M. familiaris investigated were being attended by a small (6,3 mm long) black bee, Hoplitis sp.. One nest (Nest 1) was an old two-celled nest which lacked a turret and the other (Nest 2) was a newly constructed turreted four-celled nest (Table 1). Nest 1 was clearly 360 GESS: TAXONOMY AND THE ETHOLOGY OF THE GENUS MASARINA RICHARDS an old, disused nest of M. familiaris and Nest 2 was a newly constructed nest which was being attended by both the wasp builder and the bee usurper. The wasp cells utilized by the bee had been widened by the latter prior to the construction of its petal-cells. The petals utilized were those of a purple flowered Cyanella sp.(Amaryl- lidaceae) which was growing in the vicinity. After a petal-cell had been sealed with pieces of petal the excavated cell had been sealed with compacted soil. Pollen from the provision, a mixture of pollen and nectar, was examined and found to be a mixture derived from three or more plant species. Pollen from Cyanella was examined but did not match any of the pollen derived from the bee’s cells. Pollen is therefore being collected from different plants from that from which nesting materials are taken. Male behaviour Males of M. familiaris were observed and collected on flowers of Aspalathus spp., however, no interactions between males and females were observed. Males were not seen at nests and as M. familiaris has not yet been observed at water it is not known whether the males encounter the females when they are collecting water. DISCUSSION As nothing was previously known of the nesting of any member of the genus Masarina it seems to be of interest to compare what is now known of the nesting of M. familiaris with what is known for other masarids. The selection of a vertically presented soil surface is behaviourally distinct from the selection of a horizontally presented soil surface (Gess, 1981). The choice by M. familiaris of vertical banks for nesting thus distinguishes it behaviourally from the other masarid species known to nest in soil, 15 Ceramius spp. (Gess and Gess, 1980, 1986 and 1988), Jugurtia confusa Richards (Gess and Gess, 1980), Quartinioid.es sp. (Gess and Gess, 1985 unpublished field notes) and Paragia tricolor Smith (Houston, 1984). The construction of an entrance turret is common to all masarid species recorded nesting in non-friable soil and using water in nest excavation. The only nester in soil not known to construct a turret is Quartinioides sp. which nests in friable soil. The burrow plan of M. familiaris is unusual in that access to each cell is from a separate shaft, several shafts each ending in a cell being excavated in turn directly from the nest entrance. The burrow plan of Ceramius spp., Jugurtia confusa and Paragia tricolor is a main shaft with the cells terminating secondary shafts leading from it. Further, the excavated cell of M. familiaris differs in shape from those of the other soil nesting species investigated in that the sides are parallel and the diameter is equal to that of the shaft except for a narrow neck and a slight narrowing at the inner end which is truncate. The cell shape of the other species is somewhat ovoid with the diameter over most of the length exceeding that of the shaft and the end wall is rounded. The egg of all the ground nesting masarids investigated has been recorded as being laid loose in the cell (Gess and Gess, 1980 and 1986, and Houston, 1984) whereas the egg of the aerial nesting Pseudomasaris edwardsi (Cresson) (Torchio, 1970) and Gayella eumenoides Spin. (Claude-Joseph, 1930 as quoted in Richards, 1962) is attached to the cell wall. It is therefore of interest that the egg of M. familiaris nesting in vertical banks, like that of the aerial nesters, is attached to the cell wall. 361 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 14, JULY 1988 The provision of M. familiaris is very wet and sticky, clearly having a very high nectar content. This seems to be unusual for masarids and, as the cell walls have no obvious waterproofing, surprising. The more usual provision for masarids seems to be a rather dry or tacky loaf with a low nectar content such as has been described for Ceramius spp. (Gess and Gess, 1980 and 1986), Jugurtia confusa (Gess and Gess, 1980), Paragia tricolor (Houston, 1984) and Pseudomasaris edwardsi (Torchio, 1970). The only record of another masarid using wet provision seems to be that for Gayella eumenoides (Claude-Joseph, 1930 as quoted in Richards, 1962). Clearly it is premature to draw any conclusions from these comparative comments. ACKNOWLEDGEMENTS The authors wish to thank Messrs Daan Smit of Kransvlei and Udo Martinussen of Klein Alexandershoek for giving them permission to work on their land. Gratitude to the C.S.I.R. is expressed by F. W. Gess for a running expenses grant for field work during the course of which the present study was made. REFERENCES Acocks, J. P. H. 