s a Ee picid ‘ eae aaa mA ; im avin ag Syre's aia, ead tn Serene Re yA hm hen SA me ANNALS OF THE NATAL MUSEUM KDITED BY ERNEST WARREN, D.Sc.(Lonp.), Direcror. PRINTED BY ORDER OF THE TRUSTEES. LONDON: ADLARD & SON & WEST NEWMAN, LTD. BARTHOLOMEW CLOSE 1919 — ar 7 ae. _ CONTENTS. VOLUME III. Part 1, issuED SepremsBeEr, 1914. Descriptions of some South African Marine Shells. By Epvear A. Smiru, 8.0. (With Plate I) A Case of Hybridism among Cockatoos. By Ernest Warren, D.Sc.(Lond.). (With Plate IT) On South African Enne, with Descriptions of New Species and Varieties. By Henry Criirpen Burnup. (With Plates III, IV, V, and Text-figs.) On the Development of the Planula in a Certain Species of Plumularian Hydroid. By Ernest Warren, D.Sc. (Lond.). (With Plate VI and 4 Text-figs.) Note on the Occurrence in South Africa of a Termitophilous Beetle of the Genus Corotoca. By Ernest Warren, D.Sc.(Lond. ) Part 2, 1ssu—ED Octroser, 1915. Studies on the Carnivorous Slugs of South Africa, including a Monograph on the Genus Apera, and a Discussion on the Phylogeny of the Aperidz, Tyestacellide, and other Agnathous Pulmonata. By Hueu Warson, M.A. (With Plates VII-XXIV, and 7 Text-figures) The Parthenogenetic Tendency in the Moth, Melanocera menippe (Westwood). By Ernest Warren, D.Sc. (Lond. ) New South African Arachnida. By Joun Hewrrt, Director, Albany Museum, Grahamstown. (With 9 Text-figures) Observations on some South African Termites. By CiaupE Fuuuer, Division of Entomology, Department of Agri- culture, Pretoria. (With Plates XXV-XXXV and 16 Text-figures ) Gar PAGE 103 107 269 289 329 1V CONTENTS. A Further Note on Hybrid Cockatoos. By Ernest WARREN, D.Se.(Lond.) Part 3, 1issuED May, 1917. The Plant Ecology of the Drakensberg Range. By J. W. Brews, M.A., D.Se., Professor of Botany, Natal University College. (With Plates XXXVI-XXXIX and 35 Text- figures ) Contributions to a Knowledge of the Terrestrial Isopoda of Natal. Part I. By Watursr E. Coxiines, D.8c., F.L.S., etc., Research Fellow of the University of St. Andrews. (With Plates XL—XLII) . South African Bagworms. Eprror1an Note South African Bagworms ; Notes on the Psychide and on the genera Gymnelema and Trichocossus, with descriptions of five new species. By A. J. T. Janss, F.E.S.L. (With Plate X LITT) Descriptions of Some New South African Microlepidopterous Bagworms. By E. Meyrick, B.A., F.R.S. South African Bagworms: their Transformations, Life- history, and Economic Importance. Part I. By C. B. Harpenspere, M.A., Government Entomologist in Charge of Wattle Insect Investigations. (With Plates XLIV-XLVI and 6 Text-figures) Descriptions of New South African Arachnida. By Joun Hewitt, Director, Albany Museum, Grahamstown. (With Plate XLVITI and 4 Text-figures) Issuep Aprit, 1919. Title-page of Vol. IIT . Contents of Vol. TIT Index of South African Mollusea General Index . PAGE 511 619 687 Vol. ITI, part 1, issued September 25th, 1914. SOME SOUTH AFRICAN MARINE SHELLS. ] Descriptions of some South African Marine Shells. By Edgar A. Smith, 1.8.0. {Lot | LIBRARY = \Z \ =z With Plate I. GEES a a is In a collection of shells, mostly from Natal, kindly sub- mitted to me for examination by Mr. H. C. Burnup, five species appear to be undescribed, and therefore I have drawn up the following descriptions of them, and at the same time made some remarks on two known forms of the genus Tivela. Mangilia shepstonensis 7. sp. Pl. 1, fig. I. Testa parva, turrita, cerea, circa medium anfractus ultim1 linea rufa interrupta cincta; anfractus 6; superiores leeves (? detriti), convexi, czteri planiusculi, superne rotunde tabulati, costis circiter duodecim leviter obliquis instructi, liris spiralibus (in anfr. penult. tribus, in ultimo ad_ tredecim) supra costas subnodulosis cincti, inter hras strus spiralibus et lneis incrementi minute decussati; apertura angusta, spira paulo brevior; labrum extra valde incrassatum, macula rufa ornatum, prope sinum conspicuum superiorem tuberculo parvo munitum; columella rectiuscula, callo tenui induta. Longit. 44, diam. 2 mm. Hab.—Port Shepstone (Burnup). The apical whorls are large for so small a shell. Two specimens have been examined, and one of them has the spire VOL. 8, PaRT 1. | D, EDGAR A. SMITH. rather more produced than the other. The fine decussation of the surface between the cost and the spiral lire is only visible under a powerful lens or microscope. Cerithiopsis (Seila) natalensis nm. sp. PI. I, fig. 2. Testa subulata, nigro-fusca, ad apicem pallida; anfractus circiter decem, convexi, regulariter acerescentes, sutura oblqua sejuncti, liris spiralibus ad quinque instructi, liris incrementi fortibus oblique arcuatis sculpti, imter lras spiraliter tenuiter striati; ultimus infra concentrice striatus ; apertura irregulariter rotundata, intus nigrescens; columella arcuata, callo tenui reflexo induta, ad basim_ oblique truncata. Longit. 10 mm., diam. 2°5. Apertura 1°75 longa, 1°25 lata. Hab.—Tongaat, Natal (Burnup). Very like Seila attenuata Hedley from Port Jackson, but with stronger lines of growth and more equal spiral cords, and intermediate thread-like lines. On crossing the spirals these become almost granulated. The nuclear whorls are wanting in the specimen described. Glyphis levicostata n. sp. PI. I, figs. 3-5. Testa oblongo-ovata, autice leviter angustata, mediocriter elata, alba, fusco-olivaceo radiata, circa foramen interdum palide livida costulis numerosissimis vix elatis et liris in- crementi tenuibus confertis undulatis sculpta ; foramen magnum, oblongum, intus callo livido postice truncato succinctum ; superficies interna albo-pellucida lineis opacis albis radiata, ad marginem lateraliter paulo elatum tenuiter crenulata. Longit. 35, diam. 23, alt. 14 mm. Hab.—Tongaat and Port Shepstone, Natal (Burnup). This species is remarkable for the fineness of its sculpture, and the elongate cap-shaped form. The radiating riblets are very little raised, and separated merely by impressed lines SOME SOUTH AFRICAN MARINE SHELLS. 3 or strie. The concentric lines of growth are crowded and wavy, and in the grooves between the riblets there is in places a somewhat punctate appearance. Four or five concentric waves or ridges at fairly equal distances from the foramen may indicate periods of arrested growth. ‘The subapical hole is very anterior, oblique in front of the apex, and the front slope beneath it is considerably concave. Considering the lightness of the external costulation, the margin is distinctly denticulate, the grooves between the denticles extendmg within a little distance from the edge. The denticles are about eighty in number, and those upon the anterior end are often sub-bipartite. G.tenuistriata (Sowerby) 1s rather like the present species as regards form, but the sculpture is not the same. In the largest specimen examined the colour rays are very faint, probably through fading. A smaller and fresher specimen exhibits seven rays —three on each side and one in front ; the two behind slope posteriorly, four are inclined towards the front, the seventh ray extending from near the foramen to the margin down the middle of the shell. Tivela compressa Sowerby. Pl. I, figs. 6-8. i U] ,» Hg Hab.—Port Shepstone, Tongaat, Scottburgh (Burnup) ; Isipimgo (Miss Lois Trotter) ; Isezela (Miss Chapman). ‘T’. compressa was originally described from the Cape of Good Hope, and it also occurs at Muizenberg, in False Bay, and at Port Alfred. The specimens from Port Shepstone, etc., differ from the typical form in having the ventral margin less roundly curved, so that the valves have a more triangular appearance, and the height of the shell is less in proportion to the length. The pallial smus is more acuminate and generally extends beyond the middle of the valves, whereas in the normal form it hardly reaches half-way across. The reddish-brown posterior dorsal area appears to be a constant feature in this species. It sometimes attains a length of 60 min. 4A, EDGAR A. SMITH. Tivela dunker: (Romer). Play ies. 9: 10: Venus (Cytherea) (Tivela) dunkeri Rimer, Novitat. Conch. Monog. Venus, p. 5, pl. v, fig. 1. Hab.—Port Shepstone (H. Burnup). The locality of this species has not been hitherto recorded. It is very donaciform in shape, the anterior end being longer than the posterior. It is described by Rémer as pale violet- white, tinged with violet dorsally. His specimen was 25 mm. in length, whilst the valves under examination, evidently young shells, are only 15. The other 8. African Tivelas are: T. natalensis Dkr. 1864 (=alucinans Sow. 1897), T. transversa Sow. 1897, and T. polita Sow. 1851 (=dolabella Sow. 1851). Tivela rejecta n. sp. Pl. I, figs. 11-18. Testa trigona, fere equilateralis, sordida lactea rufo macu- lata, distanter livido zonata, lineis incrementi striata; mar- gines dorsales utrinque recte declives; basalis late curvatus ; latus anticum acute rotundatum, posticum magis acuminatum ; umbones acuti, purpureo tincti; lunula vix definita, fusco- purpurea tincta ; area postica fusco-purpurea ; pagina interna fusco-purpurea, ad marginem basalem albida; pallu sinus angustus, trans medium procurrens. Longit. 15:5, alt. 11°5, crass. fere 7 mm. Hab.—South Africa. The shell here described may not be adult, but it differs from the young of all the other South African Tivelas. T. polita most nearly approaches it, but that species is higher in proportion to its length, flatter, and the pallial sinus is shorter and wider. ‘lhe dorsal margins within the valves in T.rejecta are of a deep purplish-brown with pale rays beneath them, one on each side, which join the white ventral margin. ‘The reddish-brown markings on the outer surface are irregular, but disposed more or less in transverse bands. ~ SOME SOUTH AFRICAN MARINE SHELLS. 9) There are also one or two livid zones at distant intervals. There are slight differences in the hinge-teeth, which it would be difficult to describe, but which can be observed by comparing the two forms. T. dunkeri, Romer, differs in form, the posterior end being rounded and not acuminate as in the present species. Loripes burnupi nv. sp. Pl. I, figs. 14-16. Testa irregulariter ovata, mediocriter convexa, subequi- laterals, tenuis, albo-pellucida, periostraco tenui lutescente partim obtecta, lineis incrementi tenuissime striata; lunula parva, profunda; umbones supra lunulam incurvati; dens cardinalis unicus valve dextre crassus, solidus, irregularis, infra umbonem situs; dentes duo valve sinistrae, quorum posticus angustus, retrorsum inclinatus, anterior solidus, irregularis, a postico fossa lata sejunctus; lgamentum elongatum, fere internum; pagina interna obsolete radiatim tenuiter striata, in fundo cicatricem angustam obliquam exhibens; cicatrix antica, mediocriter elongata, postica elongato-piriformis. Long. 19°75, alt. 20, crass. 10 mm. Hab.—Port Shepstone (Burnup). Although the hgament is not so completely internal in this shell as is the case in typical forms of Loripes, Ido not con- sider this a sufficient difference upon which to found a distinct genus since the position of the ligament is subject to con- siderable variation. A much larger shell from the Cape Verd Islands, described by Sowerby! as Cryptodon murchlandi, externally closely resembles the present species. ‘The hinge-teeth, however, are absent, or may have become obliterated with age. The form of the anterior scar also is somewhat different. 1 Sowerby, G. B., ‘Proc. Malac. Soc.,’ vol. vii, p. 303, pl. xxv, fig. 13. 6 EDGAR A. SMITH. EXPLANATION OF PLATE I, Illustrating Mr. Edgar Smith’s paper, “Descriptions of some South African Marine Shells.” Fic. 1—Mangilia shepstonensis 1. sp. Fie. 2.—Cerithiopsis (Seila) natalensis n. sp. Fies. 3-5.—Glyphis levicostata n. sp. Fies. 6-8.—Tivela compressa Sowerby. Fies. 9,10.—Tivela dunkeri Réimer. Fies. 11-13.—Tivela rejecta n. sp. Figs. 14-16.—Loripes burnupi x. sp. Ann. Natal Mus. Vol II. Rien, or SOUTH AFRICAN MOLLU aa c bh r - i u y ’ i i ; _— oe 3 : Sah 7 iA q ; 5 i j i 7 ; , ' 7 i " 7 ‘ 7 y t ‘4 , nen hy 3 ; i ; i Vs 1 : ; ‘ ; woe 6 ! i Y : : f i ' , = : : Hy an j : ah i c's Tt ' } : rn i i : 7 = . yr if - 3 F ; ; i c ; hee i 2 Th a! on pis C8 : vi weg tT: : a ; 7 ie : i : J ; ; ; : 4 i y ‘ i : ty fi . 5 Hy i ; J ) i \s : ws ; ‘ oo ; : ¥, } i ; i f o ‘ 1 ‘ 1 o% : : Po 1 ‘ . ‘ 4 : if , i : , pil j 7 1 7 i 1 = 7 be : : ? r 7 y y 7 t i : t ew I ; a a2) i \ ’ Mn wy) F i i) dee fe — t : 1 = te Rta e i) ; : ir) : 2 ; ff ' i 4 # et ’ at r me Ty = X r i , " : r ‘ : : iL ; : i ’ ar, : ‘ { “ 4 } 1 L. : ee t i , i vw ¥ i oe i . t f ‘ i eoe : R Pay j p ir ; ; i ’ 4 Ri : ; r : i t ® yt “ he F : ( A ie ws i R : it ‘i ‘ ' . bi a ie ' v a J ‘ y na 1? ait) r : Y , 1 1 + , : ; 7 ; 7 : ; 5 ( ; 4 oy - ; J y : ‘ " a 1 4 - ; i y _ om y 7 4 ¥ f 1 1 : 7 7 : ’> 7 ‘ | bn ; a 1 ies ; ‘ i Lr ae 7 i } i, ’ : , q f : 7 r ' ” > tt ; ; ioe : , ’ ; ' ny et : tt j ‘ 1 _ my ‘ _ " u t : mg - : : 17 » ni i ; a 1 i A ny 5 7 ‘ : y ay 1 : , : i rh : ‘ 7 a! L} - . i 1 D 1 : i y é mt ( ay : i a ; ; ‘ : 7 ee i : \ W 7 7 ; i i : Y i J : i re Ty ; 2 ; i n : ae ie a ‘ha : ~~ { R ‘ t f i: : — y i ‘ ’ , : fl : i icy ; ; ie -. : ; . : ; : i banee : i r yo iM 1 ie . Be i i ; = - ; th ¢ =o ; TY ) = i Ta 7 ; i : : : HYBRIDISM AMONG COCKATOOS. 7 A Case of Hybridism among Cockatoos. By Ernest Warren, D.Sc.Lond. With Plate II. Tere has recently occurred an interesting example of hybridism among certain tame cockatoos in Pietermaritzburg, and since the character of hybrids gives some insight into the problems of hereditary transmission it appeared to be desir- able to publish an account of the matter, and the living birds have been studied as carefully as circumstances would permit. The cockatoos belong to Lieut.-Colonel A. 'T. Wales, resident in Pietermaritzburg, who has very kindly allowed me to sketch the birds and have them photographed; my thanks are also due to his daughter, Mrs. Brown, who, having considerable control over the somewhat wild birds, rendered great assist- ance while they were being photographed. My friend, Dr. Conrad Akerman, was good enough to undertake the photo- eraphing, and from the prints of the negatives thus obtained the lithographic plate was prepared. The artist who coloured the plate was provided with a selection of the variously coloured feathers. The male-bird is the common sulphur-crested cockatoo, Cacatua galerita (Lath.), and the female is the so- called slender-billed cockatoo, Licmetis nasica (/emm.). The birds were obtained on different occasions at the coast- town of Durban, and were doubtless at the time of purchase recently imported from Australia. The birds live in a garden under fairly natural conditions ; 8 ERNEST WARREN. they are mostly confined in a capacious wired-in enclosure protected from the rain, but quite frequently they escape into the garden and climb among the trees. At the present time there are five birds in the aviary ; namely, two hybrids, the parents of the same, and a second specimen of sulphur- crested cockatoo. By wire-partitions and wire-cages the birds can be separated from one another. This is particularly necessary in the case of the elder of the hybrids, which at times exhibits a violent disposition. The bird has already killed a grey parrot, and has very severely mauled the second specimen of cockatoo, which, nevertheless, is a considerably larger and more powerful-looking bird. The female-bird was purchased some twelve years ago, and the male-bird about four years afterwards. Thus the birds have been associated together for a period of about eight years. During the first two years the birds were not confined at all; they lived among the trees in a garden in a different portion of the town. Apparently no pairing occurred during this period, and no eggs were laid. Subsequently the owner moved to another residence and the birds were lodged in their present quarters. It is difficult to say whether the closer association of the two birds arising from the confinement favoured the pairing, but for the following six years (1908-1913) eggs were laid regularly every August. It is possible, however, that the production of egos by the female cockatoo does not necessarily mean that copulation invariably took place; for in this connec- tion it may be mentioned that a tame owl (Syrnium wood- fordi) in my possession occasionally produces perfectly formed egos, although the bird is unpaired ; also, a certain unpaired, tame and unidentified parrot belonging to the Misses Black- more of this town has laid eight or nine eggs annually for several years. In the case of the birds with which we are now concerned, actual pairing has been observed and probably it. occurred every season. ‘he egos are laid about the middle of August. In 1908 two egos were produced, and the owner, under the impression HYBRIDISM AMONG COCKATOOS. s) that the cockatoos would not give proper attention to incuba- tion, placed them under a hen, but the egg-shell proved to be too delicate, and the eggs were soon broken. In 1909 another two eggs were laid, and they were placed under a pigeon ; but the pigeons were not disposed to take any notice of them, and they were almost immediately smashed. In 1910 two eggs were likewise laid in the middle of August, and the cockatoos were allowed to take charge of them. ‘lhe male- and female-birds took turns in incubation, and in due time one egg hatched, while the other was addled. Colonel Wales informs me that he did not take special note of the period of incubation, but he believes that it was about twenty-one days. The egg which did not hatch was thrown away without being examined. In 1911 the female-bird dropped two eggs trom her perch and they were smashed. In 1912 two eggs were laid and sat upon by the birds, but they were addled. In 1913 three eggs were laid, and after incubation by the cockatoos one egg hatched while the other two were addled. ‘Thus in three separate years, 1910, 1912, 1913, out ot seven eges laid in all, only two hatched, and the hybrids thus arising form the subject of the present observations. It cannot be said for certain whether the eggs failed to hatch through non-fertilisation, or through the death of the embryos owing to constitutional weakness. It is clear that the pro- _portion of eggs which hatched (two out of seven) was low, and it would appear probable that this was due to the fact that the male- and female-birds were of different species. The male-bird (Pl. II, A), was a fine specimen of C. galerita; and it may be useful to quote here the description of the species given in the British Museum Catalogue.! “ A dult.—White, crest sulphur-yellow ; base of the inner web of quills and tail-feathers pale sulphur-yellow, lke the base of the feathers of the head and neck; ear-coverts some- times tinged with sulphur-yellow ; naked skin round the eyes white; iris dark-brown; bill black; feet blackish. Total 1 * Catalogue of the Birds in the British Museum,’ vol. xx, p. 117. 10 ERNEST WARREN. length 20-18 inches, wing 15-13°5, tail 8°5-7:5, bill 2-1:7, tarsus 1*15-I-1. “ Hab.—Australia and Tasmania.” Count Salvadori further remarks: “I have not been able to examine specimens from Tasmania, which, on account of some little difference in the shape of the bill, have been specifically separated as C. licmetorhyncha (Bp.).” The present specimen when standing on perch measured 17°5 inches (445 mm.) from the crown of the head to the tip of the tail. Sulphur-yellow crest well-developed, and with recurved extremity ; the longest feather measured 4°5 inches (114 mm.). Prominent projecting forehead. ‘The ear-coverts were tinged with sulphur-yellow. Exposed portion of eye, when eye-lids were open, was somewhat oval in shape, set horizontally, and measured 31 sq. mm. Hye-lids, when wide open, exhibited at their margin a circle of well-developed, symmetrical, squarish granulations. Iris very dark brown. Naked skin round the eye oval in shape, set horizontally, and measured 144 sq. mm. It was irregularly furrowed, and both it and the eyelids were of a chalky white. The bases of many of the feathers of the head and throat were sulphur-yellow in colour. Length of tail 8 inches (203 mm.) ; base of inner web of tail-feathers resembled that of the quill-feathers in being pale sulphur-yellow. Bill, black horn-colour, with the appearance of being dusted with white powder. The upper mandible from forehead to tip of hook measured 1°68 inches (42 mm.) ; depth from fore- head to lower inner angle, 1°05 inches (27 mm.). ‘he lower edge of the upper mandible, when bill was in a normal position of rest, was only about 8° from the horizontal, and was nearly parallel to a line passing through the centre of the eye and the middle point of the junction of the fore- head and bill. There is considerable variation in this matter in different individuals, as it has been observed that in some HYBRIDISM AMONG COCKATOOS. 11 specimens the lower edge slopes upwards towards the imner corner at an angle of 10°-15°. The hook of the bill was short and perpendicular. There is also much individual variation in the length and curvature of the hook ; and from Count Salvadori’s remarks, quoted above, it would appear that in 'lasmanian specimens the hook tends to be longer and more nearly like that of the genus Licmetis. The upper edge of the lower mandible is shghtly concave for its proximal two-thirds, while towards the tip it is deeply concave owing to the curvature of the bill. Nostrils surrounded by narrow naked cere at the insertion of the bill. The female-bird (Pl. IJ, B) was a normal specimen of Licmetis nasica, and the description of the species given in the British Museum Catalogue ! is as follows : “A dult.—White ; lores and a narrow frontal band red ; the feathers of the head, neck and breast are scarlet at the base, showing through the white, especially on the breast: the under surface of the wing washed with pale brimstone- yellow, the under surface of the tail also washed with brim- stone-yellow, but brighter: iris light brown; bill white ; naked skin round the eyes light blue; feet dull olive-grey. ‘Total length 15 inches; wing 11; tail 5; bill 1:95; tarsus 0°95. “Hab.—Australia, from the Gulf of Carpentaria, through the interior to N. 8S. Wales and South Australia.” ‘he present specimen when standing on perch measured 13) inches (343 mm.) from the crown of the head to the tip of the tail. The crest is in the form of a ruff consisting of the front white feathers of the crown, which are erected when the bird is annoyed. The feathers which correspond to the long recurved sulphur-yellow feathers of the crest of C. galerita are further back on the crown, and are short feathers about 15 inches (38 mm.) in length. ‘These have red bases which are not ordinarily visible when the bird is quiescent. 1 Thid., p. 134. he ERNEST WARREN. Forehead high, but with a receding aspect, and not promi- nent as in C. galerita. Kar-coverts pale brownish-yellow or buff. Exposed portion of eye, when eye-lids were wide open, was circular in shape and measured 25 sq.mm. The edges of the eye-lids did not exhibit granulations. Iris, very dark-brown, not light-brown as stated in the description quoted above. Naked skin around the eye was irregular in shape and extended much further below the eye than above it. The area was large, measuring 343 sq. mm., and grooved and furrowed ; both it and the eye-lids were of a whitish-blue colour. Lores red; a narrow red frontal band was continued from the lores above the insertion of the upper mandible, and a rather paler red coloration extended backwards both above and below the naked area around the eye. The bases of most of the features of the head, throat, neck and upper chest were bright red in colour. Length of tail 4°75 inches (120 mm.). Under surface of tail and wing sulphur-yellow. Bill whitish, the colour of boiled bone, proximally a slightly bluish tinge and distally a yellowish tinge. The upper mandible from forehead to tip of hook measured 1°8 inches (47 mm.) ; depth from forehead to lower inner angle 0°75 inches (19 mm.). The lower edge of upper mandible, when bill was in a normal position of rest, sloped upwards towards the inner corner at an angle of about 27° with the horizontal. The hook was long and conspicuously curved, and the proximal portion of the hook projected outwards and not perpendicu- larly as in C. galerita. The upper edge of the lower mandible was nearly straight proximally for about one-third of its length, while the distal two-thirds was very deeply concave. Cere feathered, and nostrils situated further in from the insertion of the bill than in C. galerita. The three years old hybrid (Pl. II, C), when standing on HYBRIDISM AMONG COCKATOOS. 13 perch, measured 15 inches (381 mm.) from the crown of the head to the tip of the tail. A crest was present consisting of short reddish-yellow or orange feathers obviously representing the recurved sulphur- yellow feathers of the crest of C. galerita. ‘These feathers were only 2°8 inches (73 mm.) in length, and they were quite straight and exhibited no tendency towards the recurved condition seen in Cacatua. Thus, in the condition of the crest, the hybrid stands between the male- and female-parent, but somewhat nearer to the male. The short feathers of the female corresponding to the crest-feathers of the hybrid are red at their bases, and the colour of the crest of the hybrid is clearly a mixture of yellow and red. The forehead of the hybrid was more prominent than in the female- and less so than in the male-parent. The ear-coverts were pale yellowish-buff and did not differ appreciably in colour from those of Licmetis nasica. The eye was large, and when the eyelids were wide open the exposed area was 34 sq. mm., thus equalling, or even slightly exceeding, the size of the eye in the male-bird. The edge of the eyelids when open exhibited granulations, but they were smaller and less regular thanin C. galerita. Iris very dark brown. The naked skin around the eye was oval in outline, and was more or less intermediate in size (208 sq. mm.) between that of the male and that of the female; but in this character also the hybrid stands somewhat closer to the male- than to the female-parent. The colour of this skin was of the palest bluish-white, which was almost exactly intermediate between the chalky white of the male and the pale blue of the female. The red colour of the lores of the female was converted into a yellowish orange in the hybrid; while the red colour of the bases of the feathers of the head and neck was replaced by a beautiful gold colour. Length of tail 7 inches (178 mm.). Under surface of tail and wing sulphur-yellow. Bill dark horn-colour, but not quite so black as in C. 14 ERNEST WARREN. galerita; the difference is more obvious when the bill is wetted. The upper mandible, from forehead to tip, measured 1-7 inches (43 mm.); depth from forehead to lower inner angle 0°93 inch (24 mm.). The lower edge of upper mandible, when bill was in a normal position of rest, sloped upwards towards the inner corner at an angle of about 19° to the horizontal. The hook was larger than in the male-bird and was more or less intermediate in shape between that of the male and female. The upper edge of the lower mandible was, on the whole, somewhat more concave than in the male-, but less so than in the female-bird. Cere partially feathered so that the nostrils, which were situated just inside the junction of the bill and head, had the back edge feathered and the front edge naked, a condition strikingly intermediate between that seen in the male- and in the female-bird. The three months old hybrid (Pl. I, D) was not sufficiently mature to render it advisable to compare its measurements with those of the parents. In all general characters it very closely resembled the mature hybrid. The orange crest was already rather longer than in the older hybrid, but none of the feathers were recurved. The area of the lores coloured yellowish-orange was rela- tively less than in the three years old hybrid, and considerably smaller than the red areain L.nasica. Naked skin around the eye very pale bluish-white. Iris very dark brown. Bill very dark, but not quite so black as in C. galerita. The shape of the upper mandible and hook of same was somewhat nearer to that seen in the male than was the case in the older hybrid. Also, the curvature of the upper edge of the lower mandible was somewhat closer to that of the male. From a Mendelian point of view animal hybrids derived from the crossing of species are as a rule singularly unsatis- factory. They usually exhibit an intimate blending in the majority of the characters of the two parent species, and HYBRIDISM AMONG COCKATOOS. 15 they are mostly sterile. The present case is no exception ; the characters of C. galerita and L. nasica have very obviously blended, and the resulting hybrid stands between the two species, but somewhat nearer to Cacatua than to Licmetis. A simple Mendehan relation would have been shown if, for example, the sulphur-yellow of the Cacatua had been dominant and the red of Licmetis recessive, and on subse- quent crossing of the hybrid (assuming that it was possible) if 75 per cent. of the offspring had been entirely yellow and 25 per cent. entirely red. But in the present case of hybridism we do not find any definite evidence of the existence of such dominance in any of the characters examined. It is true that the colour of the hybrid is nearer to yellow than to red, and that in this character there is prepotency on the part of the male-bird; but the yellow colour of the hybrid is not the sulphur-yellow of Cacatua; the latter colour has become blended with red to form an orange crest, orange lores, and a beautiful gold under-plumage to the head and neck. Again, it would not serve any useful purpose to regard the possession of a crest as a dominant character. ‘The short, straight, orange crest of the hybrid is obviously a kind of mean between the long recurved sulphur-yellow crest of Cacatua and the short, white and red feathers of the crown of Licmetis. A similar blending of characters is seen in the position of the nostrils, the condition of the cere with reference to the feathering, the colour and shape of the upper and lower mandibles, and the shape of the forehead. The characters which do not readily admit of numerical expression are grouped together in the accompanying table (p-) 116). It will be noticed that under “ Hybrid” more of the characters are found in the column “nearer Cacatua” than elsewhere, but it must be remembered that in every case the influence of Licmetis is very distinctly obvious, and, except perhaps in the character of the possession or non-possession WARREN. ERNEST © ‘pot pot nology rogram AAC ‘paaano pur suo, yooy ‘ lepue[s AjTaAry -vardutoos adeyg ‘pedojeaop I1® 4 Ayoo.re0g | “‘SUIPO00Y ‘patoyyre “Ayjavpuorpued.aed yas ‘SUOTYO «PpNSatty ony FST Pea 12 hal & ur uryy IaT[VUIS Vale pu | Q5URIO pemopoy | wpe poloyyvoy ApQAeg tepnoao ATYPON yy entq eyed £10 A YOR] ssoy nq ‘pomnopoo yarqg poamno ‘19.dU0T qeyMomos pur JLSIB14S SSeT YOoy ‘OATSSBUL sso] edvYyG IR[NSet pur POYyAVUI-[JOM SSeryT pos quriypLig QoUVL0 -YstmoTpod TING P2Mo]Od JON uLoy youpq Ay[Avjnorpued.ted qos pur ‘qyorr.4s “‘qaoys yooy eATssvut aduyg Iepnsad PUL poy.leUl-T[O AA. qUOTIMOd | poytn Ay[eyu0z -1LOY, 49S “[VAC) op AMPLY OPITPM LO MoTjes-ny dng Mo|fod-angdpag SuUUIN[Oo JouOyIppYy 80.10] peatno[oo Jo UOISSOSsO-[ [fq Jo mojog yooy Jo UWOTYIPUOD pur saTq -IpULUT 1eMOT pure toddn ayy Jo edeyg prede Jo WOIRNUR.LY) preyatoy JO soUSUIULO.A | 2100 JO SULIAY VOT ofa punoae var peyeu jo odeyg efo punore var peyeu Jo anopoy S.LOYIVOF yoou pure prey JO soseq FO amojpog qsotod JO MOTOD ‘(ST¥OULOIT) o[BMAT “STO WOT] 100BON “3 pu P m99M4oq oT BIpeUL -I9JUT SSOT LO BLOT *BNIBIBDH JOIBON *priqa "(B14 BOBO) OB) *1OJOBIBIY) Sala valy I~ COCKATOOS. AMONG 1 lanl HYBRIDIS “& BOTISVuU SIgouLoIy | | | | | | -O(| QOUALOY I(T * "9 Woda} “HI oe oe . c& GG - “ — A “ ole 08 Cr ob L 8¢ oll “WU 6T “UG fa Iz | ‘Wut $Z El “TUT & “WU ) iP “ULU f ef 6 | ‘wu EF iC i “UU | | “WU OZT “TUT BG : ce | “wm gst | = al “UU GZ ‘ur "bs eFe | ‘urat ‘bs eey 3 cg “wut “bs oz | - Tg | ‘wut ‘bs F9 | ‘ut "bs ez | ‘wut ‘bs 6 ia>} atte com -(aeout) [ey jousuoT | ES a. CP MOPMOaASCSAYeESN 22 Se py ie 160 HHI9N SOE Pe 1920 SH XH SH I0 Ye 00 E~ ) & = 2 =- = as} mts _ S ‘ ht) S., 32 S ace Orn [oF , Pe “— Mm = C= @ aaqnmnr 2) OR BS Bea Oem & A es SS) ae Seis ae -— =o Oo Si ORersias, S145 Colts iia) yen felay 9 eS) =) See & aan — _ RT le el Ob See eR ew Rm a Me Og o a) SS) XS XS| ~~) (eo Lon tI aogmea Oo A — = aS ievienthaes tat fst Paced) gery ton Ola iy Onset) Se SM Set Om ae ax} nm ¢ _ 2a = 5 a Np An 0] 8 OS Om ONO SG ry ~ roseicapilla of | Means and addition columns C. galerita gd Hybrid L. nasica 9? HYBRIDISM AMONG COCKATOOS. 23 From this point of view a glance at Table III may be of interest. In this table the characters of the different species of Cacatua, as given in the British Museum Catalogue, are rather imperfectly summarised; but we are able to see the general range of variability in the genus, with a view to obtain some insight into the hereditary tendencies of the male-parent; and also of the female-parent, since the genus Licmetis to which the latter belongs is closely related to Cacatua. For example, the general coloration in the different species of Cacatua is nearly always white, and we find as the result of the blended inheritance that the amount of coloration in the hybrid is less than half of that of the female-parent. The recurved condition of the crest is found in less than half of the species of Cacatua, and we notice that the recurved feathers of the male-parent are not so strongly unpressed on the offspring as to counteract the influence of the female-parent. Yellow or white crests occur in the majority of the species of Cacatua, and the amount of red coloration mixed with the yellow of the hybrid is compara- tively small. The feathered and non-feathered conditions of the cere are about equally divided in the different species, and in the hybrid the condition with respect to this matter is almost exactly intermediate between the two parents. The bill is more often black than light-coloured in the various species, and in the hybrid it is much nearer to black than white. In nearly all of the species of Cacatua the naked skin around the eye is oval or circular, and in the hybrid it is roughly circular. The shade of colour of the naked skin, whether white, blue or grey, varies considerably in the different species of Cacatua, and the hybrid in this character is closely intermediate between the two parents, thus indicat- ing that there is no marked preponderating influence in either direction. The general shape of the upper and lower mandibles is very 24, ERNEST WARREN. similar in all the species of Cacatua, while it is rather peculiar in Licmetis; in the hybrid the general shape is nearer to that in the former genus than to that in the latter. Lores separable from the surrounding plumage in colour are seen apparently in only three or four of the fifteen species of Cacatua, and to be comparable with the above results the colouring of the lores in the hybrid should be less than half the amount seen in Licmetis. The amount of colouring in the hybrid is certainly considerably less than in the female- bird, but it is greater than one-half, and consequently in this character there is some prepotency on the part of the mother. With reference to the absolute dimensions of general body- length, length of tail and bill, given in the second, third and fourth columns of Table ITI, it will be seen that there is great variation in the different species, and this is also the case with the relative lengths of the tail and bill. Consequently we perhaps should not expect to find any very strong prepotent action. Asa matter of fact, the absolute dimensions in the hybrid are not so very far removed from the means of the dimensions of the male- and female-parent. ‘lhe means for the parents of the three dimensions are 15°5, 6-4 and 1°74 inches respectively, and the measurements in the hybrid are 15:0, 7:0 and 1:7 inches respectively ; the male-parent is prepotent with reference to the length of tail and beak, and the female-parent with respect to body-length. The mean leneth of tail F length of body Cacatua °45, for female Licmetis °36, and for hybrid 45. Thus, in this case the prepotency of the male-bird has been ratio or all the species is ‘28, for male strong enough to counteract the more typical proportion of the tail-length seen in the female-bird. On the whole it perhaps may be said that old-established characters appearing in the majority of the species of a genus may from this very fact be regarded as strongly inheritable characters, which if present in one parent would tend to be impressed on the hybrid to the partial exclusion of any more peculiar characters that happened to be present in the other HYBRIDISM AMONG COCKATOOS. 295 parent. Thus, the former parent (Cacatua in the present case) would appear to be prepotent over the latter parent (Licmetis). There is exhibited, as a rule, no simple Mendelian relation- ship in the hybrids between animal species. The characters of the parents tend to blend, and often they blend very thoroughly as in the cockatoos: and although the hybrid may be nearer in certain characters to one parent than to the other, yet in no character ix the direct influence of either parent apparently absent. No character observed could be described as “dominant” or “ recessive,” unless these terms are used in a popular sense, and a dominant character is simply a prepotent one. It is true that the Mendelian relationship concerns characters only and not individuals as a whole. Thus, supposing that the hybrid cockatoos were capable of crossing among themselves and that the Mendelian mode of inheritance occurred, then the offspring would not be wholly like C. galerita or wholly like L. nasica, but some of them would very strongly resemble one or the other of the parent forms in some of their characters.’ According to the Mendelian conception of hereditary transmission by definite unit-characters, every such character is capable of being replaced by an alternative, and by the discovery of the alternatives (Allelomorphs) the exact nature of the units can alone be determined. A character like stature or other dimension of an animal is obviously a complex one which could not be expected to follow the Mendelian mode of inheritance as a whole, unless groups of characters are supposed to be capable of remaining connected together in some obscure manner and of being transmitted as a single character. Blended inheritance is the antithesis of exclusive inheri- tance: and in the case of the latter the Mendehan relation- ship can be readily understood on the simple and ingenious supposition propounded by Mendel of the segregation of the germinal cells. The extreme Mendelian school tend to deny the existence of real blended inheritance ; they attempt to explain 26 ERNEST WARREN. the apparent blending as arising through the character being a compound one which requires analysing into its component parts before the Mendelian relationship can be recognised. In breeding experiments for testing the occurrence of the Mendelian relationship it is most important, as was pointed out by Prof. Weldon some years ago, that the categories into which the offspring are grouped should be clearly defined. In practice it is sometimes extremely difficult to sort out the offspring into clearly defined groups, and unfortunately it may become a matter of personal judgment as to whether an individual should be placed in one or another category. Colour, independently of its distribution, may doubtless be regarded sometimes as a simple character, and if the present hybrid were fertile, and the Mendelian relationship occurred, we should have, assuming that the yellow was dominant, 25 per cent. sulphur-yellow offspring (pure dominants), 25 per cent. red offspring (pure recessives), and the remaining 50 per cent. would be yellow impure dominants. But the character of the hybrid of the first generation does not favour the view that such a result would be obtained. Neither colour is dominant, but the yellow and red have blended to form a kind of orange in which there is more yellow than red. I have not attempted to obtain a numerical expression for the amount of the two colours in the orange of the hybrid, although possibly such could be obtained. Again, the possession or absence of a_ sulphur-yellow lanceolate recurved feathered crest as seen in C. galerita might be conceivably a simple character; but the feathers in the hybrid corresponding to the recurved crest-feathers of the male are not recurved, they are relatively wider and much shorter, and they are in fact more or less intermediate in character between the feathers in the male and female. Thus in every character examined, with the possible exception of coloured and non-coloured lores, there is a very obvious blending of the male and female characteristics ; and although the external appearance of any hybrid is not to be regarded as an absolute guide to its inherent gametic HYBRIDISM AMONG COCKATOOS. 27. character, the evidence, as far as it goes, would point to a real blending of the characters of the parents, and the probable absence of so-called gametic purity, or the segrega- tion of characters in their sexual elements. If characters do not blend, and each hybrid parent, what- ever the external appearance may be, produces germ-cells in equal numbers capable of producing either one character or the other, but not a mixture of both, then the relationship found by Mendel and others would naturally follow in the offspring by the law of chance in the fortuitous union of the male and female elements. The hybrids of Mendel, De Vries, Bateson, and other observers are more generally crosses of varieties of the same species, and the reason why the crosses between different species have not been more often employed is owing to the fact that these hybrids are frequently sterile among themselves; but it might be reasonably expected that the appearance of dominant characters would be more often observable in the first generation of hybrids of distinct species, if any approach to gametic purity actually occurred in their sexual cells. If differentiation of the sexual cells with regard to alterna- tive characters of the parents really occurs in these hybrids, it is highly remarkable that the parent-hybrid itself should usually be approximately intermediate in all its characters between the two parent-species. The fact indicates that the great majority of the characters are capable of blending. Should the Mendelian relationship possess that universality which some are inclined to attribute to it, the hybrids derived from the crossing of distinct species ought certainly to afford some decisive evidence in its support. It is becoming increasingly obvious that Mendel’s law in its original sense is of quite limited application, since very many apparently simple characters appear to blend intimately, so that there can be no question of gametic purity or segregation. Into the debatable question that the blending is only apparent, owing to the characters being complex ones which 28 ERNEST WARREN. require analysing into their component parts before the Mendelian relation becomes apparent, I cannot enter here, and can but remark that this view offers a wide field for experimentation. EXPLANATION OF PLATE II, Illustrating Dr. E. Warren’s paper, ‘A Case of Hybridism among Cockatoos.” Fic. 1.—~x 3. Side view of head and neck of Licmetis nasica, female-parent. Fie. 2.— x 3. Similar side view of Cacatua galerita, male-parent. Fic. 3.— x 4. Hybrid three years old. Fie. 4.— x #. Hybrid three months old. ree eee ‘SCIMGAH YS INHHVd ATVN'2 INAHVd AIVNAA I pp Ste Eee SSE oe nee ee ooo oe SE Wel Oe SO ae tS ON SOUTH AFRICAN ENNEA. 29 On South African Ennez, with Descriptions of New Species and Varieties. By Henry Clifden Burnup. With Plates III, IV, V, and Text-figs. CoNTENTS. PAGE INTRODUCTION. : : , . 930 Systematic Account : : = 1 ol Ennea isipingoénsis Stur. end its varieties : aol Ennea isipingoénsis Stur. f. typica. Bs) var. discrepans Stun 35 var. sturanyi. n. 36 Ennea elliptica Melv. & Pons. f. typica 38 var. Manca n. 39 var. celata n. ; . 40 Ennea farquhari Melv. & Pons. f. typica : ot a var. berth (Melv. & Pons.) 44 var. avena n. : . 46 Ennea darglensis Melv. & Pons. f. typica ; . 48 var. illovoensis n. . A9 Ennea consobrina Ancey. : : ; . 00 Ennea thelodonta Melv. & Pons. : - onl, Ennea munita Melv. & Pons. : ; eon Ennea sylvia Melv. & Pons. . : es Ennea melvilli x. sp. - eyo Ennea columnella Melv. & ae £: teen ; 58 var. vitreola (Melv. & Pons) 59 Ennea marie Melv. & Pons. . : ol Ennea mooiensis 1. sp. Ox Ennea maritzburgensis Mel. & Pont. si ics . 64 var. contracta n. . 21) WS Ennea arnoldi Stur. ; : . eos 30 AENRY CLIFDEN BURNUP. Ennea connollyi Melv. & Pons. 69 Ennea inhluzaniensis x. sp.. 71 Ennea juxtidens Melv. & Pons. 72 Ennea premnodes Stur. 73 Ennea montana Melv, & Pons. 75 Ennea ponsonbyi 2. sp. 78 EXPLANATION OF PLATES 80 INTRODUCTION. In the following pages an attempt is made to clear up some of the difficulties that stand in the way of the student seeking an intimate acquaintance with the genus Ennea, as repre- sented in South Africa. Special attention is given to the comparison of kindred forms and to the consideration of the value of the differences between one form and another, whether they amount, in the opinion of the writer, to specific or varietal distinction, or are too trivial to warrant separate treatment. The result of the investigation, as far as it goes, is the recommendation that seven species and two varieties be con- sidered synonyms, and two species be degraded to varietal rank. It has been found necessary to differentiate six varieties of known species, and four species that do not seem to have been described. The alterations involved are as follows: Ennea laby- rinthea, microthauma, hypsoma, oppugnans, and periploca of Melv. & Pons. to become synonyms of Ennea farquhari Melv. & Pons.f. typica; Ennea callista Melv. & Pons. to become a synonym of Ennea marie Melv. & Pons. ; Ennea parallela Melv. & Pons. to become a synonym of Ennea montana Melv. & Pons.; Ennea isipingoénsis Stur. var. cylindrica Stur. to becomea synonym of Ennea isipingoénsis f. typica; HEnnea isipingoénsis var. simillima Stur. to become a synonym of Ennea isipingoénsis var. discrepans Stur.: Hnnea berthe Melv. & Pons. to become a variety of Ennea farquhari Melv. & Pons.; and Ennea vitreola Melv. & Pons. to become a variety of Ennea columnella Melv. & Pons. ON SOUTH AFRICAN ENNEZ. D1 The new varieties introduced are as follows: Knnea isi- pingoénsis Stur. var. sturanyi; HE. elliptica Melv. & Pons. var.manca; EH. elliptica Melv. & Pons. var. celata; BK. farquhari Melv. & Pons. var. avena; H. darglensis Melv. & Pons. var. illovoensis; EK. maritzburgensis Melv. & Pons. var. contracta. And the species, described as new, are: Ennea melvilli, E. mooiensis, KE. inhluza- niensis, H. ponsonbyi. It has also been thought desirable to re-describe and re- figure a few forms whose original descriptions and figures appeared to be erroneous or insufficient in detail, or had been published in works not easy of access to South African readers. The difficulty presented by the writer’s inability to examine the types in the British Museum has largely been overcome by the readiness of friends on the spot to compare them with specimens and drawings sent for the purpose, and to report upon such coincidence or disagreement of character as might be observed. ‘To those ready hetpers the writer’s most cordial thanks are extended, and their special services are duly acknowledged under the heads of the several species involved. No less are sincere thanks due to those who have supplied material for study, prominent among whom are Messrs. J. H. Ponsonby, J. Farquhar, and A. J. Taynton. The types of the species and varieties described as new, together with a few of the other shells figured herein, will be deposited in the British Museum. A reference to the explana- tion of the plates at the end of this paper will show the resting-place of each shell figured. Systematic Account. Ennea isipingoénsis Sturany; and its varieties. Through the kindness of Dr. Rudolf Sturany in lending me his types of this species and its three varieties described at the same time, I have been enabled to acquire a better on HENRY CLIFDEN BURNUP. knowledge of the group than has hitherto been possible. After a careful examination of the original specimens, and a close comparison of them with about 140 other examples, collected principally by myself in various parts of the province of Natal, and three specimens collected by Mr. Farquhar in the province of the Cape of Good Hope, I am convinced that two of Dr. Sturany’s varieties are superfluous, and that another form, hitherto undescribed, requires to be differentiated as a distinct variety, in order to make the group, so far as it is yet known, fully understood. In justice to Dr. Sturany, and in order to show how the mis- takes have arisen, it 1s necessary to explain that the original material was very meagre, and insufficient for forming a just estimate of the value of the apparent differences; for Dr. Penther only found four specimens, which, appearing to show sufficient differences, were described provisionally as the type and three varieties. Moreover, the shells were very dirty, and, possibly through the fear of breaking unique specimens, had never been properly washed, for they reached my hands much obscured and disfigured by sand and leaf- mould, ‘The specimen which was described as the variety cylindrica was calcined and had a large hole in the body- whorl, and to wash it would probably have led to its destruction; so I was reluctantly obliged to leave it as it was, except that with a damp brush I carefully removed enough of the mould from its aperture to expose the small basal tooth, which is said in the description to be absent, and to show more clearly the tooth within the outer lip, and the columellar plait, which are said to be respectively weaker and not very broad, to be practically identical with those of the typical form. The type of var. discrepans had the apex crushed in, but its general condition was such as to justify washing, though not syringing: the other two specimens were sound. The fragile condition of the specimen described as var. cylindrica has prevented my re-figuring it and checking its dimensions, but new figures of the other specimens are ON SOUTH AFRICAN ENNES. ot given, and their dimensions as taken by me will be found in the text. The original figure of the typical form does not require to be replaced, except for the convenience of having all the figures to be compared of one magnifica- tion. The original description adequately differentiates the species, but, for the sake of convenience in comparison, it has been deemed advisable to re-describe it here, where also will be found an amended description of var. discrepans Stur., and an original description of a new variety with which I have much pleasure in associating the name of the learned author of the species, who has rendered me such material help in the present study. Knnea isipingoénsis Sturany, f. typica. PI. III, fies I Ennea isipingoénsis Stur., Anz. k. Akad. Wissensch. Wien (1898), No. xvi, p. 157 (reprint p. 5); Sturany, Siidafrik. Moll. (1898), p. 27, pl. i, fig. 19. Ennea isipingoénsis var. cylindrica Stur., ibid., ibid., fig. 22. Ennea isipingéensis Stur.; (sic) Connolly, Ann. S. Af. Mus., vol. xi (1912); p. 77. Shell small, cylindric-oval, tapering towards the blunt apex, nearly white, umbilicate ; whorls 74, very convex, separated by very deep, almost channelled sutures, and very strongly transversely ribbed except the embryonal whorls, which are smooth, and immediately behind the peristome where the ribs become finer ; the last whorl is thickened at the base, forming a callosity almost encircling the umbilicus, and then becomes constricted before expanding to form the peristome ; aperture small roundish and furnished with the following plaits: A strong parietal plait arising higher on the body-whorl than the peristome and passing thence, embracing the parietal wall, to the interior of the shell; a long, strong squarish- pointed, tooth-like plait, almost bifid, arising near the margin and about the middle of the labrum, extending nearly to the VOL. 3, PART l. 3 34 HENRY CLIFDEN BURNUP. centre of the aperture, and the lower edge of it penetrating furthest inwards ; near the base of the columellar lp, a small conspicuous tooth; at the base, and set further back than the two processes last named, an inconspicuous denticle ; and a deep-seated, scoop-shaped columellar plait with mamillated point on the inner edge presented forwards ; peristome white, broad, expanded and reflexed, the ends connected by a shght callus. Height 2°83, width 1:47 mm. Hab.—lIsipingo, near Durban (Penther) ; other localities : Natal: Karkloof and Nottingham Road (Taynton), Lower Umkomaas, Umbogintwini, Ntimbankulu—Mid-Illovo, Maritz- burg, Edendale, Dargle and Inhluzani (Burnup). Cape of Good Hope: The Gorge, Van Staaden’s River (Farquhar). The umbilicus is rather widely open for so small an Ennea, and, as is usual in openly umbilicated shells of the genus, has a branch of the perforation passing into the columellar fold. In the type the callus connecting the ends of the peristome is not very clearly defined, and I can readily understand the author having overlooked its presence before the shell was cleaned. In most examples it is much better developed, sometimes obliterating the sculpture underneath, but in many instances, especially in rather young, though mature shells, allowing the ribs to be seen through it. On the outside of the shell can be seen, not only a large deep cavity corresponding with the strong labral plait, but also smaller, shallower pits marking the positions of the basal denticle and the basi-columellar tooth. The distribution, including a locality so remote as Van Staaden’s River in the Cape Province, is worthy of note: the specimens from the last-named locality are quite normal. Since I have shown the type of the variety cylindrica to be identical in its characters with the type of the species except in size and slightly in form, it becomes inadmissible to recognise it as a distinct variety. The dimensions given by the author show the two shells to be identical in the ratio of ON SOUTH AFRICAN ENNEA. 23) width to height, though to me cylindrica seems per- ceptibly narrower. The following table of dimensions of a few shells that I have measured will indicate the futility of any attempt to identify a variety based on size and form alone. Specimen. | Height. Width. | a | ele ys : Te | mm. mm E. isipingoénsis (type). Isipingo 2°83 1-47 5194 | var. cylindrica (original | measurement). Isipingo 3°00 1:50 50°00 | oat 2°7. 166 60°58 | | Inhluzani : ; , | 3.97 | 169 | 5168 | | 318 185 | 09818 Ntimbankulu . : -5| 364 E86. |) okl0 | 3:94 1-73 43°91 Edendale ; . 2°99 148 | 49°50 Umbogintwini Mr hob icy! 142 | 4948 | | Lower Umkomaas . |; 3:80 1-74 | 45°79 | | | | | Average... yea * | 3:23 | 1-64. | 51:22 | | The ratios given in the fourth column show great variability even in specimens from the same locality, e. g. Ntimbankulu. Var. discrepans Sturany. PI. III, figs. 2, 3. Ennea isipingoénsis Stur. var. discrepans Stur., Anz. k. Akad. Wissensch. Wien (1898), No. xvi, p. 157 (reprint p. 5); Sturany, Siidafrik. Moll. (1898), p. 27, pl. i, fig. 20. Ennea isipingoénsis Stur. var. simillima Stwr., Anz. k. Akad. Wissensch. Wien (1898), No. xvi, p. 157 (reprint p. 5); Sturany, Siudafrik. Moll. (1898), p. 27, pl. i, fig. 21. Shell like E. isipingoénsis, typical, but with the following differences: The contour is more oval and less cylindrical ; the aperture is larger and rather oblique and triangular ; the processes of the aperture are similarly placed but differ in the following respects: The labral plait is much enlarged, being both longer and wider, and bears a distinct, though minute, denticle on the upper edge; the minute basal denticle is replaced by an inconspicuous, scarcely raised, minute ridge, 36 HENRY CLIFDEN BURNUP. whose presence is more easily detected by the corresponding depression and callosity behind the basal lip than by its appearance within; and the columellar plait is much swollen. The labral and columellar plaits are so long that they overlap, the former in front of the latter. Height, without apex, which is crushed into the shell, 2°69, width 1°45 mm. Hab.—Isipingo, near Durban (Penther). Other localities : Umbogintwini and Equeefa (Burnup). This variety, including var. simillima Stur. (fig. 3), which is identical in all essential characters, seems to offer nearly as much variation in size as the typical form, the dimensions of the only five specimens that I have seen being as follows : | | | Specimen. Height. | Width. | SS a | mm mm t _ Type of var. discrepans from Isipingo . : aed : é 2°69 1:45 53°90 | Type of “var. simillima” from | | | Isipingo . : : ,; 2°72 148 | 5441 | Specimen from Umbogintwini . 2°78 1:47 || = 5988 : : 3°49 eal gy, DUE Specimens from Equeefa . { 3°56 178 | 50-00 | ats yp RANGTABEL Baye! Via res casa EOD 159 | 52:38 | The relative width to height is much more regular in this variety than in the typical form, so far as can be ascertained from so small a series. Var.sturanyin. PI. III, figs. 4, 5, 6. Shell hike E. isipingoénsis, typical, but differing in the following respects: The labral plait is both longer and wider and is more compact, ending in a blunted or rounded point extending beyond the centre of the aperture; the columellar plate is also larger and extends to the centre of the aperture, thus passing slightly behind the point of the ON SOUTH AFRICAN ENNE®. Bi labral plait ; and the basal denticle is replaced by an incon- spicuous, scarcely raised, minute ridge like that of var. discrepans. It differs from var. discrepans in being more cylindrical, in the labral process being less massive and more pointed, and in the total absence of the conspicuous denticle on the upper edge of the labral plait. Height, 2°95, width, 1°55 mm. Hab.—Ntimbankulu, Mid-[lovo (Burnup). Other locali- ties: Karkloof (Taynton), Maritzburg, Howick and Dargle (Burnup). The following dimensions of shells that I have measured will give an indication of the extent of variation in size and proportion that this form is hable to : | Specimen. | Height. | Width. | a aee 38 = zy e ; ; | mm. | mm. Type of var. from Ntimbankulu 2°95 1:55 52°54 2°57 1:48 57°59 | 2°80 152 | 5429 | 2°86 | 148 | 51°75 |Specimens from Dargle . : Zo alee ACE as oe) 3:13 152 | 49:20 3:03 147 | 48:18 |. 325 drole We. 4Gsay vere te MAS Yon Role O94 150 5144 It will be seen from the above table that, while this variety is less stable in ratio of width to height than the variety discrepans, it does not exhibit so much variation in that respect as isipingoénsis, typical. It will also be seen that, though some individual specimens of the other forms may be smaller than some examples of this, in striking an average of the shells measured, sturanyi is the smallest, and isipin- goénsis, typical, the largest. The nature and disposition of the armature of the aperture in the several forms remain remarkably constant, no inter- 38 HENRY CLIFDEN BURNUP. mediate forms having been observed in the very large series examined. Dr. Sturany has kindly perused my MS. and compared my figures with his original shells, and concurs in the elimination of the two varieties, simillima and cylindrica. Ennea elliptica Melv. & Pons. f.typica. PI. III, fess 3) 9; .95,10; ale do: Ennea elliptica Melv. & Pons., Ann. and Mag. Nat. Hist., vol. ii (1898), p. 126, pl. vii, fig. 2; Connolly, Ann. 8. Af. Mus., vol. xi (1912), p. 73. In the original figure, which fairly well represents the characters of the peristomatal dentition, omitting, however, the internal columellar plait, the aperture is drawn out of all proportion to the rest of the shell. It is shown to occupy about 35 per cent. of the whole height, whereas in normal specimens it does not occupy more than about 28 per cent. This would make it appear either that the type is a malformed shell or that the figure is bad. As the authors had before them, at the time the description was written, four specimens agreeing in detail, it is not likely that the type is abnormal ; therefore it is to be inferred that the figure is incorrect. In any case it seems desirable to publish a new figure, which is given on PI. III, fig. 7. This drawing represents a normal shell from the original locality in the Botanical Gardens,! Maritzburg. Fig. 8, drawn from another specimen from the same locality, represents the right side of the shell and shows the striate sculpture on the last half whorl. This species is very interesting in exhibiting, through large series from different localities, the undoubtedly close relation- ship between forms which, but for the intermediates, might possibly have been taken to be distinct species. 1 The Botanical Gardens, Maritzburg, as a faunistic locality, must not be looked upon with suspicion as indicating the probability of molluses having been introduced with plants from a distance. The spot indicated is a bush-clad hill within the boundaries ‘of the gardens but not dis- turbed by cultivation. ON SOUTH AFRICAN ENNES. 39 In the bush skirting the beach near the mouth of the Tongaat River, the only place near the coast where I have met with this species, is found a form, quite typical except that it is rather small and has the peristome rather more thickened, and the processes very strongly developed. Two examples of this form are shown in figs. 9 and 10. Next comes the typical form from Maritzburg at an eleva- tion of a little over 2000 feet (fig. 7). At Dargle, at an altitude of about 3700 feet, the shells are barely larger than those at Tongaat, and the lower tooth on the columellar lip has deteriorated into a mere inconspicuous callosity, as seen in figs. 11 and 12. The next stage is seen in the shells found at Fort Notting- ham and Curry’s Post (altitude about 4000 feet), in which the height is rather greater, and the lower tooth of the columellar lip is suppressed altogether. I think this form may con- veniently be regarded as a new variety, differentiated as follows. Var. manca w. Pl. II fies: 135 TA: Shell like elliptica, typical, but generally rather larger, comparatively narrower, and without the lower tooth on the columellar lip. Height 3°76, width 1°74 mm. Hab.—Fort Nottingham ; also Curry’s Post, Natal (A. J. Taynton). In a few specimens from these localities an almost imper- ceptible callosity may be detected in the position of the lower tooth of the columellar lip of elliptica, typical ; but in by far the greater number there is no trace of it. There are also to be found in some specimens, especially those from Curry’s Post, shght traces of the sculpture distinguishing the variety cwelata, hereafter described, for the most part only infrasutural, but occasionally extending across the whorls. The Curry’s Post specimens are rather wider than those from Fort Nottingham. Yet another form, varying from the type im respects not 40, .HENRY CLIFDEN BURNUP. affecting the gradual evolution from the Tongaat form to the variety manca, that I have endeavoured to trace out above, occurs at Eshowe, Zululand. Mr. Ponsonby has placed in my hands for determination three examples of this form, which I propose to distinguish as a separate variety, celata. Var -celata mn... Pl My figs: 15; 16; 17. Shell like elliptica, typical, but rather larger, more obese, with peristome rather thinner and less reflexed, covered all over, except the earlier whorls and the body-whorl above the aperture, with most delicate, fine, oblique striz becoming stronger on the last half-whorl, where, however, they are not so strong as those in a similar position in the typical form. Height 4°06, width 2°11 mm. Hab.—Eshowe, Zululand (Ponsonby). The other two specimens are not so obese as the type of the variety, which was chosen as being the best developed speci- men. ‘Thus, while they agree better with elliptica, typical, in contour, they are more divergent in their still greater height and yet thinner peristome. In all three the distinctive sculpture is clearly seen by the aid of a strong lens. The following additional measurements will help to convey an idea of the variation as to size : Specimens from various localities. Height. | Width. ee | za petanna ed mm. | mm. ; 4-05 1°82 | F. typica, Maritzburg \ ee ree | ( 3°60 1:73 fic7 | ( 3:17 163 fiegs4 F. typica, Tongaat : 314 i Gepit saheayel (Ug ; () 311 165 | 3:38 1-64. fig. 11) | | F. typica, Dargle R 317 1-66 fig, 12, - | 3°00 1-66 | Var. manca, Fort Nottingham 4°12 1-79 fig. 14 , 3°93 1:92 Var. manca, Curry’s Post { 3-9] 1-90 4 9. Var. celata. Eshowe { ane an | ON.SOUTH AFRICAN ENNEA. Al In considering the figures illustrating the various forms of this species, it should be borne in mind that they are not all drawn to the same scale ; therefore it is the more desirable to keep in view the dimensions given above. Ennea farquhari Melv. & Pons.f.typica. Pl. IV, fig. 24. Ennea farquhari Melv. & Pons., Ann. and Mag. Nat. Hist., vol. xvi (1895), p. 478, pl. xviii, figs. 3-5. Ennea labyrinthea Melv. & Pons., ibid., p. 479, pl. xviii, figs. 7, 8. Ennea microthauma Melv. & Pons., ibid., vol. iv (1899), p. 194, pl. iii, fig. 1. Ennea hypsoma Melv. & Pons., ibid., vol. iv (1909), p. 488, pl. viii, fig. 7. Ennea oppugnans Melv. & Pons., ibid., p. 488, pl. viii, fig. 8. Ennea periploca Melv. & Pons., ibid., p. 489, pl. viii, fig. 10. Ennea farquhari M. & P.; Connolly, Ann. S. Af. Mus., vol. xi (1912), p. 74. Ennea microthauma M. & P.; Connolly, ibid., p. 80. Shell minute, ovate-cylindrical, with rounded apex, rimate and narrowly perforate, thin, vitreous, transparent and shining; whorls 53, moderately ventricose, the first two smooth, the rest sculptured with regular transverse strie, strong below the suture and evanescent lower on the whorls, except the last, on which the sculpture continues round the base; suture moderately impressed; aperture triangularly auriform with thickened, expanded and reflexed, white, rather vitreous peristome, and furnished with the following processes : a prominent blade-hke, in-running, parietal plait, a massive, blunt, lobed, labral tooth, a small internal basal tooth, an inconspicuous, broad, slightly raised tooth on, and parallel to, the columellar lip, and a deep-seated, rounded, flattish, scoop-shaped columellar plait. Height 2°11, width 1:03 mm. _ Hab.—Grahamstown (Farquhar and Langley) ; also High- lands (19 miles. N. of Grahamstown) and _ Bathurst (Farquhar), all in the Cape Province. The locality, Bosch- 42, HENRY CLIFDEN BURNUP. berg Mountain, Somerset East, given in the description of Ennea periploca, is erroneous (see p. 61). The right side of the parietal plait, which is very promi- nent, arises higher on the body-whorl than the labrum; the upper lobe of the bipartite labral tooth is a little within the plane of the aperture, the lower is more internal; the short, broad tooth on the columellar lip is nearly level with the general plane of the peristome ; and the small basal tooth stands well back from the edge. A narrow channel enters the columellar plait from the rimal cavity, and the teeth on the labrum. columellar lip and base, have corresponding excavations behind the peristome. The types of all the so-called species here placed in the synonymy, being in the British Museum, are not available to me for examination, but I have been fortunate in securing, from the collections of Messrs. Ponsonby, Farquhar, and Langley, a very large series comprising co-types of every form, except that which was described as E. labyrinthea. These I have submitted to a searching study and careful comparison, and cannot doubt that all belong to one variable species, whose divergence from a common type is not wide enough to justify the retention of any of the later names in a varietal capacity. The form described as HEnnea berthe Melv. & Pons., as represented by my co-type, ex coll. McBean, supported by seven other specimens from three localities in Natal, which undoubtedly also belongs to the same species, is specially considered hereafter. Of the form described as HE. labyrinthea, alone, I have been unable to obtain any reputed example, and so have only had the original description and figure to guide me in uniting it with EH. farquhari. Melvill and Ponsonby not being able to concur in my views, and I being unable personally to consult the types, and hesitating to make such a sweeping consignment to synonymy without reference to them, Mr. Ponsonby most kindly arranged, through Major Connolly, with Messrs. E. A. Smith and G. C. Robson, of the British Museum, that the two last- ON SOUTH AFRICAN ENNE. 4S named should examine my numerous camera drawings of all the forms under consideration (except, of course, laby- rinthea), compare them with the types in the British Museum, and generally consider the case of synonymy that I had made out. Their deliberations have resulted in the following decision, dated at the British Museum (Natural History), Cromwell Road, December 6th, 1912: “We have examined the seven species named below,! as requested by Major Connolly, and after careful consideration have come to the conclusion that, making due allowance for variability, the species in question are all indistinguishable from E. farquhari save in a few unimportant characters, the general characters of the aperture and dentition remain- ing the same in all seven. (Signed) G. C. Rosson, HK. A. Smira.” I am deeply indebted to these gentlemen, for without their co-operation my conclusions must have remained unconvincing through the possibility of the material studied by me being wrongly identified. . Major Connolly has been good enough to supply me with revised measurements of all the types, which in some cases vary considerably from those originally published; but it seems to me more useful to give here a selection of dimen- sions of specimens from the various localities, than to emend those of the specimens originally described under names which it is sought to show should drop out of use. To this resolve, however, it may be well to make exception in the cases of the original type of HK. farquhari, and the shell originally described as E. labyrinthea—the former because it is the persistent type, and the original dimensions are manifestly wrong, no specimen in the large series examined approaching such an attenuate form as they indicate ; and the latter because Major Connolly’s measurement, if correct, 1“. farquhari, E. microthauma, E. hypsoma, E. peri- ploca, E. oppugnans, EH. berthe, E labyrinthea.” 4.4, HENRY CLIFDEN BURNUP. removes from the record dimensions that are suspiciously under-estimated. His measurements are as follows : EK. farquhari M. & P. (type), height 2°80, width 1°50 mm. E. farquhari (type of “labyrinthea”), height 2:05, width 1°20 mm. ' Both of these shells were collected at Grahamstown. The dimensions in the following table are selected from those of over forty specimens that I have measured, from the various localities in which the species in its typical form is known to live: | Locality. | Largest. Intermediate. Smallest. mm. mm. mm. | Grahamstown 2 : -| 8:35>x<1:07 | 2°20x1:04 | 1:89><0:99 | Highlands, about 19 miles N. | | | of Grahamstown ; .| 2:141:05 | 2:°06x1:04 | 1941-00 | Bathurst, about 25 miles | | §.H.of Grahamstown .| 319x136 | 263x127 | 204x1-01 As these measurements are taken from the largest, the smallest, and an intermediate specimen from each locality, they indicate very fully the divergence of size and proportion of width to height to which the form is subject, the greatest development being reached at Bathurst. Var. berth (Melv. & Pons.). Pl. IV, figs. 25, 26, 27. Ennea berthe Melv. & Pons., Ann. & Mag. Nat. Hist., vol. viii (1901), p. 315, pl. ii, fig. 1. Ennea berthe Melv. & Pons.; Connolly, Ann. 8. Af. Mus., vol. xi (1912), p. 67. Shell like E. farquhari, typical, but elliptical in contour, with six, less ventricose, whcrls and shallower suture ; the strie are rather less regular, shorter, and a little further apart ; the armature of the aperture is very similar, but the labral tooth is not quite so massive, and is less divided, the internal columellar plait is somewhat broader, and the short, ON SOUTH AFRICAN ENNEA. 45 broad tooth on the columellar lip is even less conspicuous than in the typical form. Height 2°84, width 1°50 mm. Hab.—Karkloof (McBean & Taynton) ; Nottingham Road (Taynton); and Ntimbankulu, Mid-Illovo (Burnup), all in Natal. The dimensions given above are those of a co-type of “KH. berthe M. & P.,” from the original lot collected by Mr. McBean at Karkloof, now in my collection at the Natal Museum (Mus. No. 1441), and here figured. Those of the type in the original description are manifestly wrong, for they give a ratio of width to height of 35°71 per cent., while the figure represents a ratio of 50°98 per cent., the latter being well supported by my shell and new figure with a ratio of 52°11 per cent. It will be seen that if Melvill and Pon- sonby’s height dimension of 3°56 mm. be a misprint for 2°5, and the width measurement of 1°25 mm. be correct, a ratio of 50 per cent. is established, corresponding very satisfactorily with their figure ; but in this case their shell is rather smaller than any that I have examined. The suggestion of a mis- print, however, is supported by the fact of the dimension line alongside their figure being 2°6 mm. long. The other known specimens of this variety, excepting the two in McBean’s collection, whose dimensions I have not been able to ascertain, measure as follows: Karkloof: height x width, 3:11 x 1°50, 2°85 x 1:55, 2°79 x 144mm. Nottingham Road: 3:25 x 1°52, 2°86 x 1°50 mm. Ntimbankulu: 3°32 x 1°59, 3:13 x 1°57 mm. Although Messrs. Smith and Robson concur in my view that this form belongs to the species farquhari Melv. & Pons., they have expressed no opinion as to its claim to varietal distinction. I think, however, that such a distinction is not only desirable, but inevitable, since it can be distinguished from the typical form by the characters above enumerated, and inhabits a different district, the variety having only been met with in the Province of Natal, and the typical form in the Province of the Cape of Good Hope. - In three of the 4.6 HENRY CLIFDEN BURNUP. eight shells examined the distinction is not quite so fully maintained as in the other five, as the shells are longer with- out material corresponding increase in width, and the form therefore is less elliptical; but the other distinctions remain. The number of whorls increases with the additional height to 64, my largest specimen of EH. farquhari, typical, having 52 whorls. Var.avenan. PI. IV, figs. 28, 29, 30, 31. Shell small, cylindrical, with rounded apex, rimate and most narrowly perforate, thin vitreous, whitish, transparent and shining: whorls 6}, only shghtly convex, and, excepting for a few irregularly scattered faint transverse scratches, smooth all over, except the last half whorl, which is sculptured with regular distinct transverse rib-striz continued round the base; suture rather shallow; aperture somewhat broadly triangularly auriform, with well thickened slightly expanded and reflexed white porcellanous peristome, and furnished with the following processes: a prominent blade-lke, in-running parietal plait, a massive, rounded, undivided labral tooth, a small internal basal tooth, a most inconspicuous, broad, very slightly raised tooth on, and parallel to, the columellar lp, and a deep-seated, rounded, flattish, scoop-shaped columellar plait. Height 3°21, width 1°30 mm. Hab.—Maritzburg; also Pinetown and Durban (Burnup), Nottingham Road (‘Taynton), all in Natal. The axial perforation rises from a position so deep in the rima that it is difficult to detect, but the branch passing into the deep-seated columellar plait is more easily seen: depres- sions behind the lips correspond with the labral, basal, and columellar teeth. This variety differs from EH. farquhari, typical, in its generally greater size, its more cylindrical form with nearly parallel sides, less ventricose whorls, shallower suture, absence of defined sculpture except on the last half whorl, less expanded yet more thickened peristome, more open aperture ON SOUTH AFRICAN ENNER. A.7 with more superficial, undivided labral process, and columellar tooth less developed so showing more of the deep-seated colu- mellar plait behind it. From the variety berthe it is separable by its more slender, cylindrical form, narrower columellar lip with more open aperture and solid, nearly superficial labral tooth, and generally the absence of striation, or when it is present above the last half-whorl, its extreme shortness. The twenty-five specimens collected in Durban, one of which is shown in fig. 31, and two of those found in Maritz- burg, are sculptured immediately below the suture with more or less regular, rather coarse, typical striation, which, how- ever, soon disappears, leaving the remainder of the whorls almost smooth, except the last half-whorl, which is sculptured all over. With the presence of this infrasutural striation occurs a slight increase in width; otherwise the shells agree in all respects with the smoother shells from Maritzbure and those from Pinetown and Nottingham Road. There is considerable variation in the height of individuals of this variety, though not much in width unless accompanied by the variation in sculpture ; with an increase in height there seems to be invariably also an increase in the number of whorls: thus, the specimens that I have examined vary from 2°74 mm. to 3°51 in height, and the number of whorls from 54 to 7. The following dimensions show very fairly the extent to which the variety, so far as is yet known, differs in size and proportion in various individuals : The smoother form : Maritzburg : 3°32 x 1°27, 3°10 x 1:28, 2°77 x 1°31 mm. Pinetown : 3°41 x 1°33, 2°74 x 1:33 mm. Nottingham Road: 2°85 x 1°32 mm. The form with infrasutural sculpture : Durban: 3°51 x 1°45, 3°30 x 1:45 (fig. 31), 3°02 x 1-45 mm. Maritzburg: 3°37 x 1:45, 3°33 x 1:47 mm. These dimensions, which are chosen to include the most divergent forms of the shells examined, well show the re- 48 HENRY CLIFDEN BURNUP. markable stability in width of this variety. Among the smooth shells there is a difference in height, between the extremes, of ‘67 mm.; but in width there is no more than ‘06: among those with infrasutural striation there is, in height, a difference of 49 mm., and in width only ‘02. Hnnea darglensis Melv. & Pons. f. typica. Pl. IV, fig. 32. Ennea darglensis Melv. & Pons., Ann. and Mag. Nat. Hist., vol. i (1908), p. 130, pl. vii, fig. 1. Ennea darglensis Melv. & Pons.; Connolly, Ann, 8. Af. Mus., vol. xi (1912), p. 71. The original figure not showing clearly the complex character of the columellar plait and representing the labral tooth as rather too solid and the sculpture too sharply defined, the result possibly of over-magnification, a new figure is here offered. The magnification chosen, about 102 diameters, is the same as that adopted for the figures of the new variety hereunder described, so that a fairer comparison of the variety with the typical form can be made, using the new figure instead of the old. The sinuous, duplicate columellar plait is a conspicuous characteristic of the species. The upper branch is wholly internal with its point produced forward; and the lower, and larger, is less deep-seated, its base arising near the edge of the columellar lip with its point curving inward. From the rimal cavity an axial perforation passes upwards, and, at a right angle thereto, a deep, narrow passage enters each branch of the columellar fold. The labral and basal processes are also indicated by depressions behind the peristome. The labral tooth is more distinctly bifid than the original figure suggests, the upper branch being longer, and the lower more internal. To the localities cited by Connolly, in his most useful “ Revised Reference List of South African Non-marine Mollusca,”! may be added Bulwer, near which village, in 1 «Annals of the South African Museum,’ vol. xi (1912), pp. 89-306. ON SOUTH AFRICAN ENNEA. 49 the District of Polela, Mr. C. W. Alexander has collected a single specimen while these notes were being written. It is a good deal smaller than any of the specimens from the other localities, but in form is exactly proportionate to the shell here figured, and is identical in all detail. The locality Gowie’s Kloof, Grahamstown, cited in the same place, is erroneous, and should be expunged. The shell on which the locality was cited is not darglensis, but is Ennea ponsonbyi un. sp., described hereafter, p. 78. The following are the dimensions of such shells as I have measured, besides the type: From Inhluzani: height x width, 2°69 x 1:25, 2°62 x 1°18, 2°54.x 1°18, 2°53 x 1:19 (fig. 32), 2°46x1:12. From Bulwer: 2°23 x 1:07 mm. The first and second on the list are picked, as appearing to be the largest, from a series of about seventy specimens. Var. illovoensisn. Pl. IV, figs. 33, 34, 35. Shell small, rimate and narrowly perforate, elongate ovate- cylindrical, thin, glossy, transparent, whorls 63, rather convex, almost smooth except the last half-whorl and the base which are regularly rib-striate; suture rather shallow ; aperture rather quadrate and somewhat oblique, with white, thickened and reflexed peristome armed with the following plaits and teeth : A broad blade-like, squarish parietal plait, a deeply cleft labral tooth, a small, sharp in-running basal tooth-hke plait, and a complicated two-pointed colu- mellar plait with detail as in darglensis, typical. Height 2°92, width 1:37 mm. Hab.—Ntimbankulu, Mid-Ilovo (Burnup). In form and arrangement of the armature of the aperture, this variety almost coincides with the type. The upper branch of the labral tooth in the variety is squarish at the end instead of pointed, and the complex columellar process is situate a little lower; but, though these features seem constant in the specimens examined, they are but trifling divergencies such as might be looked for in individuals of the same form. VOL. 8, PART 1. 4 50 HENRY CLIFDEN BURNUP. There are, therefore, only the superior size and the smooth polished surface to distinguish the variety from the typical form; but these are sufficiently marked to warrant, in my opinion, the assigning of a varietal name to the form from Mid-Illovo, even if intermediates be eventually found. I have examined a great number of each form and find them very constant. The dimensions of other specimens from the same locality, that I have measured, are as follows: height x width, 3:07 x 1:40, 2°87 x 1:35, 2°83 x 1:34, 2°80 x 1°34, 2°72 x 1:38 mm. In most specimens there is the appearance of a “ margined suture”: but I look upon this as an optical illusion rather than a shell character, not only in this species, but in most South African Ennez where it occurs. The lower sutural line cannot be felt with the point of a fine needle, and is, without doubt, merely the base of the previous whorl seen ° through the transparent shell. As the shell becomes calcined the illusory line vanishes. Ennea consobrina Ancey. Pl. IV, fig. 36. Ennea consobrina Ancey, Brit. Nat. (1892), p. 125; Melv. & Pons., Ann. and Mag. Nat. Hist., vol. i (1898), p. 24, pl. viii, fig. 9. Through the courtesy of Mr. J. R. le B. Tomlin, in entrust- ing to me for examination the type-specimen from his collec- tion, I have been enabled to identify further examples of this little-known species, collected by Mr. Farquhar at Martindale, Bathurst, a division of the Province of the Cape of Good Hope, adjoining that of Albany, in which the type was found. I have also taken the opportunity of having the type tem- porarily in my possession to re-figure the shell, it not having been figured in the first instance, and the figure published by Melvill and Ponsonby in 1898, though bearing some general likeness to the shell, not showing enough detail for purposes of comparison. Ancey rightly compares his species with BE. thelodonta M. & P., but,.while stating that the denticles in both species. are ‘similar, though somewhat ON SOUTH AFRICAN ENNEZ. fl | different, does not specify in what respects they differ. Melvill and Ponsonby’s figure of Ennea thelodonta? is confusing, showing at the same time, to a great extent, a side view of the spire, and a front view of the aperture; besides, the detail is faulty. I therefore present a new figure of their species (fig. 37), for comparison with that of consobrina. It will be seen from these that the likeness between the two species, though striking, is only superficial, the detail of the labral tooth and columellar plait in each being of quite a distinct character. In consobrina the large labral process bears on the inner edge two distinct, little in-running plaits: in thelodonta a somewhat similar, but smaller, process is merely slightly divided into two lobes. In consobrina the deep-seated columellar plait is strengthened by a transverse rib; in thelodonta it is merely drawn forward at the lower, inner corner, into a rounded point. In neither case does it seem likely that the difference between these processes is the result of the direct evolution of one form from the other, for they seem to be constructed on different plans; therefore I consider the two species quite distinct. In addition, there is, as pointed out by Ancey, the difference in size and number of whorls, which is very considerable, though probably not: quite so great as would appear from a comparison of the descriptions ; for, while Ancey describes his shell as 8 mm. high and 4 mm. wide, or, as he says, twice as large as thelo- donta, which is described as 4 mm. high and 2 mm. wide, my measurement of Ancey’s type makes it 8 mm. high and 3°7 mm. wide, while my two specimens of thelodonta measure as follows: Height x width, 4°87 x 2:26, 4-25 x 2°15 mm. (fig. 37). These specimens were collected at Port Khzabeth by Mr. J. Farquhar. Other specimens of E. consobrina examined by me measure as hereunder : In my collection at the Natal Museum, 7°75 x 4:0; collection of J. H. Ponsonby, 7°75 x 3°67, 7°75 x 3°67 mm. ' «Ann. and Mag. Nat. Hist.,’ vol. ix (1892), pl. vi, fig. 4. on HENRY CLIFDEN BURNUP. The dimensions of the remaining specimen in Mr. Ponsonby’s collection were not taken, as, the lip at the base being chipped, they would be misleading. As regards the difference in the number of whorls, while Ancey describes his type as having 8 whorls, I can but count 73;and while Melvill and Ponsonby assign 6 as the number of whorls in their type, their figure seems to indicate about 64 or 63, and my specimens have 6} and 7 whorls respectively. Ennea munita Melv.& Pons. Pl. IV, fig. 38. Ennea munita Melv. & Pons., Ann. and Mag. Nat. Hist., vol. ix (1892), p. 86, pl. vi, fig. 5. Shell small, rimate, ovate-cylindrical, whitish, transparent, shining, rounded towards the blunt apex; whorls about 7, only shghtly ventricose, rather strongly sculptured with regular transverse strie, except the first two, which are smooth, and the area immediately above the aperture, which is nearly smooth and brightly polished; suture rather shallow; aperture rather long and ear-shaped, with thickened, much expanded, and sh¢htlv reflexed, white porcellanous peristome. furnished with the following teeth and plates: a comparatively small, blade-like, m-running parietal plait, a massive, rounded tooth on the labrum bearing a denticle on the upper edge, a small, deep-seated basal tooth, a rather broad, short, expressed tooth, tapering below and ending abruptly above, on the columellar lp, and a very deep-seated, rounded mammillate columellar plait. Height 3°58, width 1°86 mm. Hab.—Griqualand Hast; also Kowie (Ponsonby and Langley). The original description being rather deficient and the original figure indistinct and misleading, I have described and figured a specimen in my collection given to me by Mr. Ponsonby. Major Connolly, who has kindly compared for me the figure with the type in the British Museum, writes that the only differences he could find between the type and ON SOUTH AFRICAN ENNE. 93 my figure are that the columellar hp of the former is of more uniform breadth than that of the figure, and the prominence on the deep-set columellar plait is not situate quite so high. There is a certain resemblance between this species and Ennea crassidens Pfr., but the greater size, the ovoid form and conical apex, and the shortness and squareness of the labral tooth of the latter are enough to distinguish it at a glance. It also seems to bear some resemblance to Ennea tharfieldensis, M. & P., a species unknown to me except from description and figure. Connolly has kindly compared the two types in the British Museum, and con- siders the latter sufficiently distinct through the presence of a denticle at the junction of the columella and the paries, should this feature prove to be constant. The colour attributed to the type in the original description probably arises through the remains of the animal being seen through the transparent shell. The upper whorls of my shell have an ochraceous tint from the same cause. From the rima a deep narrow channel passes into the deep- seated columelar plait, but no axial perforation can be distinguished. The processes on the outer and the columellar lips and at the base have corresponding depressions on the outside of the shell. Besides the shell figured I have been privileged to examine two specimens in Mr. Ponsonby’s collection and three in Mr. Langley’s, all from Kowie. They are all much larger than the specimens from Griqualand East, but exhibit no important variation. A slight difference in the develop- ment of the peristome and its processes may be observed, but not to any material extent. The dimensions of the specimens from Kowie are as follows : Height x width, 4°94 x 2°18, 4:94 x 2°13, 4°84 x 2°29, 4°40 x 2°15, 434 x 2:14 mm. 54 HENRY CLIFDEN BURNUP. Ennea sylvia Melv. & Pons. Pl. III, figs. 18, 19, 20. Ennea sylvia Melv. & Pons., Ann. and Mag. Nat. Hist., vol. xii (1903), p. 599, pl. xxxi, fig. 4; Connolly, Ann. 8. Af. Mus., vol. xi (1912), p. 86. Another species, hereafter described as E. melvilli, from Natal, having been mistaken for, and distributed as, kK. sylvia M. & P., it has become necessary to examine critically the characters, description, and figure of the older shell before describing the new, and to re-figure the former with a magni- fication corresponding to that chosen for the figures of the latter, in order to make comparison easier. or this purpose I have been able to examine the four co-types in Mr. Ponsonby’s collection and three shells from the original lot, presented to me by Mr. Farquhar; and from the latter I have chosen my type for the new figure, the co-types being unsuited to the purpose in that one has a broken peristome, one is holed, one is abnormally large, and the fourth has a shght extraneous obstruction within the aperture which I failed to remove. Though unsuitable for figuring, their characters are quite sufficiently revealed to leave no room for doubt that my other three specimens belong to the same species. In the original description a contradiction occurs in that the teeth and plaits are said to be four, while the details of five are given. Five is the correct number, of which the deep-seated internal columellar plait is more than mammee- form, having a strong rib across it, obliquely from left to right and downwards. The small basal “ tooth” is a small in-running plait, and that on the columellar lip appears like a triangular swelling, but is well excavated from behind. From the rima a very narrow perforation rises up the axis with a wider branch into the deep-seated columellar plait. Besides the processes of the aperture detailed in the original descrip- tion there is a very inconspicuous thickening of the shell on the base forming a slightly raised ridge, hardly amounting to a plait, parallel to the basal lip and behind the basal and labral processes. A slight corrugation indicates its position ON SOUTH AFRICAN. ENNEA. 59 externally, and the outer wall bears a double pit corresponding to the bifid labral plait. If the dimensions of the type as given in the original description, viz. height 1°75, width 0°75 mm., are correct, they cannot be considered normal in view of the measure- ments of the remaining available specimens, which are as follows : Co-types in collection of J. H. Ponsonby: height x width, 2°58 x 1:06, 2°26 x 1:01. Specimens in my collection: 2°25 x 1:06, 2°25 x 1:01, 2°06 x 0°96 mm. (figs. 18-20). I regard the first of these as being abnormally high. Knnea melvilli n. sp. Pl. III, figs. 21, 22, 23. Shell minute, elongate-oval, rimate, thin, whitish, glossy, transparent, apex rounded ; whorls 6, rather ventricose, very closely, finely sculptured with delicate transverse rib-striz, except the first two, which are smooth; suture moderately deep ; aperture very small, rather roundly triangular, with white peristome slightly thickened and reflexed and joined by a narrow callus, and furnished with the following plaits and teeth; a broad blade-like parietal plait, a squarish, bifid tooth on the outer lip, a small basal in-running plait, a broad, slightly raised, rounded tooth on the columellar lip, a deep- seated, rounded, scoop-shaped plait on the columella, and a slightly raised ridge within, and nearly parallel to, the basal lip. Height 2°12, width 0°91 mm. Hab. — Nottingham Road (Taynton); also Karkloof (McBean), Curry’s Post and Fort Nottingham (Taynton), Dargle and Edendale (Burnup). The parietal plait is very prominent, the right edge standing well forward from the general plane of the peristome before entering the aperture; its left edge is shorter and more internal. The bifid labral tooth arises very near the edge of the peristome, the lower half being rather more internal. The 56 HENRY CLIFDEN BURNUP. small basal tooth-lke plait arises but little within the edge of the peristome at the base of the columellar lip and curves inwards and upwards slightly to the left. The process on the columellar lip, which I have above called “a broad, slightly raised, rounded tooth,” consists of an inward bend of the lip covered by its outward expansion, and is quite superficial. The deep-seated columellar fold, which is wholly internal, is rounded in outline, scoop-shaped, and presents its right, lower edge furthest forward. The ridge across the base is very inconspicuous, and might easily be overlooked ; its position, however, can be traced on the outside by a slight corrugation. The positions of the other processes, except the parietal, are also indicated externally by corresponding depressions. The peristome, which is comparatively but little thickened, is exceedingly sinuous on the labrum (fig. 22). From the rima there is a deep opening passing into the columellar fold, but none can be detected into the axial perforation. Besides those of the type the following dimensions of normal specimens have been taken: Dargle: height x width, 2°12 x 0°88, 2°10 x 0°88, 2:05 x 0°89. Nottingham Road : 2°04 x 0°89, 2°07 x 0°89, 2:00 x 0°91, 1:94 x 0°89 mm., and the following are the dimensions of the largest and smallest examples in my collection, both manifestly abnormalities : Nottingham Road, 2°60 x 0-91; Dargle, 1°83 x 0°93 mm. This shell has been confused with HE. sylvia M. & P., but though the armature of the aperture is arranged much on the same plan, only one plait differing very materially, this fact, I think, must be taken as an indication of parallel develop- ment rather than close affinity, for in all other important respects the two species vary widely. The form of this species is elongate-oval or fusiform; that of sylvia cylin- drical, with almost parallel sides. In sylvia there is an umbilical opening from the rima, fairly wide for so small a shell; in melvilli it seems to be closed. Melvilli has rather more than half a whorl in excess of sylvia, the whorls of the former being more ventricose and separated by deeper sutures: with the exception of the embryonal whorls, they ON SOUTH AFRICAN ENNE. 57 are most finely, delicately, closely and regularly rib-striate ; while sylvia is nearly smooth, except for a few irregularly dispersed strize, all over except on the last half-whorl, where the strize become more regular and more clearly defined. ‘The aperture is smaller, and the peristome much less thickened and expanded in melvilli than in sylvia, and the parietal plait is larger and more prominent, the labral tooth smaller and less distinctly divided, the “tooth” or swelling on the columellar lip is less angular, and the internal columellar fold is of quite a different nature, being simple instead of crossed by an oblique strengthening rib as in sylvia. The profile of the outer lip, too, is much more sinuous in melvilli. The specimen with the abnormal height of 2°6 mm. is no wider than the type, has fully 7 whorls and has the peristome scarcely thickened, leaving the swelling of the columellar lip undeveloped; the internal columellar fold is small. Enneacolumnella M. & P. (Pl. V, fig. 47), is found in company with this species at Karkloof, Dargle, and Edendale, and, both being much of the same size, a critical comparison may here be convenient. The form of columnellais rather wider and less fusiform ; the surface is almost smooth, except immediately below the suture, where irregular microscopic striation may be detected in a few places, and on the last half-whorl where somewhat regular striation appears ; the parietal plait is narrower, the labral tooth more widely bifurcate, the basal tooth smaller and more central, the shght ridge across the base of melvilli is absent, and the columellar plait is of quite different con- struction (see postea, p. 58). With the assistance of a moderately powerful hand-lens their separation is quite easy. I have pleasure in dedicating the species to Dr. J. Cosmo Melvill, who has done so much towards the elucidation of the South African non-marine molluscan fauna. This is the form referred to by Connolly! as examples from Natal, distinct from E. sylvia, whose habitat is Mae- strom Forest, Bedford, Cape of Good Hope. 1 Tbid., p. 86. 