1953. Veld types of South Africa. Mem. hot. Surv. S. Afr. 29: i-iv, 1-192. Gess, F. W. 1981. Some aspects of an ethological study of the aculeate wasps and the bees of a karroid area in the vicinity of Grahamstown, South Africa. Ann. Cape Prov. Mas. (nat. Hist.) 14 (1): 1-180. Gess, F. W. and Gess, S. K. 1980. Ethological studies ot Jugurtia confusa Richards, Ceramius capicola Brauns, C. linearis Klug and C. lichtensteinii (Klug) (Hymenoptera: Masaridae) in the Eastern Cape Province of South Africa. Ann. Cape Prov. Mus. (nat. Hist.) 13 (6): 63-83. Gess, F. W. and Gess, S. K. 1986. Ethological notes on Ceramius bicolor (Thunberg), C. clypeatus Richards, C. nigripennis Saussure and C. socius Turner (Hymenoptera: Masaridae) in the Western Cape Province of South Africa. Ann. Cape Prov. Mus. (nat. Hist.) 16 (7): 161-178. Gess, F. W. and Gess, S. K. 1988. Further ethological notes on the genus Ceramius Latreille (Hymenoptera: Masaridae) in the southern and western Cape Province of South Africa. Ann. Cape Prov. Mus. (nat. Hist.) in press. Houston, T. F. 1984. Bionomics of a pollen-collecting wasp, Paragia tricolor (Hymenoptera: Vespidae: Masarinae), in Western Australia. Rec. West. Aust. Mus. 11 (2): 141-151. Moll, E. J., Campbell, B. M., Cowling, R. M., Bossi, L., Jarman, M. L. and Boucher, C. 1984. A description of major vegetation categories in and adjacent to the Fynbos biome. South African National Scientific Programmes Report 83: i-vi, 1—29. Richards, O. W. 1962. A revisional study of the Masarid wasps (Hymenoptera, Vespoidea). London: British Museum (Natural History). Torchio, P. F. 1970. The ethology of the wasp, Pseudomasaris edwardsii (Cresson), and a description of its immature forms (Hymenoptera: Vespoidea, Masaridae). Los Angeles County Museum Contr. in Science 202: 1-32. 362 INSTRUCTIONS TO AUTHORS LAYOUT of recent issues should be examined for details of acceptable style and format. MANUSCRIPTS should be submitted on floppy disk accompanied by three print-outs (spacing U/2 or 2). In exceptional cases typed manuscripts will be accepted. These should be submitted in triplicate, typed on A4 size paper, double-spaced, on one side of the paper only and with wide margins all round. Only generic and specific names should be underlined to indicate italics. All other marking up should be left to the editor. Footnotes should be avoided as far as possible. All measurements should be metric. Systematic papers should follow the rules of the International code of zoological nomenclature or the International code of botanical nomenclature, as applicable. ILLUSTRATIONS should bear the author's name, title of paper and number in soft pencil on the reverse side. Diagrams, line drawings and photographs are all regarded as figures, should be termed as such and should be numbered in a single series. In addition to the original illustrations a photocopy of each should be submitted. Diagrams and line drawings should be reducible to a size not exceeding 14 x 16 cm (17,5 cm including captions). Photographs for half-tone blocks should be final size, of high contrast and printed on glossy paper. Colour illustrations are only printed under special circumstances and their financing must be arranged by the author. A scale should appear with all illustrations or the magnification should be given in the caption. A list of captions should be submitted on a separate sheet. TABLES should be submitted on separate sheets. Foldouts and tables printed sideways should be avoided. REFERENCES should be listed alphabetically at the end of the paper. Abbreviations of periodical titles should conform with those in the World list of periodicals (4th ed.). Layouts must be as follows: Alexander, J. 1969. The indirect evidence for domestication. In: Ucko, P. J. and G. W. Dimbleby eds, The domestication and exploitation of plants and animals. London: Duckworth, pp. 123-129. De Winter, B. 1969. Plant taxonomy today. S.Afr.J.Sci. 65 (8): 229-242. JuBB, R. A. 1967. Freshwater fishes of southern Africa. Cape Town: Balkema. GALLEY PROOFS are sent (when practicable) to authors for checking printer’s errors. Adjustments of language, style and punctuation and any other amendments of the original manuscript, if made at this stage, may be charged to authors. AUTHOR’S COPIES (50) are issued free of charge. When more than one author is involved the fifty free copies will be sent to the senior author. Additional copies ordered before printing are supplied at a reduced cost. BD8303 — CTP Book Printers, Cape r Annals of the Cape Provincial Museums Natural History Ann. Cape Prov. Mus. (nat. Hist.) Volume 16 Part 15 29th July 1988 Published jointly by the Cape Provincial Museums at the Albany Museum, Grahamstown, South Africa ANNALS OF THE CAPE PROVINCIAL MUSEUMS These Annals are published in two series, Natural History and Human Sciences, the latter series covering cultural