58 HENRY CLIFDEN BURNUP. Ennea columnella Melv. & Pons., f. typica. Pl. V, fig. 47. Ennea columnella Melv. & Pons., Ann. and Mag. Nat. Hist., vol. viii (1901), p. 316, pl. ii, fig.2; Connolly, Ann. S. Af. Mus., vol. xi (1912), p. 69. In the original figure the parietal plait appears to be wholly internal, which is quite wrong, as its right side arises above, and in front of, the aperture. Hence a new figure has become necessary in order that comparison with kindred forms may be facilitated. In the original description the true nature of the very distinctive columellar plait is not very clearly set forth. It is not, as is usual in the South African Enneex, wholly internal; but, from the lower inner edge of a flattish, somewhat scoop-shaped internal surface, a strengthening rib curves outwards almost to the outer edge of the labium. By this feature the species can be readily dis- tinguished from other species of similar size and superficial likeness. I have not found the peristomatal processes of the many examples that I have examined exhibit among themselves con- siderable variation, as Melvill and Ponsonby did in the original lot ; but, as one of my “co-types” belongs to a different species, described in this paper as Ennea melvilli, the divergence from the type that they observed seems easily explained. The figure here presented is drawn from a co-type in my collection from the original lot found by McBean in the Kar- kloof Bush. The shell measures as follows: Height 2°15 mm., width 1:00 mm., and is the largest that I have examined. The following measurements of shells collected by me at Dargle have also been taken : Height x width: 2:08 x 0°89,.2:06) x 0:93, 1°99 x_0-92; 1:99 x 0°90, 1:98. x 0:90; 1°97 x 0:°94,.1587)x10:92 mm; The original width dimension of *75 mm. is probably wrong, as the original figure, taking the height as 2 mm., represents the width to be 1 mm., and Connolly’s measurement of the type as communicated in lit., gives height 2, width 0°9 mm. ON SOUTH AFRICAN ENNEA. 59 Since the above was written, my attention has fallen upon two shells collected by me in the year 1900 in the Beach Bush at Lower Umkomaas, which could not be identified at the time but undoubtedly belong to this species. Their dimen- sions are as follows : Height x width: 1:94 x 0°79, 1°82 x 0°83 mm. ‘Though among the smallest measured, their ratio of width to height corresponds with the typical form rather than with the more obese variety vitreola, considered below. In fact, the higher of the two is the most attenuate of all. This newly cited locality is interesting, as but few species, at least in typical form, are found both on the coast and in the higher altitudes of the midlands. Var. vitreola (Melv. & Pons). Ennea vitreola Melv. & Pons., Ann. and Mag. Nat. Hist., vol. i (1908), p. 180, pl. vii, fig. 3; Connolly, Ann. 8. Af. Mus., vol. x1 (1912), p. 87. I cannot separate this form from the preceding, except on the grounds of its more ovoid contour with shallower sutures, and the slightly weaker development of the peristome and peristomatal processes. ‘The general appearance of the shells in other respects, their sculpture, and the arrangement of the peristomatal processes, are almost identical. The type having been reported lost, I sent the better of the co-types to take its place; and this, 1 believe, is the specimen now in the British Museum. The remaining co-type is in my collection at the Natal Museum.! Although the exact locality at Hilton Road has been dil- gently searched several times since the original shells were discovered, no further examples of the variety, and no specimen of the typical form at all, have been found until the present time (November, 1913), when, a final effort being made, one specimen was obtained confirming the varietal ' Connolly’s statement, in loc. cit., that the type is in Maritzburg, is incorrect, 60 - HENRY CLIFDEN BURNUP. features to a marked extent, it beg even shorter than any of the others, with a greater relative width. In my co-type of this variety and in the newly found specimen the two parts of the labral tooth are more distant than the original figure shows to have been the case in the type, so bringing them shghtly nearer to the typical form of E. columnella. The strengthening rib connecting the columellar plait with the columellar lip is rather more tor- tuous than in the type (of vitreola), but I do not consider that an important feature. The following are the dimensions of the only four specimens known : Height. | Width. Type of E. vitreola (lost) . ne 1:87 = -0°94anm: Co-type in British Museum . CC SZ: tesO: Goa x , Natal Museum . SALES OOo es Seen ‘collected’ ObuNow: LOTS 4e1226 0:9225) The ratio of width to height in above varies from 50°27 to 52°27 per cent., while that of the measured specimens of columnella, typical, from Dargle, varies from 42°31 to 49°20 per cent., and that of the igher shell from Lower Umkomaas is 40°72 per cent. I am indebted to Messrs. H. A. Smith, Connolly, and Preston for their views on the relationship of these two forms, E. columnella, typical, and var. vitreola, arrived at during a discussion over the type of the former and co-type of the latter in the British Museum. It is at their suggestion, in which I now fully concur, that I have retained the distinction of the Hilton Road form under the name vitreola used in a varietal sense. Although it will be seen that one of the Lower Umkomaas specimens of EK, columnella is even shorter than the type of vitreola, it agrees in every respect except size with columnella, typical, its relative proportions being about equal to an average example of the Dargle specimens. The most stunted example from Dargle approaches the proportions of the variety vitreola, but can readily be distinguished ON SOUTH AFRICAN ENNER. 61 by its more cylindrical, less ovoid, form, and its deeper suture. Ennea marie Melv. & Pons. PI. V, fig. 48. Ennea marie Melv. & Pons., Ann. and Mag. Nat. Hist., vol. ix (1892), p. 92, pl. vi, fig. 12. Ennea callista Melv. & Pons., ibid., vol. iv (1909), p. 486, pl. viii, fig. 1. As neither the original figure of E. mariz nor that published with the description of “EK. callista” is drawn in true proportion, or expresses clearly the characters of the species, I take this opportunity of publishing a new figure. For this purpose I have been fortunate enough to secure the loan from Mr. Ponsonby of one of the original specimens collected by Miss M. Bowker at Somerset East. I have also examined one specimen, collected by the late Mr. Crawford, in Mr. Ponsonby’s collection, three specimens (co-types of “E. callista”) in Mr. Ponsonby’s collection, and three specimens collected by Mr. Farquhar at the same time as the type of “KE. callista,”’ two of which are in his collection andoneinmine. All these specimens were found at Somerset Kast, the locality, “Dassy Krantz, Grahamstown,” given in the description of “ E. cal lista,” being an error arising through the interchange of the contents of two boxes. Thus Somerset Hast remains the only known locality for E. marie. The other shells affected by the interchange were those described as “E. periploca”; so Somerset East cannot be accepted as a locality for E. farquhari, of which “periploca” is a synonym, on our present knowledge of its distribution. The tooth on the columellar lip is not connected with the internal columellar plait, as stated in the original description of E. marie, but is quite distinct as defined in the description of “EK. callista.” A narrow axial perforation rises from the umbilical slit, a branch entering the columellar fold; and the three peristomatal processes have corresponding depressions behind the lips. | AG\CAL v0" oe 62 HENRY CLIFDEN BURNUP. ' The dimensions of the various specimens examined are as follows : Height. Width. : mm. mm. Type of new figure (fig. 48), Miss M. Bowker 3:03 . 1:19 Mr. Crawford’s specimen (in coll. Ponsonby) 2°78 . 1:18 fee) Coos eily, Co-types of “callista” (in coll. Ponsonby) Se begs Pale Paar pa ESI) Original specimens of “callista” (in coll. (2°70 . 1:24 Farquhar). 270) tes Original specimen of “callista” (in Natal 2°55 . 1:09 Museum). My examination of these well-authenticated examples leaves no room for doubt that they belong to one species; indeed, there is little individual variation among them. “EK. callista” must, therefore, become a synonym of EH. mariz. I am indebted to Major M. Connolly, who has kindly examined the types in the British Museum for me, for the following confirmation of my views: “In E. mariz the trace of a cusp on the upper part of the labral tooth is slightly more prominent than in EH. callista. The rest of the dentition, together with the form and sculpture, is similar in both types, and the shells are undoubtedly conspecific.” He also furnishes me with the following revised dimensions of the types as taken by him: “Type of E. marie, height 3:00, width 1°35 mm. » H. callista, height 2°80, width 1:30 mm.” 9 Ennea mooiensis n. sp. Pl. V, figs. 49, 50, 51. Shell minute, elongate-elliptical, deeply rimate, thin, glossy, transparent, apex blunt; whorls 53, rather ventricose, almost smooth except immediately behind the peristome where a few transverse strize appear; suture not deeply impressed ; aper- ture roundly triangular, with white peristome thickened and reflexed, the ends joined by a thin, narrow callus, and furnished with the following processes: a broad. blade-like, ON SOUTH AFRICAN ENNEA. 63 in-running parietal plait, a sub-central moderately large labral tooth slightly bifid, a small pointed tooth near the base of the columellar lip, a flattish scoop-shaped internal columellar plait whose lower right edge advances forward almost as a mammillate point, and a shghtly raised ridge within the basal lip extending from behind the labral tooth downwards and forwards towards the base of the columellar lip. Height 2°55, width 1:24 mm. Hab.—Game Pass, Upper Mooi River, Natal (Burnup). The details of the several processes as set forth under the description of EK. melvilli, herein, would apply almost equally well to those of this species, but the following differences exist: the tooth near the base of the columellar lip is smaller, less erect, and less plait-like im form in this species; the broad, shehtly raised, rounded tooth higher on the columellar hip of melvilli is here absent; and the ridge on the basal wall is less parallel with the peristome. Although so much of the description of the one species apphes to the other, the greater dimensions, especially in width, the less fusiform contour, the larger aperture and the smoothness of the surface of this species distinguish it very readily from melvilli. The labral tooth and that near the base of the columellar lip have corresponding depressions on the outside, but the ridge on the basal wall does not seem to be excavated, though its position within can be traced on the outside by a white scar. The opening from the rima to the columellar plait is exceed- ingly narrow, while that to the axial perforation is even narrower, only being discerned by the aid of a very strong lens. This species may also be compared with E. columnella M. & P., whose peculiar columellar plait, however, will always readily distinguish it; besides, the latter is smaller, has traces of minute strize below the suture, and much more numerous and stronger striz on the last half-whorl, has the labral tooth much more widely bifid, and the basal tooth, or tooth-lke plait, much more central and in-running. 64 HENRY CLIFDEN BURNUP. The figures of E. sylvia M.& P. (PI.III, figs. 18, 19, 20) and the comparison between that species and EH. melvilli (p. 56) may here be referred to, for there is also some general like- ness between the former species and E. mooiensis, which I think an examination of the figures and a perusal of the text, above indicated, will show to be only superficial. The greater height and narrower, more cylindrical form of E. marizw Melv. & Pons., and the prominent tooth situate subcentrally on its columellar lip, prevent the possibility of it being confused with this species. The following are the dimensions of a few specimens of E. mooiensis which, in addition to the type, I have mea- sured: height x width, 2°65 x 1°25, 2°65 x 1:16, 2°55 x 1°21, PAB X AD, 2:33 *% 11850 2°25 * P12) 221 x to, 2S 118 mm. This species is plentiful in its locality, almost to the exclu- sion of other species of the genus, but has not yet been found elsewhere. Whilst visiting the district twice, during the Christmas holidays 1910-1911 and 1911-1912, I collected 224 specimens of Hnnez, 220 of which belong to this species and 4, all distinct inter se, to different, larger species not yet identified. Ennea maritzburgensis Melv.& Pons f.typica. PI.IV, figs. 39, 40, 41, 45, 46. Ennea maritzburgensis Melv. & Pons., Ann. and Mag. Nat. Hist., vol. xii (1893), p. 107, pl. iii, fig. 11. The original description seeming to require some shght revision, and the type being unavailable to me, I have chosen a normal specimen from the original locality as my type of the following emended description and accompanying new figures (Pl. IV, figs. 39, 40, 41). Shell small, rimate and narrowly perforate, thin, whitish, subdiaphanous, shining; whorls 7, slightly convex, of which the first 2, constituting the protoconch, are smooth all over, the next 44 are delicately, transversely striate immediately ON SOUTH AFRICAN ENNES. 65 below the suture, and smooth beyond, except for a few irregular striz occasionally extending lower, or even passing from suture to suture, and the last half-whorl is regularly, similarly striate all over; aperture ovate; peristome white and shining, expanded and reflexed, and furnished with four processes, as follows: a moderately large, squarish, blade-like, in-running parietal plait, a large labral tooth widely bifurcate, the lower branch being much the larger, a small tooth near the base of the columellar lip, and a columellar plait, wholly internal and bearing a conspicuous, strengthening transverse rib. Height 4°46, width 2°07 mm. Hab.—Maritzburg, Natal, typical ; and, aberrant, Notting- ham Road (Taynton) and Rosetta (Hickey, per Ponsonby). The axial perforation is very narrow and is easily over- looked; from it a branch passes at right angles into the strengthening rib of the columellar plait. The process within the labrum, originally described as two simple teeth, is, I think, better considered as one bifurcate tooth, as the two points spring from a common base, which is indicated on the outside by a single depression. The bifurcations, being solid, appear in this cavity as two white porcellaneous spots. The basi- columellar tooth is barely excavated, and also appears as a white porcellaneous spot behind the peristome. The dimensions given in the original description seeming to indicate a shell of smaller size and much narrower form than this species usually assumes, and not agreeing with the pro- portions of the original figure, Mr. Robson has kindly re- measured the type very carefully for me, and finds that its dimensions are, height 3:9, width 1:8 mm., almost exactly of the same relative proportions as the type of the new figure, but considerably smaller. On finding the emended dimen- sions so much smaller than those of any of my shells collected within recent years, I looked up my earliest examples, found about the same time as the type, and have picked out two nearly as small, measuring respectively 4°05 x 1°83 and 3°94 x 1°95 mm. This difference in size between the old and the new specimens might suggest that recent seasons have VOL. 3, PART 1. 5 66 HENRY CLIFDEN BURNUP. been more favourable for the growth of the shells, but it seems to me more likely to arise through the position in which they were collected. Since I have found the species plentifully under boxes of plants in the Alexandra Park, watered daily in dry weather, I have ceased to look for them in natural surroundings, where they are much less plentiful and their supply of moisture is restricted to the rain and dew. It is noticeable that, as far as is known, the typical form has not been found beyond the precincts of the borough of Maritzburg. The aberrant forms figured on Pl. IV, figs. 45, 46, will be more conveniently discussed after the introduction of the following variety. Var. contracta n. Pl. IV, figs. 42, 43, 44. Shellhke E. maritzburgensis, typical, with the following differences: Shell much smaller, less fusiform and more ovate ; whorls 64, more finely and more obscurely sculptured; aper- ture rounder, proportionately much smaller, and rather less obstructed, with peristome much less expanded, and labral tooth smaller and less widely bifurcate. Height 3°32, width 1°58 mm. Hab.—Nottingham Road (A. J. Taynton), very plentiful ; also Karkloof and Curry’s Post (Taynton), all in Natal. The pinched appearance of the relatively small aperture, taken in conjunction with the much reduced size of the shell, would, without further consideration, readily induce one to suppose that this form is quite distinct from E. maritzbur- gensis Melv. & Pons. ; but aclose study shows both forms to be built upon exactly the same plan, such differences in detail as exist being merely a matter of degree. The sculpture is of the same nature, and the disposition of the apertural processes is identical, with the one exception of the labral tooth in contracta being less widely cleft. In view of such important comcidence, | should have felt disposed, even if no further evidence were forthcoming, to’ consider the new form a variety of the older; but with the support of the ON SOUTH AFRICAN ENNES. 67 aberrant forms from Karkloof and Rosetta, the close relation- ship seems manifest. I think the Nottingham Road form, which is also found at Karkloof and Curry’s Post, showing the extreme divergence from the type and being well established in numbers, may conveniently be distinguished by a varietal name, contracta being chosen in reference to the pinched appearance of its aperture, while the aberrant forms remain undistinguished in name, simply as forms intermediate between maritzburgensis, typical, and its variety. These intermediates have not so far been found plentifully ; from Karkloof I have four specimens corresponding with fig. 45, taken with twenty-seven specimens of var. contracta ; and from Rosetta I have two specimens only, corresponding with fig. 46, without representatives of contracta. These, how- ever, are from a small parcel, of whose numbers I have no record, given to Mr. Ponsonby by Miss Hickey ; but it 1S justly presumable that they are a fair sample of the lot, and that var. contracta was not collected by Miss Hickey. The shell represented by fig. 45 is but little removed from maritzburgensis, typical, the peristome being only shghtly less expanded, but the labral tooth is much smaller and its bifurcation much less distinct. The Rosetta form, fig. 46, on the other hand, is much nearer var. contracta, the shell being almost equally small and the peristome much narrowed ; but the aperture is relatively large, so avoiding the pinched appearance of the named variety. The dimensions of other specimens are as follow : Intermediate form from Karkloof: height x width, 4°47 x 1:89, 4°31 x.1:95, 4:09 x 1:97 (fig. 45), 4°03 x 1:94 mm. Intermediate form from Rosetta: 3°68 x 1°76, 3-46 x 1°76 mm. (fig. 46). Var. contracta from Nottingham Road: 3°73 x 1:73, 3°51 x 1°76, 3°38 x 1:66, 3°13 x 1:63 mm. Var. contracta from Karkloof: 3°60 x 1°66,3°46 x. 1°66, 3°43 x 1:68 mm. Var. contracta from Curry’s Post: 3°51 x 1°62, 3°40 x 1-62 mm. 68 HENRY CLIFDEN BURNUP. In general appearance, without regard to detail, var. contracta is much like a diminutive E. obovata Pfr., but the arrangement of the apertural armature is quite distinct. Ennea arnoldi Sturany. Text-figs. Ennea arnoldi Stwrany, Anz. k. Akad. Wissensch. Wien (1898), No. xvi, p. 158 (reprint p. 6); Sturany, Siidafrik. Moll. (1898), p. 28, pl. ii, figs. 26-30; Connolly, Ann. 8. Af. Mus., vol. xi (1912), p. 67. From Miss Phyllis Radford I have received two shells, ») My ) ~ —— SSS A. Specimen from Tongaat. B. Specimen from East London. collected by her at East London, which at first appeared new to me, but a closer examination, and comparison with the shells in the Natal Museum, have convinced me that they can but belong to this species; and the differences between them and typical shells do not seem to be of sufficient importance to warrant a varietal name. The new locality, Hast London, then, may be added to the localities cited by Connolly in his Revised Reference List. The chief differences are the less cylindrical form of Miss Radford’s shells, with sharper apex, the more oblique aperture, and the absence of the intense constriction behind the peri- stome at the base. The species varies considerably in size and in relative width, two of the specimens of which Sturany gives the dimensions being even shorter, and one of them ON SOUTH AFRICAN ENNE. 69 relatively wider, than the Hast London shells ; therefore, the average smaller size and greater obesity of the East London shells cannot be considered of importance, since they might not be maintained in a series. The obliquity of the aperture which inclines from the right, above, to the left, below, is conspicuous, and is not equalled in any of Sturany’s figures, nor in any Natal example that I have examined. The absence of the deep constriction behind the labrum is probably the most important feature, but is not so easily detected. The accompanying figures are designed to illustrate this divergence from the type. The dimensions of the East London shells are as follows : height x width, 2°69 x 1°36, 2°64 x 1:37 mm. It will be seen that while the width of these shells exceeds that of even Sturany’s largest measured example, the ratio of width to height is greater in the specimen of which he gives the dimensions as 2°5 mm. high and 1°3 mm. wide. Knnea connollyi Melv. & Pons. Pl. V, fig. 52. Ennea connollyi Melv. & Pons., Ann. and Mag. Nat. Hist., vol. iv (1909), p. 486, pl. viii, fig.2; Connolly, Ann. S. Af. Mus., vol. xi (1912), p. 69. As Connolly indicates, loc. cit., the original figure does not adequately represent the shell. Not only is it too wide, but the sutures are too shallow, and certain apertura! processes are not shown. ‘Therefore this opportunity is taken to publish a new figure. The dimensions set forth in the description, given in “yound numbers,” probably give almost as erroneous an idea of the true proportions of the shell as the figure does, for in my series, kindly supplied by the discoverer, I have no example nearly so attenuate as the recorded dimensions of the type would indicate it to be. The following table of dimensions and ratios will indicate the amount of variation in those respects which may be 70 HENRY CLIFDEN BURNUP. looked for in this species, and also the extent of the dis- crepancy between the authors’ original dimensions and those indicated by their figure. To ascertain the dimensions repre- sented by the figure, the height is taken from the artist’s dimension line, and the width is calculated therefrom, pro- portionately to the height and width of the figure. | | | , | | Specimens from various localities. | Height. | Width, | Width x 100 | | Height | | : x : | mm. mm. | Type as per authors’ dimensions — . SOON 1 wlG00 3335 Type as per original fig. . : 325 | 1:56 48-00 | | Type of new figure from Amajuba . 3°02 TeQO ea 271 (| 319 7 Ne esis)! | Other specimens from Amajuba .)|) 3:16 131 | 41-46 | (, 3:06 Tes 43-14 ( | 328 1:28 39°02 . Pats Sot | 3:28 1:22 3720 | Specimens from Hennop’s River. j 3-09 | 1:99 39:48 | ie EO or a LGAs ie eels ey Although my examples from Hennop’s River tend towards a more attenuate form than those from Amajuba, some of the narrowest of the latter are, either actually or relatively, narrower than the widest of the former; so a larger series than I have been able to examine would be necessary before it could be seen if the narrower average would be main- tained. Coincident with the more attenuate form, a deeper, more oblique suture is observable; but no imcrease in the number of convolutions, beyond, perhaps, a quarter of a whorl, can be traced; therefore it would appear that the increase in height is mainly attained by a looser coiling of the whorls. In the Hennop’s River shells the aperture is less erect than in those from Amajuba, sloping, in varying degree, from right, above, to left, below; and the upper tooth of the columellar lip is uniformly situate higher, being close up to the junction of that lip with the wall of the body-whorl. Melvyill and Ponsonby draw attention to the alliance of this ON: SOUTH AFRICAN ENNER. fia! species with E. arnoldi Stur., to which it bears a strong resemblance; but the former is easily distinguished by its wider, less obstructed aperture, and the two teeth on the columellar lip which are absent in Sturany’s species. Ennea inhluzaniensis n.sp. PI. V, figs. 53, 54, 55. Shell small, cylindro-elliptical, deeply rimate, thin, shining, semi-transparent, apex rounded ; whorls 63, shghtly convex, the first 24 smooth, the next 3} delicately transversely striate immediately below the suture, with occasional strize passing partly or wholly across the whorl, and the last }-whorl more strongly, regularly striate across its breadth, round the base, and into the rimal depression ; suture moderately impressed ; aperture erect, oblong, with white peristome — slightly thickened and moderately expanded, the ends joined by a distinct white callus, and furnished with the following pro- cesses: A rather narrow, sharp, m-running parietal plait arising a little above the suture and in front of the plane of the peristome, a rather long, narrow, simple, sub-central labral tooth, and a small, rounded, internal columellar plait. There is also a slight, sub-central swelling on the columellar lip. Height 3°60, width 1°47 mm. Hab.—Inhluzani Hill, Dargle, Natal (Burnup). Among the forty-five specimens collected, scarcely any deviation from the type, except in size, is observable. The “columellar plait” of the description is merely the rounded termination of the axial column without contortion or thickening. The labral tooth is more solid than is usual, the excavation behind the lip beg very shallow; but the position of the tooth is there very clearly shown by a con- spicuous white spot. The swelling on the columellar hp is scarcely thickened, the corresponding excavation being very deep. In the study of this shell I feared, from Melvill and Ponsonby’s description and figure of Ennea juxtidens, and from two very poor specimens, immature, calcined, and 72 HENRY CLIFDEN BURNUP. holed, in my collection, that the new form might possibly be inseparable from their species. Mr. E. A. Smith, therefore, most kindly undertook the comparison of my type with the type of juxtidens in the British Museum, and has furnished me with the following detailed report : “Knnea juxtidens, type, is a larger shell than yours, the sculpture is hardly apparent, or at all events much less visible than in your shells: the peristome is stronger, especially on the columellar side, where, and at the base, it is rather more expanded than in inhluzaniensis. The parietal lamella, beyond being stronger, offers no difference of any importance—it arises higher on the body-whorl than the suture, just the same as in your shells. The columellar fold is quite the same as in inhluzaniensis. The tooth on the labrum is of the same character. In juxtidens there is a feeble denticle low down on the columellar side, but not quite basal ; it is well within the edge of the peristome. The swelling about the middle of the collumellar lip in your shells is higher up than the faint tooth in juxtidens. This is the most distinguishing feature in M. & P.’s species in comparison with yours. “On the whole I think it would be safer to regard inhlu- zaniensis as a distinct species rather than as a variety of juxtidens.” Since seeking Mr. Smith’s opinion, Mr. J. H. Ponsonby has generously given me a mature co-type of E.juxtidens in excellent condition, from his own collection. Thus I am enabled to present new figures of that species (Pl. V, figs. 57, 58, 59) for comparison with those of inhluzaniensis. The dimensions are—height 4°53, width 2°05 mm. In con- sidering the difference in size, it should be borne in mind that the figures of inhluzaniensis are magnified by 9-2 diameters, and those of juxtidens by 7°45. Mr. Farquhar has suggested a _ relationship between inhluzaniensis and elliptica Melv. & Pons., and though I think the differences are too many and too great to admit of close affinity, the size and general form of the » ON SOUTH AFRICAN ENNEA. i) shells, together with the unusual freedom from obstruction in the aperture, warrant a comparison. The flatter whorls, shallower suture, much smoother and more highly polished surface, rounder aperture, with more widely expanded peristome, shorter parietal plait, and the two teeth on the columellar lip (or one in var. manca) should always readily distinguish elliptica from the present species. Elliptica, too, shows a conspicuous opening from the rima to the axial perforation, while none is visible ininhluzanien- sis. The swelling on the columellar lip of inhluzaniensis cannot be looked upon as representing one or other of the teeth on the columellar lip of elliptica, for its position does not correspond with that of either. The dimensions of such specimens of inhluzaniensisas | have measured in addition to the type, are as follows: height x width, 3°59 x 1°52, 3:41 x 1:49, 3°35 x 1:49,3°17 x 1°42, 2°84 x 1:42 mm. From the dimensions of the two specimens figured it would appear that juxtidens, besides being larger than inhluza- niensis, 1s also proportionately a wider shell, the ratio of width to height of these specimens bemg respectively 45:25 and 40°84 per cent.; and Melvill and Ponsonby’s dimen- sions of their type uphold the suggestion. The appear- ance of my calcined co-types of juxtidens, which are too brittle to admit of actual measurement, also seems to support it; buta larger series would be required before the fact could be established, for my smallest example of inhluzaniensis, manifestly a dwarfed shell, is, with a ratio of 50 per cent., proportionately much wider than the figured juxtidens. Ennea premnodes, Sturany. PI. V, fig. 56. Ennea premnodes Stur., Ann. Hofmus. Wien, xvi (1901), pub. 1902, p. 69, fig. 5; Connolly, Ann. S. Af. Mus., vol. xi (1912), p. 84. 1901 ” (sic). Through the kindness of Dr. Sturany in giving me one of his co-types, I am enabled to offer a description and figure of 74 HENRY CLIFDEN BURNUP. this httle-known species, which will be easier of access to South African readers than the original. Shell small, elongate, cylindrical, narrowly umbilicate, thick, white, opaque, apex rounded ; whorls 74, slightly ventricose, the first 3 smooth, the next 4 nearly smooth, except im- mediately below the suture, where they are clearly, rather coarsely transversely rib-striate, and the last half whorl simi- larly rib-striate all over, except immediately behind the lip, where it again becomes smooth; suture moderately deep ; aperture rather large, rounded, with peristome much thickened, expanded and reflexed, with well-developed callus connecting the extremities, and furnished with the following processes: a strong blade-like in-running parietal plait, a large prominent, bluntly pointed labral tooth with corre- sponding pit outside, a small tooth near the base of the columellar lip, and a medium-sized deep-seated columellar plait, flattish and drawn to a small mammillated point at the lower right corner. In addition to these regular processes there is a very minute denticle on the outer, upper edge of the labrum, close to the suture, not shown in the fig. Height, 5°89; width (including expanded labrum), 2:21 mm. Hab.—Albany District, Cape of Good Hope (Penther). The width of the shell gradually increases with each successive whorl after the rapid increase of the apex, so giving the spire a somewhat bluntly elongate-conical form. To what extent the opaqueness of the shell may be considered characteristic of the species, or to what extent it may be due to incipient erosion, could only be decided were more ex- amples available for examination; but its solidity suggests that in maturity it would always be opaque: my shell, though in good condition, does not appear to have been collected alive, so shght erosion in it is probable, though it is not conspicuous. ‘The minute denticle on the outer, upper edge of the labrum, which is very imconspicuous, may not be characteristic, and is scarcely worthy of note except as regards a comparison of this shell with one of the forms of Hnnea montana Melv. & Pons. (Pl. V, fig. 60), hereafter discussed. “I Oe ON, SOUTH AFRICAN ENNEZ. Ennea montana Melv. & Pons. Pl. V, figs. 60, 61, 62, 63, 64, 65, 66. Ennea montana Melv. & Pons., Ann. and Mag. Nat. Hist., vol. xii (1903), p. 599, pl. xxxi, fig. 15. Ennea parallela Melv. & Pons., ibid., vol. iv (1909), p. 489, pl. viil, fig. 9. On first studying this species the material at my command was small, and showed so much variation, especially between my own two shells, ea coll. Farquhar, reputed to represent respectively montana and parallela (figs. 61, 62, 63, and fig. 60), but also between these two shells and two of the co-types of parallela lent to me by Mr. Ponsonby (figs. 65 and 66), and between those two co-types iter se, that | was not only prepared to admit the validity of the two species, but was endeavouring to define the limitations of intermediate forms. As, however, neither of my shells and none of Mr. Ponsonby’s co-types of parallela seemed to agree accurately with Melvill and Ponsonby’s descriptions and figures, a com- parison with the types in the British Museum became essential. Mr. Edgar A. Smith most kindly undertook the investigation, and, being supplied by me with certain shells and drawings, has furnished the following report : “Fig. 65 represents the true parallela. Of the shells, that figured as 65 and the one unfigured agree with M. & P.’s type. I do not regard the third example (fig. 66) as a distinct species. In my opinion it is an older shell with more developed teeth and a shghtly modified labrum. The shell figured as 60 I also consider as parallela with well-developed lip and teeth ; the labral tooth is strong, and does not show the tendency to be double, as in other specimens. “T have compared the types of parallela and montana, and cannot regard them as distinct. The latter is an old shell with lip and teeth more thickened. Your figure (re- drawn as fig. 61) shows a minute columellar tooth which does not exist in the type.” Mr. Smith does not comment upon the presence, in the shell 76 HENRY CLIFDEN. BURNUP. drawn as fig. 60, of the minute tooth on the owter edge of the labrum, close under the suture, so I infer that he attaches no significance to it, and it is probably an accidental development. It is interesting, however, to note the presence of an equivalent tooth in my specimen of Ennea premnodes Stwr. (fig. 56), a somewhat similar but much larger and heavier-built shell. The position of this tooth in premnodes is a little further forward, and so it is lost in perspective and shade in the figure. As Sturany does not mention this tooth in his descrip- tion of premnodes, it is to be inferred that it isabsent from the type, and its presence in my specimen is also accidental. The minute columellar tooth shown in fig. 61, and, fide Smith, non-existent in the type of montana, is, curiously enough, shown in the original figure, though not very pro- minently, but is not referred to in the original description. In no specimen that I have examined is it so conspicuous as in that shown in figs. 61, 62, 63. In some it is but slightly raised (e.g. fig. 60); in others it is merely represented by a shght swelling (e. g. fig. 66), but in the greater number it 1s entirely absent. Since submitting the matter to Mr. Smith, Mr. Farquhar has kindly sent me his whole series for examination, con- sisting of four specimens of montana collected at and near Mountain Drive, Grahamstown, one of them being from the original lot, five specimens collected with the original lot of parallela near the Asylum, Grahamstown, and ten specimens, unnamed, collected at Gowie’s Kloof, near the same city. I have also been favoured by an opportunity to examine the only specimen in the Alexander McGregor Museum, Kimberley, also from Gowie’s Kloof. This large series amply confirms Mr. Smith’s views that parallela is inseparable from montana. It seems to be a species very constant in its elongate cylindrical form, in its sculpture con- fined (except on the apical whorls, which are smooth, and the last half-whorl, in which the sculpture crosses the whole width) to the area immediately below the suture, and in the position of the important apertural processes, but most variable in the ON SOUTH AFRICAN ENNE®. TAT development of those processes and in the extent to which the peristome is thickened and expanded. The divergencies from a conceivable normal type being apparently more a matter of degree in development than a modification of plan, and no series of specimens exactly sup- porting any definite line of departure from such a normal type, it would not seem desirable to attempt to define named varieties, and the species must remain one difficult to deter- mine. Itis hoped that the series of figures chosen to illustrate the species may serve as a guide to the amount of variation to be looked for. From these a summarised definition of the species may be drawn up as follows: Shell small, elongate-cylindrical, rounded above and below, deeply rimate and narrowly perforate, semi-transparent, polished ; whorls about 7, moderately convex and impressed at the suture, sculptured immediately below the suture with transverse rib-striz of varying strength, except the first 2} whorls, which are smooth, and the last half-whorl, in which the striz are continued across the width of the whorl and around the base; aperture somewhat roundly trigonal, broader than high, with more or less thickened and ex- panded peristome reflexed on the columellar and basal lips, furnished with at least four processes varying much in development, viz. a moderate, sometimes heavy, in-running parietal plait, a moderate or large, simple or more or less divided labral tooth, a small tooth about the base of the columella, and a wholly internal columellar plait, large or small, more or less thickened below, rounded or produced to a mammillate point at the lower right margin, the colu- mellar lip sometimes bearing a minute subcentral tooth or swelling. Height 3°03, width 1:19 mm. (average of 22 specimens measured), Hab.—Environs of Grahamstown, Cape of Good Hope. The following are the dimensions of the figured specimens and a few others: 78 HENRY CLIFDEN BURNUP. Height Width mm. mm. In my collection (figs. 61, 62, 63) Ah por oad Ae tleo, Ditto (fig.60) . : ; see oO Sens peleglG In collection of J. H. Ponwonliay (co-type of parallela) (fig. 65) eae eS pices OILY Ditto, ditto (fig. 66). 2 PaO Orla In collection of J. Farquahar (part of original lot of parallela) rake 1:24 Ditto, ditto. Peek’) 1-15 In collection of J. Papaaher (part of his series of montana). : Se sO) Tt Ditto, ditto . 3°02 ISIE) In collection of J. Farquhar, fom Cart 1e’s Kloof . : 30S) 45. Sales Ditto, ditto : eon In collection of Alexander McGregor Museum, Kimberley . : : ee 19) eam Ft 2} In collection of J. H. Ponsonby, from Hamilton Reservoir, Grahamstown (fig. 64) ; ; ao Oae 2A) 124 Hnnea ponsonbyi n.sp. Pl. V, figs. 67, 68, 69. Shell small, ovate-cylindrical, with blunt apex, vitreous, semi-transparent, whitish, deeply rimate and narrowly but distinctly perforate; whorls 6, rather convex, the first 2 smooth, the rest regularly, distinctly, rather strongly, sculptured with transverse rib-striz, except a small area on the body-whorl, the whole width of the whorl, immediately above the columellar lp, which is smooth; suture rather deeply impressed; aperture rather erect, roundly quadrate with whitish peristome much thickened, widely expanded and strongly reflexed, the ends connected by an indistinct callus, and the apertural processes being as follows: Ww JCreen, dees SOUTH AFRICAN ENNE&. Ann.Natal Mus.Vol lll. Pl.Y. Burnup del. J.Green lith. SOULH APRICAN ENNEA: . ' z | ®t ha ’ 7! a DEVELOPMENT OF PLANULA IN A PLUMULARIAN HYDROID. 83 On the Development of the Planula in a Certain Species of Plumularian Hydroid. By Ernest Warren, D.Se.(Lond.). With Plate VI and 4 Text-figs. ie Tue development of the planula in this hydroid is note- worthy in that the egg never becomes charged with yolk. The ovum remains small and segments in the midst of a feed- ing or placental tissue. Ultimately the embryo grows into a well-developed planula with dimensions very greatly exceeding those of the original egg. Some brief notes on the subject were communicated to the South African Association for the Advancement of Science at the annual meeting held in Lourenco Marques, July, 1913. In the present paper the necessary details with drawings are given, and the species of hydroid is described. The hydroid was found in January, 1911, in a rock-pool on the north side of the mouth of the St. John’s River, Pondo- land. Only two clusters were gathered, and they both occurred on the shells of living oysters much incrusted with the calcareous tubes of Serpula. The upright pinnate stems were about 4 in. in height. TropHosoME.—Hydrorhiza consists of an irregularly branched stolon, which in the present specimens was creep- ing among the tubes of Serpula covering oyster-shells (text- TExT-FIG. 1. DEVELOPMENT OF PLANULA IN A PLUMULARIAN HYDROID. 85 fig 1, A). The perisare is not so distinctly divided into an outer diffuse layer and an imner compact layer as in some species of Plumularia. Diameter of hydrorhiza about 0°19 mm.; thickness of perisare 22°7 mu. Hy drocaulus.—Monosiphonic, the hydrorhiza carries up- right pinnate stems of various heights. The proximal portion of the main-stem, not bearing pinne, is about half the length of the distal portion carrying them, and it is somewhat irregularly divided into internodes by transverse nodes (text-fig. 1, B). The distal portion of the stem is regularly divided by oblique nodes. Between the two portions of the main-stem there are two very oblique joints (text-fig. 1, B; C, n., n.). All the internodes of the main-stem in the distal pinnate portion carry a hydrotheca with a mesial sub-calycine nema- tophore and a pair of supra-calycine ones. The pinne alternate, and on anordinary stem there are nine to eleven on each side. They usually carry only two hydrothece, but sometimes there may be three (text-fig. 1, B). The pinnee are borne on a short process springing from the main-stem at the back of the hydrotheca (text-fig. 1, C). The proximal internode is very short, bears no structures, and the upper and lower nodes are transverse (1.1, .;) ; the rest of the pinna generally consists of only two alternating non-thecate and thecate internodes separated by oblique nodes. These non- thecate internodes carry a median nematophore. Frequently the pinne, especially the proximal ones, bear a pinnule which is carried on a short process springing from the level of the proximal hydrotheca. It consists of a basal internode (text-fig. 2, p.v.p.) with transverse nodes (text-fig. 1, C, 2.9, 2.9), followed by a non-thecate and thecate internode. TEXT-FIG. 1.—A (nat. size). Colony on oyster shell. Bx6. Hydrorhiza bearing pinnate stem with pinne carrying three hydrothece. C x 55. Piece of pinnate stem with male and female gonangia. D x 55. Side view of female gonangium carrying nematophores at the base. 86 ERNEST WARREN. It appears that the pinnules invariably spring from the lower side of the pinne. They are set at an angle of about 30° to the pinna, and project downwards and outwards. These short pinnules are characteristic of the species. Length of the pinnate-stem with 11 pairs of pinnz about 12 mm., proximal non-pinnate portion 4 mm., pinnate part 8mm. Diameter of main-stem in non-pinnate region 0°16 mm., thickness of perisare 21°58 n. Length of thecate internode about 0°30 mm., width 0°14 mm. The average length of TEXT-FIG. 2. = Se ae Internodes of pinna and base of pinnule, x 200. pinna with two hydrothecz is about 0°91 mm., and the average length of pinnule is about 0°51 mm. The ccenosare of the internodes exhibits an ectoderm channel on the posterior surface (text-fig. 3, A, p.e.c.). Hydrotheca.—Cup- shaped, rather deep, expanding, margin somewhat everted, adcauline surface not adnate for whole length, but free above for about one-half of its length. Plane of mouth of hydrotheca is set at an angle of about 35° to main-stem or pinna. The hydrothece are placed on the front or anterior surface of the pinne (text-fig. 1, C). Height of hydrotheca about 0°16 mm., width at mouth about the same. DEVELOPMENT OF PLANULA IN A PLUMULARIAN HYDROID. 87 Nematophores all free and have the typical plumularian structure ; there is a median sub-calycine nematophore, and a pair of larger supra-calycine ones whi¢h scarcely project beyond the level of the mouth of the hydrotheca (text-fig. 2). The sarcothece are bithalamic, canaliculate, and tend to be narrow at the base. Hydranth.—About twenty tentacles in a single verticil. Polyp constricted below the whorl of tentacles, dividing the coelenteron into an upper and lower division ; the upper part is lined by narrow columnar cells, the lower portion by vacuolated digestive endoderm (text-fig. 3, A). Nematocysts in the ectoderm of the tentacles are small; they measure about 3°2 win length and 1:06 u in breadth. In the nematophores the nematocysts are of considerable size, having a length of 14°8 4 and a breadth of 3°2 wu. (ronosomME.—The male gonangia are borne on the more proximally placed pinne. The female gonangia occur on the proximal portion of the pinnate region of the main-stem ; they arise just below the hydrotheca and on one side of the sub- calycine nematophore. The nutritive substances are naturally more abundant in the main-stem than in the lateral pinne, and from analogy with other hydroids we should expect to find the female gonangia in the former rather than in the latter position. ‘This is in accordance with the view that the development of the female sex is partly induced by a rich nutritive supply. Male Gonangium.—The gonotheca when fully grown is cylindrical, and is about 0°36 mm. in length and 0:12 mm. in width (text-fig. 1, C). The perisare is quite thin, and there is no distinct operculum. At the base where it springs from the pinna there is no jointed stalk, and there are no nematophores arising from it. In the developing male gonangium the spermatic cells, which will subsequently form the generative mass of the single reduced gonophore, may be seen embedded in the endoderm of the hydrocaulus (text-fig. 8, A, sp.c.). In the figure the spermatic cells are entering the blastostyle (b.G.) of SS nies ERNEST WARREN. the growing gonangium. Subsequently the genital cells will separate themselves from the endoderm on one side of the blastostyle, and this may be regarded as representing the formation of a gonophore (fig. B, t), which is covered by a thin layer of ectoderm cells continuous with the ectoderm of the blastostyle. At the apical region the ectoderm (c.p.) of the blastostyle consists of elongated cells constituting the “ Deckenplatte ” of Weismann. Female Gonangium.—The gonotheca when mature is ovoid and flattened. The plane of flattening is at right angles to the antero-posterior plane passing through the main-axis. The female gonotheca arises on the main-stem just below the hydrotheca and on one side of the sub-calycine nemato- phore. It possesses a short stalk-segment with transverse nodes (text-fig. 1, C, .,, 1.3). The stalk and the basal portion of the gonotheca are so curved that the main-axis of the structure is set at an angle of about 30° to the main-stem (text-fig. 1,C and D). The oldest gonangium is the most proximal one, and the gonangia are successively younger on passing distally. The perisarc is not very thick, but there isa well-defined large operculum, and there the perisarc is stouter. At the base of the gonotheca itself and on the flattened surface there is a nematophore on each side (text-fig. 1, D). Length about 0°81 mm., greatest lateral width 0°51 mm., thickness at right angles to plane of flattening 0°12 mm. The female gonangium bears in an obscure manner a single gonophore which becomes provided with a single ovum. Before the ovum commences segmentation the gono- theca is very hollow, and the blastostyle occurs as a thin tubular structure running through the mid-axis. Systematic Postrion.—The hydroid possesses the typi- cal characters of the genus Plumularia with the exception of the presence of the downwardly directed off- shoots from the pinne, and the somewhat unusual occurrence of the main-stem bearing hydrothece. These pinnules are similar in every way to the pinne, and they originate from 89 NT OF PLANULA IN A PLUMULARIAN HYDROID. 4 7 vi} DEVELOPMI st as the latter do from the main-stem: they the pinne ju bear a short basal internode with transverse nodes and no nematophore. Bale! deseribes both Plumularia campanula Busk and 9 v0. + Te TEXT-FIC ” a : <— <> ig AOS yp 062° a) 2o%, \ ee OL peaeBioegeePone o202% 20%? 3 60,2909 20°06 20% fElale/e O° 99 O65 SE Te ene ne ee es pes BOs |e plo © oje[ejo le 9 i Paik eae” \si8 RO Pao eps eA Pe eA ere yl ~tez hi Re” Se BCOOLC: ae, y Ie lePle Pea py ast Ne cleltefel = foleplelefeteys zs |" | ORT aie oP ea “ 0. 5 x 2 f—} 5 tions of male gonangium, Longitudinal sec the oy hydrothecz on similar Plumularia buskii Bale as_ bearing in condition a figures Cc ° an~ Allm and Plumularia armata Allman. . ’ main-stem ' Bale, W. M., ‘Catalogue of the Australian Hydroid Zoophytes,’ 1884, pp. 124 -5, Allman, G. J., “Challenger” Reports,’ vol. vii, 18 3, pl. iv, fig. 4. 8 9 90 ERNEST WARREN. In the species Plumularia aglaophenoides Bale, Bale! describes the occurrence of pmnules in the following terms: “Hydrocaulus polysiphonic, flexuous, pinne alternate, dis- tant, given off from each flexure of the stem, with two distinct oblique joints near the base, and obscurely jointed for the rest of their length, bearing hydrothece as well as pinnules; pin- nules alternate, approximate, one on each internode of the primary pinne, both series borne. towards the front and supporting a hydrotheca on each internode.” According to the definition of the genus Plumularia as given by Nutting, the pinne are without accessory branches of any kind; and the only other genera to which the present species could be referred are Polyplumularia G. O. Sars (modified by Nutting) and Schizotricha Allman (modified by Nutting). Nutting defines Polyplumularia as possess- ing pinne furnished with a hydrothecate ramulus or pinnule which springs from the first internode, and is more slender than the pinna from which it grows; while in Schizotricha the pinne bifurcate beyond the first internode at least in the mature colony. In the present species the ramulus or pinnule springs not from the first internode, but from the first hydrothecate inter- node. ‘The pinnule arises from the pinna by a short internode bearing no structures and with transverse nodes, just as the pinna itself arises from the main-stem. This condition is shown in Allman’s® figure of Schizotricha unifurcata Allman. Consequently, the lower of the two branches is to be regarded as the pinnule, and the upper as the distal portion of the pinna, as in the case of the species being described. It is not very clear that this production of pinnules forms a very satisfactory basis for founding a genus. The present species 1s transitional in this respect between Plumularia 1 Bale, W.M., loc. cit., p. 126. > Nutting, C. C.,‘ American Hydroids. Part I: The Plumularide,’ 1900, pp. 54, 78, 83. * Allman, G. J., loc. cit., pl. vii, fig. 2. DEVELOPMENT OF PLANULA IN A PLUMULARIAN HYDROID. 9] with no pinnules according to definition, and Schizotricha with many. In the present species the formation of the pin- nules tends to be limited to the more proximally placed pinne, while in Schizotricha it would appear that practically all the pinne bear them. Asa provisional designation, however, the name Schizotricha simplex is proposed. ie The Development of the Planula. The material was fixed in a warm alcoholic solution of cor- rosive sublimate and acetic acid. Several pinnate stems with female gonangia were sectioned in different planes, and the sections were stained with Delafield’s haematoxylin followed by orange. Eachstem bore gonangia of varying ages, and consequently all the different stages of development would appear to have been observed. Although such was the case, yet the youngest ovum that could be definitely identified as such was already in the endoderm situated at a short distance below the level of the mesial sub-calycine nematophore. Presumably, according to August Weismann and from analogy with observations on other hydroids, the ova first arose in the ectoderm, and then migrated into the endoderm ; but in the present case the young ova were not definitely located in the ectoderm. The youngest ovum seen was small, measuring about 14 pu in diameter, and surrounded by ordinary endoderm cells. The presence of the ovum causes a slight swelling to project into the lumen of the internode (PI. VI, fig. 1, O.). The ectoderm immediately above the area where the ovum is imbedded in the endoderm early becomes slightly modified in that the cells are more columnar and regular than ordinary ectoderm cells (fig. 1, b. G.). This is the first beginning of the future gonangium. The perisare situated just above (d. p.) has become markedly 92 ERNEST WARREN. thinner, apparently through some dissolving action of the subjacent ectoderm cells. In the next stage the plate of ectoderm has grown out into an ovoid swelling which is covered by a very delicate layer of perisare continuous with the general perisarc. The endoderm follows the ectoderm, and the portion of endoderm (fig. 2, end,) carrying the ovum (0.) passes into the gonangium, and becomes located on one side of the blastostyle. ‘The diameter of the ovum has increased to 17 wu. It may be mentioned here that all the figures in the plate are drawn from sections cut in the plane of flattening of the gonangium, with the exception of the following: figs. 7 and 18 represent sections at mght angles to the plane of flattening, and fig. 3 is a transverse section. At this time the mesogloea between the ectoderm and the area of endoderm in which the ovum occurs becomes thin and evanescent. ‘lhe ovum gradually separates from the endo- derm, and the ectoderm above becomes modified, divides into smaller cells and forms a kind of cap over it. This cap and the layer of endoderm (fig. 2, end., and fig. 4, g. e.) immediately below the ovum may be regarded as representing a rudimentary gonophore. The distal portion of the ectoderm of the young gonangium consists of tall columnar cells with short pear-shaped granular cells wedged between. ‘This is the beginning of the covering plate or “ Deckenplatte” (fig. 2, ¢. p.). The ccelenteron (C.) expands distally and becomes somewhat T-shaped. The ovum now enters the ectoderm (fig. 3, O.), and the covering cap becomes divisible into an inner columnar layer (fig. 4, 7. 1.) in immediate contact with the ovum, and an outer layer (o. l.) of flat cells. The shght crack-like space (fig. 3, c. U.) between the two layers is the equivalent of the umbrella- cavity of a medusa, or more complete gonophore, which arises as a split in a distally placed thickening of ectoderm. The ectoderm at the sides of the gonangium gradually contracts away from the perisarc of the gonotheca (fig. 4, sp.,, sp.9), and outside the flat ectodermal layer (0. /.) of the gonophore there DEVELOPMENT OF PLANULA IN A PLUMULARIAN HYDROID, 93 is a narrow space (figs. 3 and 4, sp.). The long diameter of the ovum has now increased to about 20 pe. The gonangium continues to expand, and the ovum grows shghtly and reaches a diameter of about 23, but there is no obvious yolk passed into it. The covering cap of ectoderm of two layers also grows, the outer layer (fig. 5, 0. l.) becomes thinner, and the inner (7./.) thicker. A delicate mesoglcea 1s formed between the ectodermal cap andthe endoderm, The endoderm cells on which the ovum is seated are granular and in an active physiological condition ; they stain more readily than the rest of the endoderm. The gonangium expands further, and the distal plate of tall columnar ectoderm cells (fig. 5, c. p.) begins to secrete a thicker layer of perisarc; this is the beginning of the operculum. The blastostyle, except in the region of the operculum, has become quite separated from the chitinous gvonotheca, and lies in the mid-axis. The ectoderm lning the umbrella-cavity of the gonophore, consisting of the two layers, becomes separated from the ovum, so that a very definite cavity occurs above the ege (fig. 6, c.0.). This ‘avity may be regarded as a space in the mesoglcea of the gonophore. The umbrella-cavity (c. U.) of the gonophore is very narrow and almost obliterated. The ovum is probably fertilised just, before this stage, and, presumably after such fertilisation, it appears to secrete a 7, 8, v. m.) which is thicker on the outer than on the inner side. It is possible, kind of vitelline membrane (figs. 6, however, that this membrane is really of the nature of mesogloea, and is not actually formed by the egg. ‘The endo- derm below the ovum becomes shehtly pushed outwards into a blunt process (fig. 8, d.C.) consisting of narrow granular cells, and the egg is seated on this out-pushing, which may be regarded as the spadix of the reduced gonophore. The ovum now attains its maximum diameter of about 34 yx. In text-fig. 4 a comparison is made between a typical gonophore (A) and the Inghly modified gonophore~ (B) carried by the blastostyle of the present species. 94, ERNEST WARREN. In Bthe diverticulum of the ccelenteron (C.) is directly comparable with the endodermal cavity of an ordinary gonophore, the granular layer (q. e.) being equivalent to the manubrial endoderm. Radial endodermal canals are not represented. The cavity (c. O.) around the outer surface of the egg is the space between the manubrial éndo- derm and ectoderm ; the manubrial ectoderm (Jf. e.) in the modified gonophore consists of a hemispherical layer of cubical cells. The umbrella-cavity (c. U.) is a narrow slit-like space. The outer ectoderm-lining (e.1. U.) to the umbrella- TrxtT-FIG. 4. Comparison of a typical gonophore (A) and the modified gonophore (B). cavity consists of a layer of flat cells. Outside this ecto- dermal lining of the umbrella there is a space (sp.) equivalent to the space between the umbrella-ectoderm and the outer ectoderm layer of the gonophore. The outermost ectoderm (e.l.) of a gonophore is indicated by a few flattened cells adherent to the tall attenuated cells lining the upper portion of the gonotheca. In fig. 7 of the Plate the gonangium is shown in vertical section at right angles to the plane of flattening. The terminal portion of the ccelenteron of the blastostyle is seen at C. and the out-pushing to form the spadix occurs at d. C. The gonangium now grows to its full size, and the cells of DEVELOPMENT OF PLANULA IN A PLUMULARIAN HYDROID. 95 the covering-plate and of the distal endoderm of the blasto- style become very attenuated and full of vacuoles, and will subsequently disappear. ‘lhe space (fig. 8, ¢. O.) between the egg and the ectoderm of the gonophoreé in- creases in size, and the membrane around the ovum (vr. m) is very pronounced. The operculum consists of thicker perisarc than that of the rest of the gonotheca, and its lower edge is sharply marked off from the thinner perisarc below. The gonotheca has now assumed its definite form, and it does not further increase in size; its greatest width is about 0°51 mm. and length 0°85 mm. At a somewhat later stage the outermost gonophore-layer of flat ectoderm cells disappears and the inner layer of cubical or columnar cells breaks up to form an irregular cluster around the ovum (fig. 9, e. G.). The covering plate of ectoderm cells below the operculum and the endoderm layer of the terminal portion of the blastostyle are still faintly visible, but they are on the point of disappearing. The out-pushing of endoderm (d. C.) or spadix of the gonophore consists of cells which remain in an actively living condition. The ovum does not grow, the protoplasm is finely granular, and there is a large nucleus with nucleolus (fig. 9). There is still a pronounced membrane around the egg; but it is much less developed on the mner side against the granular cells of the endoderm than on the outer side. In the next stage it will be seen that considerable changes have occurred in that the whole of the covering plate of ectoderm and the distal horizontal portion of the endoderm of the blastostyle have entirely disappeared. The out-pushing of endoderm (fig. 10, d. C.) or spadix of the gonophore has grown upwards, and has pushed the egg, together with the surrounding ceils derived from the breaking-up of the columnar ectoderm cells of the gonophore, into a more or less central position in close contact with the operculum. The vitelline membrane or mesoglcea-layer round the ovum be- comes less distinct. The cells around the ovum arrange 96 ERNEST WARREN. themselves in such a manner that they extend as columnar cells from the perisarc to the endoderm on which the ovum is still seated. Subsequently the ovum becomes quite central at the apex of the gonangium, and is closely surrounded by the tissue derived from the ectoderm of the rudimentary gonophore. The endoderm is now in the form of two symmetrically placed lateral lobes, one being the terminal portion of the endoderm of the blastostyle (fig. 11, t.p.), and the other is the endodermal spadix of the gonophore (d.C.). The egg is seated symme- trically between the two lobes. The ovum segments and a rounded cluster of about sixteen loosely attached blastomeres is formed (fig. 11, MW.). The cells of the endoderm immediately under the young embryo (fig. 12, end.), become taller and more granular than those of the rest of the endoderm. It is considered probable that the nutritive substances for the growth of the embryo are passed by this endoderm into the ectodermal tissue in which the embryo is imbedded. The blastomeres increase in number and become smaller. A segmentation-cavity soon appears (fig. 13, Sg. C.), and this is not central, but is nearer to the apex of the gonangium. Accordingly the embryo is two or three cells in thickness below, while above it is mostly only one cell thick. The cells of the placental tissue (pl.t.) increase in size and height; the tissue is continuous with the thin layer of ectoderm which covers the sides of the blastostyle (e. B.). The embryo is very obviously supphed with nourishment by the cells in which it is imbedded. The diameter of the embryo is two and a half times that of the original ovum, which never had any appreciable quantity of yolk. The embryo continues to expand and the constituent cells increase in number and diminish in size. The segementation- cavity becomes larger (fig. 14, B/.), A differentiation in the cells of the embryo can be seen on the lower side where an inner layer of flattened cells now becomes evident (end.,). It may be noticed that this inner layer of endoderm lines the DEVELOPMENT OF PLANULA IN A PLUMULARIAN HYDROID. 97 segmentation-cavity only on the lower half of the embryo. In the earliest stages of development the inner layer would appear to arise rather by differentiation than by delamination. The outer cells or ectoderm (ect.,) form a layer one cell thick above and several cells thick below. The cells of the placental tissue in the space between the cup of the endoderm and the lower two-thirds of the embryo also show differentiation. Some of the cells are rounded or somewhat stellate ; they are further distinguishable by staining much more readily, and they appear to consist of protoplasm denser than that of the other cells. It is possible that these cells are the more active agents in transferring nutriment to the growing embryo. The embryo continues to grow, and the cells cf which it is composed increase greatly in number, but their size remains the same as before. The embryo becomes pear-shaped (fiz. 15). The inner layer of flat endoderm cells (end.,) is beginning to creep upwards, but the upper half of the embryo is still only one cell thick and consists of the outer ectoderm layer. ‘The embryo as a whole is sinking down- wards, and the endodermal cup formed from blastostyle and gonophore is becoming deeply flask-shaped. The effective placental tissue ( pl.t.) now lies only between the embryo and the endoderm-cup. The outer portion of this tissue (e. H.), is extending downwards and consists of extremely clongated cells. At the mouth of the flask the ectoderm layer has separated slightly from the perisarc and forms a definite epithelium adhering to the endoderm, and leaves a slight crack-like space (d. S.), extending from the perisare to the pointed apex of the embryo. In the next stage the embryo has sunk still further towards the centre of the gonangium (fig. 16). The ectoderm of the embryo has become more than one cell thick all round, and the endoderm also is several cells thick below and has com- pletely surrounded the blastoccel (Bl.).. The placental tissue (pl. t.) 1s well-developed, while the outer portion (e. #.) retains its transparent, attenuated condition. The flask- shaped endoderm layer (gy. e.) derived from the blastostyle VoL. 3, PART I, 98 ERNEST WARREN. and gonophore is more granular than the rest of the endoderm. The outer surface of the ectoderm of the embryo (fig. 17, ect.,), is not sharply marked off from the placental tissue (pl. t.), and there appears to be the most intimate contact. Both rounded and stellate cells of dense protoplasm are seen in this placental tissue. Between the endoderm (end.,) and the ectoderm (ect.,) of the embryo a faint line of division can be seen, and this is the begimning of the mesogloea. Such is the appearance of the embryo in a section cut in the plane of flattening of the gonangium. In a vertical section at right angles to this plane the planula bears a different aspect. The sides of the embryo are squeezed flat against the perisare of the gonotheca with no gonangial ectoderm between (fig. 18). Laterally the embryo is very thin-walled, and both the ectoderm and endoderm layers are one cell thick, while on the other hand these layers are two or more cells thick at the distal and proximal ends. The blastoccel (Bl.) is roughly rectangular in section. Thus the embryo at this stage is of a curious shape. It is thick-walled everywhere (fig. 19), except laterally (fig. 18), over the flattened sides of the gonotheca, where the walls are very thin. The material has not permitted an examination of a later stage, but it is probable that the planula in the above con- dition is nearly ready to burst out of the gonotheca, for the gonotheca which was immediately below, and therefore shehtly older, was empty, and the operculum absent, doubt- less through the bursting out of the planula. Summary.—Here we have an interesting case of the nourishing of the developing embryo. In the great majority of Invertebrata the ovum becomes charged with sufficient food or yolk-material to enable development to proceed far enough for the young creature to be able to provide itself with food. In the present hydroid the egg remains quite small and is never provided with a perceptible quantity of yolk, but it segments, and development takes place, in a kind of maternal placental tissue which supples the embryo with food during DEVELOPMENT OF PLANULA IN A PLUMULARIAN HYDROID. 0}8) the whole development. his placental tissue arises as a modification of a specialised portion of the ectoderm of the blastostyle. This portion, as we have seen, forms a kind of cap over the young ovum, and doubtless may be regarded as representing the manubrial ectoderm of a rudimentary gono- phore which bears one egg. EXPLANATION OF PLATE VI, Illustrating Dr. Ernest Warren’s paper, “On the Develop- ment of the Planula in a certain Species of Plumularian Hydroid.” EXPLANATORY REFERENCES FOR PLATE AND TEXT-FIGURES. B. Blastostyle. Bl. Blastocel. 6. G. Beginning of gonangium. C. Celenteron. c¢. p. Covering plate (* Deckenplatte’) of ectoderm of gonangium. ¢. O. Cavity above ovum, which may be regarded as a cavity between the manubrial ectoderm and endoderm of gonophore. c. U. Cavity of umbrella. d. C. Diverticulum of ccelenteron representing the endoderm of gonophore. d.p. Dissolved perisare. e. B. Ectoderm of blastostyle. e. G. Scattered ectoderm cells of gonophore. e. H. Outer enveloping ectoderm derived from the scattered ectoderm cells of gono- phore. e./. External ectoderm layer of typical gonophore. e./. U. Keto- derm layer of umbrella. ect.,. Ectoderm of planula. end. Endoderm of blastostyle and of gonophore. end,. Endoderm of planula. g.e. Granular endoderm. 7./. Inner or manubrial ectoderm layer of gonophore. M. Morula. M.e. Manubrial ectoderm. n., x. Oblique nodes at base of pinnate portion of main-stem. #., 7., Transverse nodes at base of pinna. iy, Ny» Transverse nodes at the base of pinnule. O. Ovum, o. 1. Outer layer of gonophore, equivalent to the layer of ectoderm lining the um- brella-cavity. op. Operculum. jp. e. ¢. Posterior ectodermal space of hydrocaulus. p. 7. p. Proximal internode of pinnule. pi. ¢. Placental tissue. Sy. C. Segmentation-cavity. s.p. Space between the external ectoderm layer of gonophore and the ectoderm of the wmbrella-cavity. Sp.\, Sp» Spaces between perisare of gonotheca and blastostyle. sp. e¢. Spermatic cells in endoderm of hydrocaulus. #¢. Spermatic cells (testis) of male gonangium. ¢. p. Terminal portion of the ccelenteron of blasto- style. v. m. Membrane around ovwn. Fia. 1.— x 220. Longitudinal section through the main-stem imme- diately below a hydrotheca showing the beginning of the gonangium 100 ERNEST WARREN. (b. G.), the superimposed attenuated perisare (d. p.), and the youngest observed ovum (Q.) embedded in the endoderm. Fie, 2.— x 220. Longitudinal section through a young gonangium showing the ovum (0.) being thrust out into a special thickening of the ectoderm, which early becomes differentiated into an outer (o. 1.) and an inner layer (7. /.) and represents the ectodermal portion of a reduced gonophore. The slight space between the two layers is equivalent to the umbrella-cavity of a medusa. Fie. 3.—x 260. Transverse section through a young gonangium, showing the ovum surrounded by the ectodermof the gonophore. sp. is the space which arises between the outer ectoderm of the gonophore and the ectoderm of the reduced umbrella-cavity. Fie. 4.—~x 220. Longitudinal section through a slightly older gonangium. The gonophore, consisting entirely of ectoderm, is more developed. The outer layer (0. l.) is now well-defined, while the inner layer (¢..) closely surrounds the ovum (O.). The ectoderm of the gonan- gium at the distal surface has thickened considerably and forms the “covering cap” (c.p.). The space (sp.) is clearly seen. Spaces (sp.,, sp.s) are appearing between the perisare of the gonotheca and the general ectoderm. The endoderm immediately under the gonophore ectoderm is more granular than the remainder of that layer. Fie. 5—x 220. Longitudinal section through an older gonangium. The outer layer of ectoderm (o0.1.) has become more fiattened through pressure exerted by the expanding inner layer (7. /.). Beneath the gono- phore the endoderm is beginning to grow and to form a diverticulum of the ccelenteron which represents the ccelenteron cavity of the reduced gonophore. Fic, 6.— x 220. In this older stage the longitudinal section shows a well-defined space (c. O.) between the ovum and the ectoderm (7. /.) of the gonophore. The ovum is surrounded by a membrane which is more marked externally than on the inner surface, which is in contact with the endoderm. Fie. 7.—x 220. Longitudinal section of gonangium at right angles to the plane of flattening. Here the cavity (c. O.) of the gonophore has expanded, and the membrane around the ovum (v. 7.) is thickened. Fie. 8.—x 220. Longitudinal section of gonangium in the plane of flattening. This is in the same stage of growth as in fig. 7. Fie. 9— x 220. Longitudinal section at a later stage. The ectoderm layer of the gonophore has broken down (e. G.), and forms an irregular mass of tissue around the ovum. The endoderm below is granular and in an active condition of growth. The cells of the distally placed DEVELOPMENT OF PLANULA IN A PLUMULARIAN HYDROID. 10] endoderm and those of the ‘covering cap” have become very elongated and attenuated. Fie. 10.— x 220. Longitudinal section at the next stage. The distal endoderm and the “ covering cap” have completely disappeared and the ovum is surrounded by cells which have arisen from the scattered ecto- derm of the broken-down gonophore. The ovum is being pushed into a central position by the growth of the diverticulum of the colenteron (d. C.) which is the endodermal spadix of the gonophore. Fic. 11.— x 220. Longitudinal section showing embryo (M.) derived from the segmentation of the ovum. It consists of about sixteen cells, and is central in position immediately below the middle point of the distal surface of the perisare of the gonotheca. The embryo is sur- rounded by the placental tissue. The diverticulum of the colenteron (d. C.) or gonophore spadix has become equal in size to the terminal portion (f. p.) of the ccelenteron of the gonangiun. Fira. 12.—~x 480. Enlarged view of morula imbedded in placental tissue (pl. t.). Fie. 13.—x 220. Later stage. The embryo, the placental tissue, and the celenteron pouches have grown. The embryo has developed a segmentation space (sg. C.). Fic. 14.— x 220. In the next stage, a differentiation of the cells of the embryo into an outer ectoderm and an inner endoderm is seen, Distally there is only ectoderm (ect.,) which is very thin; proximally the ectoderm is thicker and is more than one cell thick. Fie. 15.—x 220. The embryo becomes pear-shaped, and the endo- derm is not continued distally. The maternal ectoderm (e H.), derived originally from the placental tissue, is continued around the outside of the cwlenteron pouches, and the cells become very elongated and reach the lateral perisare of the gonotheca. Fie. 16.—x 229. A later stage where the embryo has grown con- siderably and has extended inwards towards the centre of the gonangiun. The calenteron pouches, consisting, on one side, of the spadix of the gonophore, and on the other, of the terminal portion of the colenteron of the blastostyle, form a deep cup containing the embryo and the true placental tissue (pl. ¢.). The enveloping ectoderm (e. EH.) has also ereatly extended. Bl. is blastocceel. The endoderm (gq. e.) of the gonangium at the base of the embryo is more granular than elsewhere. The endoderm of the embryo now forms a continuous layer. It is mostly more than one cell thick, as also is the outer ectoderm. Fie. 17.—~x 440. Enlarged view of a piece of embryo, the surround- ing placental tissue and the endoderm of gonangium. The ectoderm LO? ERNEST WARREN. (ect.,) of the embryo is in intimate contact with the placental tissue (pl. t.). Fie. 18.— x 220. Longitudinal section of gonangium at right angles to the plane of flattening. The planula is seen wedged between the perisare of the gonotheca. The wall of the planula is very thin laterally, but much thicker proximally and distally. Fie. 19.—x 220. The planula in section in the plane of flattening of the gonotheca. The wall of the planula is of considerable thickness all round. Both the endoderm (end.,) and ectoderm (ect.,) are several cells in thickness. — Ann. Natal Mus. Vol. ee bs = — Ve Ty, 19 xe20. 16 x ero. 17 x 440. aeenac, 14 x 220, Warren del. Huth, London oe ; hfe A TERMITOPHILOUS BEETLE OF THE GENUS COROTOCA. 103 Note on the Occurrence in South Africa of a Termitophilous Beetle of the Genus Corotoca. By Ernest Warren, D.Sc.Lond. AmonG the most remarkable and interesting of the termito- philous insects are the beetles of the genera Spirachtha and Corotoca, belonging to the family Staphylinide. They have a much swollen abdomen, with the greater portion of it permanently turned forwards over the dorsal surface of the thorax. It extends as far as the middle of the pronotum. These beetles were first described and figured by J. C. Schiddte in 1854 from specimens obtained from termite nests in Brazil. In 1905, Dr. Ivar Tragardh, of the Upsala University, visited South Africa on a collecting trip, and he discovered in the nest of a species of Hutermes, in a limited area in Zulu- land, a great number of specimens of anew Staphylinid, which he has called Termitomimus. This genus is characterised by the second and third segments of the abdomen being very greatly swollen, and the remainder of the abdomen, which is somewhat less swollen, is turned forward over the thorax. Dr. Tragardh examined many nests of termites in Natal, but no other physogastric Staphylinids were discovered. Recently, numerous nests of the common termite Hutermes trinervius (Rambur) have been examined in another connec- tion by Dr. Conrad Akerman and myself in the neighbourhood of Pietermaritzburg, and in March, 1913, Dr. Akerman found in one of the nests a single specimen of a physogastric Staphylinid. Subsequent examination proved that it un- 104. ERNEST WARREN. doubtedly belongs to the genus Corotoca, and is closely allied to, although specifically distinct from, Corotoca melantho Schiodte of Brazil. Since this discovery, about 120 nests in different localities have been examined with a view to obtaining more material and ascertaining the life-history of the species, but only three other specimens have been found. In certain nests a beetle-larva with some resemblance to the larva figured by Schiddte as being that of Corotoca melantho has been found in fair numbers, but all attempts to rear it have proved fruitless. The termite Hutermes trinervius does not live for any length of time under arti- ficial conditions, and the beetle-larva is even more delicate. The antenne and mouth-parts are quite lke those of the larva of C. melantho; the abdominal segments also bear lateral processes, but in the present species they carry bristles. On the whole it seems probable that the larva is that of the Corotoca, but it is remarkable that it is comparatively common, while the beetle appears to be excessively rare, and this rarity explains the circumstance that apparently it has not been previously found. A possible explanation is that the termites interfere injuriously with the beetle when in the pupal condition, so that only a very few ever arrive at the perfect state. It is hoped that when further material has been collected, a comparison of the anatomy of this genus may be made with that of Termitomimus entendveniensis T'rag., which was carefully described and figured by Dr. Tragardh! in 1907. The occurrence of this highly specialised termitophilous genus Corotoca both in Brazil and South Africa is interesting in indicating its antiquity, and in connection with the fact that the similarly modified termitophilous dipteron, T’erm1i- toxenia, is found both in South Africa and India. 1 Tragdirdh, Ivar, ‘ Description of Termitomimus, a New Genus of termitophilous physogastric Aleocharini, with Notes on its Anatomy, Upsala, 1907. A 'TERMITOPHILOUS BEETLE OF THE GENUS CoROTOCA. 105 J. S. Schiddte! has given a careful definition of the genus Corotoca, and also concise diagnoses of two species, C. melantho and C. phylo, both from Brazil. These diagnoses are quoted below in order that they may be compared with the brief description of the new species which is now given, pending a detailed account of the anatomy of the insect, which it is proposed to publish later. I have pleasure in associating this imteresting species with its discoverer, Dr. Conrad Akerman. “1, Corotoca melantho Schiodte. “Fusca, fronte foveolata, pronoto multifoveolato, disco bituberculato : tibiis posterioribus fusiformibus, fuscis: scutis ventralibus segmenti quarti quintique transversis.— Mas, Fem. “Long. a fronte ad apicem segmenti secundi abdominis, 23-3 millim. “2. Corotoca phylo Schidte. ‘“Fusea, vertice foveolato, pronoto multifoveolata, disco trituberculato: tibiis posterioribus lnearibus, nigrofuscis : scutis ventralibus segmenti quarti quintique subquadratis.— Fem. “Long. a fronte ad apicem segmenti secundi abdomuinis, 24-3) millim.” 3. Corotoca akermani sp. n. Brown, forehead or vertex not pitted, occiput flat or she¢htly concave, not tuberculated ; pronotum smooth, except for a very shallow and rather wide and inconspicuous median groove extending along the posterior three-quarters of its length ; posterior tibie rather slender, straight or shghtly curved, not fusiform, but rod-shaped; sternites of the third, fourth, fifth, sixth and seventh abdominal segments appear as five isolated, uniform, narrow, transverse, brown strips of chitin ' Schiddte, J. C., ‘ Corotoca og Spirachtha, Kjgbenhavn, 1854, pp. 8-9. VOL. 3, PART 1. 8 106 ERNEST WARREN. on the upper surface, all of which are widely separated from one another by intersegmental membranes caused by the swelling of the abdomen ; the morphologically dorsal surface of the abdomen rather deeply concave, fitting over the thoracic region, which is not flattened or otherwise much modified. Length from forehead to hinder end of upturned abdomen, 3°2-3°6 mm. Loc.—Environs of Pietermaritzburg: Durban Road (Dr. Akerman, March, 1913); Botanical Gardens (W. Culling- worth, July, 1913); Sweetwaters Road (July, 1913); Bisley Road (December, 1913). ol. IIL, part 2, issued October 22nd, 1915. a a a a en THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 107 Studies on The Carnivorous Slugs of South Africa, INCLUDING A Monograph on the Genus Apera, and a Discussion on the Phylogeny of the Aperide, Testacellide, and other Agnathous Pulmonata. By Hugh Watson, Wi.A. With Plates VII-X XIV, and 7 text-figures. CONTENTS. INTRODUCTION . : ; . 108 THE History OF THE GENUS APERA ; _ 109 A DESCRIPTION OF THE GENUS APERA pela fs External Characters : ets The Structure of the Skin . 124 The Shell, the Mantle-cavity, and the Mantle leary The Pedal Gland : : : 5 Als} The Nervous System. . 1386 The Structure of the Upper Mereacies ond the Mente cole Retractors : . 153 The Digestive System. . 155 The Vascular System. ; vst The Excretory System . . 186 The Reproductive Organs . 188 Distribution. : . 193 THe DIFFERENTIAL CHARACTERS OF THE SPECIES AND SUB- SPECIES OF APERA . . 193 Apera gibbonsi (Binney) ; ; . 193 rubella x. subsp. : . 197 gracilis n. subsp. . 199 lupata n. subsp. . 200 VOL. 5, PART: 2. 9 108 HUGH WATSON. , PAGE Apera parva %. sp. ; : : : . 202 Apera dimidia n. sp. . : : : . 204. Apera purcelli Collinge : . 207 Apera burnupi Smith : . 210 Apera sexangula n. sp. : : . 213 THe PHYLOGENY OF THE SPECIES OF APERA : SPAly/ THE OCCURRENCE AT CAPE TOWN OF TESTACELLA MAUGEI Fer. : ; ; ; : . 220 THE AFFINITIES 0F APERA AND TESTACELLA ; SO, The Evolution of Carnivorous Characters . : SOR The Phylogeny of Apera and other naked Carnivorous Slugs 231 The Phylogeny of Testacella and its possible Allies . 238 APPENDIX—BIBLIOGRAPHY OF TESTACELLA MAUGBEI Fér. . 256 EXPLANATION OF PLATES . : ? ; SoS INTRODUCTION. Many snails and slugs of great interest inhabit South Africa, but very little is known about their anatomy. I have therefore been asked to describe some of these animals, beginning with those in which the shell is degenerate or absent. The present article deals with the slugs belonging to the carnivorous group, and principally with those contained in the genus Apera; and as this genus is a very remarkable one and seems to be confined to South Africa, I am endeavour- ing to describe it in some detail. Unfortunately, however, the material at my disposal has been limited to a small number of preserved specimens, some of which have proved to be immature. More than half of these specimens were collected and sent to me by Mr. H. C. Burnup, of Pieter- maritzburg, to whom I am very deeply indebted. I am also elad to have this opportunity of expressing my thanks to the Rev. Prof. H. M. Gwatkin, to Dr. Doncaster of Cambridge, Dr. Pérmguey of Cape Town, Dr. Pilsbry of Philadelphia, Dr. Simroth of Leipzig, and Dr. Warren of Pietermaritzburg, to Major M. Connolly, and to Messrs. Keppel H. Barnard, Walter KH. Collinge, John Farquhar, John H. Ponsonby, G. C. Robson, Edgar A. Smith, and William ams. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 109 VEE ASTOR OR “RE h GHNWS ACP ERA. Nearly forty years ago, when Mr. J. 5. Gibbons was visit- ing Natal, he found a strange slug under a stone in a wood. This «animal puzzled him greatly ; it seemed more hke a Testacella than anything else he knew; yet clearly it was not a T'estacella, for where the shell should have been there was merely a little hole—the respiratory orifice.! Gibbons therefore sent the slug to Mr. W. G. Binney, with a description of its appearance when alive. Binney. examined its radula, and found that its teeth were lke those of Glandina. He also discovered that it had no jaw, but that it possessed an internal shell, which was broken when he found it, but which he wrongly assumed to be hexagonal when perfect. And Binney? established a new genus for this peculiar animal, and published a description of it im 1879, together with Gibbons’s account of its appearance when alive, a rough drawing of the animal, and a figure of its radula. And he named the slug Chlamydephorus gibbons. In 1884 Binney? repeated his description of the radula of this slug and reproduced his figure. In the same year Tryon 4 reproduced the rough drawing of the animal, and gave a brief description of the genus, which he placed in the Testacel- lidee; and he also amended the spelling of the generic name, changing it to Chlamydophorus. In the following year, however, he reverted to the original spelling, when he des- eribed the genus and species in the ‘ Manual of Conchology,’ ® and again reproduced the rough drawing of the animal. Now Binney named the genus Chlamydephorus because he said that the mantle covered the whole back. But in this Binney was wrong, as Heynemann ° was the first to point out, 1 Vide Collinge, W. E., ‘ Ann. Natal Mus.,’ 1910, vol. ii, p. 164. * * Bull. Mus. Comp. Zool. Harvard,’ vol. v, p. 331, pl. ii, figs. a, b. 3 Notes on the Jaw and Lingual Dentition of Pulmonate Mol- lusks,” ‘Ann. N. Y. Acad. Sci.,’ vol. iii, p. 81, pl. xvii, fig. a. ‘ «Struct. and Syst. Conch.,’ vol. iii, p. 13, pl. ci, fig. 47. > Vol. i (Qnd ser.), pp. 7, 17, pl. ui, fig. 95. §* Jahrb. d. Deutsch. Mal. Ges.,’ vol. xii, p. 17, pl. ii, figs. 5-7. 110 HUGH WATSON. when, in 1885, he published a description and three figures of the external characters of a specimen in the British Museum, labelled “ Cape Colony.” He thought that the animal was nearly related to Testacella; and as he showed that it had no mantle on its back, he suggested that the generic name should be changed to A pera. In 1890 Prof. T. D. A. Cockerell! published a short descrip- tion with measurements of what appears to be the same specimen, and he said that Heynemann’s name would pro- bably have to be used, as Chlamydophorus (differing from Binney’s name only by one letter) had been employed by Harlan for a genus of Mammalia in 1825. In 1892 Mr. E. A. Smith 2 described the external characters of anew species of Apera, very different from A. gibbonsi, which Mr. H. ©. Burnup had found in Natal, and he named it Apera burnupi. He also thought that Apera was the best name to use, as Chlamydophorus had been employed in the Mammalia, but he attributed the latter name to Agassiz (‘ Nomenclator Zoologicus, Mammals,’ 1842, p. 8), instead of to Harlan. Prof. Cockerell* then discovered that Apera was preoccupied in botany, and therefore thought that it might be better to use Binney’s name with the original spell- ing. He placed the genus in the Testacellide. On the other hand, Dr. Simroth,* after studying Heynemann’s article, suggested that the genus might be allied to the Janellide, its carnivorous characters having arisen independently of those of Testacella. He thought that the fragmentary condition of the shell, as found by Binney, might be natural ; but had difficulty in explaining how it was that Apera had two pairs of tentacles and Janella only one. In 1895. Apera and Chlamydephorus were treated as though. they were separate genera by Dr. A. H. Cooke.° 1 * Ann. Mac. Nat. Hist.’ (6th ser.), vol. vi, p. 390. ‘Ann. Mag. Nat. Hist.’ (6th ser.), vol. x, p. 466. 3 * Conchologist, 1893, vol. 11, p. 206. 4 * Nacktschnecken, Deutsch-Ost-Afrika,’ 1895, ex. vol. iv, p. 20. 5 «The Cambridge Natural History.’ vol. iii, ‘Molluscs and Brachio- pods, pp. 333, 334, 440. we YHE CARNIVOROUS SLUGS OF SOUJH AFRICA. 11] ‘Two years later Mi. W. E. Collinge! described and figured the alimentary canal, pedal gland, and reproductive organs of a specimen of Apera which he had received from Mr. E. A. Smith. Unfortunately this specimen was not in good condition for dissection, and Collinge’s description is inaccurate in several respects, and his figures bear but little resemblance to the natural condition of the organs that they are intended to portray. Moreover the specimen was named Apera burnupi, whereas it really belonged to a quite distinct species. Collinge stated that he considered that Apera had affinities with the Testacellida, but he thought that there was a wide gap between it and either Testacella or Daudebardia, and a still wider gap between A pera and Schizoglossa. In the following year Dr. Pilsbry published a short review of Collinge’s article in the ‘ Nautilus.’ ? In 1899 Paul and Fritz Sarasin” brought forward the theory that Apera was ancestral to Atopus. Now Atopus has a large mantle covermg the whole body, and the brothers Sarasin therefore rejected Heynemann’s interpretation of the morphology of Apera, and advocated the use of Binney’s name Chlamydephorus, In 1900 Collinge * received through Mr. J. H. Ponsonby an example of Apera burnupi Smith, found by the Rev. J. R. Ward at Richmond, Natal, and belonging to the South African Museum, Cape ‘own. He at once saw that this specimen differed from the species which he regarded as Apera burnupi, and accordingly he named it Apera natalensis wv. sp. He made a superficial dissection of some of the internal organs, but the specimen does not appear to have been in a much better condition than that of the other species which he had previously dissected, and his figures of its anatomy are in consequence equally faulty. In the same 1 ‘Ann. Mag. Nat. Hist.’ (6th ser.), 1897, vol. xx, pp. 221-225, pl. v, figs. 1-6. 2 Vole xi, ps 12: * * Die Land-Mollusken von Celebes,’ p. 112. * Ann. 8S. Afr. Mus.,’ vol. ii, pp. 3-5, pl. i, figs. 3, 4, pl. ii, figs. 14, 15 QGl S¥5° ay Oe uf LIBRARY |S \ \ =oe /y a o 1 HUGH WATSON. paper Collinge published two coloured figures of both species by Mr. F. J. Partridge. In the following year Collinge! had an opportunity of examining a better specimen of the species which he supposed to be Apera burnupi, and he discovered how misleading were his former figures of its anatomy. He _ therefore published a more accurate figure and description of the reproductive organs of that species. At the same time he criticised the theory of P. and F. Sarasin, and said that Apera was probably closely allied to the Testacellidew, while Atopus on the other hand seemed to be far removed from that family. In December, 1901, Collinge* described the external features of a new species of Apera found by Mr. R. M. Lightfoot on the slopes of Table Mountain, Cape Town, and published coloured figures of it by Partridge. This species he named A pera purcelli. Dr. Simroth,’® in the same month, published a copy of Collinge’s figure of Apera natalensis (= A. burnupi Smith), in his interesting article on carnivorous snails and slugs. In this paper he unfortunately stated that Apera was without a shell, and again suggested that the genus might have been derived from the Janellide, or possibly from some other ancient family of herbivorous slugs found in the Southern Hemisphere, although he admitted that too little was known of the genus to give any value to his hypo- thesis. In 1902 Collinge? reproduced on a larger scale his second figure of the reproductive organs of the species which he erroneously supposed to be Apera burnupi. In the fol- lowing year Dr. von Moellendorff ° removed the genus irom the Testacellide, and placed it in a family by itself, which he 1 * Journ. of Malac.,’ vol. viii, pp. 71, 72. 2 “Ann. 8. Afr. Mus.,’ vol. ii, p. 2380, pl. xiv, figs. 1, 2. 3 *‘Naturwiss. Wochenscbrift,’ vol. xvii, p. 111. fig. 7. 4 «Journ. of Malac.,’ vol. ix, pl. vi, fig. 66. » *Conch.-Cabinet (Agnatha),’ p. 5. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 1 ae: named the Aperide. In 1906 Heynemann! devoted a para- graph to the genus in his article on the geographical distri- bution of slugs. In 1907 Dr. Pilsbry * followed Moellendorff’s example and placed the genus in the family A peridz, which he included in his super-family Agnathomorpha; and he suggested that it might prove to be more nearly related to the Rhytididz than to the other carnivorous families. In 1909 and 1912 Simroth *® included in Bronn’s ‘ Tier-Reich’ some of the information given in Collinge’s articles, and published copies of four of the figures which had illustrated his first paper in the ‘Annals of the South African Museum.’ In 1910 Collinge* gave a short account of the external features of the three species of Apera then known to inhabit Natal, together with a definition of the genus, and some remarks on its habits, distribution, affinities, and history. This paper contains valuable quotations from letters of Gibbons and Burnup, but the error about Apera burnup1 is repeated. In treating of the affinities of the genus, Collinge again stated that he considered that P. and F. Sarasin were wrong in regarding A pera asin any way related to Atopus; but he now expressed the opinion that it was not related to Testacella either, and thought that it should probably stand by itself as a remnant of some very ancient stock of carnivorous Pulmonates. Accordingly he placed it in a separate family, which he termed the Aperaide fam. nov., being apparently ignorant of the classification of both Moellendorff and Pilsbry. Lastly, in 1912, Major Connolly * included A pera in his valuable “ Reference List of South African Non-marine Mol- lusea,” following Collinge’s nomenclature of the species. Such is the story of the treatment that A pera has received, 1+ Abhandl. Her. v. d. Senkenb. Naturf. Gesell. (Frankfurt), vol. Xxx, p. 23. > *Manual of Conchology’ (2nd ser.), vol. xix. pp. 1x, xi. ’* Klassen und Ordnungen des Tier-Reichs III, Gastropoda Pul- monata,’ pp. 143, 611, 612, fig. 42c¢, pl. iv, figs. 9-11. + * Ann. Natal Mus.,’ vol. ii, pp. 163-170. > *Ann. §. Afr. Mus.,’ vol. xi, pp. 62, 63. 114 HUGH WATSON. and it is largely a story of mistakes. It will have been seen that four species have been described, three from Natal and one from the Cape, but that very little is known about the anatomy of any of these, as only two reliable figures have hitherto been published of their internal features, namely, Binney’s figure of the radula of Apera gibbonsi, and Collinge’s second figure of the reproductive organs of the species which he supposed to be A. burnupi. Nevertheless this ignorance does not seem to have prevented malacologists from speculating about the affinities of the genus; for, as we have seen, it has been suggested in turn that it may be related to the Testacellide, to the Janellide, to Atopus, or to the Rhytidide, forms which differ immensely from one another. It is hoped that the information contained in the present article may give greater value to such speculations in the future. In the following pages I am attempting to describe this remarkable genus in as great detail as the hmited number of specimens at my disposal will permit, giving numerous figures of its anatomy. Secondly, I am giving a concise description of the distinctive characters—both external and internal—of each species; that is to say, of Apera gibbonsi, A. pur- celli, A. burnupi, the species that Collinge erroneously supposed to be A. burnupi, which I am naming A. sex- angula, and two new species—A. dimidia from Natal, and A. parva from Grahamstown. And at the same time | am pointing out that there are some very distinct races of A. gibbonsi, which may also prove to be separate species, but which I am provisionally regarding as sub-species until more material is obtainable. After dealing with the relation of these forms to one another, | am describing briefly the species of Testacella that occurs at the Cape, and am then dis- cussing at some leneth the probable relations of Apera and Testacella to each other and to the remaining genera of carnivorous snails and slugs. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 1 LS} A DESCRIPTION OF THE GENUS APEHRA. EXTERNAL CHARACTERS. GENERAL SHarpe.—The slugs contained in the genus A pera are subcylindrical, tapering gradually towards the head and more abruptly at the hind end, the broadest region being usually behind the middle of the animal. he height is equal to the breadth or sometimes slightly less, while the width of the foot varies from three-fifths to scarcely one-half of the greatest breadth of the body. Most of the species are rather slender, especially A. parva and some forms of A. gibbonsi; and it is interesting to notice that the latter do not contract when about to rest, after the usual manner, but merely bend their bodies at various angles, in which state, Mr. Burnup intorms me, they much resemble the rhizomes of terns. S1zE.—NSo few specimens of Apera have been examined that the mean dimensions attained by each species cannot be stated with accuracy. It seems probable, however, that the average length of each of the three larger species—A. gibbonsi, A. burnupi and A. sexangula—hes between 60 and 65 mm., in the case of specimens preserved in alcohol, although occasional examples may reach a length of nearly 80 mm. On the other hand adult specimens of A. dimidia preserved in spirit are only about 33 mm. long, and A. parva and A. purcelli seem to be still smaller species. ‘TenracLes.—As usual inthe Stylommatophora, there are two pairs of tentacles, the upper pair being retractile and bearing the eyes.'! he lower tentacles are small, and beneath them there is another pair of short processes of a paler colour, which doubtless correspond to the labial feelers found in most of the carnivorous genera. GeniraL Opentnc.—The common opening of the reproductive organs is on the right side of the head, two or three milli- metres behind the right lower tentacle, and about half that distance from the margin of the foot. 1 See pp. 153, 154 for structural details. LAG HUGH WATSON. Resprratory Oprenrnc.—The aperture of the mantle-cavity or “respiratory orifice” is situated towards the hind end of the back, just to the rignt of the mid-dorsal line. Its posterior position, about four times as far from the head as from the hind end of the slug, is due to the lengthening of the anterior part of the body and the retrogression of the pallial organs, as in Selenochlamys and Testacella. The dorsal posi- tion of the opening is explained by the absence of any external mantle or shield in Apera.? The orifice is always a little nearer the right side than the left, a fact which previous writers seem to have overlooked, although it is of considerable morphological importance, as may be seen from Pl. X, fig. 33 (a transverse section through the respiratory opening of Apera gibbonsi rubella). This illustration also shows that the opening is surrounded by two. lips, one within the other. In Apera gibbonsiand A. parva the outer hp projects over the inner lip so as almost to con- ceal it; but im the other four species the outer lip is much narrower, and leaves a wide circular opening, within which the light-coloured inner hp is exposed to view (Pl. VII, fig. 12; Pl. VEL, figs. 24, 25, 26).. The right anterior edge of the inner lip is cleft by an oblique slit, running forwards and outwards from the central aperture. Kerets.