history, ethnology, anthropology and archaeology. They are issued in parts at irregular intervals as material becomes available The primary objective of these Annals is to disseminate the results of research work undertaken by staff members of the Cape Provincial Museums. However, a manuscript by any author is eligible for publication provided it is based wholly or partly on the collection/s of one or more of these museums or if the material on which the work is based is wholly or partly deposited in one of these museums. Parts are obtainable from the Librarian of the Albany Museum. Correspondence in connection with these Annals should be addressed to the Editor, Albany Museum, Grahamstown 6140. Editor Dr F. W. GESS: 1978- Editorial Assistant Mrs S. K. GESS: 1980- Historical records of the Wattled Crane Bugeranus carunculatus (Gmelin) in the Cape Province and the Orange Free State, South Africa R. K. BROOKE (Percy FitzPatrick Institute of African Ornithology, University of Cape Town, 7700 Rondebosch) and C. J. VERNON (East London Museum, P.O. Box 11021, 5213 Southernwood) CONTENTS Page Abstract 363 Introduction and methods 363 Findings 364 Discussion and conclusions 367 Acknowledgments 368 References 368 Appendix 1 370 ABSTRACT Records from the 19th century of the Wattled Crane Bugeranus carunculatus (Gmelin) and place names containing the element “kraan” (= crane) in the Cape Province and the Orange Free State, South Africa, are reviewed. Summer breeding is shown to have occurred in the western Cape. Breeding probably occurred in the eastern Cape. There is at least one breeding record from the Transkei. Breeding occurred along the northwestern and northern borders of the Orange Free State and probably more widely before the disturbance caused by the campaigns of Shaka Zulu. The decrease in the southern populations to the point where there is one breeding pair left is attributed to the effects of disturbance by humans and their cattle, starting with the arrival of cattle in the southwestern Cape over 1 800 years ago. INTRODUCTION AND METHODS West (1976) and Brooke (1984) pointed out that the Wattled Crane Bugeranus carunculatus (Gmelin) was said by early writers to have bred in the western and southern Cape Province, South Africa, but that it no longer did so. There was no attempt to document the former range until the sketch by Vernon and Boshoff (1986 and 1987). This paper attempts to do so in greater’ 363 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 15, JULY 1988 detail by reviewing all pre-20th century records from the Cape Colony and the Orange Free State. In addition, 10 places, the names of which include the element “kraan” (= crane in Afrikaans and Nederlands), have been visited to assess the possibility that they were named after the Wattled Crane (Lelkraanvoel in Afrikaans). FINDINGS Western Cape Province The species was made known to science by Gmelin in 1789, following Latham’s work on material from the Cape of Good Hope (Walkinshaw, 1973), but the date of collection and exact provenance of the material is unknown. Thereafter, the first to record the Wattled Crane was Andrew Smith in Roberts (1936) who said “This bird is often seen in marshy ground in different parts of the extensive flats, and appears at a distance to be larger than the Paradisea [the Blue Crane Anthwpoides paradiseus (Lichtenstein)]. It is commonly seen only alone or only with another. . . . Moults during the months of November and December. Lays during the months of November and December, one or two large oval eggs. Builds its nest of grass or reeds in the middle of fleys of water.’’. Unfortunately, the document is undated so that one cannot be sure what areas (Kirby, 1965) Smith had visited when he wrote. As Roberts (1936) remarked, most of the notes refer to the western Cape at the end of the 1820s but before that Smith had served in the eastern Cape, based in Grahamstown. The authors believe that his remarks apply to both areas and, therefore, these have been quoted under the western Cape which is dealt with first. Layard (1867: 303), apparently recognizing that this already rare species was threatened by unnecessary approach, was imprecise in his remarks “This great crane is only found in a few favoured localities, scattered over the country. A single pair take up their haunts and maintain it for years, breeding constantly in the same nest, which is repaired as occasion requires. I had the pleasure of watching a pair, through my binoculars, engaged in this proceeding: both birds contributed to the work, stopping now and then to do a little courting, . . . Still ‘something came of it,’ for my friend Hugo took two glorious eggs out of that selfsame nest, and presented them to me, and they now form part of the treasures of the South African Museum.’’