—In most of the species of Apera the back is rounded, but A. burnupiand A. sexangula both have four longitudinal keels, two on each side of the body, which may possibly serve to increase the rigidity of the skin. The upper keels of A. sexangula are nearly parallel throughout the greater part of theirlength, the distance between them measur- ing about three-fifths of the breadth of the body. When followed backwards, however, they diverge very slightly just in front of the respiratory opening, and then gradually approach each other behind it, until they unite to form a single median keel, 3 or 4mm. in length, at the hind end of the animal. In A. burnupi the upper keels are further apart, and become very prominent towards the hind end, diverging considerably and 1 See pp. 130, 131. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. Viture then curving round to meet in a very obtuse angle just above the posterior extremity of the foot. The lower keels of Apera sexangula extend on each side along the whole length of the animal, about half-way between the upper keels and the edges of the foot. The slug is thus roughly hexagonal in transverse section, but when it con- tracts the surfaces between the keels become concave. In A. burnupi the lower keels are much nearer the foot than in A.sexangula, and become obsolete towards the hind end of the slug. The other species of Apera are without any traces of either median or lateral keels. They are bluntly pointed at the posterior end. ‘here is no caudal mucous pore in A pera. Dermat Grooves.—In all the species of Apera the back and sides of the animal are covered with a network of grooves, dividing the skin into numerous polygonal ruge. These grooves are specially deep in A. burnupi. The centres of the ruge are often raised in well-preserved specimens, and the skin has therefore a granular appearance. Certain of the grooves are larger than the others, and run in more definite directions, forming as it were the main channels from which branch the smaller grooves that form the network. First there are the two dorsal grooves which run along the middle of the back from the respiratory opening to the head. These are most strongly developed in Apera burnupi, and least conspicuous in A. dimidia, in which species the main grooves are scarcely more distinct than the other grooves ot the network. The distance sepa- rating the dorsal grooves varies from about one-seventh of the breadth of the body in A. dimidia to less than half that proportion in some forms of A. gibbonsi (Pl. VII, fig. 4). In A. burnupi, A. sexangula, and A. pur- celli these grooves are separate throughout their entire length, but in A. gibbonsi, A: parva, and A. dimidia they are united posteriorly, and arise from the respiratory opening as a single groove which divides between 3 and 118 HUGH WATSON. 9 mm.in front of the opening (Pl. VII, figs. 1, 4; Pl. VIII, figs. 22, 24). A single median posterior groove runs backwards from the orifice. It is usually deep and conspicuous in front, but breaks up into finer irregular grooves before reaching the hind end of the slug. Well-marked radial grooves diverge laterally from the region of the respiratory opening. About four of these on each side arise from the opening itself, and two or three others branch from the median posterior groove behind the opening. In Apera purcelli, the most anterior grooves, instead of arising from the respiratory opening, branch from the dorsal grooves in front of it (Pl. VIII, fig. 25). As a rule the radial grooves extend down the sides of the body, but in A. gib- bonsi some of them usually bifurcate before reaching the edge of the foot. The most anterior grooves of A. gibbonsi, A. parva, and A. dimidia always divide within a short distance of the respiratory opening, the lower branch sloping downwards towards the foot, while the other continues forward. The latter branch soon divides again in a similar manner, and this process is repeated several times, so that an irregular lateral groove is formed, which runs forward parallel to the dorsal groove and gives off oblique sub-lateral grooves parallel to one another. The lateral grooves are seldom well developed in A. gibbonsi and A. dimidia, but the oblique sub-lateral grooves are rather conspicuous in the former species (Pl. VII, figs. 2, 3, 5,6). On the other hand, in A. parva, A. burnupi, and A. sexangula the lateral grooves are more definite and better developed than the sub-lateral grooves, of which there are only a small number (Pl. VIII, fig. 23; Pl. VII, figs. 11,13). The lateral grooves of the two latter forms are lower down than those of the other species, and branch from the most anterior pair of radial grooves at a greater distance from the respiratory opening. ‘They run along between the upper and lower keels as far as the head, the right lateral groove ending in the genital opening as in Testacella and other forms. It is probable that the greater THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 119 development and lower position of the lateral grooves in these two species is due to the presence of the keels. ‘here are no definite lateral grooves in A. purcelli. The foot-frnge is bounded above by a peripodial groove, which is usually rather deep (PI. XI, fig. 35). In Apera burnupi and A. sexangula there is a second groove, parallel to the peripodial groove, but separated from it by a narrow row of ruge, and the radial grooves do not extend below this upper groove, but terminate in it. As a rule the other species have only the single peripodial groove with which the radial grooves unite, but in A. dimidia the upper limits of the lowest row of ruge are sometimes arranged so as to form an irregular groove parallel to the peripodial groove on each side. The foot-fringe itself is narrow and is crossed by a number of small vertical grooves, which are often continued across the edges of the foot-sole. In contracted specimens many of these transverse grooves extend right across the sole, although they are never so numerous in the middle as at the edges. A conspicuous but irregular longitudinal groove often runs along the centre of the foot-sole, but this also is probably due to the contraction of the specimens. In some examples two grooves diverge from a point at the hind end of the foot-sole and extend forward nearly parallel to the edges of the foot, thus dividing the hinder part of the sole into median and lateral areas. Occasionally these grooves may be traced forward for about two-thirds of the length of the animal, but they are more usually confined to the hind end, as shown in Pl. VII, fig. 9, and in many specimens they are entirely absent. In Natalina I have sometimes found similar grooves towards the hind end of the foot; but none of the grooves on the foot-sole seem to have so much systematic value as those on the back and sides of the slug, for they may be present or absent in different individuals of the same species. CoLour.—The skin of Apera is coloured, and the colour is almost entirely due to pigment. It is true that specimens of the more darkly coloured species when preserved in spirit 120 HUGH WATSON. sometimes have a slightly bluish or bluish-green tinge, and that this is probably due not to pigment, but to minute particles in the outer layer of the skin which intercept the blue rays. It is doubtful, however, whether this interference effect would be as noticeable in living specimens. Apera possesses two chief kinds of pigment: the first is dark and relatively stable, the second is hghter, very unstable, and superposed upon the first. The dark pigment is usually brownish-grey, varying from dull brown in typical examples of Apera gibbonsi to dark grey in A. purcelli. It does not spread uniformly over the back and sides of the animal, but is more concentrated in some parts than in others. Inthe first place the dark pigment tends to be aggregated into small irregular patches, which give the slug a mottled appearance. This tendency is least apparent in-Apera purcelli, and most marked in A. dimidia, A. burnupi, and some varieties of A. gibbonsi. In these animals the patches of colour are well defined, so that the skin is conspicuously mottled. When the dark patches are examined through a strong lens they are seen to contain minute lighter specks, which are probably caused by the dermal mucous glands. Secondly, the dark pigment generally shows a tendency to become concentrated along the sides of the dermal grooves. Thus in A. gibbonsi the oblique sub-lateral grooves are usually rendered more conspicuous by the way in which the patches of pigment tend to coalesce along their courses. Thirdly, the dark pigment is more abundant on some regions of the body than on others. In Apera gibbonsi it is concentrated dorsally, although a lighter band is gene- rally left along the very centre of the back (PI. VIII, figs. 14-19). On the sides of the animal the mottling becomes sparser, and it dies out before reaching the foot-fringe. In the keeled species, as well as in A. dimidia and A. purcelli, the dark pigment tends to be concentrated laterally, so as to form an irregular longitudinal band on each side of the body. Between the bands there is a lighter THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 1 dorsal area, palest at the edges, but a little darker in the middle; below them the colour shades off gradually towards the foot. This arrangement of the pigment is best developed in Apera dimidia (Pl. VIII, figs. 20 and 21), but it is also fairly conspicuous in A. burnupi, where the dark mottling is concentrated to form a band on each side below the upper keels. In A. sexangula and A. purcelli the bands, though present, are less noticeable and the colour is more evenly distributed, although the edges of the keels of A. sexangula are almost free from pigment. It will be seen that in these four species the pigment tends to form a pattern similar to that found in the genus Arion. On the other hand in Apera gibbonsi the scheme of colora- tion is different, and the narrow lighter band which this slug usually has along the centre of the back recalls the similar band possessed by many species of Veronicella. It should be noticed that this difference in the position of the greatest concentration of the pigment is not correlated with the difference in the position of the lateral grooves; for in A. dimidia the grooves are situated as in A. gibbonsi, while the dark bands are lower down and occupy a similar position to those of the keeled species. Apera dimidia is smgular in having small patches of the dark pigment on the foot-sole (PI. VII, fig. 9). But im some specimens of A. sexangula and in at least one example of A. gibbonsi minute reddish-brown specks are scattered over the body, and in the former species these extend on to the foot. The hghter pigment is usually of some shade of orange, but varies from pale orange-yellow to dull red. It is generally diffused over the whole surface of the body, including the foot-sole, but is often darker above than below, the backs of Apera burnupi and A. gibbonsi rubella being especially deeply coloured. But this pigment is very unstable, and specimens which have been preserved in spirit seldom show more than very shght traces of it. In the case of an example of A. gibbonsi rubella, which was 122 HUGH WATSON. drowned by Mr. Burnup, the red pigment came out of the skin when the animal was dead, and coloured the water in which the slug was immersed. Notwithstanding the dark grey mottling of this form, the lhving animal is of a reddish colour, changing to orange below, as the unstable colouring matter is sufficiently abundant to mask the darker pigment underneath. The coloration of a typical specimen of Apera gibbonsi increases its resemblance to the rhizome of a fern, and the chestnut-brown colour of A. burnupi may help to render the slug inconspicuous among dead leaves. Even the reddish colour of A. gibbonsi rubella may serve to conceal the animal, although the bright pigment of the red variety of A rion ater is supposed to act as a warning colour. Mr. Burnup found a specimen of this Apera in the bush at Durban, and he noticed that a few of the dead leaves! under which the slug was found were of a dark brownish-red colour, strongly veined, and curled up into rolls, and these were so very like the slug that it might easily have been mistaken for one of them. On the other hand, it seems possible that the unstable red and orange pigments found in Apera and other slugs which are often carnivorous, as well as the similar bright colours so characteristic of most of the snails belonging to the Strep- taxide, may be partly due to their animal food. It is known that this is the case with the red pigment found in the flamingos, the scarlet ibis, and the roseate spoonbill ; for the colour tends to disappear when the birds are deprived of their natural food, which consists chiefly of mollusks and other invertebrates. M. F. Woodward? has described the rich orange colour of the edge of the mantle in specimens of Natalina caffra which were fed on snails and beefsteak, and Miss Davies*® mentions the brilhant orange-red colour of 1 Dr. J. Medley Wood, Director of the Natal Herbarium, has kindly examined one of these leaves, and thinks that it probably belongs to Combretum krausii. 2 «Proc. Mal. Soc.,’ 1895, vol. i, p. 271. 3 «Proc. Roy. Soc. Victoria,’ 1913, vol. xxi, p. 222. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 123 the mantle-border and foot in Paryphanta atramentaria (Shuttl.), another carnivorous snail. Again, the climate has a marked influence on the colour of slugs, moisture being favourable to the development of dark colours, while heat increases the production of red pig- ments. Now in Apera gibbonsi rubella both the dark and the red pigments are specially well developed, and this form is at present only known from the coastal region of Natal, which is both moist and warm; whereas the typical form of A. gibbonsi, as well as the other species inhabiting the Province, extend further inland. Dr. Simroth! has maintained that in Testacella the skin probably plays a more important part in respiration than the lune, and it is suggested that in such slugs as these the chief function of the pigment is to assist in the taking up of oxygen. But it does not seem likely that this is the case in Apera. For im this genus the respiratory tissue of the lung is well developed ; while, on the other hand, the structure of the integument and the arrangement of the blood-vessels which it contains do not suggest that the skin plays an important part in respiration. On the whole it seems probable that the pigments of these slugs, hke those of many other animals, may be looked upon as by-products of metabolism, which are deposited in the skin usually in such a way as to help to render the animal incon- spicuous or to protect it from harmful rays of light, the development of the pigments depending partly on the food, but being also influenced by the climate and surroundings. Yet it must not be forgotten that the coloration of any slug is due not simply to the climate, the food, and the colour of the surroundings, but to the power of reacting to these influences which the animal possesses through inheri- tance. '*Nova Acta Acad. Cesar Leop.-Carol. Germ. Nat. Cur.,’ 1891, vol. Ivi, p. 245, VOL. 3, PART 2. 10 124, HUGH WATSON. THE STRUCTURE OF THE SKIN. Errpermis.—The outermost layer of the skin on the back and sides of the animal consists of a well-marked cubical epithelium, which in Apera dimidia attains a thickness of rather less than ‘Ol mm. (PI. XII, fig. 37). The cells composing this epithelium are approximately square in section; they stain rather deeply, and each contains a large rounded nucleus. The outer walls of these cells are thickened to form a cuticle, while the remaining walls are much thinner. In this respect, therefore, the epidermis of Apera seems to resemble that of Testacella rather than that of Daude- bardia, in which all the walls of the epidermal cells are thickened,! or of Paryphanta, in which a cuticle is said to be absent.” The epidermis does not change its character on the dorsal side of the foot-fringe, as in such forms as Daudebardia ?* and Cystopelta,* but on the sole of the foot it becomes quite ditferent. Here the epithelium is columnar rather than cubical, and the inner ends of the cells taper and interdigitate with the underlying connective tissue, so that the limits of the epithelium are very ill-defined (Pl. XII, fig. 38°). Excepting towards the edges of the foot-sole, the outer end of each cell is produced into a number of delicate cilia, which in A. dimidia measure about *0035 mm. in length. Inner Layers or tHe Sxry.—In the outer part of the con- nective tissue, which everywhere underlies the epidermis, are embedded the dermal mucous glands and the pigment- granules. The mucous glands (Pl. XII, fig. 37) are oval or pear-shaped, and reach a considerable size, although each is formed from a single cell. The protoplasm of the cell is 1 Plate, L. H., ‘ Zool. Jahrb.,’ 1891, vol. iv, p. 529. * Beutler, B., * Zool. Jahrb.,’ 1901, vol. xiv, p. 374. 3 Plate, op. cit., p. 527. 4 Davies, Miss O. B., ‘Proc. Roy. Soc. Victoria,’ 1912, vol. xxiv, p. 332. » Tam indebted to Mr. W. Tams for kindly taking this photomicro- graph. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 125 chiefly confined to a thin layer lining the base of the gland and containing the rather large nucleus; the rest of the interior is filled with mucous secretion, which can be extruded through a narrow pore between the cells of the epidermis. Dr. Plate has separated the unicellular mucous glands of Testacella! and Daudebardia? into two classes, according to the nature of this secretion. A similar classification seems applicable to the dermal glands of A pera, for in sections some of the glands are seen to be entirely filled with a vacuolated colourless secretion, while in others the mucus has become concentrated into a compact, deeply stained mass in the centre of the gland. No dermal glands of either kind occur in the foot-sole. In this respect Apera resembles Paryphanta,® and contrasts strongly with Testacella* and other forms. Possibly the great development of the pedal gland may have rendered the unicellular glands unnecessary, although in Paryphanta the pedal gland is not much more highly developed than it is in Testacella. The minute granules of dark pigment occur in the con- nective tissue between the mucous glands, being chiefly aggregated just beneath the epidermis. Some of the granules are contained in pigment-cells, others appear to be irregularly scattered about, but possibly these may occur inside delicate branches of the pigment-cells. Beneath this glandular zone the connective tissue contains blood-vessels, nerves, and numerous muscle-fibres, longi- tudinal, radial, oblique, and circular, the last-mentioned being especially abundant near the inner surface of the skin.’ In Apera dimidia there are more muscle-fibres in the foot than in the connective tissue of the back and sides. his OW. cit... p.i530. 2 Op. cit., pp. 527, 528. > Beutler, op. cit., p. 375. * Lacaze-Duthiers, H. de,‘ Arch. Zool. Expér.’ (2nd ser.), 1887, vol. v, p. 526. * Tuse the term “skin” in a wide sense to include the whole body- wall, for no sharp line can be drawn separating the subcutaneous tissue with its muscle-fibres from the dermis or cutis vera. (S mem ‘xn at IBRARY 126 HUGH WATSON. zone of connective tissue is extremely thick, especially towards the hind end of the animal; indeed, the most remark- able thing about the skin of A pera is its great thickness and strength. Only in the region of the head, where flexibility is essential, is the skin moderately thin. Its thickness further back in A. gibbonsi is well shown in Pl. X, fig. 32. On the other hand, A. dimidia has a somewhat thinner skin than most of the other species, though even in this form its thickness is not inconsiderable, as will be seen from PI. XI, figs. 34 and 35. This unusual thickness of the skin can easily be explained. In Apera the shell is so degenerate that it can no longer fulfil its proper functions. Yet the slug still needs protection from other animals seeking to prey upon it, and from the pressure of the soil if it burrows in the ground lke Testacella, and also from excessive evaporation. And it still needs a firm attachment for its powerful retractor muscles. Therefore the skin has taken over the functions of the shell, and has become adapted to fulfil these functions adequately. The Rey. Dr. Cooke! has recently called attention to a case among the Polyplacophora in which the protection once afforded by the shell has been transferred to the integument. But in that case the mantle has taken the place of the shell, while in the present instance it is the skin itself which has been strengthened. The blood-vessels contained in the skin are irregularly disposed, though most of them run in a longitudinal direction. They communicate with the body-cavity or hzemoccele by minute pores, which can be seen puncturing the inner surface of the skin (Pl. X, fig. 32). In all parts of the skin these vessels are provided at rather frequent intervals with power- ful sphincters composed of circular muscles (Pl. XIII, fig. 40). Similar sphincters have been described by Dr. Simroth and P. and F. Sarasin? in Veronicella (a genus in no way allied to Apera); and it has been suggested that they serve to prevent the reflux of the blood when the animal moves. 1 * Proc. Mal. Soc.,’ 1913, vol. x, pp. 321, 322. * «Die Land-Mollusken von Celebes,’ 1899, p. 90, pl. xiii, figs. 122, 123. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 127 Perhaps in Apera they also assist the circulation of the blood in a similar manner. But it seems possible that their chief function may be to enable the animal to increase the turgidity of some or all of the dermal vessels when it contracts, and thus to give greater rigidity to the skin either as a whole or in part. For if the slug by closing the sphincters could thus increase the pressure in the dermal blood-vessels, the skin would become much more rigid in consequence, Just as the strength of the stem of an herbaceous plant is largely due to the turgidity of its cells. THE SHELL, THE MANTLE-CAVITY, AND THE MANTLE. THE SHeLL.—The small shell of Apera les deep in a little pocket im the skin just behind and to the left of the respiratory opening. As a rule its length is only about one-fourteenth of that of the animal (in alcohol). When viewed from above the somewhat oval contour of the shell is seen to be inter- rupted by a broad sinus which extends backwards from the front eud more than half way along the right side. The posterior part of the shell is thus broader than the anterior half, which projects forward on the left side of the respiratory opening. ‘lhe sinus is deepest in Apera gibbonsi, and shallowest in A. purcelli and A. dimidia. It will be seen from Pl. XIII, figs. 42, 45, 48, and 51, that the shell of Apera is often much less flattened than is usually the case with the internal shells of slugs. In A. burnupi and A. sexangula it is very convex, and culminates im a conical apex, which overhangs the posterior margin of the shell. In A. purcelli the apex projects still further, but it is rounded instead of conical, as is also the case in A. dimidia, in which the apex is less prominent. Lastly, in A. gibbonsi and A. parva the shell is more depressed, and the apex is in front of the posterior margin. The shell is extremely thin and trauslucent in Apera purcelliand A. dimidia. In the other species itis thicker 128 HUGH WATSON. and more calcareous, especially in full-grown specimens, though it 18 usually very brittle. Indeed in some examples of A. evibbonsi its thickness is sufficient to suggest that the chief function of the shell in this species and its allies may be to enable the animal to get rid of any excess of calcium carbonate in its blood. It is difficult to see what other function it could fulfil; and if it were to be regarded asa purely vestigial structure with no function at all, one would have expected it to have become much smaller as well as flatter in this species than in the more primitive forms such as A. purcelli. Yet, while the shape has degenerated, the relative size has remained the same and the thickness has greatly increased. THe SHELL-sAc.—The pocket in which the shell lies has no communication with the exterior, but is close to the imner surface of the skin (Pl. X, fig. 33). In those species in which the shell has a prominent apex, the floor of the sac is raised towards the hind end to form a papilla fitting into the hollow of the shell. Where the margin of the shell rests upon the floor of the pocket, there is a fold, which separates the central part of the floor underlying the interior of the shell from a narrow rim beyond it. The shell-sac is lined by a well-defined epithelium. On the root of the cavity this epithelium is very thin, being formed in Apera dimidia of a compact layer of small cells which are slightly broader than they are high. Underneath the shell the epithelium is thicker in the same species, and composed of larger cells, whose height exceeds their breadth. These larger cells are continued over the edge of the fold mentioned above, the transition from the thicker to the thinner epithe- lium taking place on the outer side of the fold. THe ManriE-cavity or Luna. The mantle-cavity occupies the region below the shell and the respiratory opening, and also extends a short distance further forward. It is, nevertheless, mainly behind instead of above the body-cavity or hemoceele, from which it is separated by a muscular diaphragm sloping obliquely downwards and backwards. There is, however, a THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 129 funnel-shaped prolongation of the body-cavity which extends backwards for a variable distance beneath the mantle-cavity, and raises its floor along the middle. This feature is well shown in Pl. XI, figs. 34,35, and Pl. XII, fig. 36, for the prolongation is unusually large in Apera dimidia and A. purcelli. In horizontal section the mantle-cavity or lung is roughly triangular, with the large pericardium occupying the centre of the broad anterior part, which is thus divided into a nght and a left wing (PI. IX, figs. 27-31). The right division opens widely behind into the posterior part of the mantle-cavity, but the left division is cut off by the kidney, which extends obliquely backwards from the pericardium and fuses with the left wall and the roof of the cavity. ‘Thus the left anterior division only communicates with the rest of the mantle-cavity by a narrow space round the top and_ front of the pericardium and kidney. It is, therefore, not sur- prising that the walls of this part of the cavity are without respiratory tissue. But the right anterior division, which communicates more freely with the air, and the greater part of the posterior half of the cavity have their walls richly supplied with pulmonary veins. These vessels branch repeatedly, and in the larger species the fine branches form a complicated network which divides the surface into numerous alveoli. The thinness of the epithelhum which separates the blood in these vessels from the air in the mantle-cavity will be seen from Pl]. XII, fig. 39. The only portion of the posterior half of the cavity that is devoid of respiratory tissue 1s a part of the roof! and the immediate neighbourhood of the respiratory opening and the anus just below it. At the hind end the cavity tapers to a point. In Apera gibbonsi this point is below the extreme hind end of the shell-sac; but in species such as A. sexangula and A. dimidia it is within the papilla which is formed by the floor of the shell-sac projecting into the hollow apex of the shell. Hence in these species the hind end of the mantle-cavity is inside the shell itself. 1 See p. 187. 130 HUGH WATSON. I have only been able to examine the minute structure of the epithelium lning the mantle-cavity in one species, namely A. dimidia. In this species (and probably also in the others), that part of the posterior half of the cavity where there is no respiratory tissue is lined by moderately large cells, shightly higher than they are broad, and strongly ciliate near the respiratory opening. These cells are continuous through the aperture with those of the epidermis. The whole of the remainder of the cavity, including both of the anterior divisions, is lined by a pavement-epithelium of cells so thin that the nucleus of each usually forms a knob-hke projection in the centre of its surface. Tue Mantie.—A pera has no external mantle—a remark- able fact that was first pointed out by Heynemann.' Yet it cannot be entirely without a mantle, for, as we have just seen, it possesses both a mantle-cavity and a shell; and all Gastro- pods that have a shell also possess a mantle, as_ the shell is always secreted by the mantle. I therefore regard that part of the imtegument which forms the roof of the mantle-cavity beneath and around the shell as being undoubtedly homologous with the mantle. Moreover I beheve that the mantle is not entirely concealed from view in all the species of A pera, for it seems probable that it forms the mner hp of the respiratory opening, and that the upper surface of this hp may be regarded as part of the upper surface of the mantle (Pl. X, fig. 33). This theory is confirmed by the occurrence of an oblique cleft in the inner hp of the opening in a similar position to the cleft which runs from the opening to the edge of the mantle im an ordinary slug. We may suppose that Apera has been evolved from a snail, having on its back a mantle covered by a shell. It might well be an advantage to such an animal to get rid of this projection, especially if it burrowed in the ground in search of worms. Accordingly both mantle and shell seem to have sunk into the back and to have become covered by folds of the surrounding skin which have grown over it on each side. 1 * Jahrb. d. Deutsch. Mal. Ges.,’ 1885, vol. xii, p. 19. THE CARNIVOROUS SLUGS OF SOU'TH AFRICA. Toil The upward growth of the skin on the right side would tend to cover the respiratory opening, which would be in its usual position in the right margin of the mantle. To obviate this, the mantle edge, with the opening, would have to move up towards the middle line, and the growth of the right side of the shell would be retarded in consequence. And at the same time the fold of the skin would be likely to grow more slowly immediately opposite the opening than before and behind it, thus forming a notch in the edge of the fold. Eventually this fold of the skin would unite with the other in the middle line excepting at the notch, which would be converted into an opening above the original respiratory opening in the mantle-edge. This hypothesis may not be correct, but it has at least the merit of explaimimg not only the absence of an external mantle in Apera, but also the dorsal and shghtly asym- metrical position of the respiratory opening, the fact that the opening has an inner and an outer hp, the former with an oblique cleft, and the pecuhar sinus in the mght side of the shell. So far as I am aware the morphology of the mantle and shell of Apera differs from that of all other Pulmonates. THE PEDAL GLAND. The pedal or supra-pedal gland attains a greater develop- ment in Apera than in any other genus of snails or slugs known to science. It consists essentially of a long duct with glandular walls, which opens in front between the mouth and the anterior end of the foot, and ends behind in a small vesicle. THe Position AND LenatH or THE GLAND.-—In the herbivorous genera of snails and slugs the pedal gland is usually more or less embedded in the muscular tissue of the foot, and this is also the case in Daudebardia,! Plutonia,? and the Tri- gonochlamine*; but in Apera it lies freely in the lower ' Plate, L. H., ‘ Zool. Jahrb.,’ 1891, vol. iv, p. 525. 2 Simroth, H.,‘ Nova Acta Acad. Cesar Leop.-Carol. Germ. Nat. Cur.,’ 1891, vol. lvi, p. 229. ’ Simroth, H., * Festschrift Leuckarts, 1892, pp. 53, 55; ete. 132 HUGH WATSON. part of the body-cavity, as in Natalina, Rhytida, Pary- phanta, Schizoglossa, Testacella, Streptaxis, Atopus, and a few other carnivorous genera, and is only held in position by the pedal artery which supphes it with blood. This artery confines the gland to the anterior part of the body-cavity, for it runs back above the gland, giving off branches to it as it goes, and, after becoming attached to the posterior vesicle for a short space, it dips down and unites with the upper surface of the foot about the middle of its length. The pedal gland, however, is many times longer than this free portion of the pedal artery and is consequently thrown into numerous foldsand loops. In Apera sexangula the gland pursues a comparatively regular zig-zag course, bending alternately to the mght and the left about half a dozen times (Pl. XIII, fig. 57). But in A. burnupi and A. purcelli, and in the posterior part of the gland in the remaining species, the loops are deeper and more complicated, twisting about in all directions (figs. 52-56). The gland attains its greatest length in A. burnupi, m which its folds not only extend up the sides of the body-cavity, but even curve over on to the top of some of the other organs (PI. IX, fig. 30). THe GranpuLaR TissuE.—The pedal gland of A pera is not only unusually long, it is also often exceptionally broad, as, for example, in A. burnupi and in the anterior half of its length in A. dimidia (Pl. XIII, figs. 56, 54). In A. gibbonsi and A. parva it is more slender (figs. 52, 53), possibly because the great size of the odontophore in these species leaves little room for the growth of any other bulky organ in the anterior half of the body-cavity. The breadth of the gland is due to an enormous develop- ment of glandular tissue on the floor and sides of its duct. In Apera purcelli, A. burnupi, and A. sexangula this glandular tissue extends along the whole length of the duct as far as the terminal vesicle ; but it is less abundant towards the posterior end and the gland becomes rather narrower behind in consequence. In A. gibbonsi rubella the THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 138 olandular tissue becomes still scarcer towards the hind end, and disappears entirely just before the vesicle is reached, while in the other races of A. gibbonsi and in A. parva the last few millimetres of the duct are quite destitute of gland-cells. Finally, in A. dimidia the glandular tissue only extends along half the length of the duct, the posterior half being entirely without it, and therefore much narrower than the anterior half (Pl. XIII, figs. 54, 59, and 60). Pl. XIV, fig. 65 shows the structure of the glandular tissue near the front end of the pedal gland in A. gibbonsi rubella. It will be seen that among the large oval gland- cells, which taper towards the duct, there are scattered con- nective-tissue cells with conspicuous rounded nuclei, and there can be little doubt that it is from these cells that the glands are formed. Similar cells occur near the front end of the pedal gland in A. dimidia, but further back, near the hind end of the anterior half of the gland, no such cells occur, and a transverse section has the appearance shown in Pl. XIII, fig. 59. Possibly the absence of connective-tissue cells from this region may be correlated with the absence of gland-cells still further back. The posterior part of the gland will probably be the first to be formed in development, and if all its cells, apart from those which form the duct itself, are converted into glands, there will be nothing left to replace these when they die and are discharged into the duct ; for according to André! the gland-cells perish after a certain time and need to be replaced. Mr. Collinge? has stated that the pedal gland of Apera sexangula contains “a large series of microscopic chitinous (7) dart-like bodies,” but his figures leave little doubt that these are merely the solidified contents of the glands. In specimens preserved in alcohol this congealed secretion appears through a simple lens as numerous white specks, which often give a glistening appearance to the pedal gland. 1 «Revue Suisse de Zoologie,’ 1894, vol. ii, pp. 310, 311. 2 «Ann. Mag. Nat. Hist.’ (6th ser.), 1897, vol. xx, p. 223, pl. v, figs. 4,5. 134 HUGH WATSON. THe Duct.—The duct of the pedal gland is semi-trans- lucent, and can be seen running along the centre of its upper surface. Near the opening of the gland the duct usually lies more deeply (Pl. XIV, fig. 63), but elsewhere the glandular tissue is restricted to its sides and lower surface, as shown in Pl. XIII, fig. 59. Where the glandular tissue is less abundant the exposed surface of the duct often shows a tendency to have alternate swellings and constrictions. In Apera gibbonsi this culminates in a close convolution of the posterior part of the duct (Pl. XIII, fig. 58), which reminds one of the somewhat similar convolution of part of the duct in Testacella.! On the upper surface of the anterior half of the gland in A. dimidia, two brownish lines can often be seen running along, one on each side of the duct, at its junc- tion with the glandular tissue. In section the duct is approximately circular, without any folds on its floor (Pl. XIII, figs. 59, 60, Pl. XIV, fig. 63). In this respect it resembles that of Testacella, and differs from that of Daudebardia, for in the latter genus, as in most herbivorous snails and slugs, the floor of the duct has two longitudinal folds enclosing a median groove, and the gland- cells only open into this groove, instead of all round the duct.” In Paryphanta hochstetteri (Pfr.) traces of the folds and median groove seem to occur, but only towards the anterior end of the duct.* The duct is lined by an epithelium composed of small cubical cells, and this is surrounded by two thin muscular layers, the fibres of the inner layer running round the duct, while those of the outer layer are longitu- dinal. Both layers of muscles occur inside the glandular tissue. THe TERMINAL VeEsicLy.— After becoming gradually narrower the gland widens again at the hind end to form an oval or pod- shaped vesicle (Pl. XIII, figs. 52-57). In length this vesicle 1 See de Lacaze-Duthiers, ‘ Arch. Zool. Expér.’ (2nd ser.), 1887, vol. v, pl. xl, fig. 95. * Plate, L. H., ‘ Zool. Jahrb.,’ 1891, vol. iv, p. 525, pl. xxxui, fig. 16. 3 Beutler, B., ‘Zool. Jahrb.,’ 1901, vol. xiv, p. 376, pl. xxvi, fig. 9. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 1395) varies from about 1 mm. in the smaller species to more than 5mm. ina large specimen of Apera burnupi, and its breadth is usually about one-third of its length. The walls of the vesicle are thin and semi-translucent like those of the duct, with which they are continuous. A few gland-cells occur along the ventral! surface in A. burnupi, A. sexangula, and A. purcelli, but none at all in the other species. The structure of the walls of the vesicle is similar to that of the walls of the duct, excepting that the cells of the epithelium become somewhat higher on the dorsal side, and among the ordinary epithelial cells there are scattered a few rather small vacuolated cells. From the dorsal wall of the vesicle there projects into its cavity a remarkable fold, resembling a typhlosole, which fills up a large portion of the interior, so that the cavity is cres- centic or horse-shoe-shaped in section (Pl. XIII, fig. 62). If the wall of the vesicle in Apera sexangula be cut open, it will be seen that the fold has a spongy and somewhat lami- nated appearance. Serial sections through the vesicle of A. dimidia show that in this species the fold is rounded excepting at the ends, where it becomes wrinkled and divided by furrows (cf. Pl. XIII, figs. 61, 62, Pl. XIV, fig. 66). Fig. 66 shows the structure of the fold in A. dimidia. It will be seen that it 1s entirely composed of vacuolated tissue, which is not bounded by a distinct epithelium. The pedal artery, which always becomes intimately connected with the dorsal surface of the vesicle, gives off a branch into the middle of the fold (Pl. XIII, fig. 62). In A. gibbonsi the structure of the fold is different, for in this species it is divided into a compheated series of narrow lamellae. Fig. 58 shows the appearanee of these as dimly seen through the wall of the vesicle, while Pl. XIV, fig. 65, shows them in section. Each ' T use the terms “ dorsal ” and “ ventral ” in a morphological sense, the dorsal side being that to which the artery is attached. Generally, however, the vesicle lies on its side, owing to the twisting of the duct, and the attachment of the artery to the floor of the body-cavity beyond the vesicle. 136 HUGH WATSON. lamella seems to be composed of a double layer of peculiar rounded or polygonal cells with thick walls. The hind end of the pedal gland is also enlarged to form a vesicle in Rhytida,!' Paryphanta,? and Schizoglossa,® and indeed this is probably the case in all the genera of the Rhytidide, for I have found such a vesicle in Natalina as well. Moreover, Beutler! has shown that in Paryphanta it contains a large fold similar to that in Apera, but appa- rently less highly specialised in structure. In Toctarelly there is also a dorsal fold’ projecting into the duct at the hind end of the pedal gland, but the structure of the fold appears to be very different from that of A pera, and there is no terminal enlargement of the gland to form a vesicle. Various possibilities suggest themselves with regard to the function of the terminal vesicle of the pedal gland, and the fold which it contains ; but so little is known about the physio- logy of the gland in the carnivorous genera that such specu- lations can have little value. It is better, therefore, simply to state that the function of these structures is at present unknown. THE NERVOUS SYSTEM. The central nervous system of Apera consists of three compact ganglionic masses joined by connectives. The first of these is formed of the cerebral ganglia, the second of the buccal gangla, and the third of the pedal, pleural, and visceral ganglia. THE CeReBRAL GANGLIA AND CoMMISSURES.—As in nearly all carnivorous snails and slugs, excepting Daudebardia® and 1 Collinge, W. E.,‘Ann. Mag. Nat. Hist.’ (7th ser.), 1901, vol. vii, pp. 67, 68, pl. i, figs. 12, 13. 2 Collinge, op. cit., p. 70, pl. ii, figs. 21, 24; Murdoch, ‘Trans. N.Z. Inst.,’ 1904, vol. xxxvi, p. 158, pl. vi, fig. 7. . f Riedoch: R., ‘ Proce. Mal. Soe.,’ 1901, vol. iv, p. 170, pl. xvii, fig. 10. ‘Zool. Jahrb.,’ 1901, vol. xiv, p. 376, pl. xxvi, fig. 7. 5 Plate, L. H., ‘ Zool. Jahrb.,’ 1891, vol. iv, p. 525, pl. xxxii, figs. 17, 18. 6 Plate, op. cit., pp. 589, 593. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 137 the Trigonochlamine,! the cerebral ganglia are situated very near together on the dorsal side of the buccal mass. Indeed, in Apera gibbonsi they are so closely bound to each other by connective tissue that they appear as one (Pl. XIII, fig. 52); and it is only in the smaller species, such as A. parva (fig. 53) and A. purcelli (fig. 55), that they can be clearly seen to be separate gangha without the help of any reagents. When freed from their surrounding tissue, the ganglia are seen to be oval structures, nearly twice as long as they are broad, and rather nearer together behind than in front (Pl. XV, figs. 70-72). Their average length in A pera is about # mm. ‘l'owards the outer side of the front end of each ganglion there is a slight eminence formed by the accessory lobe. ‘This is best seen in A. dimidia. In this species the apex of each lobe is prolonged into a filament, which is very like a slender nerve (fig. 71), but is probably the remains of the embryonic cerebral tube which originally connected the accessory lobe with the exterior. Similar accessory lobes have been figured by de Lacaze-Duthiers ? in Testacella, by Wiegmann? in Ennea (Edentulina), and by Beutler* in Paryphanta, and the last author has described them at some length. The cerebral ganglia are joined by two commissures, namely, the cerebral commissure above the buccal mass, and the sub-cerebral commissure below it. The cerebral com- missure is very short and broad, and is situated somewhat towards the posterior ends of the gangha (Pl. XV, fig. 71). The sub-cerebral commissure, on the other hand, is very long and extremely slender. It arises from about the middle of the outer sides of the gangha, and encircles the buccal mass just in front of the cerebro-pedal connectives and the pedal ganglia, being closely bound to these by connective tissue ' Simroth, H., ‘ Festschrift Leuckarts,’ 1892, pp. 53, 55. 2 “Arch. Zool. Expér.’ (2nd ser.), 1887, vol. v, pl. xxxviii, figs. 70, 71. 3 * Mitt. Zool. Samml. Mus. Nat. Berlin,’ 1898, vol. i, pl. iii, fig. 7. * «Zool. Jahrb.,’ 1901, vol. xiv, pl. xxix, figs. 58, 61, and pp. 400-402 ; see also Pelseneer, ‘Mém. Acad. Roy. Belg.,’ 1901, ex. vol. liv, pp. 31-37. 38) HUGH WATSON. (figs. 73 and 74). It thus forms the most anterior element of the nerve-collar. Amaudrut! has described a similar sub- cerebral commissure in Achatina panthera (Fér.), Buli- mus funki (Nyst.), Nanina cambodyjiensis (Reeve), and Helix aspersa Mill.; but hitherto this slender commissure has been noticed in only a very few Pulmonates and in none of the carnivorous forms. I have found it, however, in Natalina quekettiana (M. & P.), the only member of the Rhytididz whose ganglia I have examined microscopically; and I beheve that if malacologists would examine the central nervous system of snails with greater care they would find it in many other pulmonate genera. For among the marine Euthyneura a sub-cerebral commissure has been observed not only in the Pleurobranchidz and in numerous Nudi- branchs,” but also in so primitive a form as Acteon.? Moreover, I am inclined to regard the sub-cerebral com- missure as homologous with the important labial commissure found in the Aspidobranchia, and in Vivipara and Ampullaria among the Pectinibranchia, as well as in the Amphineura, Scaphopoda, and Cephalopoda.* In most of these forms the cerebro-buccal connectives arise from the labial commissure instead of from the cerebral ganglia themselves, and in Apera the cerebro-buccal con- nectives arise from the ganglia very close to the ends of the sub-cerebral commissure. A little further back arise the thick cerebro-pedal con- nectives; and behind this again, nearly at the posterior end of the ganglia, the almost equally broad cerebro-pleural con- nectives arise. In the specimen of Apera dimidia that I 1 ‘Bull. Soc. Philom. Paris’ (7th ser.), 1885-86, vol. x, pp. 107-117; ‘Ann. Nat. Sci., Zool.,’ 1898, vol. vii, p. 127. 2 Pelseneer, P.,‘ Mém. Couronné Acad. Roy. Belg., 1893, ex. vol. ii, p. 69. 3 Bouvier, E. L.,‘ Bull. Soc. Philom. Paris’ (8th ser.), 1893, vol. v, p. 67. 4 Pelseneer has shown that the so-called “ labial commissure ” which Pleurobranchea and afew other forms possess in addition to the sub-cerebral commissure, is merely an anastomosis of two of the nerves to the lips (op. cit., p. 38). THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 139 examined, these two connectives were joined on the right side by a short transverse filament not far from their origin (PLC tio dl). THe Nerves oF THE CEREBRAL GANGLIA.—From each cerebral ganglon arise two large nerves which are as thick as the cerebro-pedal and cerebro-pleural connectives. These are the olfactory nerve to the upper tentacle, and the nerve to the lower tentacle. The olfactory nerve springs from the upper surface of the anterior end of the ganglion, towards the inner side (Pl. XV, figs. 70-72). Near its origin there also arise three slender nerves, namely, the two peritentacular nerves, and the optic nerve, which is distinct from the olfactory nerve even from the ganglion. The nerve to the lower tentacle arises nearer the outer edge of the ganglion than the four nerves just described, and external to the accessory lobe. Near it arise the two labial nerves, which are very much narrower than the nerve to the lower tentacle, but not quite so slender as the optic and_ peritentacular nerves. As inmost genera the nerve to the penis also springs from this region in the right cerebral ganglion ; it is of about the same thickness as the labial nerves. Beutler! states that in Paryphanta hochstetteri (Pfr.) the penial nerve does not arise separately from the cerebral ganglon, but seems to branch from one of the other nerves. On the other hand, in the Streptaxide?” the nerve to the penis arises directly from the pedal ganglion, instead of its fibres first traversing the right cerebral ganglion, as is usually the case. Near the origin of the sub-cerebral commissure a very slender nerve arises on each side, and runs down with it in the connective tissue on the anterior side of the cerebro-pedal connective. Above this connective, on the dorsal surface of each ganglion, there arises another very slender nerve, which at first goes obliquely backwards, but quickly curves outwards, and runs down to the ventral group of ganglia between the two connectives. This is almost certainly the nervus 1 Op. cit., p. 400. » Wiegmann, F., op. cit., p. 62. VOL. 3, PART 2, 11 140 HUGH WATSON. otocysticus, although I was unable to find the otocysts themselves even in sections of the pedal ganglia. Lastly, a slender nerve arises from each ganglion between the two connectives, usually nearer to the cerebro-pleural connective than to the other.