. It is clear from a study of Layard (1867) that most of his fieldwork took place between Cape Town, Swellendam and Bredasdorp (c/. Winterbottom, 1968). Unfortunately, RKB has not found a reference to these eggs in the old egg catalogues of the South African Museum. The authors agree with Dr G. R. McLachlan (pers. comm.), however, that they are probably those of an unlabelled 19th century clutch in the collection. The collection site was probably Soetendalsvlei, south of Bredasdorp, in the midst of the greatest area of marshland in the southwestern Cape. Layard (1867) refers to field work at Soetendalsvlei on pages 35, 67, 311, 342 and 355. The date is given as November 1865 on page 355, indicating that this may have been Layard’s only visit. That Hugo was Mr Hugo of Franschhoek, southwestern Cape, and that he also took eggs of the rare Bittern Botaurus stellaris at Soetendalsvlei is recorded on page 311. Sclater (1906) mentions specimens from Caledon and Somerset West (the latter taken in 1857 — South African Museum catalogue) and the only record thereafter is of a vagrant seen by R. Hallack on 21 August 1938 at Villiersdorp, Worcester District (Winterbottom, 1979). The authors have been unable to trace the source of the record from Carnarvon mapped in Snow (1978) and have not included it in Fig. 1. It is suggested that it may be a misinterpretation of the Carnarvon Farm, Sterkstroom, record mentioned below. 364 BROOKE & VERNON: HIST. REC. OF WATTLED CRANE BURGERANUS IN S.A. Fig. 1. Map of South Africa showing the approximate location of 19th century Wattled Crane records in the Cape Province, Transkei and Orange Free State (solid triangles) and place names containing the element “kraan” (= crane), solid squares for probable Wattled Crane sites, open squares for probable Blue Crane sites and sites not visited. The shaded area is the post-1970 breeding and nonbreeding range after Brooke (1984). There are many wetlands in the southwestern Cape apparently suitable for the Wattled Crane and it appears from Layard (1867), supported by Andrew Smith in Roberts (1936), that some of them supported a pair of Wattled Crane. A study of farm and place names sometimes throws light on former distributions {vide Skead, 1962; Boshoff and Vernon, 1980). Appen- dix 1 lists all the place names containing the element “kraan” (= crane in Afrikaans and Nederlands) that the authors have come across in the Cape Province and the Orange Free State: there are none in the Transvaal (P. le S. Milstein pers. comm.). The great majority lie in areas where only the Blue Crane regularly occurs (Fig. 1) and presumably refer to that species. 365 ANN. CAPE PROV, MUS. (NAT. HIST.) VOL. 16, PT. 15, JULY 1988 However, two (Appendix 1) occur in upland valleys in the southwestern Cape which is typical habitat for the Wattled Crane {vide Konrad, 1981) but not for the Blue Crane, viz. Kraanfontein near the source of the Duiwenhoks River in the Riversdale District and Kraanvoelfontein by the Leeu River marshes in the Kouebokkeveld of the Ceres District, the visiting of which precipitated this contribution (Brooke, 1987). In addition, there is a possible Wattled Crane site at Kraanvlei in the wheatlands west of Moorreesburg (Appendix 1). Eastern Cape Province and Transkei William Burchell seems to have been the first to collect the Wattled Crane in the eastern Cape. Davies and Hull (1983) list two Burchell specimens of this species: Burchell nos B303 and B404; Oxford University Museum nos B/1954 & 5 respectively. Burchell’s numbering is sequential or nearly so and B308 was obtained at Graaff-Reinet (Davies and Hull, 1983 as read with McKay, 1943). It is concluded that B303 was obtained somewhere between Middelburg and Graaff-Reinet between 22 March and 12 May 1813 (McKay, 1943). The authors have found no clue to where in the eastern Cape B404 was collected or when but the species with Burchell numbers within five of B404 were taken in an area of vleis, forests and kranses. He spent most time (25 May to 21 June 1813) at Somerset East (McKay, 1943) so the second specimen may have been obtained there. However, there can be no certainty about this as Burchell remained in the eastern Cape until February 1814. Ludwig Krebs collected two specimens near Uitenhage, probably in 1821 (ffolliott and Liversidge, 1971: 42, 188, 194, 198). ffolliott and Liversidge (1971: 231) record another specimen in Kreb’s 12th collection but this is a misinterpretation of Lichtenstein’s use of the name Cracula carunculata Gmelin which is a junior synonym of Creatophora cinerea (Meuschen), the Wattled Starling. However, they interpret Lichtenstein correctly on page 238/9. The two specimens were received in Berlin in 1823 (ffolliott and Liversidge, 1971: 195) and one went to the Naturhistorisches Museum, Vienna, in 1824, being described as from “Caffernland” (H. Schifter in litt., 1987). Krauss (1973) saw Wattled Crane at Cape Tsitsikama in March 1839 and collected a specimen