- In Apera dimidia, A. purcelli, A. burnupi, and A. sexangula these nerves run back to the anterior end of the buccal retractor, which they innervate. Kach gives off two branches close to the cerebral ganglia ; one of these anastomoses with the cerebro-buccal connective not far from its origin; the other runs down between the cerebro-pedal and cerebro-pleural connectives, and divides into two about half way to the ventral group of ganglia. One of the divisions of this branch leaves the nerve collar and innervates the retractor of the lower tentacle.! The other division continues down to the ventral ganglia; but whether it fuses with the pleural ganglion, as one might expect from Amaudrut’s observations,” or whether it merely crosses the dorsal surface of the ventral group of ganglia and leaves it again in the wall of the buccal artery, I was unable to discover with the lhmited amount of material at my disposal. The examination of these slender nerves, deeply embedded amongst the connective tissue and blood-vessels which surround the larger nerve-cords and ganglha, is a matter of considerable difficulty ; but it does not even require a com- pound microscope to see that the anterior end of the buccal retractor in A. dimidia, A. sexangula, and their allies is innervated by nerves arising from the sides of the cerebral ganglia. This is one of the very few respects in which Apera resembles Daudebardia more than any other carnivorous genus with which I am acquainted.® In A. gibbonsi the buccal mass with the odontophore is much 1 In Natalina quekettiana (M. & P.) and Rhytida capillacea (Fér.) the retractors of the lower tentacles are also innervated by very slender nerves issuing from the nerve-collar about half-way down each side, and this is probably the case in several other forms as well. 2 «Ann. Nat. Sci., Zool.,’ 1898, vol. vii, pp. 1238-126. 3 See Plate, L. H., ‘ Zool. Jahrb.,’ 1891, vol. iv, p. 591. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 141 larger, and the buccal retractors are more highly developed, and it might therefore have been expected that these nerves would have been larger and more conspicuous than in the other species of Apera. ‘I'he opposite is the case. In this form the corresponding nerves are extremely slender. They are joined for some distance with the cerebro-buccal connec- tives (Pl. XIII, fig. 52), and then leave these and unite with the sheath of the odontophore at the anterior end of that organ. Bourne, G. C., ‘Proc. Zool. Soc. Lond.,’ 1908, pl. lv, figs. 36, 57, pl. lvi, fig. 38; and 1911, pp. 791, 792, pl. xxxviii, figs. 49, 50. 148 HUGH WATSON. carnivorous form in which this has hitherto been observed is Rhytida inequalis (Pfeiffer)! In Selenochlamys,’ Testacella, Euglandina, Streptostyla,? Salasiella,* and Paryphanta’ the three visceral ganglia remain distinct, and I have found that this is also the case in Natalina and in Rhytida capillacea (Fer.). In Daudebardia the abdominal is fused with the right parietal ganglion.’ Lastly, in the Streptaxide, the three visceral ganglia are all more or less fused to form a single mass, which is separated from the pleural ganglia by long connectives.’ Some of the nerve-cells in the ventral group of ganglia are remarkably large, especially those in the posterior parts of the visceral ganglia (Pl. XIV, fig. 67). One of the cells in the right parietal ganglion of a specimen of Apera gibbonsi rubella measures ‘18 mm. xX ‘143 mm., and its nucleus is no less than "16 mm. x ‘11 mm. Fig. 69 represents a photo- micrograph of a section through one of these large cells in the right parietal ganglion of A. dimidia, and shows the granular appearance of the large nucleus and its prominent nucleolus. THe NERVES oF THE VENTRAL GaANGLIA.—The numerous pedal nerves arise in an irregular longitudinal line along the ventral surface of each pedal ganglion (Pl. XV, figs. 73, 74), 1 Fischer, P., ‘Journ. de Conchyl.’ (3rd ser.), 1873, vol. xiii, p. 8, pl. iii, fig. 6. ? Simroth, H., ‘Festschrift Leucharts, 1892, p. 55, pl. vi, fig. 15. According to Simroth, the three visceral ganglia also remain separate in Phrixolestes, but in the other genera of the Trigonochlamine they tend to fuse with one another, and even with the pleural ganglia in Trigonochlamys itself (see Bronn’s * Klassen u. Ordn. Tier-Reichs III, Gastr. Pulmonata,’ 1910, p. 257, fig. 80). % Crosse and Fischer, ‘ Mission scientifique au Mexique,’ 1878, pl. iv, fig 5. 4 Strebel, H ,‘ Beitrag z. Kenntniss d. Fauna Mexik. L.-u. Siisswasser- Conchylien,’ 1878, vol. ii, pl. x, fig. 7. > Beutler, op. cit., p. 405, pl. xxix, fig. 59. 6 Plate, op. cit., p. 590, pl. xxxvi, figs. 94, 97. 7 Wiegmann, F., ‘ Mitt. Zool. Samml. Mus. Nat. Berlin,’ 1898, vol. i, p. 61. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 149 and are distributed to all parts of the foot. The most posterior nerves, which arise from the extreme hind end of the ganglia, are the largest, and run back for almost the whole length of the animal, as will be seen from fig. 75. This figure also shows that the other pairs of pedal nerves are not arranged quite symmetrically. The pedal gland is inner- vated by a pair of nerves which branch from the inner sides of the most anterior pair of pedal nerves. The size of these nerves to the pedal gland is in direct proportion to the size of the gland itself; thusin Apera dimidia, with its highly developed pedal gland, the nerve is larger than in A. sexangula, as will be seen by comparing figs. 73 and 74. From the sides of the ventral group of ganglia a large number of nerves radiate to the skin on each side of the body and head, passing below the tentacular retractors, but above the vagina and the adjacent part of the vas deferens on the right side (Pl. XIII, fig. 57). These are the nerves that de Lacaze-Duthiers termed the ‘ most posterior of these nerves can be seen running back on ‘nerves of the neck.” The each side along the inner surface of the body-wall sometimes nearly as far as the diaphragm, tlius confirming the hypothesis that all this region of the skin corresponds to that of the neck of ordinary suails and slugs. The posterior ends of the common retractors of the tentacles also receive nerves belong- ing to this group. All these lateral nerves appear to arise by repeated branching from three main roots on each side of the ventral ganglia. ‘The most anterior root springs from the outer edge of the pedal ganglion a little behind the cerebro-pedal connective. The second arises from the junc- tion of the pedal and pleural ganglia—that is to say, from the very short pleuro-pedal connective. The third and most posterior root arises from the pleural ganglion itself. Fischer’s figures of the nervous system in Rhytida inequalis (P/7.) and the Oleacinidz also show nerves arising from the pleural ganglia, but according to de Lacaze-Duthiers, Plate, Wiegmann, and Beutler, no nerves arise from the pleural ganglia in Testacella, Daudebardia, Ennea (Edentu- 150 HUGH WATSON. lina) and Paryphanta.! It must be remembered that at least some of these authors believed that the pleural ganglia never did give rise to nerves in the Pulmonata, a belief that Amaudrut® has shown to be altogether erroneous. At the same time it is unlikely that all these careful observers would have figured the “nerves of the neck” as arising solely from the pedal ganglia, if in all carnivorous forms they arise partly from the pleural ganglia, as they undoubtedly do in Apera. Moreover the fact that the central roots of these nerves arise in Apera exactly from the junction of the pedal and pleural ganglia also suggests that the more posterior “nerves of the neck” arise indifferently either from the outer side of the pedal ganglia or from the pleural ganglia. This seems to uphold the theory that the outer dorsal portions of the so-called pedal ganglia of most Gas- tropods are in reality parts of the pleural nerve-centres which have become united with the pedal ganglia, and that all the ‘‘nerves of the neck” therefore spring from the pleural division of the central nervous system. The composite nature of the pedal ganglia has been shown to be specially evident in some of the most primitive Gastropods, such as Pleurotomaria;* and the fact that the so-called pedal commissure contains, as we have seen, nerve-fibres emanating from the pleural ganglia affords additional evidence in favour of this theory.4 The nerves arising from the visceral ganglia are fewer ' For references, see p. 148. 2 «Ann. Sci. Nat., Zool.,’ 1898, vol. vii, p. 128. * Bouvier and Fischer, ‘ Journ. de Conchyl.,’ 1899, vol. xlvii, pp. 109- 145. * It is interesting to note that in Aplysiella and some species of Aplysia, as well as in the Gymnosomata, the cervical nerves also arise partly from the pleural and partly from the pedal ganglia, although they anastomose to form a single plexus (Pelseneer, *Mém. Couronné Acad. Roy. Belg.,’ 1893, ex vol. liii, p. 27, pl. x, fig. 81); while in the Auriculide these nerves may arise either from the pleural ganglia or from the pleuro-pedal connectives (Bouvier, ‘Comptes Rendus Soc. de Biologie’ (9th ser.), 1892, vol. iv, p. 990). THE CARNIVOROUS SLUGS OF SOUTH AFRICA. Pol than in most genera, probably because they innervate a region of the body which is greatly reduced in Apera. ‘The origin of these nerves is shown in Pl. XV, figs. 75 and 74, and their distribution in A. dimidia will be seen from fig. 75. One important nerve arises from the right parietal or supra- intestinal ganglion, not far from its junction with the other visceral ganglion, and innervates the right wall of the mantle- cavity. This may be regarded as the right pallial nerve. The corresponding left pallial nerve is more slender, and arises from the outer or parietal portion of the left visceral ganglion. This nerve innervates the left wall of the mantle- cavity. From the median or abdominal portion of the left visceral ganglion two large nerves arise close together. That to the left becomes closely united for some distance with the buccal retractor in Apera dimidia, and then runs back on the right side of the rectum, and eventually branches around the anus and respiratory orifice. The right abdominal nerve is joined to the aorta for almost the whole of its length, but at its hind end the nerve leaves the blood-vessel and enters the diaphragm, which it probably innervates. Beyond this the nerve becomes attached to the adjacent wall of the peri- cardium, and can be traced backwards to a point just behind the opening of the reno-pericardial duct. The other nervesfrom the visceral gangla are also attached to the aorta for some distance, the left pallial nerve being the first to become free. One or two small additional nerves can be seen with a strong objective to leave the visceral ganglia and run along in the tissue surrounding the aorta, but these are so extremely fine that I was unable to trace them far. Thus only four nerves of any importance arise from the visceral ganglia in A pera— two pallial and two abdominal ; and inasmuch as the abdominal portion of the left ganglion is probably itself formed from the union of the sub-intestinal ganglion with the original median abdominal ganglion,! it follows that there is but one nerve corresponding to each original ganglion on the visceral loop. ' Pelseneer, P., ‘Mém. Acad. Roy. Belg.,’ 1901, ex vol. liv, pp. 47, 48. lay HUGH WATSON. Tue Nervous System as a Wuore.—It will have been seen from the above account that the central nervous system of Apera may be regarded as formed of five sections, each TEXT-FIG. 2. MM Cerebral. E Pedal. BH Buceal. C=] Pleural. Visceral. Diagram of the central nervous system of Apera. with a single pair of ganglia. These five sections are dia- grammatically represented in the accompanying text-figure. There is first the buccal section, which is the most anterior only in A. burnupiand A. sexangula. ‘This imnervates the anterior part of the alimentary canal, including the odontophore and salivary glands. Secondly, there is the THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 153 cerebral division, which is connected with the sense-organs of the head,! and is the only section in which the nerve-cells are ageregated to form ganglia on the dorsal side of the alimentary canal. Thirdly, there is the pedal section, which innervates the foot and pedal gland. Fourthly, there is the pleural section, innervating the sides of the body and head. And fifthly, there is the visceral section, which innervates the neighbourhood of the mantle-cavity. And just as there are five pairs of ganglia, so also are there five commissures ventral to the alimentary canal, as will be clearly seen from the diagram. Now some authors have thought that there were only three such commissures in the Pulmonata, and very few have found more than four. Must we then regard Apera as exceptional in this respect? I think not. For it seems probable that further researches will show that there are normally five commissures ventral to the alimentary canal in the Euthyneura, although the slender sub-cerebral commissure may have entirely disappeared in some Pulmonates, just as in the Streptoneura the corre- sponding labial commissure seems to have disappeared in many of the Pectinibranchia. THE STRUCTURE OF THE UPPER TENTACLES, AND THE TENTACULAR RETRACTORS. In Pl. XVI, the figs. 80-86 represent photomicrographs of serial sections through the left upper tentacle of A pera dimidia in its retracted condition. It will be seen that the olfactory organ is large and extends forwards in front of the eye, thus reducing the width of the cavity in the retracted tentacle. The eye itself appears to be of the usual structure, the small size of the lens being possibly due to contraction. The retina is very deeply pigmented. ‘The ' It is not improbable that the motor nerve to the buccal retractor should be regarded as belonging to the pleural section, for it arises close to the origin of the cerebro-pleural connective. 154 HUGH WATSON. nerves from the different parts of the olfactory organ converge and unite behind the level of the eye to form the large olfactory nerve, which runs back for a considerable distance in the centre of the posterior part of the tentacle. The much smaller optic nerve is situated laterally, being embedded in the muscle-fibres on the outer side of the olfactory nerve. These longitudinal muscle-fibres make their appear- ance just behind the sense-organs, and are chiefly restricted to the peripheral part of the retracted tentacle ; but as they are traced backwards they gradually increase in number, and become united into muscular strands, which coalesce with one another to form the retractor of the tentacle. In the spaces between these strands, before they all coalesce, large cells occur, as will be seen from fig. 84. These large cells stain rather deeply, and have rounded nuclei, containing prominent nucleoli. They are evidently homologous with the somewhat similar cells that Beutler! found in the tentacles of Pary- phanta hochstetteri (Pfr.). In this region the tentacle is rather darkly coloured; yet the pigment-granules are not contained in these large cells, but are confined to narrow filaments which surround the cells and the different strands of muscle. Moreover the pigment extends a little further back, even shghtly beyond the point shown in fig. 86, where the olfactory and optic nerves emerge from the tentacular retractor and curve over to the cerebral ganglion (Pl. XIII, figs. 52-55). Not more than 2 or 3 mm. further back the retractor of the upper tentacle is joined by that of the lower tentacle, which is narrower and les more ventrally. The common retractor thus formed is of no great length. It arises from the side of the body-wall between a quarter and a third of the distance from the head to the hind end of the animal. The common tentacular retractors of each side have no connection with each other or with the buccal retractor ; on the contrary, the three retractors have become unusually widely separated in their origin. On the right side the retractors of the upper 1 * Zool. Jahrb.,’ 1901, vol. xiv, p. 404, pl. xxix, figs. 63, 64. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. oo and lower tentacles are separated by the penis, which projects between them (PI. XIII, figs. 52-57; see p. 190). In one specimen of Apera sexangula I found that the retracted upper tentacles were joined by a transverse connec- tion slightly anterior to their junction with the olfactory and optic nerves (Pl. XV, fig. 76). The transverse connection was pigmented in the same way as the tentacles themselves. This remarkable abnormality reminds one of the more striking monstrosity figured by Forbes and Hanley,! in which the upper tentacles of a specimen of Agriolimax agrestis (Lin.) were united throughout their entire length. The posterior ends of the common tentacular retractors receive nerves arising from the pleural ganglia (or possibly from the short pleuro-pedal connectives) ; and the retractors of the lower tentacles are innervated, as already described, by nerves issuing from the sides of the nerve-collar between the cerebro-pedal and cerebro-pleural connectives, but pro- bably emanating from the cerebral ganglia by the same roots as the nerves to the buccal retractors. Beneath the tentacular retractors a few smaller muscles arise on each side and run forward to the anterior part of the head. These are best seen in Pl. XIII, fig. 57. The largest of these minor cephale retractors is inserted in the upper lip of the pedal gland. It is doubtless the contraction of these muscles, together with the combined action of the buccal and tentacular retractors, which causes the front of the head to be invaginated in some specimens. THE DIGESTIVE SYSTEM. THe Mourn anp tHE RapuLa-sac.—As in most carnivorous genera, the mouth has three main lips, one dorsal and two lateral, arranged like the sides of an inverted isosceles 1 “Hist. Brit. Mollusca,’ 1853, vol. i, pl. JJJ, fig. 4. VOL. 3, PART 2. te 156 HUGH WATSON. triangle. But it will be seen from Pl. XV, fig. 77, and Pl. XVI, fig. 88, that the buccal cavity has four sides, and only appears somewhat triangular owing to the narrowness of the ventral wall. These figures also show the great thick- ness of the sides of the buccal mass—a feature which is still better seen in Pl. XVI, fig. 87. Externally there is a layer of longitudinal muscles, which is thin in Apera dimidia and A. sexangula, but somewhat thicker in A. gib- bonsi. Within this there is a very thick layer consisting chiefly of circular muscles, but also containing some radial and afew longitudinal fibres. The last become more numerous towards the inner surface, so as almost to form a third muscular layer in which the longitudinal predominate over the circular fibres. Lastly, the cavity is lined by a compact cubical epithelium which is continuous over the lips with the epidermis of the skin. The cells composing this epithelium are small, and their nuclei are situated towards their outer ends—that is to say, the ends turned away from the cavity. Towards the opening the epithelial cells become taller, and the translucent cuticle which they secrete is consider- ably thickened on the lips. There is, however, no jaw in Apera. At its hind end the buccal cavity extends into a long tubular diverticulum, which opens into the centre of the posterior wall of the cavity. This is the radula-sac, and is shown in longitudinal section in Pl. XVII, fig. 92, and in trans- verse section in Pl. XVIII, figs. 116-120. It will be seen that a broad fold or cushion projects like a typhlosole from the upper wall into the cavity, and renders the latter crescentic in section. Towards the hind end of the radula-sac the edges of this cushion are spirally rolled (fig. 119). Further forward a narrow longitudinal ridge occurs dorsally at each side of the cushion. The epithelium on the sides and floor of the radula- sac consists of small cubical cells, compactly arranged, and not unlike those lining the buccal cavity itself; but the epi- thelium on the cushion is of a quite different character, being composed of very tall and remarkably irregular cells a) THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 15 (Pl. XVI, fig. 89). The centre of the cushion is loosely filled with connective tissue. Immediately surrounding the whole radula-sac is a thin layer of circular muscles. Around the periphery of the hind wall of the buccal cavity there is a circular groove, which is deepened dorsally and ventrally (Pl. XVIII, figs. 115,116). The dorsal pocket is of no great depth, and is connected with the opening of the radula-sac by a median vertical groove. The ventral pocket, on the other hand, is much deeper, and contains the anterior end of the radula, which is thus folded back below the remaining part in the radula-sac (Pl. XVII, fig. 92). Between the opening of the radula-sac and that of the ventral pocket there is a slight projection formed by the front end of the odontophoral support covered by the radula, and it is this projection that is the first to appear when the odontophore is protruded through the mouth. THe Raputa.—The general appearance of the radula, when freed from the surrounding tissue, is shown in Pl. XVII, figs. 97, 98, 99, 102,107, and 114. Itis long and relatively narrow, its breadth being sometimes less than one-sixth of its length. The teeth are thorn-shaped with narrow bases, and they are arranged in rows which diverge obliquely forwards on each side of the middle line. In other words, the radula of Apera is of the same general type as is found in the Rhytidida, the Oleacinidx, the Testacellide, and, indeed, in all the familes of carnivorous snails and slugs excepting the Streptaxide, in which the teeth are of a slightly different shape. Nevertheless, the radula of A pera is excep- tionally interesting, and shows a diversity which is scarcely surpassed in any genus of the Pulmonata containing so few species. In the first place, the size of the radula varies enormously. In Apera gibbonsi and A. parva the radula is extremely large ; when flattened out it measures about a third of the total length of the animal, and in the former species it is pro- portionately broader than in the other members of the genus. In A. dimidia the radula is scarcely a sixth of the length 158 HUGH WATSON. of the animal. In A. purcelli it is about a ninth, while in A.burnupi and A. sexangula it is only about a twelfth of the length of the slug. The diversity in the size of the teeth is even greater. In Apera burnupi and A. sexangula the animal is about 550 times the length of the largest tooth initsradula. In A. dimidia and A. purcelli the proportion is 335: 1. Ina typical example of A. gibbonsi it is about 115: 1; while in A. parvaand A. gibbonsi lupata the animal is only about 66 times as long as its largest tooth, which attains a length of more than 0°6 mm. in the latter form. But the diversity found in the radula of Apera extends to other features besides its dimensions. Even in the closely related forms which I am provisionally regarding as subspecies of A. gibbonsi, there is a remarkable variation in the form and number of the teeth (Pl. XIX, figs. 123, 124, Pl. XX, 125, 126, and text-fig. 3, a-n).. The typical form has the most primitive type of radula, and the others show a pro- gressive adaptation to vermivorous habits. A slug that feeds on worms—and it is known that this species does so—uses its radula, not for rasping off pieces from its victim, but for catching hold of it and drawing it back through the mouth into the crop. For this purpose the teeth must be large, with long sharp points for piercing the worm’s skin and firm cuticle; and secondly, the teeth must be shaped so as to retain their position in the skin of the worm when they are pulling it back into the mouth. Accordingly we find, first, a progressive increase in the size of the larger teeth compared with the size of the animal, and especially in the length of their cusps. And in order to make room for these the number of rows is diminished, and the smaller teeth towards the outer edges of the radula, as well as those down the centre, become still less and eventually disappear (cf. Pl. XX, fig. 125, and Pl. XIX, fig. 123, and see also the numbers of the rows, and of the teeth in each row, given on pp. 195-201). Again, we find that the attachment of the teeth to the basal membrane of the radula also increases in length and forms a projecting apophysis, thus TEXT-FIG. 3. . gibbonsi, s.s. rubella. . gracilis. . lupata. . lupata v. duplex. 08 ge GQ Ie Representative teeth from the radule of A pera. 160 HUGH WATSON. making it more difficult for the teeth to be pulled out of posi- tion and keeping them rigidly in place. Lastly, the teeth become modified in one of two ways in order to prevent the worm that they have transfixed from escaping. In A. gib- bonsi rubella the cusps are strongly curved so as to form veritable hooks (text-fig. 3, B), and they are also slightly broadened towards the points. On the other hand, in A. gibbonsi lupata the cusps are barbed on the lower side, as in Testacella—a remarkable case of parallel evolution (cf. Pl. XXIV, fig. 157, and text-fig. 3, p). Still more highly specialised is the var. duplex of the same form. In this variety all the larger teeth are doubly barbed, there being a small barb on the upper side of the cusp nearer to the point than that on the lower side (text-fig. 3, #). I do not know of any other Pulmonate in which the radula has become so highly specialised along these lines as it has in this variety. We have to go to the marine carnivorous genera, such as Conus, to find such formidable doubly barbed teeth, and here also we notice that one barb is nearer the point of the tooth than the other. In Conus, however, the barbs are pointed instead of being merely square as in Apera, and the distal barb appears to have been developed first, for the other is often absent or merely represented by a flange. And of course the bases and arrangement of the teeth in Conus are quite different. The radula of Apera parva bears a general resemblance to that of A. gibbonsi, but the bases of the teeth are not so narrow, the cusps of the inner teeth are slightly shorter, and the teeth become relatively larger towards the edges of the radula (text-fig. 4, a). None of the transverse rows contain more than thirty-five teeth in this species, and some have only thirty-four, as the vestigial central tooth is absent from some of the rows. On the whole the radula of A. parva is less unlike the type found in the genus Rhytida than are the radulz of the other species of Apera. In Apera purcelli and A. dimidia the central tooth also degenerates. It is present, though small, in A. purcelli ee 162 HUGH WATSON. (text-fig. 4, c), but no trace of it remains in A. dimidia (Pl. XX, fig. 127, and text-fig. 4, B). The outer teeth of these two species have rather long, shghtly curved cusps, and are not very unlike the teeth of the typical form of A. gib- bonsi ona smaller scale; but the cusps of the inner teeth are remarkably short and broad, and the teeth have therefore a quite different appearance. In the more primitive species, A. purcelli, only the first three or four teeth on each side are thus modified, but in A. dimidia half the teeth are of this shape. Hence the teeth of these species, unlike those of nearly all other carnivorous forms, are differentiated into laterals and marginals, although there are one or two on each side which might be regarded as transitional. The figures of the radula of A. dimidia (Pl. XX, fig. 127, and text-fig. 4, B) show that in this species the corresponding teeth on each side are not opposite to each other, the right half of the row being in front of the left. A similar displacement occurs in A. parva (text-fig. 4, a), and it is also often observable in A. gibbonsi (Pl. XX, fig. 125); but it occurs occasionally in other carni- vorous genera, for | have noticed it in Testacella halio- tidea Drap., T. scutulum Sow., Huglandina truncata (Gmel.), E. corneola (Binn.), and Rhytida franklandi- ensis (Forbes). In Apera sexangula and A. burnupi the differentia- tion into lateral and marginal teeth is still more marked (PI. XX, fig. 128 ; text-fig. 4, p, x). The marginal teeth are more than twice as numerous as the laterals, and their cusps are long, slender, and nearly straight, those of one row over- lapping the bases of the teeth in the row behind. The lateral and central teeth have short broad cusps, and the central tooth—which is only shghtly smaller than those on each side of it—is not unlike the lateral teeth of A. purcelli and A. dimidia, though perhaps a little shorter. The cusps of the lateral teeth, however, are not only very short, but they are bifid and end in two points, the inner of which is the longer (text-fig. 4, p-y). Beutler’ has shown that in Pary- 1 * Zool. Jahrb., 1901, vol. xiv, p. 380, pl. xxvii, fig. 23a. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 163 phanta hochstetteri (Pfr.) one of the teeth in each row may abnormally develop a double cusp; but I believe that | am right in saying that in no carnivorous genus excepting Apera are some of the teeth normally provided with double cusps, for I regard the bicuspid teeth of Pseudosubulina lirifera (Morelet) as something quite different. What isthe reason for this remarkable modification of the inner teeth in these two species? Perhaps it may be found that Apera burnupi and A. sexangula live chiefly on some special kind of food, and have their radule specially modified in conse- quence. Or possibly we have here an instance of carnivorous slugs reverting to a partly vegetarian diet, and the lateral teeth are used in eating vegetable and the marginals in eating animal food. The fact that Collinge has found both animal and vegetable remains in the intestine of A. sex- angula is in favour of the latter view. In case this hypo- thesis should prove to be correct, it 1s interesting to compare the type of radula found in these species with that which is characteristic of the Limacide, Zonitidew, and other omnivorous families. ‘These also have developed thorn-shaped marginal teeth, very similar to those in Apera, and here, too, we find that the lateral and central teeth are shorter and broader with one or two less pointed cusps. Nevertheless, these inner teeth are very different to those of Apera burnupi and A. sexangula, for they are of the ordinary type with broad oblong bases, such as is found in the great majority of herbivorous snails, whereas in the species of Apera the teeth have narrow bases and have obviously been derived from the thorn-shaped type. ‘This difference, how- ever, could be easily explained by supposing that in the Zonitidex, etc., we have a group of herbivorous snails which have become adapted to a partially carnivorous diet, while in these species of Apera we are dealing with carnivorous slugs which have become partially herbivorous. Further details about the radula of Apera will be found under the head of the different species. I need only add here that the teeth towards the hind end of the radula are more or 164 HUGH WATSON. less tinged with brown, as is so commonly the case among the carnivorous genera. THe OponropHors.—The hind end of the buccal mass is prolonged beyond the opening of the cesophagus to form the cylindrical muscular odontophore containing the radula-sac. In most carnivorous snails the hind end of the odontophore is curved abruptly downwards, but in Apera,as in Testacella, it has become quite straight, probably owing to the degenera- tion of the spiral shell. The size of the odontophore—and, to a less extent, the size of the buccal mass in front of it—is proportionate to the size of the radula. In A. burnupi (Pl. XXI, fig. 183) and A. sexangula (fig. 134) it is quite small, being only about 4 mm. long; in A. purcelli (Pls XW fig) 55) and A. dimidia (Pl XX te. 152) at as much larger; finally, in A. gibbonsi and A. parva (figs. 129-131) the odontophore attains an enormous size, being one of the largest organs in the body. Yet the structure of the odontophore is remarkably constant through- out the genus, the differences between the arrangement of its muscles in the various species being quite insignificant. The odontophoral support or cartilage is a semi- cylindrical structure which extends along almost the entire length of the odontophore. It is crescentic in section, the convex surface being ventral. In the hollow formed by its upper surtace lie the radula-sac and the muscles immediately surrounding it, while the ventral pocket containing the front end of the radula is situated beneath the anterior part of the support, the radula being folded over its anterior end (PI. XVII, fig. 92, and Pl. XVIII, figs. 115-122). The upper edges of the support are parallel throughout the greater part of its length, but anteriorly they curve inwards towards each other, and then diverge again and slope downwards at the front end, thus leaving an oblique oval aperture where the opening of the radula-sac is situated (Pl. XVII, figs. 95, 100, 105, 112). As might have been expected, this aperture is largest in Apera gibbonsi, and the front end of the support is rounded in this species and in A. parva. On the other a THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 165 hand, in A. burnupi and A. sexangula the aperture is very small, and the front end of the support is pointed (see fig. 112). The hind end of the odontophoral support is rather abruptly truncated. Along the inner surface of the support there runs a median longitudinal furrow, which can be best seen in transverse sections of the odontophore of A. dimidia (Pl. XVIII, figs. 116-121). In the other species it is not usually so well marked. The odontophoral support is flexible, but much firmer than ordinary muscular tissue, and it is usually semi-translucent ; yet it does not contain any cartilage, but is entirely com- posed of long narrow cells stretching radially from the inner to the outer surface (Pl. XVI, fig. 90). The nuclei of these cells are also lengthened, and are rather: more numerous towards the outer than the inner surface of the support, doubtless because the cells are on an average slightly broader towards the outside. Most authors have regarded these long cells which occur in the odontophoral support of carnivorous snails and slugs as being muscle-fibres; but Beutler! has disputed this view, and has maintained that in Paryphanta hochstetteri (Pfr.) this tissue 1s not muscular. Now, I should certainly not be inclined to consider these cells as being of the nature of ordinary muscular fibres in Apera, but they mght perhaps be regarded as muscle-fibres which have become modified for a supporting or skeletal purpose, and, if this is so, it 1s possible that they have become more modified in genera such as Paryphanta and Apera than in the types studied by Plate. In Testacella and Daude- bardia there are longitudinal muscle-fibres and other cells intercalated among the radial elements of the support,?and I have found that this is also the case in Huglandina venezuelensis (Preston); but these are entirely absent in Apera, as in Paryphanta?® and Natalina.* ! Op. cit., pp. 380, 381. Plate, L. H.,‘ Zool. Jahrb.,’ 1891, vol. iv, pl. xxxiii, figs. 30, 31, 39, 41, pl. xxxiv, fig. 51. * Beutler, op. cit., p. 380. 4 Woodward, M. F., ‘ Proc. Mal. Soc.,’ 1895, vol. i, p. 273. 166 HUGH WATSON. A longitudinal muscle runs along the upper edge of the support on each side towards its anterior end (Pl. XVI, fig. 90; Pl. XVIII, figs. 116-118), and probably serves to bend up the front of the support, as shown in Pl. XVII, fig. 108, and may therefore be termed the flexor muscle of the odonto- phoral support. In Apera gibbonsi there is also a very thin layer of longitudinal muscle-fibres covering the ventral surface of the support, and possibly the flexor muscles might be regarded as a special thickening of this layer. The most important muscles in the odontophore are the powerful retractors, which arise from the odontophoral support and are inserted in the radula-sac. These may be divided into lateral retractors, median retractors, and terminal retractors. ‘The numerous lateral retractors are attached in front to the radula-sac, chiefly around its anterior end, and pass obliquely backwards and outwards on each side, curving over the edges of the support and becoming attached to its outer sides (Pl. XVII, figs. 92, 98, 96, 101, 106, 109, 111, 113, and Pl. XVIII, figs. 116-121). The median retractors are a pair of specially thick muscles, which are attached for some distance to the floor of the radula-sac towards its anterior end, and pass backwards to the extreme hind end of the support (Pl. XVII, figs. 92, 106, and Pl. XVIII, figs. 118-122). They belong to the same series as the lateral retractors, of which they might be regarded as forming the innermost pair. No muscles are attached to the posterior part of the radula- sac, excepting at its hind end, where there is another pair of retractors, which I am calling the terminal retractors. ‘These are not so thick as the median retractors, and pass backwards above them to the hind end of the support (Pl. XVII, fig. 92, and Pl. XVIII, figs. 121, 122). All these retractor muscles have the same function—that of pullimg back the radula. In A. gibbonsi some of the more posterior strands of muscle are attached to the inner surface of the support, close to its hind end, while others are not attached to the support at all, but to the hind end of the sheath of the odontophore opposite to the insertion of the extrinsic buccal retractors. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. iNOe/ Numerous muscles arise in two rows, one on each side of the mid-dorsal line of the outer sheath of the odontophore, and, passing downwards outside the lateral retractors, are . attached to the outer sides of the support just below them (Pl. XVIL, figs. 91, 104, and Pl. XVIII, figs. 115-121). These muscles I am terming the suspensor muscles of the support. They are nowhere of any great thickness, but are decidedly thinner behind than in front. A thin layer of ventral muscles arises from the hind end of the odontophoral support, and is inserted in the anterior part of the radula, or rather in the walls of the ventral pocket which contains it (Pl. XVII, figs. 91, 92, 101, 104, and Pl. XVIII, figs. 117-121). These muscles do not only run longi- ‘ tudinally beneath the centre of the support, but the more lateral strands diverge radially from the ventral pocket of the radula and unite with the suspensor muscles of the support on each side. They serve to pull back the anterior part of the radula over the front edge of the support, and thus to pull forward the radula-sac when its retractors are relaxed, so that the radula assumes the form shown in Pl. XVII, figs. 92,97,102,114. Then, when the teeth have become imbedded in the skin of the prey, these ventral muscles will be relaxed, and the powerful retractors will draw back the radula until it has the form shown in fig. 98. A slender median dorsal muscle is inserted in the sheath of the odontophore, between the suspensor muscles, a_ short distance behind the opening of the cesophagus, and_ is attached posteriorly to the front end of the terminal retractors, its sides also becoming connected with the neigh- bouring lateral retractors (Pl. XVII, fig. 92, and Pl. XVIII, figs. 116-120). Thisis evidently homologous with the muscle that Amaudrut terms the “papillaire supérieur.” It may assist the ventral muscle in pulling forward the radula- sac, but its chief function is probably to pull back the posterior lip of the opening of the cesophagus when the retractor muscles are brought into play, thus making a wide passage towards the crop. 168 HUGH WATSON. The sheath of the odontophore is composed of three layers of muscle-fibres. Externally there is a thin layer of longi- tudinal muscles; within this hes a layer of circular muscles, which is equally thin throughout the greater part of the length of the odontophore, but becomes a little thicker towards the front end; lastly, there is a very thin layer of longitudinal muscles hning the whole of the sheath excepting the dorsal part which hes between the attachment on each side of the suspensor muscles of the support. At its hind end the sheath of the odontophore becomes intimately united with the support and the posterior ends of the retractor muscles, but throughout nearly the whole of its length its only connection with the underlying tissues is by means of the suspensor muscles. I have examined the arrangement of the muscles of the odontophore in all the species of Apera excepting A. purcelli, and the amount of variation is so small, notwith- standing the diversity of the radule, that it must be admitted that these muscles are of considerable systematic importance. It will, therefore, be of special interest to compare the arrangement of the odontophoral muscles of Apera with that which is found in other families of carnivorous snails and slugs. As an example of the Rhytididze we may choose Natalina. Nineteen years ago M. F. Woodward! published an account of the anatomy of Natalina caffra (Fér.), with special reference to the structure of the odontophore; but unfortunately his account shows evidence of inaccuracy. I shall therefore ignore his description and figures, and compare the odontophoral muscles of Apera with those of a species of Natalina, nearly allied to N. caffra, which I have been able to examine myself, namely N. quekettiana (M. & P.). On the whole the odontophore of this species resembles that of Apera very closely; it differs, however, in the following features.—(1) Instead of being straight the odonto- phore is slightly curved towards the left; moreover its hind 1 «Proc. Mal. Soc.,’ 1895, vol. i, pp. 270-277, pl. xvii. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 169 end is abruptly bent downwards, and the outer longitudinal muscles of the sheath form a pad extending from the recurved end a short distance forwards along the ventral surface. (2) The lateral retractors towards the posterior end of the odontophore are attached to the odontophoral support just within its edges; further forward they are attached to the outer sides of the support as in Apera. (3) The median dorsal muscle is extremely slender, and the terminal retractor is also very weak, being almost devoid of muscular fibres. (4) The circular muscles of the sheath are much thicker laterally and ventrally, and the suspensor muscles of the support are also much thicker than in Apera. In all other respects the structure of the odontophore is strikingly similar to that found in Apera, and especiaily to that of A. gibbonsi; for in Natalina quekettiana a thin layer of longitudinal muscles covers the lower surface of the odonto- phoral support, and some of the posterior retractors are not attached to the support, but arise from the sheath of the odontophore opposite to the insertion of the upper branches of the extrinsic buccal retractors—features which are also foundin A. gibbonsi. The chief differences mentioned above may be easily explained. The curvature of the odontophore is doubtless due to the fact that Natalina possesses a large spiral shell. The greater thickness of the constrictor muscles running in a circular direction, both those in the sheath and those beneath it, is probably connected with the circumstance that Natalina feeds on snails rather than on worms; for, as Woodward has pointed out, the contraction of these muscles will press the radula against the body of its prey, as it tears the teeth through its victim’s flesh, a thing that a vermivorous form does not do. ‘The reduction of the median dorsal muscle and the terminal retractor, which might be regarded as its continua- tion, may be due to the fact that Natalina rasps off portions of its prey, and therefore does not require to enlarge the opening of the cesophagus to such an extent as an animal that swallows worms whole. We see, therefore, that the 170 HUGH WATSON. differences between the structure of the odontophore in Apera and Natalina are of little morphological impor- tance, and are certainly less striking than the differences between the radule of the two genera, for the radule of Natalina and Apera have become highly specialised in different directions. Turning now to the Testacellide, we find slugs which are vermivorous hke Apera gibbonsi and have radule remarkably similar tothe var. lupata of that species. Yet an examination of the odontophore of Testacella maugei Fer. reveals many features in which the arrangement of its muscles is quite different from anything that we have foundin Apera or Natalina (Pl. XXIV, figs. 155, 156). The structure of the odontophore of Testacella maugei differs from that which I have described as characteristic of Apera in the following respects.—(1) There is no terminal retractor (unless it is represented by the small connections between the floor of the radula-sac and the median retractor). (2) Perhaps in conse- quence of this the radula-sac is shorter than the ventral pocket, instead of being much longer. (3) The median dorsal muscle is short, very oblique, and divided into three or four fine strands. (4) The median retractors curve together above the radula-sac to form a tube which is continued to the hind end of the odontophore. (5) The muscles which form the top of this tube are not only firmly attached to the lateral retractors on each side, but also to the sheath of the odontophore; and they are continuous with the extrinsic buccal retractors, which join the posterior half of the odontophore on each side of the mid-dorsal line. (6) The lateral retractors are scarcely divided into separate strands. (7) There seem to be no flexor muscles along the upper edges of the support, but on its inner surface there is a slight median longitudinal ridge which dis- appears towards the hind end and is connected with the median retractors by a few slender strands. (8) he ventral muscles are rather stouter than in Aperaand more definitely split up into separate strands. (9) Anterior to these the suspensor muscles are free, but directed obliquely downwards THE CARNIVOROUS SLUGS OF SOUTH AFRICA. LAL and backwards from the sides of the roof of the odontophore to the support, parallel to the lateral retractors; further back, however, they are either absent or completely fused with the lateral retractors. Taking into consideration the similarity of the radula of Testacella to that of Apera, it is surpris- ing that there should be so many differences in the muscles which control its movements. In the Oleacinidz the odontophore resembles that of Testacella in some of its features, and that of Natalina in others, although it often possesses more primitive characters than can be found in either of these genera. Strebel! has given a detailed account of the odontophore of Euglan dina sowerbyana (P/r.), from which it will be seen that in this species the median retractors enclose the hind end of the radula-sac and become continuous posteriorly with the extrinsic buccal retractor. In this respect, therefore, Hug- landina sowerbyana resembles Testacella. But this is not the case with many of the other forms, such as H. venezuelensis (Preston) or Streptostyla shuttle- worthi (Pfr.),? for in these the radula-sac is long, and its posterior part has not yet become entirely surrounded with muscles, but projects slightly from the hind end of the odonto- phore. In most genera the posterior end of the odontophore is curved downwards as in Natalina. The median dorsal muscle resembles that of Apera, and the ventral muscles are also like those of Aperaand Natalina. On the other hand, the suspensor muscles are like those of Testacella, and I found that Huglandina venezuelensis also resembled that genus in having the retractors fused with the sheath of the odontophore towards the hind end. No median ridge or furrow occurs along the inner surface of the support in E. venezuelensis, and I did not find any flexor muscles along its edges. Possibly the longitudinal muscle-fibres within the support take the place m Testacella and ' «Beitrag z. Kenntniss d. Fauna Mexikanischer L. u. Siisswasser- Conchylien,’ 1878, vol. iii, pp. 40-42, pls. xvii, xviii. * Strebel, op. cit., pl. v, fig. 7a. VOL. 8, PART 2. , 13 M72 HUGH WATSON. Huglandina ofthe flexor muscles in Aperaand Natalina. It is interesting to notice that while the odontophore in the Oleacinidz is usually curved as in the Rhytidide, in several features of its internal structure it appears to resemble the odontophore of Testacella more than does that of Apera. Tue Buccat Rerracrors and Prorracrors.—In all the species of Apera excepting A. gibbonsiand A. parva, the extrinsic buccal retractor consists of a long, narrow muscle, arising from the right side of the floor of the body-cavity at the entrance to its funnel-shaped prolongation beneath the mantle-cavity or lung. The origin of the buccal retractor is thus nearly as far back as the heart, and is slightly asym- metrical (Pl. XV, fig. 75). In front the muscle bifurcates just before reaching the odontophore, and the two branches are inserted one on each side of its posterior extremity (Pl. XX1I, figs. 132-134). As might have been expected, the retractor is narrower in A. burnupi and A. sexangula thanin A. dimidia with its larger odontophore. In Apera gibbonsi and A. parva the retractor is split up into a number of powerful muscles, which radiate from the hind end of the odontophore, and are attached to a considerable area of the floor and the lower part of the sides of the body-cavity (Pl. XXI, figs. 129-131). In A. gibbonsi these muscles originate towards the hind end of the body-cavity, though further forward than the origin of the single buccal retractor in the other species (PI. IX, fig. 27), but in A. parva the muscles have become shorter and arise from about the middle of the cavity (fig. 28). These muscles are often forked and neighbouring strands occasionally anastomose, but there is frequently a shght gap dividing those on the right side from those on the left. It is easy to see that the same carnivorous habits which have led to the enlargement of the radula and odontophore in A. gibbonsi and A. parva would produce a corresponding development and multiplication of the buccal retractors; and it is also evident that it would be an advantage if the ends of HE CARNIVOROUS SLUGS OF SOUTH AFRICA. IBS these muscles were distributed over a considerable area, so that the strain produced by the contraction of the muscles was not concentrated on a small patch of the body-wall. But although it is not difficult to derive the many radially disposed retractors of A. gibbonsi and A. parva from the single retractor of the other forms, I would emphasize the fact that the difference is a very great one, and not likely to have been completed within a short space of time. So far as I am aware the only other form which has developed radial buccal retractors at all similar to those of Apera is Selenochlamys.! Since the buccal retractor is inserted in the hind end of the odontophore, it might have been expected that it would receive nerves from the buccal ganglia, which inner- vate the odontophore and the posterior part of the buccal mass. Yet this is not the case; for while the posterior end of the retractor receives nerves emanating from the right pedal ganglion, the front end is imnervated by a pair of nerves from the cerebral ganglia in all the forms with a single retractor. This apparently anomalous inner- vation may be explained by assuming that the buccal retractor belongs to the same series of muscles as the tentacular retractors, with which it is united posteriorly in the great majority of snails; and that it was originally inserted, like the tentacular muscles, far forward, and in the region innervated by the cerebral gangha. Later the anterior part, which consists of a right and left division, would become applied to the sides of the odontophore and fused with them, so that the retractor would appear to be inserted in the hind end of the odontophore. This theory is supported by the facts (1) that the nerves to the buccal retractor and to the retractors of the lower tentacles arise from the cerebral ganglia by the same roots, and (2) that the longitudinal muscle-fibres forming the outer layer of the sheath of the odontophore are directly continuous on each side with those of the buccal retractor behind (Pl. XVII, ' Simroth, H., ‘ Festschrift Leuckarts,’ 1892, p. 55. 174 HUGH WATSON. ‘fig. 122). This view would also explain how it is that in A. gibbonsi the nerves which innervate the buccal retractor in ~ the other species merely unite with the sheath of the odontophore towards the anterior end of that organ. For we may suppose that in A. gibbonsiand A. parva the odontophore has grown back further than in the other species between the two halves of the buccal retractor, which have thus become applied to its sides, and consequently the free radial retractors of these forms are homologous with only the posterior half of the long retractor found in the remaining species. We have seen that the origin of the buccal retractor in such forms as Apera dimidia is to the right of the middle line, and that its posterior end is innervated exclusively by branches of the right posterior pedal nerve (Pl. XV, fig. 75). Now the foot is a symmetrical organ, and the anterior part of the retractor is certainly symmetrical, as is proved by its innervation ; it might, therefore, have been expected that the posterior part of the buccal retractor would have been symmetrical also, unless it had shifted somewhat to the left because of the pressure of the reproductive organs on the right side, as in the case of Daudebardia saulcyi (Bgt.) and Testacella gestroi (Issel)! Why, then, does the buccal retractor spring from the rght side of the floor of the body-cavity in Apera? When an animal possesses a feature which is asymmetrical for no apparent reason, the explanation of that asymmetry is usually to be found by a consideration of the animal’s phylogeny. It is_ highly probable that Apera has been evolved from a snail possess- ing a depressed or heliciform dextral shell. Now when such a shell is in its natural position upon the back of the animal, its columella is to the right of the aperture. Consequently the columellar muscle, from which the buccal retractor springs, will be towards the right side of the animal. Then, when the shell degenerates, all the upper part of the colu- mellar muscle will disappear, and only the lower part, or 1 Plate, L. H., ‘Zool. Jahrb.,’ 1891, vol. iv, p. 596. ee THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 175 “tail muscle,” in connection with the foot, will remain as the direct continuation of the buccal retractor, but it will still be situated towards the right side of the animal. In Schizo- g@lossa we have a slug in which the degeneration of the shell has only partially taken place, and the chief muscle-fibres are still attached to the reduced columella; and it would be dificult to draw a diagram illustrating the theory which I have just set forth more aptly than does Murdoch’s figure of the retractor muscles of Schizoglossa novoseelandica (Pfr.).! Several pairs of buccal protractors stretch from the outer layer of the walls of the buccal mass to the skin of the anterior part of the head. The longest and most important of these are inserted on each side just in front of the odonto- phore and below the opening of the cesophagus (Pl. XXI, figs. 129-134). These protractor muscles, together with the intrinsic muscles of the buccal mass, serve to protrude the odontophore, which seems to be protrusible in all the species of Apera. Butevenin A. gibbonsi the protractors are not very thick ; and this is not surprising, for it is evident that less powerful muscles will be required to protrude the odonto- phore than to retract it after the teeth have become fixed in the skin of a worm. THe CisopHacus AnD Crop. In Apera gibbonsi and A. parva the cesophagus is very short and broad, and merges imperceptibly into the crop (Pl. XXI, figs. 129-131). The latter is fusiform in these species, and after increasing in width it gradually tapers again towards the openings of the hepatic ducts, its broadest part being about twice as far from these ducts as from the opening of the cesophagus into the buccal mass. In the remaining species the cesophagus is long and narrow, as will be seen from figs. 132 to 154. At its posterior end it enlarges rather abruptly to form the crop, which is vsually widest near the front end, but remains fairly broad almost as far back as the ducts of the liver, and then narrows 1 “Proc. Mal. Soc.,’ 1900, vol. iv, pl. xvii, fig. 10. 176 HUGH WATSON. rather suddenly; but, as might have been expected, the crop is much more distended in some specimens than im others. The walls of the cesophagus and crop are very similar in structure (Pl. XV, figs. 78, 79, and Pl. XXII, fis. 135): The epithelial lning is normally thrown into longitudinal folds, and consists of columnar cells without any cilia. Out- side the epithelium of both cesophagus and crop there are three layers of muscle-fibres. The fibres of the inner and outer layers are longitudinal, while those of the intermediate layer run in a circular direction. THE Sromacu, INrestine, AND Recrum.—Beyond the crop in Apera parva the alimentary canal enlarges to form a small stomach, into which the hepatic ducts open (Pl. XXI, fig. 131). In the other species of Apera (with the possible exception of A. purcelli) there is no true stomach, but the crop passes directly into the intestine at the openings of the hepatic ducts. At first the intestine bends abruptly upwards and towards the right side of the animal. In A. gibbonsi it then describes a curve on the upper surface of the liver like a reversed § (PI. IX, fig. 27; Pl. XXI, figs. 129 and 130). The posterior curve towards the left side is somewhat shallower than that towards the right, and beyond it the alimentary canal continues backwards low down on the right side of the posterior end of the liver. In A. parva the intestine, after curving over to the right side of the liver, passes straight backwards, the second curve towards the left side being absent (PI. [X, fig. 28; Pl. XXI, fig. 131). In Apera burnupiand A. sexangula the first curve is much deeper, and forms a loop which extends the whole way down the mght side of the liver, in which it is partially embedded (Pl. IX, figs. 80, 31, and Pl. XXI, figs. 133, 134). The anterior portion of this loop hes further forward than the openings of the hepatic ducts. The posterior curve is shallow in these species, and the alimentary canal is continued on the right side, as in A. gibbonsi and A.parva. ‘The 'HE CARNIVOROUS SLUGS OF SOUTH AFRICA. yz intestine is considerably narrower in A. burnupi and A. sexangula than in the other species. In Apera dimidia and A. purcelli the first curve of the intestine hes in a more nearly vertical plane, and does not extend so far towards the right side (Pl. IX, fig. 29, and Pl. XXI, fig. 132). Then, after curving down on the left side, the intestine continues backwards on that side of the liver instead of on the right (cf. Pl. X, fig. 32, and Pl. XI, figs. 34, 35). Beyond the region of the liver the alimentary canal con- tinues backwards as the rectum beneath the mantle-cavity. It is at first surrounded by the funnel-shaped prolongation of the body-cavity, but further back it curves towards the right side of the animal, and bends up the right wall of the mantle-cavity on reaching the level of the respiratory opening. As will be seen from PI. X, fig. 35, the anus is immediately below the respiratory opening, and the space between them is partially separated from the rest of the mantle-cavity by an oblique fold of the wall of the cavity, which projects forward on the left side of the space. The epithehum lning the intestine is composed of very irregular columnar cells, the appearance of which is well shown in Pl. XXII, fig. 157. The inner walls of these cells are produced into a delicate fringe of extremely minute cilia. Outside the epithelium there is a thin layer of circular muscles, and external to thisa thin layer of longitudinal muscles. The structure of the rectum is very similar to that of the intestine, excepting that the epithelium is normally thrown into deep folds (Pl. XXII, fig. 1386). At the anus, however, the epithelial cells become more regular and oblong in section, and their cilia become far longer and less delicate. THE Sativary GLanps and Ducrs.—The salivary glands are fairly compact, and are situated above and on each side of the crop near its Junction with the cesophagus. There are always two glands, but in Apera gibbonsi and A. dimidia, and to some extent in A. parva and A. purcelli, they are united above the crop, as in so many of the carnivorous snails 178 HUGH WATSON. (Pl. IX, figs. 27-29, and Pl. XXI, figs. 129-132). In A. burnupi and A. sexangula the glands are separate, but they are joined to each other by blood-vessels in such a way as to suggest that they may possibly have been more closely united in the ancestors of these species (Pl. IX, figs. 30, 31, and Pl. XXI, figs. 133, 134). The glands are never united underneath the crop, as in many of the Oleacinide, but remain widely separate below (Pl. XV, fig. 78). The salivary ducts are long and very slender, and discharge into the buccal mass on each side of the opening of the cesophagus. The structure of the ducts is shown in PI. XXII, fig. 1388. It will be seen that the epithehal cells are of a peculiar shape and provided with long cilia. The epithelium is immediately surrounded by a layer of circular muscles, and outside these there is a layer of longitudinal muscles in which is embedded the salivary nerve. Within the gland the duct splits up into numerous branches, and in these also the epithelium is surrounded by circular muscles. One of these branches is shown in fig. 139, which also shows the appear- ance of the glandular cells of which the gland itself is almost entirely composed. Tse Liver.—The greater part of the posterior half of the body-cavity is occupied by the liver or digestive gland (PI. IX, figs. 27-31). In Apera gibbonsi, A. parva, A. burnupi, and A. sexangula the liver consists of two very distinct divisions, one anterior, dorsal, and to the right, the other posterior, ventral, and to the left (Pl. XXI, figs. 129- 131, 155 and 154). Not only do these divisions discharge their secretions into the alimentary canal by quite separate ducts, but they are supplied with blood by different arteries, the right division being supplied by one or two branches from the anterior aorta, while the left division is suppled by the so-called posterior aorta. The right division of the liver is divided by the intestine into three main lobes, one lying within the anterior loop of the intestine, another in front of it, and the third behind it, as shown in the figures. Of these the anterior lobe, lying above the hind end of the crop, is the THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 179 most distinct. Hach lobe is again subdivided into a number of lobules, the arrangement of which varies in different individuals. he left division of the liver les entirely to the left of the imtestine, and is about equal in size to the right. It also is divided and subdivided into lobes and lobules, but not so distinctly as is the right division. In Apera dimidia the liver is also divided into two main divisions, one right and anterior, the other left and posterior ; and, as in the other species, these divisions have separate ducts. They are not, however, quite so distinct from each other as in the species mentioned above, and they both receive their blood from a single branch of the anterior aorta, the so-called posterior aorta being absent. Moreover, the disposition of these divisions with regard to the intestine is quite different to that which I have described as character- istic of A. gibbonsi, A. parva, A. burnupi, and A. sex- angula. ‘The anterior or right division les above the crop, entirely in front of the intestine, while the posterior or left division, which is divided into two lobes, occupies a position corresponding to that held in the other species by the two posterior lobes of the right division of the liver (Pl. XXI, fig. 132). This fact suggests that possibly the whole of the liver of A. dimidia corresponds to the right division of the liver in the other species. If this theory be correct, it would explain, not only the anomalous position of the liver with regard to the intestine in A. dimidia, but also its still more surprising blood-supply. We would have to suppose, however, that the origin of the branch of the right hepatic duct coming from the anterior lobe had shifted inwards as far as the intes- tine itself; but this might have been caused by the greater separation of the lobe due to the lengthening of the right division of the liver as it came to occupy also the position of the left division. And this separation of the two principal branches of the hepatic duct would be merely a case of rever- sion, for H. Fischer! has shown that each division of the liver was probably originally subdivided into two or more lobes, 1 <« Bull. Sci. France et Belg.,’ 1892, vol. xxiv, p. 331. 180 HUGH WATSON. which opened separately into the almentary canal. But it is more difficult to explain why half the liver should disappear in thisway. Itis known, however, that the reduction of that part of the body-cavity which usually occupies the spire of the shell sometimes produces a corresponding reduction of the left lobe of the hver which it contains.!. It seems pos- sible, therefore, that in A. dimidia the left division of the liver, instead of merely pushing forward the right, as in the other species of Apera, has disappeared altogether, Just as the right division has disappeared in some of the Pectini- branchia. On the other hand, it is possible that the two divisions of the liver in A. dimidia are actually homologous with the two divisions in the other species. Blood-vessels are always very subject to variation, and if by some mutation the so- called posterior aorta had disappeared, the arteries supplying the right division of the lver would probably extend to the left also. Further, the posterior aorta passes to the left of the intestine, and may to some extent hold it in position (Pl. IX, fig. 27). When, therefore, this vessel is not present, there is nothing to prevent the intestine shitting to the left and coming to occupy the position that it holds in A. dimidia. Possibly an examination of the hver of Apera purcelli may show which of these views is correct. In this species the intestine seems to occupy a similar position to that of A. dimidia, but unfortunately the liver of the only specimen of A. purcelli which I have been able to examine was in a state of partial disintegration. Sections through the liver of Apera dimidia show that the food is not confined to the alimentary canal itself, but passes up the hepatic ducts into the lobes of the liver, and probably digestion takes place chiefly inside this organ in Apera as in Atopus.? Perhaps the disappearance of a stomach in these genera may be partly explained by this fact. 1 Pelseneer, P., ‘ Mém. Acad. Roy. Belg., 1901, ex vol. liv, p. 55. Simroth, H., ‘ Naturwiss. Wochenschr.,’ 1901, vol. xvii, p. 121. HE CARNIVOROUS SLUGS OF SOUTH AFRICA. 18] THE VASCULAR SYSTEM. Tue Heart and Putmonary Veins.—The genus Apera is opisthopneumic, the greater part of the respiratory tissue being behind the heart. Unlike Daudebardia, the auricle has moved round with the pulmonary veins so as to lie obliquely behind and to the right of the ventricle (PI. 1X, figs. 27-31), but it has not rotated so far as in Testacella. The pulmonary veins form a complicated network, variable in its structure; but it is generally possible to distinguish three main vessels converging towards the auricle, namely, the right anterior pulmonary vein, from the corner of the mantle-cavity to the right of the pericardium ; the right posterior pulmonary vein, from the region between the pericardium and the anus ; and the left posterior pulmonary vein, from the part of the mantle-cavity behind the kidney. Owing to the absence of the respiratory tissue from the left anterior corner of the mantle-cavity there is no left anterior pulmonary vein. The walls of the auricle are very thin (Pl. XI, fig. 35), but those of the ventricle are extremely thick and muscular (fig. 34). The cardiac muscle-fibres are arranged in bundles passing in various directions, and leaving spaces between them which communicate freely with the central cavity of the ven- tricle (Pl. XXII, fig. 140). The muscles are thus well supplied with blood without the intervention of any cardiac arteries. THe ArrertaAL System.—The distribution of the principal arteries is shown in PI. IX, figs. 27-29 and 31. The single aorta leaves the posterior end of the ventricle, passes through the diaphragm, and then bends abruptly to the right and curves downwards, so that it comes to lie on the right side of the liver. In Apera gibbonsi, A. parva, A. burnup1, and A. sexangula the so-called posterior aorta branches off to the left within 3 or 4 mm. of the diaphragm. This vessel passes forwards on the right side of the left division of the liver, keeping to the left side of the intestine. It divides into two or three branches supplying the left division of the liver, but the most anterior branch first passes through the 182 HUGH WATSON. hermaphrodite gland, to which it also supphes blood. In A. burnupi, but not in the other species, this branch also gives off a small artery to the hind end of the right division of the liver. In A. parva the posterior aorta divides immediately into three branches at the point where it separates from the anterior aorta, but in the other species the division of the vessel takes place further forward. No posterior aorta is present in A. dimidia, and there is probably none in A. purcelli either. I would suggest that perhaps this so-called posterior aorta may be merely the left hepatic artery which has become dis- placed, and that the Gastropoda (with the possible excep- tions of Haliotis and Fissurella') resemble the Amphi- neura and the more primitive members of the Pelecypoda in having no true posterior aorta. The posterior position of this artery would be accounted for by the posterior position which the left division of the liver has come to occupy owing to the torsion of the visceral hump. Its comparatively large size in most genera might be due to the fact that the left division of the liver is generally much larger than the right, since it extends up the spire of the shell. And as the herma- phrodite gland is usually more or less embedded in the left division of the liver, it would be hkely to be supplied with blood by the same artery. This view wili be made clear by the accompanying diagrams A to bp. Diagram a represents a primitive symmetrical arrangement, and c the condition after the torsion has taken place; B shows a hypothetical stage between a and c, while D represents the most usual arrange- ment of the arteries and liver in Apera. The anterior aorta continues forward on the right side of the liver, crossing over the intestine in A. gibbonsi, A. burnupi, and A. sexangula. In A. gibbonsi, A. dimidia, and probably in A. purcelli, it gives off a single artery to the liver on reaching the level of the anterior loop 1 See Wegmann, H., * Arch. Zool. Expér.’ (2nd ser.), 1884, vol. ii, pp. 352, 353, pl. xviii, figs. 1-4; and Boutan, L., ibid., 1885, ex vol. iii bis suppl., pp. 34, 180, pl. xxxiil, figs. 5, 6. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 183 THXT-FIG. 5. Diagrams illustrating the hypothetical evolution of the “posterior aorta.” (Diagram D shows the arrangement in Apera gibbonsi lupata.) L.L. = Left division of liver. R. = Rectum. ee Right division of liver. V.= Ventricle. 184 HUGH WATSON. of the intestine. This hepatic artery soon divides into two branches. In A. gibbonsi the anterior branch chiefly supplies the anterior lobe of the right division of the liver, while the remainder of the right division is supplied by the posterior branch. In A. dimidia the anterior branch supplies the anterior division of the liver, while the posterior branch supphes not only the posterior division, but also the hermaphrodite gland which it traverses. In Apera burnupi and A. parva the blood-supply of the right division of the liver is similar to that in A. gibbonsi, excepting that the two branches arise separately from the anterior aorta, though very close together. In A. sexangula these arteries also arise separately from the anterior aorta, and their origins are much further apart. The anterior aorta passes through the loop formed by the intestine in Apera burnupi, A. sexangula, A. parva, andin A. gibbonsi gracilis and A. g. lupata (Pl. IX, figs. 28, 31, and text-fig. 5, p); but in A. dimidia, A. pur- celli, A. gibbonsi s. s., and A. g. rubella, the aorta passes straight forward on the right side of the loop (figs. 27, 29). It is very remarkable to find this important difference separating forms which seem to be so closely related that I have not ventured to regard them as distinct species. For the difference amounts to this: that in the first group the aorta, on its way from the dorsally situated heart to the ventral ganglia, passes on the left side of the alimentary canal, while in the second group the aorta passes on the right.! The aorta then runs forward between the digestive and re- productive organs, being often closely applied to the left side of the albumen gland. It gives off to the left three or four arteries to the crop and salivary glands, and to the right at least one artery to the albumen gland and another to the common duct, a branch of the latter going forward to the receptaculum seminis. The aorta then becomes more ventrally situated and 1 A similar variation in the course of the aorta has been found by Kohler inthe genus Siphonaria (‘ Zool. Jahrb., 1893, vol. vii; compare fig. B, p. 27, with figs. A and c, p. 32). 'HE CARNIVOROUS SLUGS OF SOUTH AFRICA. 185 runs forward towards the ventral group of gangha. Just before reaching these a pair of arteries is given off to the tentacular retractors (Pl. XIII, figs. 52 and 54-57). In most of the species these arteries unite with the common tenta- cular retractors just behind the points where the latter divide ; but in Apera dimidia they unite with the upper tentacular retractors, and give off small branches to the retractors of the lower tentacles, which join the nerves from the sides of the “nerve-collar” (fig. 54). Finally the aorta fuses with the ventral group of ganglia. From this point radiate numerous short vessels to the various parts of the central nervous system, etc., as well as two longer arteries, namely, the buccal and pedal arteries. The buccal artery passes upwards and divides into two branches, one running forwards and the other backwards. In Apera dimidia the posterior branch is about twice as long as the other (Pl. XVII, fig. 103) ; it runs back to the posterior extremity of the odontophore, which it enters, and then bends round the hind end of the odontophoral support (Pl. XVIII, fig. 122). The anterior branch runs forward and divides into a pair of arteries which apply themselves to the sides of the bueeal mass, dividing again as they do so into anterior and posterior vessels. In A. burnupi and A.sexangula the buccal artery branches in a similar manner, but owing to the small size of the odontophore, the posterior branch is no longer than the anterior branch (Pl. XVII, fig. 110). In A. gibbonsi and A. parva, on the other hand, the posterior branch is very long, and in the former species the right and left anterior arteries do not branch from a single median vessel, but arise separately (Pl. XIII, fig. 52). In Testacella the artery supplying the odontophore with blood behaves differently, for, instead of running straight to the hind end of the organ, it divides into lateral vessels which pass through the sides of the sheath as shown in de Lacaze- Duthier’s admirable figures.!. On the other hand, in Nata- lina quekettiana (M. & P.), this artery behaves in the ’ ‘Arch. Zool, Expér.’ (2nd ser.), 1887, vol. v, pls. xxxv, xxxvi. 186 . HUGH WATSON. same way as it does in Apera. In EHuglandina we have an intermediate arrangement, as will be seen from Strebel’s figures!; for the structure that he described and figured as a median odontophoral nerve from the ventral group of ganglia, is, in reality, the buccal artery, as I have found from an examination of HE. venezuelensis (Preston). The pedal artery runs backwards above the pedal gland, to which it gives off branches on each side. In Apera gib- bonsi, A. dimidia, and A. sexangula, I have founda very slender vessel uniting the anterior aorta with the pedal artery about half way along the pedal gland (Pl. XIII, figs. 52, 54, and 57). After uniting with the terminal vesicle of the pedal gland, the pedal artery continues backwards and downwards, and becomes more or less embedded in the muscular con- nective tissue of the foot. The venous system is largely lacunar, apart from the pulmonary veins already described. Indeed, the only well- defined vessel containing venous blood is the body-cavity itself. I have already dealt with the dermal veins and their curious sphincters, when treating of the structure of the skin.’ THE EXCRETORY SYSTEM. The pericardium communicates with the mantle-cavity by means of a single renal ccelomoduct, divided into an ascend- ing glandular portion, which forms the kidney, and a descending non-glandular portion, which forms the ureter. It will be seen from Pl. IX, figs. 27-31, Pl. XI, figs. 34, 35, and Pl. XII, fig. 36, that the kidney is applied to the left side of the pericardium, and also extends over its posterior part towards the right side, thus broadly connecting the peri- cardium with the wall of the mantle-cavity, with which the kidney fuses a little further back. Ventrally the kidney is prolonged backwards in the floor of the mantle-cavity on ' «Beitrag z. Kenntn. d. Fauna Mexikan. L.- u. Siisswasser-Conchylien,” 1878, vol. iii, pl. xvi, fig. 5, pl. xviii, figs. 1, 6, 10; pl. xix, fig. 1, pl. xx, fig. 1. ? See pp. 126, 127. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 187 the left side of the rectum and even beyond the anus (PI. X - fig. 33). The walls of the kidney are very much folded internally (see especially Pl. XI, fig. 35), and are lined by an epithelium consisting of glandular excretory cells with conspicuous round nuclei (Pl. XXII, fig. 141). The reno-pericardial duct is very narrow, but projects some distance into the kidney. Its position is shown in Pl. XI, fig. 35. The epithelial cells lining the duct are provided with very long cilia. The kidney communicates with the ureter by a minute pore situated in the centre of a slightly raised papilla at the right end of its dorsal extension (Pl. XII, fig. 36). The ureter, as in other sigmurethrous Stylommatophora, is doubled back upon the kidney, and curves round its inner posterior side. Ventrally it is continued backwards beside the ventral prolongation of the kidney, and extends to the extreme end of the mantle-cavity, where it bends abruptly upwards and opens. ‘I'he ureter is lined by a_ cubical epithelium, the appearance of which is shown in Pl. XXII, fig. 141. This epithelium becomes thinner over the papilla in which the opening of the kidney is situated ; elsewhere it is remarkably like the epithelium which lines the lower surface of the shell-sac. As in Testacella and the Rhytidide, there is no secondary ureter; but a rather irregular groove runs for- wards from the opening of the ureter along the roof of the mantle-cavity. This groove reaches nearly to the point where the ureter begins, and then bends round and runs obliquely backwards and to the right, to end in the cleft in the inner lip of the respiratory opening. The epithelium lining this groove is similar to that lining the ureter itself, and the groove is probably to be regarded as an incipient secondary ureter, such as is found in many other forms. The somewhat peculiar course of the groove might be explained on the assumption. that the respiratory orifice was originally situated further forwards. VOL. 3, PART 2. 14 188 HUGH WATSON. THE REPRODUCTIVE ORGANS. The genital or reproductive organs of Apera are of a simple generalised type, at least so far as their external structure is concerned (Pl. XXIII, figs. 145, 144, 146-149). As up to the present only a very few fully mature specimens have been collected, I have thought it advisable to preserve the reproductive organs of these almost intact, and not to cut them up in order to examine their histology. The following account, therefore, deals chiefly with the external morphology of the genital system. The reproductive organs of Apera purcelli are unknown, the only dissected specimen being immature. THe HermMapHropite AND Femate Orcans.—The herma- phrodite gland or ovotestis 1s partially embedded in the right side of the posterior division of the liver (see especially Pl. IX, fig. 31). It is a racemose gland composed of a cluster of oval or pear-shaped follicles. These follicles are smallest in Apera burnupiand A. sexangula (Pl. XXIII, figs. 148, 149). In the other species they are larger and tend to be fewer in number, until in A. parva the entire hermaphrodite gland consists of only about four follicles (fig. 146). The very slender ducts from the different follicles unite to form the hermaphrodite duct, which is long, and usually more or less convoluted and slightly swollen towards the middle of its length, especially in A. dimidia. But the convolution is not so pronounced as in many other genera. In front the hermaphrodite duct becomes embedded in the albumen gland. At the extreme anterior end it is swollen so as to form a minute spherical vesicle, which is doubtless to be regarded as a rudimentary vesicula seminalis. The albumen gland varies enormously in size according to the exact state of maturity of the individual (cf. Pl. IX, figs. 27, 31). In some specimens it reaches a very large size indeed (Pl. XXIII, fig. 144). The common duct, or ovi- spermatoduct, is very long and much convoluted and twisted in Apera burnupi and A. sexangula (Pl. XXIII, figs. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 189 148, 149); in A. gibbonsi, A. parva, and A. dimidia it is shorter and generally broader (figs. 145, 144, 146, 147). In some specimens the female or oviducal portion is greatly swollen, as in the example of A. dimidia of which the repro- ductive organs are shown in fig. 147. The free oviduct is rather long and narrow, and is directly continuous with the vagina, which is also rather narrow but usually somewhat shorter. The vestibule or genital atrium is extremely short. THe RecepracuLum Seminis AND irs Duct.—The sperma- theca or receptaculum seminis is an oval vesicle, situated, in A. parva, beside the anterior end of the free oviduct, the receptacular duct in this species being extremely short (fig. 146). In the other forms the receptaculum seminis hes against the anterior end of the common duct, and the length of the receptacular duct is therefore about equal to that of the free oviduct, since the anterior opening of the receptacular duct marks the junction of the free oviduct with the vagina. In Apera gibbonsi the receptacular duct is a simple narrow tube (figs. 144, 145). In A. sexangula it becomes swollen towards its union with the vagina, so that the anterior part of the duct is fusiform (fig. 149). In A. burnupi the anterior part of the duct is still more swollen (fig. 148). Lastly, in A. dimidia the swollen end of the duct has become attached to the adjacent part of the body- wall by a number of muscles on each side (fig. 147). And as we pass up this series the receptaculum itself becomes progressively smaller, and it seems as though the receptacular duct were taking its place. If this tendency were to proceed still further we should have a vesicle which would be indis- tinguishable from a sessile receptaculum seminis, although in reality it was the swollen anterior end of the receptacular duct, the rest of the organ having disappeared. Possibly this may have happened in some of the genera which are usually described as having a sessile receptaculum. Tue Marte Orcans.—The vas deferens arises from the anterior end of the common duct, and passes forwards close to the free oviduct and vagina as far as the vestibule. It then 190 HUGH WATSON. bends round and runs up the side of the penis. In Apera gibbonsi the vas deferens is rather thick excepting near the vestibule, where it has the more usual form of a very slender duct. In this species, and also in A. dimidia, before the vas deferens unites with the posterior end of the penis, it is slightly swollen for a distance equal to about one-third of the length of that organ, thus forming a rudimentary epiphallus. In A. sexangula this swollen part is very much shorter, and bears at its anterior end a minute flagellum, which, like the vas deferens itself, is closely applied to the side of the penis (Pl. XXIII, fig. 149). The penis passes between the retractor muscles of the right upper and lower tentacles (Pl. XIII, figs. 52-57). In this respect Apera resembles Testacella, Trigono- chlamys, Poiretia, Streptostyla, Strebelia, Rhytida, and Paryphanta, and differs from Daudebardia, Plu- tonia, Selenochlamys, Natalina, Delos, and the Strep- taxidee, in which the reproductive organs are on the right of both retractors, as was first shown by Pfeffer, Simroth, and Murdoch. In Apera dimidia the penis is fairly short, but a little swollen posteriorly (Pl. XXIII, fig. 147). In A. parva it is about 4 mm. in length, and somewhat fusiform (fig. 146). In typical specimens of A. gibbonsi from Zulu- land, the penis is also short (fig. 143), but in the other races of that species it is very long and twisted in a peculiar manner (fig. 144). In A. sexangula the penis is long and very narrow, excepting at the posterior end, where it is broad and somewhat flattened (fig. 149). In A. burnupi the anterior part of the penis is also narrow, but it is much shorter than in A. sexangula, and more or less twisted. The posterior broader portion, on the other hand, is longer, and appears to be double, owing to the presence of a deep longitudinal groove on one side; and the vas deferens dis- appears into this groove about 5 mm. in front of the end of the penis (fig. 148). Possibly this peculiar arrangement may be due to the broad sides of the penis having arched over the end of the vas deferens and joined each other above it ; but THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 191 this hypothesis must remain very doubtful until a sufficient number of specimens of this rare species has been collected in order to justify the dissection of the penis. In a full-grown example of A. gibbonsi rubella from Durban, the penis was found to be quite vestigial. As so few full-grown specimens of Apera have been collected, I have only felt justified in examining the internal structure of the male organs in two forms, namely, A. dimidia and A. gibbonsi rubella. No penis-papilla is present in either of these species, but in both forms the inner surface of the muscular wall of the penis is completely covered by numerous small papille, as shown in Pl. XXIII, fig. 145. ‘These papille are diamond-shaped in transverse section, and each ends in a short point. The structure of the papille is very similar in the two species, and is shown in Pl. XXII, fig. 142, which represents a photomicrograph of the inner half of a single papilla from the penis of A. gibbonsi rubella. In this form the papillae become considerably smaller around the opening of the epiphallus, and within the opening they become much narrower in proportion to their length, changing, in fact, from papille into filaments. These fila- ments occur throughout the entire length of the epiphallus. They do not, however, arise from all parts of the inner surface of the wall of the epiphallus, but are arranged in longitudinal rows, of which there are about fourteen in the specimen that I examined. Between the rows the wall of the epiphallus is extremely thin. In Apera dimidia the epi- phallus has a different structure, being lined by numerous minute papille, closely arranged, and not unlike the papille lining the penis, only much smaller. It is somewhat remark- able that while the inside of the penis is so very similar in the two species, the inside of the epiphallus should present such a marked contrast. The retractor of the penis is attached to its posterior end, and arises from the roof of the body-cavity near the mid-dorsal line. In Apera burnupi, A. sexangula, A. purcell1, and A. parva, it is very long and narrow, arising in A. 192 HUGH WATSON. sexangula and A. parva less than 4 mm. in front of the diaphragm, and only a very short distance further forwards in the other two species (Pl. IX, figs. 28, 30, 31). On the other hand, in A. gibbonsi and A. dimidia it is broader and very much shorter, the exact position of its origin depending on the length of the penis (figs. 27 and 29). Probably the retractor originally arose in the neighbourhood of the diaphragm ; but owing to the posterior position which the diaphragm has come to occupy, the origin of the retractor has moved forward. If this view be correct, the condition found in A. sexangula and A. parva would be the most primitive. THE SperMavozoa.—Pl. XXIII, fig. 151, shows the anterior end of a spermatozoon of Apera dimidia magnified 1200 diameters. It will be seen that the head is curved into the shape of a hook, and is about °0035 mm. in length. The ‘““middle-piece,” or proximal portion of the tail, is surrounded by a spiral filament or flange, very loosely coiled, the whorls being about ‘0045 mm. apart. The posterior portion of the tail is extremely long. In Apera sexangula and A. gibbonsi rubella the head of the spermatozoon is straighter and broader, though pointed in front. Fig. 150 shows the anterior end of a spermatozoon of the latter variety. It will be seen that in this form the whorls of the spiral filament are a little nearer together than in A. dimidia, being separated by a distance equal to the length of the head, namely, ‘0037 mm. The spermatozoa of Testacella are very different from those of Apera, as may be seen from Pl. XXIV, fig. 160. In Testacella the head is globular, and there are two spiral filaments surrounding the proximal portion of the tail. Moreover the filaments are much more closely coiled than in Apera, adjacent whorls being only about ‘0009 mm. apart in T. maugei. In Rhytida capillacea (Fér.) I find that there are also two filaments, but they are not so closely coiled, the whorls being separated by about ‘00165 mm. ; and in this species, as well as in Natalina quekettiana (M. & P.), the ee THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 1938 head is not unlike that of Apera sexangula and A. gvibbonsi rubella. On the other hand, in Streptaxis funki Pfr. the head is even more strongly curved than in Apera dimidia; and, as in the Testacellidew and Rhy- tididw, I have found two spiral filaments encircling the proximal portion of the tail, the distance separating adjacent whorls in this species being about ‘00115 mm. The embryology of A pera is unknown. DISTRIBUTION. The genus Apera is only known to inhabit the maritime provinces of South Africa. One species—A. purcelli— occurs on the slopes of Table Mountain close to Cape ‘own; the other forms have only been found in Natal and in the eastern districts of the Cape Province. A. parva is at present only known to occur near Grahamstown; A. sexangula extends from that neighbourhood as far east as Pieter- maritzburg ; the various races of A. gibbonsi are known to inhabit an area stretching from Pondoland to Zululand ; while A. burnupi and A. dimidia have only been found in Natal. The animals are usually found in the bush, either amongst dead leaves or under stones or logs; but none of the species are at all common. Apera sexangula and A. gibbons} seem to be the least rare. Of the others only six examples ot A. dimidia, four of A. burnupi, two of A. purcelli, and one of A. parva are known to have been collected. It 1s to be hoped that South African naturalists will gather further specimens of this interesting genus, which not improbably contains more species than are at present known. THE DIFFERENTIAL CHARACTERS OF THE SPECIES AND SUBSPECIES OF APEHRA. Apera gibbonsi (Binney). Chlamydephorus gibbonsi Binn., Bull. Mus. Comp. Zool. Har- vard, 1879, vol. v, p. 331, pl. ii, figs. a,b; Binney, Ann. N.Y. Acad. 194, HUGH WATSON. Sci., 1884, vol. iii, p. 81, pl. xvii, fig. a; Tryon, Man. of Conch. (2nd ser.), 1885, vol. i, pp. 17, 251, pl. ii, fig. 95; Cockerell, Ann. Mag. Nat. Hist. (6th ser.), 1890, vol. vi, p. 390. Chlamydophorus gibbonsi Binn.; Tryon, Struct. and Syst. Conch., 1884, vol. iii, p. 13, pl. ci, fig. 47. Apera gibbonsi (Binn.); Heynemann, Jahrb. d. Deutsch. Mal. Ges., 1885, vol. xii, p. 17, pl. ii, figs. 5-7; Collinge, Ann. 8. Afr. Mus., 1900, vol. ii, p. 4; Collinge, Ann. Natal Mus., 1910, vol. ii, p. 165; Connolly, Ann. S. Afr. Mus., 1912, vol. xi, p. 68. Exrrernat Cuaracters (PI. VII, figs. 1,2; Pl. VIIL, figs. 14, 15).—Animal slender, tapering to an acute angle at the hind end. Back rounded, without keels. Outer lip of respiratory opening almost concealing the inner lip. Skin coarsely reticulated. Dorsal grooves usually about 1 mm. apart, separated by a single row of ruge, but united for about 4 or 5 mm. in front of the respiratory opening. Lateral grooves irregular, rather near the dorsal grooves, and seldom extend- ing as far as the head. Radial grooves conspicuous, some being forked. Oblique sub-lateral grooves also conspicuous, and rather numerous; these and the radial grooves termi- nating in the single peripodial groove. In the typical form the body is mottled with brown, the small patches of colour being chiefly concentrated dorsally, but leaving-a paler band, very sparsely mottled, along the centre of the back from the head to the respiratory opening. Laterally the brown pigment becomes confined to the sides of the grooves in the skin, and dies out completely some distance above the peripodial groove. The ground-colour is pale yellow in specimens preserved in spirit, but living examples are tinged with dull orange, the colour being deepest on the back. The type was described by Mr. Gibbons as being rather more than 3 in. long when fully extended. The following are the dimensions (in alcohol) of two specimens that I have examined, but the reproductive organs of the smaller one were not quite fully developed. a THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 195 Length . é : ; 2) 623% mm o-.2 66:5. mm: Breadth in middle . ; fh Ose +, 2 PC eee Breadth at respiratory opening 7 , . 9 ,, Breadth of foot-sole . ; 2S) PARA Rh Pees toy ido Greatest height ’ : ALO Ton ae Linas 9 Parse Distance from respiratory open- ing to hindend . :