either there or elsewhere (Sharpe, 1894; cf. Schiiz, 1966). Holub (Holub and von Pelzeln, 1882) saw his first and only Wattled Crane on the Great Fish River north of Cradock, almost certainly in early 1879 when he spent some time in the area during which a drought broke and the river came down in flood (Holub, 1976). Layard and Sharpe (1884) record that one Wattled Crane was seen by Rickard near East London in about 1880 and they imply that Rickard had seen this species elsewhere in the eastern Cape and the Transkei. Two pairs bred in the years 1885-7 at the source of the Goqwana River (= Gqukunqa River in 3128 BA?) (Godfrey, 1923; Vernon and Boshoff, 1986, 1987), the only certain 19th century record from the Transkei. Another specimen was obtained from Cradock in 1893 according to an old catalogue in the Port Elizabeth Museum. In the 1890s Wood (1896) was told of its presence in the Aliwal North District but did not meet it himself (cf. the number of “kraan” names in this district in Appendix 1). It was collected on Carnarvon Farm east of Sterkstroom, in 1899 (Schdnland, 1900; Skead, 1967). Unfortunately, this record was erron- eously mapped in Brooke (1984) as from Douglas in the northern Cape. It was also collected at East London in 1899 and this specimen is now in the Albany Museum, Grahamstown. Davies (1907) mentions a young Wattled Crane taken from a nest near Lusikisiki, Transkei, a few years before. While this is not unequivocal evidence for a 19th century breeding site, it is highly suggestive since the Wattled Crane is site tenacious (Tarboton, 1984). It is also possible 366 BROOKE & VERNON: HIST. REC. OF WATTLED CRANE BURGERANUS IN S.A. that some of the 20th century records listed in Vernon and Boshoff (1986 and 1987) imply 19th century breeding sites but this suggestion is not pursued further in the absence of positive indications. About 1960 W. R. Siegfried (pers. comm.) saw a very large, old nest in a marsh near Swartberg {cf. Shephard, 1962), then in the Transkei but now in Natal. The site was probably in use in the 19th century. A survey of some “kraan” place names in the northeastern Karoo by CJV has revealed (Appendix 1) three places apparently named for the Wattled Crane, viz. Kraanvoelfontein by Kraankuil Railway Siding, Hopetown District; Kraanvoelvlei on the Seekoei River, Richmond District; Kraanvoelvlei at the junction of the Ongers River and the Brak River, Victoria West District. The last is the most westerley suspected Wattled Crane site in the eastern Karoo. The four degree squares 2930/2324 contain many “kraan” place names (Appendix 1) and it has not proved practical to show all of them in Fig. 1. Northern Cape Province and Orange Free State A Wattled Crane and two eggs were collected on the Vaal River at Oppermansdrif in the Bloemhof area in the winter of 1838 (Skead and Brandt, 1968; ffolliott and Liversidge, 1971: 123, 248/9, 256). However, the eggs are attributed to Ardea paradisea Lichtenstein (= the Blue Crane Anthropoides paradiseus — note that Anthropoides is masculine contrary to the usage of South African books) by ffolliott and Liversidge (1971: 258) which casts doubt on the identification. Leyland (1972) saw 60-80 Wattled Crane flying by the Vaal River in September 1850, probably between Warrenton and Barkly West. Windsorton lying between these two towns on the Vaal River and Windsorton Road on the main railway line are linked by a farm Kraanvoelvlei (Appendix 1) which is probably an allusion to the presence of Wattled Crane in the area. G. A. Phillips found a nest with two eggs about to hatch in a very large lagoon near the Vaal River (which side?) (Gurney, 1868). The South African Museum has a specimen from Harrismith taken on 4 November 1917. Snow (1978) mapped a record from west of Mafikeng but the authors have not traced the source of that record and have not mapped it in Fig. 1. Brooke (1984) mapped a record from Douglas but, as noted above, this was in error and the specimen was actually from east of Sterkstroom in the eastern Cape. DISCUSSION AND CONCLUSIONS The data presented above show that the Wattled Crane bred in the southwestern Cape, the Transkei and the northwestern Orange Free State in the last century. The eastern Cape records might all be of vagrants but this is taking an extreme position. It is much more likely that it also bred in that area as one pair still does near Ugie (Vernon and Boshoff, 1986 and 1987) and as implied by Andrew Smith in Roberts (1936). Winterbottom’s (1968) dictum in his western Cape checklist that it was only a vagrant in the western Cape is not supported. Contrary to the belief of Collar and Stuart (1985) it did occur on the peripheries of the Orange Free State historically and perhaps more widely than is presently known. However, the evidence for historical occurrence in the Orange Free State is found in obscure sources which Collar and Stuart ( 1985) could easily have overlooked when working on an Afrotropical basis. The fringe of records in the Vaal valley and in the northeastern Karoo suggests that before 1800 AD the Wattled Crane occurred in suitable habitat throughout the Orange Free State and Lesotho, linking up with the populations occurring around the main Drakensberg massif in the northeastern Cape Province, Transkei, Natal and the eastern Transvaal. There was massive 367 ANN. CAPE PROV. MUS. (NAT. HIST.) VOL. 16, PT. 15, JULY 1988 disturbance of the Orange Free State and Lesotho early in the 19th century arising from the campaigns of Shake Zulu. It seems likely that as a by-product Wattled Crane populations were exterminated or driven away so that virtually none were reported by the earliest naturalists to travel in those areas. Drainage and human disturbance (Brooke, 1984; Tarboton, 1984; detailed in Vernon and Boshoff, 1986 and 1987) have driven the Wattled Crane from all three areas studied in this paper but earlier European settlement in the western Cape probably ensured that that area was the first to lose its breeding Wattled Crane populations, probably by 1875 or 1880. However, it is possible to look further back into the history of decrease. There are many wetlands in the southwestern Cape which seem suitable habitats for the Wattled Crane. Konrad (1981) showed that its principal foods were rhizomes and bulbs of Cy penis, Eleocharis and Nymphaea. Southern Cape Province wetlands contain great numbers of Cyperus and lesser numbers of other rhizomatous plants including Eleocharis, Aponogeton and Nymphaea (Bond and Goldblatt, 1984). Thus food was not, apparently, a limiting factor in its historical distribution. However, the few published references to the Wattled Crane in the early period show that it was then rare, as stated by Latham in Walkinshaw (1973) when first describing the species in English in 1785. In the southwestern Cape the major wetlands are in coastal areas whereas in the mountains the wetlands are smaller, being often large enough for only a single pair each. Wattled Crane usually laid eggs in November and December (Andrew Smith in Roberts, 1936, cf. the argument for the Soetendalsvlei breeding site above) at a time when the winter- rain-induced vegetation of the lowlands was drying out. The peak in egg laying between May and August (Maclean, 1985) refers only to the summer rainfall areas where it now lives. Cattle had been owned by the Khoi since c 100 AD (Deacon, 1986) and these used to be driven into the marshes in summer to forage. The Khoi also camped by the marshes and foraged freely for tubers etc. in and around them (Schapera, 1930). The resulting disturbance presumably made most coastal wetlands unacceptable to the Wattled Crane (Tarboton, 1984). Thus the decrease in the Cape Province’s population of Wattled Crane probably started some 1 800 years ago and the 19th century merely saw the virtual end of a long process of local extinctions induced by sensitivity to disturbance by cattle herders and food gatherers. ACKNOWLEDGMENTS Geographical data were made available by the South African Bird Ringing Unit, University of Cape Town, and by the Surveyor-General, Cape Town, and the authors are obliged to both. REFERENCES Bond. P. and Goldblatt. P. 1984. Plants of the Cape Flora — a descriptive catalogue. J. S. Afr. Bot. Suppl. vol. 13: 1-455. Boshoff. A. F. and Vernon, C. J. 1980. The past and present distribution and status of the Cape Vulture in the Cape Province. Ostrich 51: 230-250. Brooke, R. K. 1984. South African red data book — birds. S. Afr. Natn. Sci. Progr. Rpt 97: 1-213. Brooke, R. K. 1987. Former breeding sites of the Wattled Crane (207) in the CBC atlas area. Promerops 177: 4-5. Collar, N. J. and Stuart, S. N, 1985. Threatened birds of Africa and related islands. Cambridge: International Council for Bird Preservation. Davies, C. G. 1907. Notes on birds observed and collected in the Districts of Port St. Johns, Lusikisiki, Flagstaff, and Bizana, Pondoland. during the years 1904 to 1906 and the beginning of 1907. J. S. Afr. Orn. Union 3: 180-206. Davies, K. C, and Hull, J, 1983. Burcheli’s South African bird collection (1810-1815). Arch. Nat. Hist. 11: 317-342. Deacon, J. 1986. Human settlement in South Africa and archaeological evidence for alien plants and animals. In: Macdonald. I. A. W., Kruger, F. J. and Ferrar. A. A. eds. The ecology and management of biological invasions in southern Africa. Cape Town: Oxford University Press, pp. 3-19. 368 BROOKE & VERNON: HIST. REC. OF WATTLED CRANE BURGERANUS IN S.A. FFOLLiOTT, P. and Liversidge, R. 1971. Ludwig Krebs — Cape naturalist to the King of Prussia 1792-1844. Cape Town: Balkema. Godfrey, R. 1923. The birds of Somerville. S. Afr. J. Nat. Hist. 4: 69-11 . Gurney, J. H. 1868. A ninth additional list of birds from Natal. Ibis ser. 2, Vol. 4: 460-471. Holub, E. 1976. Seven years in South Africa (2 vols) (Frewer, E. transl.). Johannesburg: Africana Book Society. Holub, E. and Von Pelzeln, A. 1882. Beitrdge zur Ornithologie Sudafrikas. Vienna: Holder. Kirby, P. R. 1965. Sir Andrew Smith, M.D., K.C.B. — his life, letters and works. Cape Town: Balkema. Konrad, P. M. 1981. Status and ecology of Wattled Crane in Africa. In: Lewis, J. C. and Masatomi, H. eds. Crane research around the world: Proceedings of the International Crane Symposium at Sapporo, Japan in 1980 and papers from the World Working Group on Cranes, International Council for Bird Preservation. Baraboo: International Crane Foundation, pp. 220-237. Krauss, F, 1973. Travel journal (Spohr, E, O. ed.). Cape Town: Balkema. Layard, E. L, 1867. The birds of South Africa. Cape Town: Juta. Layard, E. L. and Sharpe, R. B. 1884. The birds of South Africa. London: Quaritch. Leyland, j. 1972. Adventures in the far interior of South Africa. Cape Town: Struik. Maclean, G. L. 1985. Roberts' birds of southern Africa. Cape Town: Trustees of the John Voelcker Bird Book Fund. McKay, H. M. 1943. Sketch map of Burchell’s trek. J. S. Afr. Bot. 9: 27-78. Roberts, A. 1936. Some unpublished field notes made by Dr (Sir) Andrew Smith. Ann. Transvaal Mus. 18: 271-323. Schapera, I. 1930. The Khoisan peoples of South Africa. London: Routledge and Sons. Schonland, S. 1900. List of the principal accessions. In: Report of the Committee of the Albany Museum for the year 1899. Cape Colonial Gvt G.Vl900: 7. ScHuz, E. 1966. Ornithological relations between the Stuttgart Natural History Museum and South Africa. Ostrich suppl. 6: 9-12. ScLATER, W. L. 1906. The birds of South Africa, Vol. 4. London: Porter. Sharpe, R. B. 1894. Catalogue of the Fulicariae and Alectorides in the collection of the British Museum, Cat. Birds Brit. Mus. 23: 1-353. Shephard, J. B. 1962. Check list of birds of Swartberg District. 5. Afr. Avif. Ser. 6: 1-17. Skead, C, j. 1962. The use of early historical references in the study of bird and mammal distribution. Ann. Cape Prov. Mus. 2: 138-142. Skead, C. J. 1967. Ecology of birds in the eastern Cape Province. Ostrich suppl. 7: 1-103. Skead, D. M. and Brandt, H. D. 1968. Further notes on the birds of the Bloemhof District, Transvaal, Ostrich 39: 228-230. Snow, D. W. 1978. Gruidae. In: Snow, D. W. ed.. An atlas of speciation in African non-passerine birds. London: Trustees of the British Museum (Natural History), pp. 157-159. Tarboton, W. R. 1984. The status and conservation of the Wattled Crane in the Transvaal. In; Ledger, J. ed.. Proceedings of the Fifth Pan African Ornithological Congress. Johannesburg: Southern African Ornithological Society, pp. 665-678. Vernon, C. J. and Boshoff, A, F. 1986. The status and conservation of the Wattled Crane Grus carunculatus , in the Cape Province and Transkei. Bontebok 5: 37-41. Vernon, C. J. and Boshoff, A. F. 1987. The status and conservation of the Wattled Crane in the Cape Province. Pelea 6: 38-46. (A slightly simplified version of Vernon and Boshoff, 1986.) Walkinshaw, L. 1973. Cranes of the world. New York: Winchester Press. West, O. 1976. Notes on the distribution and status of the southern population of Wattled Crane in Africa. In: Lewis, J. C. and Masatomi, H. eds. Proceedings of the International Crane Workshop. Norman: Oklahoma State University Press, pp. 347-349. Winterbottom, j. M. 1968. A check list of the land and fresh water birds of the western Cape Province. Ann. S. Afr. Mus. 53: 1-276. Winterbottom, J. M. 1979. A guide to the birds of the S. VF. Cape. Cape Town: Cape Bird Club. Wood, J. 1896. The birds of Aliwal North. Part II. Sci. Afr. 1: 87-89. 369 ANN. CAPE PROV. MUS. (NAT, HIST.) VOL. 16, PT. 15, JULY 1988 APPENDIX 1 Annotated list of Cape Province and Orange Free State place names containing “Kraan”: spelling in the Afrikaans mode. Place Name Administrative District Location Groot Kraanvoelkuil Beaufort West 32° 31'S, 22° 53'E Klein Kraanvoelkuil Beaufort West 32° 27'S, 22° 57'E Kraanberg Aliwal North 30° 47'S, 26° 34'E Now known as Kramberg: there are other Kraanberg in 3026 AD, 3026 DA and 3026 DC, sometimes also called Kramberg. Kraanfontein Riversdale 33° 59'S, 21° 07'E Many small marshes along the headwaters of the Duiwenhoks River in the southern foothills of the Langeberg. Not Blue Crane habitat but possible for the Wattled Crane. Kraanfontein Ladismith 33° 41'S, 21° 29'E A hilltop subdivision of Kroonfontein! Kraankuil Calvinia 31° 30'S, 19° 30'E Now dry open karoo. Not a Wattled Crane site but probably a former Blue Crane site. Kraankuil Calvinia 3020 BD Kraankuil Elliot 31° 25'S, 27° 31'E Kraankuil Siding Hopetown 29° 53'S, 24° 12'E Apparently called after the virtually adjacent Kraanvoelfontein q.v. Kraankuil Philipstown 30° 17'S, 24° 13'E Also known as Kraanvoelkuil. Kraankuilsdam Hopetown 29° 53'S, 24° ll'E Kraanvlei Malmesbury 33° 07'S, 18° 32'E Now a seasonal marsh after 200 years of wheat cultivation but large enough, when perennial, for the Wattled Crane. Kraanvoel Siding. Bloemfontein 29° 13'S, 26° 34'E Kraanvoelfontein Ceres 32° 49'S, 19° 16'E An upland valley with extensive marshes along the Leeu River and an undoubted Wattled Crane site. Kraanvoelfontein Hopetown 29° 43'S, 24° 13'E Now a large vlei with a dam, presumably covering the original kuil (= pool). A probable Wattle Crane site. Kraanvoelfontein Philipstown 29° 50'S, 24° 20'E Kraanvoelkuil Aberdeen 32° 26'S, 23° 47'E 370 BROOKE & VERNON: HIST. REC. OF WATTLED CRANE BURGERANUS IN S.A. Place Name Administrative District Location Kraanvoelkuil road halt Beaufort West 32°31'S. 22° 53'E Kraanvoelkuil Britstown 30° 19'S, 23° 50'E Now heavily eroded but the geomorphoiogy does not suggest a former wetland on the Honderblafspruit able to support the Wattled Crane. A probable former Blue Crane site. Kraanvoelkuil Pearston 32° 37'S. 25° 04'E Kraanvoelkuil Philipstown 30° 16'S. 23° 49'E Kraanvoelkuil Philipstown 30° 17'S. 24° 13'E Also known as Kraankuil. Kraanvoelpan Prieska 29° 43'S, 22° 02'E A small pan in currently overgrazed veld. A probable former Blue Crane site but not big enough for the Wattled Crane. Kraanvoelsviakte Winburg 2827 CA Kraanvoelvlei Dewetsdorp 29° 38'S. 26° 32'E Kraanvoelvlei Ventersberg 28° 04'S, 27° 08'E Kraanvoelvlei Kimberley 28° 21'S, 24° 46'E Not visited but a probable Wattled Crane site: see text. Kraanvoelvlei Richmond 31° 23'S. 24° 19'E A section of the Seekoei River the geomorphology of the valley of which suggests that there used to be vleis along it of a size suitable for the Wattled Crane. Kraanvoelvlei Victoria West 31° OS'S. 23° 13'E The junction of the Ongers River and Brak River where the geomorphology suggests the existence of a former wetland large enough to support the Wattled Crane. Kraanvoelvlei Victoria West 31° 18'S. 23° 27'E Now flat eroded shale karoo in which the Wattled Crane could never have lived. A probable former Blue Crane site. 371 INSTRUCTIONS TO AUTHORS LAYOUT of recent issues should be examined for details of acceptable style and format. MANUSCRIPTS should be submitted on floppy disk accompanied by three print-outs (spacing V-ji or 2). In exceptional cases typed manuscripts will be accepted. These should be submitted in triplicate, typed on A4 size paper, double-spaced, on one side of the paper only and with wide margins all round. Only generic and specific names should be underlined to indicate italics. All other marking up should be left to the editor. Footnotes should be avoided as far as possible. All measurements should be metric. Systematic papers should follow the rules of the International code of zoological nomenclature or the International code of botanical nomenclature, as applicable. ILLUSTRATIONS should bear the author’s name, title of paper and number in soft pencil on the reverse side. Diagrams, line drawings and photographs are all regarded as figures, should be termed as such and should be numbered in a single series. In addition to the original illustrations a photocopy of each should be submitted. Diagrams and line drawings should be reducible to a size not exceeding 14 x 16 cm (17,5 cm including captions). Photographs for half-tone blocks should be final size, of high contrast and printed on glossy paper. Colour illustrations are only printed under special circumstances and their financing must be arranged by the author. A scale should appear with all illustrations or the magnification should be given in the caption. A list of captions should be submitted on a separate sheet. TABLES should be submitted on separate sheets. Foldouts and tables printed sideways should be avoided. REFERENCES should be listed alphabetically at the end of the paper. Abbreviations of periodical titles should conform with those in the World list of periodicals (4th ed.). Layouts must be as follows: Alexander, J. 1969. The indirect evidence for domestication. In: Ucko, P. J. and G. W. Dimbleby eds, The domestication and exploitation of plants and animals. London: Duckworth, pp. 123-129. De Winter, B. 1969. Plant taxonomy today. S.Afr.J.Sci. 65 (8): 229-242. JuBB, R. A. 1967. Freshwater fishes of southern Africa. Cape Town: Balkema. GALLEY PROOFS are sent (when practicable) to authors for checking printer’s errors. Adjustments of language, style and punctuation and any other amendments of the original manuscript, if made at this stage, may be charged to authors. AUTHOR’S COPIES (50) are issued free of charge. When more than one author is involved the fifty free copies will be sent to the senior author. Additional copies ordered before printing are supplied at a reduced cost. BD8304 — CTP Book Printers, Cape 913 YF 10/03/ ie asioi B15750 i I S'-:, r i i I ) I I \ f : I I ,1 I. i ’ ’ ' (^' ' I f ( t I ' I I i !' 1 I ^ ,.V7 SMrTHSONIAN INSTtTUTION LIBRARIES 3 9088 01 740 7321