Re XXIII, figs. 145, 150). Hasirat.—Equeefa, Natal (Burnup) ; Durban (under dead leaves in the bush which fringes the shore; Burnup). ; Type.—The type of subspecies is in the Natal Museum, Pietermaritzburg. Another specimen will be found in the British Museum. A¥FInirigs, Evc.—The chief characters which distinguish this form from Apera gibbonsi s.s. are its broader shape and darker colour, its radula, and its penis (although it is not known whether the penis of the typical form is always as short as it isin the Zululand specimens). It is possible that this may bea geographical variety of A. gibbonsi, as the two forms are not known to occur in the same district, but the radule are so different as to suggest that this may prove to be a distinct species. The single specimen froin Durban differs from those found at Equeefa in being without a paler dorsal band, in having more teeth in each row in its radula, and in the penis being abnormally reduced to a minute vestige. The resem- blance of this specimen to certain dead leaves amongst those under which it was found has already been referred to (see p. 122). Mr. Burnup informs me that this subspecies contracts after the usual manner when about to rest, instead of merely a THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 199 bending its body at various angles like the typical form of A. gibbonsi. Apera gibbonsi gracilis n. subsp. ExternaL Cuaracters (Pl. VII, fig. 6).—Animal even more slender than the typical form of Apera gibbonsi, tapering to a very acute angle at the hind end. Outer lip of the respiratory opening not concealing so much of the inner lip as in the other forms of A. gibbonsi. Reticulation of skin coarse, as in the typical form; dorsal grooves united for 6 mm. in front of the respiratory opening. Body rather sparsely mottled with brown, as in A. gib- bonsi, s.s.; the patches of colour, however, are not concen- trated on each side of the mid-dorsal area, but extend across it, so that there is no median paler band. The following are the measurements of the only specimen that has been found : Length . 44mm. Greatest height . 6 mm. Breadth in middle 6_,, Distance of resp. Breadth at resp. opening from hind opening. i Abe, end © . 1S ee Breadth of foot-sole 5 ,, ) InrernaL CHaracters.—Shell (Pl. XIII, fig. 43) rather longer and narrower at the posterior end than in the other forms of A. gibbonsi. Radula (Pl. XIX, fig. 123; text- fig. 3, c, p. 159) with fewer teeth than that of any of the other forms; central tooth absent; first on each side extremely small; outer teeth larger than in the preceding forms; cusps only slightly curved, not broadened or barbed; apophyses very prominent ; formula: (19 + 0 + 20) x 34. Anterior aorta passing through loop of intestine; branches of pul- monary veins not very numerous. Penis rather long and twisted. Other characters agreeing with those of A. gib- boms1/ sas. Hasirar.—Hqueefa, Natal (Burnup). 200 HUGH WATSON. Type.—The type of subspecies is in the Natal Museum, Pietermaritzburg. AFFINITIES, ETC.—The single specimen of this form differs greatly from the last subspecies, although it was found in the same locality. In some respects it is intermediate between Apera gibbonsi s.s. and the next subspecies, but it seems impossible to associate it with either of these forms, and it must therefore stand alone until further examples have been collected. It will be noticed that owing to the length of the posterior extremity the respiratory opening is proportionately further forward than in the other forms. Apera gibbonsi lupata n. subsp. ExTernaL Cuaracters (Pl. VIII, figs. 18, 19)—Animal intermediate between A. gibbonsis.s. and A. g. rubella both in shape and in the reticulation of the skin, being usually a little broader and less coarsely reticulated than the typical form of A. gibbonsi, and a little narrower and less finely reticulated than A. g. rubella. Colour variable; sometimes dark like A. g@. rubella, sometimes light as in A. gibbonsi s.s. In the typical form the mottling is inconspicuous and there is only a very slight trace of the paler median dorsal band. This form (and possibly also the last subspecies) may be slightly smaller than A. gibbonsi s.s. The following are the measurements of a specimen which seemed to be practically mature : Length ‘ . 46mm. Greatest height . 7 mm. Breadth in middle 7°5 ,, Distance of resp. Breadth at resp. opening from hind opening . + [Ona end? 2: : See Breadth of foot-sole 4 ,, InrerNAL CHaracters.—Shell much depressed. Radula (Pl. XVII, fig. 99; Pl. XIX, fig. 124; text-fig. 3, p, p. 159) with a comparatively small number of very large teeth; central tooth typically absent ; cusps somewhat curved and THE CARNIVOROUS SLUGS OF SOULH AFRICA. 201 conspicuously barbed on the lower side; apophyses promi- nent; formulee of two specimens (the second slightly imma- ture): (20 + 0 + 20) x 44, (21 + 0+ 21) x 38. Anterior aorta passing through loop of intestine (text-fig. 5, D, p. 183); pulmonary veins not very numerous. Penis long and twisted. Other characters agreeing with those of A. gibbonsi s.s. Hasrrar.—Port Shepstone, Natal (Burnup). Var. duplex n. var. Cuaracters.—Back distinctly mottled, with a well-marked paler median band, the coloration being very similar to that of Apera gibbonsi s.s. Cusps of teeth rather longer than in the typical form of the subspecies, and having an additional barb on the upper side near the poimt (excepting in the smaller teeth) ; apophyses very prominent; a vestigial central tooth is present in most of the rows (text-fig. 3, BE, p. 159) ; formula: (22 + (1) + 22) x 42. Otherwise similar to the Port Shepstone specimens of A. gibbonsi lupata. Hasrrar.—Port St. John’s, Pondoland (one specimen; Mrs. K. Warren). Tyezr.—The types of subspecies and variety are in the Natal Museum, Pietermaritzburg. Another specimen of the subspecies will be found in the British Museum. AFFINITIES, ETC.—This subspecies is especially characterised by its large barbed teeth. ‘he course of the anterior aorta also serves to separate it from the typical form of Apera gibbonsi, as well as from A. g. rubella. But although the average shape of this subspecies is slightly different from either of these forms, the differences are so small that in practice I find that it 1s quite impossible to distinguish, without dissection, light specimens of the present subspecies from A. gibbonsi s.s., and dark examples from A. g. rubella, unless, perchance, the odontophore is sufficiently protruded to enable one to see the teeth. It follows, there- fore, that none of these three forms can be identified with 202 HUGH WATSON. certainty from an examination of their external characters only. It will be noticed that this race of A. gibbonsi seems to have a more westerly distribution than the preceding forms. Apera parva n. sp. ExternaL Cuaractrers (Pl. VIII, figs. 22, 23).—Animal very slender, tapering to an acute angle at the hind end. Back rounded, without keels. Outer lip of respiratory open- ing almost concealing the inner lp. Skin coarsely reticu- lated. Dorsal grooves about 4 mm. apart, separated by a single row of ruge, but united for 3 mm. in front of the respiratory opening. Lateral grooves near the dorsal grooves and extending forwards to the head, the right one ending in the genital opening. Radial grooves well-marked, but fewer than usual, only three arising from the respiratory opening on each side (in addition to the two median grooves). Oblique sub-lateral grooves less numerous than in A. gibbonsi, being absent from the anterior third of the animal. Both radial and sub-lateral terminate in the single peripodial groove. Mr. Farquhar informs me that the living animal is fawn- coloured ; dark dorsally, but much lighter below. In alcohol the specimen has become yellowish brown, as shown in the figures. The following are the dimensions of the slug in alcohol; but when alive and moving Mr. Farquhar tells me that it attained a length of about 35 mm. Length : . 25 mm. Greatest height . 3 mm. Breadth in middle 38 ,, Distance from resp. Breadth — at resp. opening to hind opening. . sg 2eAae end . ( Sb base Breadth offoot-sole 19 ,, INTERNAL CHaracters.—Skin very thick. Shell (Pl. XII, fig, 44, 45) 1°6 x -9 mm., thick, depressed, with a flattened THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 203 apex in front of the posterior margin; sinus of moderate depth ; left posterior angle obliquely truncate. Pedal gland (PI. XIII, fig. 53) —Slender; glandular tissue gradually dying out a little behind the middle of the gland ; duct less convoluted posteriorly than in A. gibbonsi. Nervous system.—Cerebral gangha clearly separate ; buccal ganglia situated posteriorly and jomed to the cerebral ganglia by long connectives; cerebro-pedal and cerebro- pleural connectives rather long. Digestive System (Pl. XVII, figs. 100-102 ; text-fig. 4, A, p. 161; Pl. XXI, fig. 131).—Radula very large, 8 mm. long; central tooth very small, absent in some rows; others scarcely differentiated into laterals and marginals, becoming extremely large towards the edges of the radula; cusps simple, very slightly curved, short towards the centre of the radula, longer towards the edges; bases broader than in A. gibbonsi; corresponding teeth not opposite to each other; formula: (17 + 1 + 17) x 48. Odontophore also very large; posterior end joined by numerous short radial buccal retractors to the sides and floor of the body-cavity about the middle of its length ; odontophoral support rounded in front. (Hsophagus short and broad, merging into a fusiform crop. Between the crop and the wide intestine is an oval stomach, into which the hepatic ducts open. Second loop of intestine absent, the alimentary canal curving over to the right side of the liver and then passing straight backwards. Salivary glands contiguous above the crop, but less closely fused than in A.gibbonsi. Liver in two approximately equal divisions, the intestine only traversing the right. Vascular System (Pl. IX, fig. 28).—Pulmonary veins not very numerous. “ Posterior aorta” dividing at its origin into three branches which supply the left division of the liver and the hermaphrodite gland; right division of liver supplied by two arteries from the anterior aorta, which arise separately but close together; anterior aorta passing through loop of intestine ; anterior branches of buccal artery leaving the posterior branch as a single median vessel. VOL. 3, PART 2, 15 204 HUGH WATSON. Reproductive System (Pl. XXIII, fig. 146). — Her- maphrodite gland composed of only about four follicles, which are relatively very large; albumen gland in the specimen examined extending to the left side of the body-cavity, passing between some of the posterior buccal retractors ; common duct rather short, curving twice to the right and once to the left; vagina rather narrow ; receptaculum seminis further forward than in the other species; receptacular duct broad and extremely short. Vas deferens not enlarged to form an epiphallus ; penis straight, rather short, and shghtly swollen in the middle; penial retractor long and narrow. Hasitar.—Near Fern Kloof, about three miles south of Grahamstown, Cape of Good Hope (in a stony place with very little grass, at least 200 yards away from the bush; Farquhar). Type.—The type isin the Natal Museum, Pietermaritzburg. Arrrnities, etc.—The only example of this species at present known was discovered by Mr. John Farquhar, of Grahams- town, on October 12th, 1915. In many respects the slug bears a close resemblance to Apera gibbonsi, and it might easily be mistaken for a young specimen of that species. Nevertheless, it differs from all the forms which I have included under A. gibbonsi in so many characters that there can be very little doubt that it is specifically distinct. The following are the features in which some of the most important differences occur: the size; the number of the radial and sub-lateral grooves; the distance between the cerebral ganglia; the size of the teeth towards the edges of the radula, and the width of the bases of all the teeth; the position of the origin of the buccal retractors ; the stomach ; the course of the intestine; the division of the arteries to both portions of the liver; the hermaphrodite gland; and the length of the receptacular duct and of the penial retractor, these last features being perhaps the most striking of all. Apera dimidia n. sp. Externat Cuaracters (Pl. VII, figs. 7-9; Pl. VIII, figs. 20, THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 205 21, 24).— Animal sometimes shghtly swollen towards the middle, tapering to a blunt point at the hind end. Back rounded, without keels. Outer lip of respiratory opening narrow, exposing the inner lip. Skin covered with a network of grooves, among which the main grooves are less conspicuous than in the other species of Apera. Dorsal grooves about 1 mm. apart (notwithstanding the small size of the species), usually separated by a double row of ruge, but united for about 3 mm. in front of the respiratory opening. Lateral grooves rather near the dorsal grooves, irregular, but sometimes extending as far as the head. Oblique sub-lateral grooves also irregular, seldom reaching the peripodial groove, and less numerous than in A. gibbonsi, being usually absent from the anterior third of the animal. Radial grooves fairly well marked dorsally, but like the sub-lateral grooves they cannot always be traced as far as the peripodial groove. The upper limits of the lowest rug oneach side often form an irregular longitudinal groove near the peripodial groove and parallel to it. Body conspicuously mottled with greyish-brown pigment, which is chiefly concentrated to form an irregular and some- what discontinuous dark band along each side of the animal. Between these bands is a much paler dorsal area only sparsely mottled, darkest around the anterior border of the respiratory opening and lightest at the edges next to the dark bands. Below these the skin is darker than it is dorsally, being more thickly mottled, especially next to the dark band, though the pigmentation extends down to the peripodial groove. Both dorsally and laterally there is a tendency for some of the pigment to be concentrated in the dermal grooves, thus emphasizing the reticulation. Small patches of the same pigment are sparsely scattered over the foot-sole. The ground-colour of specimens preserved in spirit 1s yellowish white. The following are the dimensions (in alcohol) of two specimens, in both of which the reproductive organs were fully developed: 206 HUGH WATSON. Length : 4 , 32 Damm: } Aoo sR: Breadth in middle ¢ 6°5 Breadth at respiratory Son = : VD opening . : bs 3) 4. (90 3) Breadth of foot-sole . Soe, Ar Dd. Greatest height . Os ae. Ooms Distance from respira- tory opening to hind end ; é Osa es. soy p AZZ Oss InrerNAL CHaracters.—S kin not quite so thick as in most of the species (Pl. XI, figs. 34, 35; Pl. XII, figs. 36-39; PI XID fies 240). Shell. (PE XI, ne 46) 3 x48 mms thin and translucent, convex; apex rounded, quite at the posterior end; sinus rather shallow. Pedal gland (PI. XIII, figs. 54, 59-62; Pl. XIV, fig. 66).— It has the glandular tissue entirely concentrated in the anterior half, which is very broad in consequence; fold of terminal vesicle not divided into narrow lamelle. Nervous System (Pl. XIV, figs. 68, 69; Pl. XV, figs. 71, 73, 75; Pl. XVI, figs. 80-86).—Cerebral ganglia with prominent accessory lobes; buccal gangla normally situated posteriorly, jomed to cerebral gangha by comparatively short connectives ; nerves to pedal ganglia large. Digestive System (Pl. XVI, figs. 88-90; Pl. XVII, fies. 103-107 ;* Pl. XVM figs: 115-1225) Pl KX ties Ie text-fig. 4,8, p.161; Pl. XXI, fig. 132; Pl. XXII, figs. 136, 137). —Radula 4°8 mm. long ; central tooth absent; about thirteen laterals with very short curved cusps, and the same number of marginals with longer narrower cusps, on each side; corresponding teeth not opposite to each other; formula: (25 + 0+ 26) x 61. Odontophore rather large, with a single buccal retractor. Cisophagus narrow and as long as the salivary ducts; crop swollen anteriorly ; loops of wide intestine nearly vertical, mainly on the left side of the posterior division of the liver. Salivary glands united above anterior end of crop. Anterior division of liver above crop and in front of intestine. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 207 Vascular System (PI. IX, fig. 29; Pl. XXII, fig. 140).— Pulmonary veins not very numerous; no “ posterior aorta ”’ ; two main arteries of liver arising from anterior aorta as a single vessel, posterior artery supplying hermaphrodite gland as well as posterior division of liver ; anterior aorta passing to right of intestine. Reproductive System (PI. XXIII, figs. 147, 151).— Hermaphrodite gland cblong, with moderately large follicles; hermaphrodite duct rather more convoluted and swollen in the middle than in the other species; common duct rather short, curving twice to the right and twice to the left; vagina broader than in A. gibbonsi; receptaculum seminis small ; receptacular duct much swollen anteriorly, the swollen part being attached to the adjacent body-wall by short transverse muscles on each side. Vas deferens becoming’ slightly swollen near the posterior half of the penis to form an epiphallus ; penis short, broadened posteriorly ; penial retractor also short. Hasrrar.—Port Shepstone, Natal (Burnup) ; Town Bush, Pietermaritzburg (ona mossy stone; Mrs. Warren) ; Equeeta, Natal (Burnup). Type.—The type is in the British Museum. Other specimens will be found in the Natal Museum, Pietermaritzburg. AFFINITIES, ETC.—This species may be distinguished at a glance from young specimens of any of the forms of Apera gibbonsi by its dark lateral bands, while the absence of keels separates it from the other species found in Natal. Internally it differs widely from A. gibbonsi and A. parvain its radula, buccal retractor, cesophagus, liver, receptacular duct, etc.; while in no other species is the pedal gland so broad in front and so slender behind. ‘This species was first discovered by Mr. Burnup in March, 1903. Apera purcelli Collinge. Apera purcelli Clige., Ann. 8. Afr. Mus., 1901, vol. ii, p. 230, pl. xiv, figs. 1,2; Connolly, Ann. S. Afr. Mus., 1912, vol. xi, p. 63. 208 HUGH WATSON. Exrernat Cuaracters (Pl. VIII, fig. 25; and ‘Ann. 8. Afr. Mus.,’ vol.ii, pl. xiv, figs. 1,2)—Animal rather broad anteriorly, tapering toa blunt point at the hind end. Back rounded, with- out keels. Outer lip of respiratory opening narrow, exposing the inner lip. Skin with well-marked reticulation, the rugee being rather large for the size of the slug. Dorsal grooves less than 4+ mm. apart, separated throughout their length by a single row of ruge, not umited in front of the respira- tory opening. Radial grooves well-developed, and extending down to the single peripodial groove, which is not very deep. The most anterior radial grooves branch from the dorsal erooves near the respiratory opening, instead of arising from the opening itself. There is not even a trace of lateral erooves, excepting perhaps for a very short distance; but irregular oblique grooves occur in front of the radial grooves, and these correspond to the sub-lateral grooves of other species. Body dark grey or bluish black, the pigment being chiefly concentrated in the dermal grooves and thus emphasizing the reticulation. On each side there is a suggestion of a slightly darker lateral band, the intervening dorsal area being a little paler, although it tends to be darker towards the centre than next to the obscure lateral bands. Below these the skin becomes much lighter, the pigment dying out completely before it reaches the peripodial groove. The foot and the ground-colour of the back is whitish yellow in the case of specimens preserved in spirit, but in living examples the foot is said to be ferruginous. According to Mr. Collinge the type (in alcohol) is 25 mm. long, and its foot-sole is 4 mm. in breadth. I have only seen a young specimen, of which the following are the measure- ments: Length . : . 16 mm. Greatest height . 3°5mm. Breadth in middle 3°75,, Distance from resp. Breadth at resp. opening to hind opening. woes end . : sora 9 Breadth of foot-sole d) THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 209 InrerNAL CuHaracters.—Skin thick. Shell (Pl. XIII, figs. 47, 48) 1-2 x *6 mm. (in the immature specimen whose measurements are given above), thin and translucent, very convex ; apex rounded, overhanging the posterior margin to a considerable extent ; sinus shallow. Pedal gland (PI. XIII, fig. 55).—Glandular throughout its entire length. Nervous system.—The various ganglia more distinctly separate than in most species of Apera; buccal gangha situated posteriorly, joimed to cerebral ganglia by compara- tively short connectives. Digestive System (Pl. XIIL fig. 55; text-fig. 4, c, p. 161). —Radula (of immature specimen) 1°8 mm. long; teeth less numerous than in most species; central tooth small; three or four laterals with very short cusps, and about four times as many marginals with longer narrower cusps, on each side ; formula (of immature specimen): (18+1+18) x41. Odonto- phore with single buccal retractor. Cisophagus narrow ; loops of intestine nearly vertical, mainly on the left side of the liver. Salivary glands more or less united. Anterior aorta passing to the right of the intestine, “nosterior aorta” probably absent. Penial retractor long, as in the next species. Hasirat.—Table Mountain, Cape Town (scarce; Lightfoot). ''ype.—The type is in the South African Museum, Cape Town. An immature specimen will be found in the Cambridge University Museum of Zoology. AFFINITIES, ETC.—This rare species is the only member of the genus that is known to occur in the west of the Cape Province. In some respects it resembles Apera dimidia more than any of the other species, but it differs from both A.dimidia and A. gibbonsi in having the dorsal grooves separate throughout their entire length, in the glandular tissue extending to the terminal vesicle of the pedal gland, and in the long penial retractor. In these features the present form resembles the keeled speciesof Apera. Unfor- tunately the reproductive organs of this species are practically 210 HUGH WAI''SON. unknown, and it is to be hoped that further specimens will be discovered so that these organs may be described. Apera burnupi Smith. Apera burnupi Smith, Ann. Mag. Nat. Hist. (6th ser.), 1892, vol. x, p. 466. Apera natalensis Collinge, Ann. 8. Afr. Mus., 1900, vol. ii, p. 3, pl. i, figs. 3, 4, pl. il, figs. 14, 15; Simroth, Naturwiss. Wochenschr., 1901, vol. xvii, p. 111, fig. 7; Simroth, Bronn’s Klass. u. Ordn. d. Tier- Reichs ITI, Gastr. Pulm., 1909, p. 143, fig. 42, pl. iv, figs. 9, 10; Collinge, Ann. Natal Mus., 1910, vol. ii, p. 167; Simroth, Bronn’s Klass. u. Ordn. d. Tier-Reichs III, Gastr. Pulm., 1912, p. 611; Connolly, Ann. 8. Afr. Mus., 1912, vol. xi, p. 65. EXTERNAL, CHaracters (Pl.-VIL, figs. 10; 11; > Pl. Vill fig. 26; and ‘Ann. 8. Afr. Mus.,’ vol. 11, pl. i, figs. 3, 4).— Animal broad and flattened towards the hind end, tapering anteriorly, with four longitudinal keels, two on each side of the body. Notwithstanding the two pairs of keels the animal is squarish in section, excepting posteriorly, as the lower keels are not far from the edges of the foot. Upper keels prominent though blunt, nearly parallel to each other in the middle of the body and separated by about seven-eighths of the breadth of the animal. Anteriorly they converge a little towards the head. In the posterior three- eighths of the slug the upper keels become very prominent, at first diverging and then curving round to meet in a very obtuse angle above the posterior extremity of the foot (but the angle is probably less obtuse in young than in adult specimens). They thus enclose a large flattened oval area, sloping downwards and backwards, and having the respira- tory opening slightly in front and to the right of its centre. Lower keels less prominent, twice as far from the upper keels as from the edges of the foot, extending for nearly three quarters of the length of the animal but not reaching the hind end. Foot rather broad, but tapering behind as in the other species, so that the posterior dorsal expansion of the body overhangs the edges of the foot on each side. Outer lip of respiratory opening rather narrow, exposing the THE CARNIVOROUS SLUGS OF SOUTH AFRICA. Dial inner lip. Skin coarsely reticulated, the rugee being especi- ally prominent on the anterior part of the back. Dorsal grooves very conspicuous, about 1 mm. apart (or rather more in the centre), separated by a row of rugee which is usually double in the middle of the body, converging as they approach the respiratory opening, but not uniting with each other. Lateral grooves deep but rather irregular, extending along each side of the body as far as the head, a little nearer the upper than the lower keels, the right one terminating in the genital opening. Below each lateral groove another more irregular groove branches from the most anterior radial groove and runs forward above the lower keel until it unites with the lateral groove on the side of the head. There are two or three irregular sub-lateral grooves on each side, and also three or four equally irregular supra-lateral grooves, parallel to the most anterior radial grooves and connecting the dorsal grooves with the lateral grooves on each side of the back. The upper keels are notched by these supra-lateral grooves as they cross them, but the keels are more deeply notched by the radial grooves, which are rather numerous and very conspicuous. Both radial and sub-lateral grooves terminate below in a longitudinal groove, parallel to the peri- podial groove, but separated from it by a narrow row ot ruge. In addition to the usual transverse grooves there is an obscure longitudinal groove on the foot-frmge, also parallel to the peripodial groove. Body mottled with greyish-brown patches of colour, which sometimes have a greenish-blue tinge. The mottling is chiefly concentrated on the sides of the body between the keels, but patches of colour also occur between the lower keels and the peripodial groove, and more sparsely on the dorsal area. The ground-colour of specimens preserved in spirit is usually pale yellow slightly tinged with reddish-brown dorsally ; but in living examples the skin is more deeply coloured, the back being reddish-brown, and the sides and foot-sole tinted with orange-red. The following are the dimensions (in alcohol) of the type DAP, HUGH WATSON. of A. natalensis according to Collinge, and of the still larger specimen belonging to the South African Museum shown in Pl. VII, fig. 10: Length . ; : : Oo) imine es < 7-7 onan Breadth between upper keels inmiddle. : : BELO a Bo he Breadth between upper keels behind respiratory opening . 13°5_,, 17 5b Breadth between lower keels inmiddle_. : : oo sllar pee Breadth of foot-sole 3 BS ne Oy tie ilies Greatest height (15 mm. in front of respiratory opening) = STATES. Distance from respiratory open- ing to hindend . ‘ oo Bagley a InrerRNAL CHaracrers.—Skin extremely thick, especially towards the hind end (Pl. IX, fig. 30). Shell (Pl. XIII, fig. 49) 5°5 x 3:2 mm., of moderate thickness, convex, with a conical apex projecting shghtly beyond the posterior margin ; sinus rather deep. Pedal gland (Pl. XIII, fig. 56).—Exceptionally large, its loops wrapping round the other organs; glandular tissue extending throughout its entire length. Nervous system.—Buccal ganglia situated anteriorly ; connectives short. Digestive System (Pl. XVII, figs. 108, 109; Pl. XX, fig. 128; text-fig. 4, p, p. 161; Pl. XXI, fig. 133).—Radula (of large specimen) 6mm. long; central tooth present, with a very short single cusp ; eleven laterals with short double cusps, and more than twice as many marginals with long, narrow, nearly straight cusps, on each side, the transition between laterals and marginals being abrupt ; formula (of large speci- men): (25+ 11 +14 11+ 25) x77. Odontophore small, with a single slender buccal retractor; odontophoral support bluntly pointed in front. (sophagus narrow and as long as the salivary ducts; crop cylindrical ; intestine narrow, first loop deep, posterior continuation on right side of liver. Salivary THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 213 glands separate. Liver of two approximately equal divisions, the intestine only traversing the right. Vascular System (Pl. IX, fig. 30)—Pulmonary veins forming a network. ‘Posterior aorta” supplying left division of liver and hermaphrodite gland, and sending a small branch to posterior end of right division of liver ; remainder of right division supplied by two arteries from the anterior aorta, which arise separately, but close together ; anterior aorta passing through loop of intestine. Reproductive system (Pl. XXIII, fig. 148).—Herma- phrodite gland composed of relatively smaller follicles than in A.dimidia; hermaphrodite duct convoluted; common duct very long, rather narrow, and much convoluted and twisted ; vagina bent at ght angles towards its anterior end; recepta- culum seminis small; receptacular duct greatly swollen towards the anterior end. Penis moderately long; anterior part narrow and more or less twisted, posterior part broad and appearing as though it were double owing to the presence of a longitudinal groove along one side; vas deferens dis- appearing into this groove about 5 mm. from the posterior end of the penis ; pemal retractor long and narrow. Hasrrat.—Chase Bush, Pietermaritzburg (Burnup) ; Rich- mond, Natal (nearly in the centre of a rotten log; Ward). Type.—The type of A. burnupi is inthe British Museum, of A. natalensis in the South African Museum, Cape Town. Another specimen will be found in the Natal Museum, Pieter- maritzbure. AFFINITIES, ETC.—This striking form differs widely from all the preceding species not only on account of its conspicuous keels, but also in the double cusps of the lateral teeth of the radula and the separate salivary glands. Apera sexangula n. sp. Apera burnupi “Smith”; Collinge, Ann. Mag. Nat. Hist. (6th ser.), 1897, vol. xx, p. 221, pl. v, figs. 1-6; Pilsbry, Nautilus, 1898, vol. xii, p. 12; Collinge, Ann. S. Afr. Mus., 1900, vol. ii, p. 4, pl. i, figs. 5, 6; Collinge, Journ. of Mal., 1901, vol. viii, p. 71, fig. 1; Collinge, Journ. 214, HUGH WATSON. of Mal., 1902, vol. ix, pl. vi, fig. 66; Simroth, Bronn’s Klass. u. Ordn. d. Tier-Reichs III, Gastr. Pulm., 1909, pl. iv, fig. 11; Collinge, Ann. Natal Mus., 1910, vol. ii, p. 166; Simroth, Bronn’s Klass. u. Ordn. d. Tier-Reichs III, Gastr. Pulm., 1912, p. 611; Connolly, Ann. S. Afr. Mus., 1912, vol. xi, p. 62. ExrernaL Cuaractzers (Pl. VII, figs. 12, 13; and ‘Ann. 8. Afr. Mus., vol. 11, pl. 1, figs. 5 and 6).—Animal rather slender, approximately hexagonal in section, owing to the presence of four prominent longitudinal keels, two on each side of the body. Upper keels separated by about three-fifths of the breadth of the slug, and nearly parallel to each other, excepting pos- teriorly, where they diverge very slightly just in front of the respiratory opening, and then gradually converge behind it, until they unite to form a single median keel, 3 or 4 mm. in length, at the hind end of the animal. Lower keels extend- ing along the whole length of the slug, about half way between the upper keels and the edges of the foot, but sloping downwards towards the hind end. The areas between the keels are nearly flat when the animal is in motion, but when it contracts they become deeply concave. Body, as seen from above, tapering to an acute angle posteriorly, but when viewed from the side the hind end appears blunter owing to the short median keel (though not always so rounded as in fig. 13). Outer lip of respiratory opening narrow, exposing the inner lip. Rug minutely subdivided. Dorsal grooves well marked, usually about 1 mm. apart, separated by a row of rugz which is single in front and behind but often becomes irregularly double towards the middle, converging slightly as they approach the respiratory opening, but not uniting with each other. Lateral grooves also conspicuous, extending along each side of the body as far as the head about half way between the upper and lower keels, the right one ending in the genital opening. Sub-lateral grooves few and irregular, only two or three being traceable on each side. Radial grooves well marked, forming slight notches in the keels where they cross them. Both radial and sub-lateral grooves terminate below in a longitudinal groove, parallel to THE CARNIVOROUS SLUGS OF SOUTH AFRICA, Ole) the peripodial groove, but separated from it by a narrow row of ruge. Body mottled with greyish-brown pigment, excepting along the edges of the keels. ‘The colour tends to be more concen- trated towards the hind end, and along the sides of the animal just below the upper keels, although it extends down to the peripodial groove. As a rule, minute reddish-brown specks are also thickly scattered over the body, and these, unlike the greyer patches of colour, occur also on the foot- sole, though much less abundantly than on the back. The ground-colour of specimens preserved in spirit is yellowish- white, very shghtly tinged with dull red dorsally ; but in life the skin is more deeply stained with dull red pigment, the general colour of the living animal being usually chestnut or reddish-brown, paler on the keels and foot.! The following are the dimensions (in alcohol) of a small but mature specimen, whose reproductive organs are shown in Pl. XXIII, fig. 149, of a shghtly larger example in which the genital system is not quite fully developed, and of an unusually large specimen in the Natal Museum, of which Mr. Burnup has kindly given me the measurements. mm. mm. mm. Length . : , Se POa Ol 2a. OO Scher if Breadth between upper “Weise in middle . : J ASOD, or, yO WOR: Breadth between upper Grek at respiratory opening. : ay Ae ae ntay (Aipon iar) O°O Breadth between lower keels in middle . : , : Hel sor ye LOWES | 16 Breadth of foot- ae : : pote we a Aonts.. LOD Greatest height (in middle). ar ie 5 els: Distance from respiratory opening to hind end . : : : a a a 5 AK InrerNAL Cuaracters.—S kin thick, especially towards the hind end. Shell (Pl. XIII, figs. 50, 51) 4 x 2 mm., of 1 For a description of the living animal, see ‘Ann. Natal Mus.,’ 1910, vol. ii, pp. 166, 167. 216 HUGH WATSON. moderate thickness, convex, laterally compressed, with a prominent conical apex projecting beyond the posterior margin; sinus of moderate depth. Pedal gland (fig. 57)—Much smaller than in the last species, curving alternately to the right and to the left in a comparatively regular manner; glandular tissue extending throughout its entire length. Nervous system (Pl. XV, figs. 72, 74, 76).— With buccal evangha situated anteriorly; connectives short; nerves to pedal gland rather slender. Digestive: System (Pl. XVI, fig. 67; Pl. XVI, figs: 110-114; Text-fig. 4, u, F, p. 161; Pl. XXI, fig. 134, Pl. XXII, fig. 135).—Radula 4 or 5 mm. long; central tooth present, with a very short, broad, single cusp ; eight or nine laterals with short double cusps, and more than twice as many marginals with longer narrower cusps, on each side, the transition between laterals and marginals being rather abrupt; formula (of a Port Shepstone specimen): (23 +9 +1+9+4 23) x 63. Odontophore small, with a single slender buccal retractor. (sophagus narrow, shorter than the salivary ducts ; crop swollen between the salivary glands; intestine narrow, first loop deep, posterior continuation on right side of liver. Salivary glands separate. Liver of two approximately equal divisions, the intestine only traversing the right. Vascular System (PI. IX, fig. 31).—Pulmonary veins forming a network. ‘ Posterior aorta” supplying left division of liver and hermaphrodite gland; right division supphed by two arteries from the anterior aorta which arise at some distance from each other ; anterior aorta passing through loop of intestine. Reproductive System (Pl. XXIII, fig. 149). — Her- maphrodite gland consisting of a cluster of small follicles ; hermaphrodite duct only very shghtly convoluted ; common duct very long and much convoluted and twisted; vagina rather broad; receptaculum seminis rather small; receptacular duct fusiform, swollen towards the anterior end. Penis long and narrow excepting at the posterior end, where it is THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 217 broadened; end of vas deferens swollen to form an epiphallus about 2°5 mm. long, which bears anteriorly a minute flagellum; penial retractor long and narrow. Hasrrat.—Port Shepstone, Natal (Burnup) ; Hilton Road, near Pietermaritzburg (Burnup); Grahamstown, Cape of Good Hope (Farquhar, French). Type.—The type is in the British Museum. Other specimens will be found in the Natal Museum, Pietermaritzburg, in the South African Museum, Cape Town, and in the Academy of Natural Sciences, Philadelphia. AFFINITIES, ETC.—This species resembles Apera burnupi in many respects, and was at one time confused with that species. Nevertheless, the two forms can be very easily dis- tinguished by their external characters alone, for in the present species the upper keels meet at an acute angle behind, where they form a short median keel, and the lower keels are equally prominent and extend the whole way to the hind end of the animal. The colour of the two forms is also different. Internally the most conspicuous features in which the present species differs from the last are to be found in the pedal gland and the reproductive organs. Hitherto this species has usually been known as ‘“‘A pera burnupi,” but an examination of the type-specimen of A. burnupi leaves no doubt that that name should be apphed to the preceding species, which Collinge subsequently named A. natalensis, and it has therefore been necessary to give a new name to the present form. THE PHYLOGENY OF THE SPECIES OF APERA. It is evident that Apera burnupi and A. sexangula are closely related to each other. Of these it is probable that the former is the more highly specialised; for in the pedal oland, the reproductive organs, and the keels, A. sexangula seems to be rather more primitive than A. burnupi. A wide gap separates these two species from those without keels ; nevertheless, A. purcelli resembles A. burnupi and A. Dies HUGH WATSON. sexangula in not a few characters, notably the pedal gland, the long penial retractor, and the complete separation of the dorsal grooves. Moreover, it is probable that the ancestral form from which A. burnupiand A. sexangula have been evolved resembled A. purcelli still more closely, since the keels and the double cusps of the lateral teeth of these species cannot be regarded as primitive characters. Yet it is not likely that in A. purcelli we have the direct ancestor of these two species, for A. purcelli has in some respects become modified in a different direction, its central tooth having become smaller and the course of the intestine having become quite different from what we must regard as the more primitive arrangement found in A. burnupi, A. sex- angula,and A. gibbonsi. It is therefore probable that A. purcelli must be regarded as having diverged somewhat from the branch which gave rise to the keeled species. Apera dimidia in some respects resembles A. purcelli, but it is much more highly specialised. The glandular tissue of the pedal gland has become entirely concentrated around the anterior half of the duct; the central tooth of the radula has disappeared, and the laterals have become more numerous ; the penial retractor has become quite short; the dorsal grooves are united posteriorly: in fact, while A. dimidia has probably been derived from the same branch as A. purcelli, it has diverged very far from both A. purcelli and the keeled species. Apera gibbonsi and A. parva differ widely from the other species in their enormous radula and numerous radial buccal retractors, as well as in the cesophagus, receptacular duct, shell, coloration, etc., and apparently in the structure of the terminal vesicle of the pedal gland. In some respects they resemble A. dimidia most closely, especially in the absence of glandular tissue from the posterior end of the pedal gland, the short common duct, the union of the dorsal grooves in front of the respiratory opening, and the position of the lateral grooves; and it might be possible to regard A. gibbonsi and A. parva as being derived from the THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 219 same branch as A. dimidia, but having become still more highly specialised in many ways. On the other hand, we find that these species seem to be more primitive than either A. dimidia or A. purcelli not only in the uniformity of the teeth of the radula and the presence of a well-developed central tooth in typical examples of A. gibbonsi, but also in the arrangement of the intestine and liver. Now the nature of these characters renders it improbable that if they had once become so much modified they would revert to their original condition ; and I am therefore inclined to regard the resemblances between A. gibbonsi and A. parva on the one hand, and A. dimidia on the other, as due to parallel evolution. Apera gibbonsi and A. parva are probably even more closely related to each other than are A. burnupi and A. sexangula. But A. parva is decidedly more primitive than A. gibbonsi in its nervons system and long penial retractor, and in possessing a true stomach. On the other hand, in a few features, such as the buccal retractors and hermaphrodite gland, we must regard A. parva as being more highly specialised than A. gibbonsi or any other member of the genus. If these views be correct, the phylogeny of the species of A.gibbonsi A.burnupi A.dimidia A.parva A.sexangula Apera might possibly be somewhat as shown in the above genealogical tree. VOL. 3, PART 2. 16 bo bo i) HUGH WATSON. THE OCCURRENCE AT CAPE TOWN OF TESTACELLA MAUGEI FER. This well-known species! is the only carnivorous slug, in addition to those belonging to the genus Apera, which is known to occur in South Africa, for it is probable that Ceratoconchites schultzei (Simroth) is the larva of a fly? The genus T'estacella can be distinguished at a glance from all the other genera of slugs known to inhabit South Africa, because in this form alone the hind end of the back is covered by an external shell. In some respects the anatomy of Testacella resembles that of Apera, but it isso well known that it is unnecessary for me to describe it here. More than a hundred years ago Cuvier® gave an account of the internal organs of one member of the genus; and since then the re- searches of Gassies and Fischer, de Lacaze-Duthiers, Plate, Simroth, and others, have given us a fuller knowledge of the anatomy of Testacella than of almost any other carnivorous snail or slug. The followmg are the principal characters by which Testacella maugei can be distinguished from the other species of the same genus: (1) the comparatively large size of the shell, and its length and convexity; (2) the widely separated origin of the lateral grooves; (3) the presence of a central tooth in the radula; (4) the small number of the buccal retractors; (5) the unusual length of the tentacular retractors, and the fact that they both arise to the left of the middle line; (6) the swollen anterior end of the receptacular 1 See Appendix for the more important references to Testacella maugei. 2 Simroth, H., ‘Zool. Anz. Leipzig,’ 1907, vol. xxxi, pp. 794, ete. ; Simroth, H., ‘Deutsche Siidpolar Exped.,’ 1910, vol. xii, Zool. iv, p. 172; Bottger, O., ‘Abhandl. Senckenb. Naturf. Ges. Frankfurt,’ 1910, vol. XXxxii, p. 433. 3* Ann. Mus. d’Hist. Nat.,’ 1804, vol. v, pp. 435-444, pl. xxix, figs. 6-11. THE CARNIVOROUS SLUGS OF SOUTH AFRICA, 221. duct, its comparatively great length, and its lateral junction with the receptaculum seminis; (7) the convolution of that part of the vas deferens which lies next to the free oviduct ; (8) the length of the narrow anterior part of the penis, and its broader posterior end, without any flagellum. It will be seen from Pl. XXIV that the specimens of Testacella from Cape Town which I have examined possess all these characters, and I have, therefore, no hesitation in assigning them to T’. maugei. Nevertheless the South African examples of this genus have usually been named “‘l, aurigaster Layard.”' Now, although Major Connolly® believes that Layard published a description of this supposed species, I have not been able to find any such description, nor any type-specimens, and I cannot therefore state definitely that T. aurigaster is a synonym of T. maugel, as it is conceivable that more than one species of Testacella may have found its way to Cape Town during the last sixty years. At the same time, all the specimens that I have seen labelled “'T’. aurigaster” have proved to belong to T. maugei, and I think that T. aurigaster may be safely expunged from the list of South African Mollusca. The slug occurs in gardens at Cape Town, and I have little doubt that it has been introduced into South Africa by man, notwithstanding Dr. Simroth’s arguments in favour of the possibility of Testacella being a native of that country. | believe that the natural distribution of the genus Testacella is limited to the western part of the Palearctic region, from Great Britain and Hungary to the Canary Islands, and that ' The following is a copy of a manuscript note written by Layard himself, for which I am indebted to Major Connolly: “ Testacella aurigaster Layard. I only found this shelled siug in the Gardens round Cape Town. It was common in the grounds of the South African Museum, which was built at the lower end of the Botanical Garden in Cape Town. The belly of the animal was a rich golden yellow, hence the name selected. It devours large worms, & will afsimilate individuals far larger than itself !” 2 “Ann. S. Afr. Mus.,’ 1912, vol. xi, p. 64. * ‘Zool. Anz. Leipzig,’ 1907, vol. xxxi, p. 796. 999) HUGH WATSON. its occurrence in Philadelphia, Cape Town, and Auckland is due to the agency of man. Testacella maugei occurs not only in the south-west of England, but also in Portugal, Madeira, and the Canaries; and it seems to me that it might easily have been transported to South Africa from any of these localities, amongst the roots of plants. For the species of Testacella are especially common in the rich soil of gardens where worms are abundant; and it is known that under adverse conditions a specimen can surround itself with a sort of cocoon of hardened slime which protects it from drought, and may remain dormant in this state for several weeks. It will be seen from Pl. XXIV, fig. 159, that in some of the South African specimens the posterior end of the penis is curved round in a pecular manner, and the retractor muscle is attached to the vas deferens, a short distance in front of it.? Possibly this unusual arrangement may have been partly caused by the way in which the animal contracted when it was killed, for it only occurred in some shrivelled specimens, which, in other respects, seemed to be identical with the rest. Figs. 157 and 158 represent two of the shells labelled “«T. aurigaster” in the MacAndrew collection at Cam- bridge. The first shows the usual shape of the shell in T. maugei, while the second may be assigned to the variety aperta Taylor, a form which was originally described from the Azores. THE AFFINITIES OF APERA AND TESTACELLA. THe Evontution oF CARNIVOROUS CHARACTERS. Both Apera and Testacella undoubtedly belong to the group of the Stylommatophora comprising the carnivorous 1 The junction of the retractor muscle with the vas deferens is also shown in Webb’s figure of the genital organs of T. maugei (‘ Journ. of Malac.,’ 1897, vol. vi, pl. vi, fig. 3). 2 *Monog. L. and F.-W. Mollusca Brit. Isles,’ 1902, vol. ii, p. 24, pl. i, fig. 15. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 223 snails and slugs, a group which is usually known as the Agnatha, since very few of these forms possess a jaw. But to state this is to say very little about the affinities of these genera, for it is doubtful whether the various carnivorous snails are at all closely related to one another, some authorities asserting that the tribe Agnatha is polyphyletic. Certain eminent malacologists have held that the Stylom- matophora may be divided into two groups, one mainly herbivorous, the other mainly carnivorous; and they would account for the diversity of the carnivorous forms by supposing that they had evolved in a manner parallel to the gnathophorous families, which show a similar diversity. This view was well expressed by Fischer, who, writing so long ago as 1873, said: “Je suis persuadé qu’on pourra con- stituer, dans quelques années, une série d’Agnathes ou Testa- cellidee ayant une grande extension, et dont les genres seront représentatifs de ceux des Helicidze, ainsi que, chez les Mammui- féres, certains groupes des Marsupiaux représentent d’autres groupes des Monodelphes.”’! On the other hand, many modern authorities maintain that the various families of carnivorous snails and slugs have been evolved independently from different families of the Gnathophora, and that they are in reality no more nearly related to one another than are the Huropean and ‘Tasmanian wolves. ‘The characters which the agnathous families undoubtedly possess in common they would explain as being due to convergence, brought about by their common acquisition of carnivorous habits.» Now there is no a priori objection to this second view, for we know that many gnathophorous forms occasionally devour animal food, and there seems to be no reason why this should not have become the usual diet in some cases. ‘he question, therefore, can only be settled by a careful examination of the comparative anatomy of the various families included in the 1+ Journ. de Conchyl.,’ vol. xxi, p. 12. * See an interesting paper by Dr. Simroth (‘ Naturwiss. Wochenschr.,’ 1901, vol. xvii, pp. 109-114, 121-127, 137-140), in which he has ably advocated the polyphyletic origin of the carnivorous snails and slugs. 224 HUGH WATSON. Agnatha. If we find that two agnathous families only resemble each other in those features which are likely to have been acquired through carnivorous habits, and are very similar to different gnathophorous families in their other characters, then we may assume that they are probably of different origin ; but if we find that the resemblances between two families of agnathous snails cannot all be explained in this way, and that these families are in their most important features more similar to each other than to any families of the Gnathophora, the probability is that they are closely related. It is evident, however, that before we can discuss the affinities of any particular genus, we must have a clear idea as to how a snail is likely to become modified if it adopts carnivorous habits, and which of the features characteristic of Apera, Testacella, and the other genera included in the Agnatha, are hkely to be due to their animal food. Snails and slugs find their food chiefly by means of their sense of smell, and one might expect this sense to be especi- ally well developed in the carnivorous forms, because animals which move slowly enough for snails to catch them must be very much more difficult to find than plants. Now im many of the Agnatha, and more especially in the Rhytididez and the Oleacinide, the olfactory organs at the tips of the upper tentacles are so large that the eye comes to occupy a position some distance behind the extremity.' Moreover Plate has shown that Testacella, unlike most of the Stylommatophora, retains a pallial olfactory organ. The unusual development of these sense-organs might be expected to lead to a corresponding development and concentration of the sensory nerve-centres, and accordingly we find that in nearly all the carnivorous forms the cerebral gangha are large and close together, Phrixolestes being perhaps the most notable exception. Carnivorous snails and slugs prey chiefly upon the herbi- ' See Strebel, H., ‘Beitrag z. Kenntn. d. Fauna Mexikan. L.- u. Siiss- 5 wasser-Conchyl.,’ 1878, vol. iii, pl. xv, fig. le; Suter, H., ‘Journ. of Mal.,’ 1899, vol. vii, pl. iii, fig. la. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 225 vorous forms and on worms, for most arthropods move too quickly and are too well protected by their chitinous exo- skeleton to fall a prey to snails. Now worms burrow in the soil, and snails hide themselves in crevices and retire deeply within their shells when attacked; it is therefore evident that a carnivorous snail will find the presence of a bulky un- yielding shell on its back a great inconvenience when it is trying to get near its victim. In order to obviate this diffi- culty the shell has become modified in various ways. In the first place we find that in Paryphanta it has become more or less flexible owing to the degeneration of the inner cal- careous layer. Secondly, the shape of the shell has become altered in many of the carnivorous genera. In Diplom- phalus, for example, the shell has become greatly flattened. This will enable the animal to penetrate into crevices, but it is obvious that the breadth of the shell will have to be reduced as well as its height, if the snail is to crawl into narrow holes. Now the only way in which both the height and breadth of a depressed or heliciform shell can be reduced is by the curvature of the axis or columella, until its direction corresponds more nearly with that in which the animal moves; and this is what has occurred in that remark- able genus of carnivorous snails, Streptaxis. If, however, the shell has a raised spire, the columella naturally takes up a position parallel to the direction in which the animal moves, and in this case it is only necessary for the height of the spire to be increased in order to reduce the diameter of the shell. This is what has taken place in the large genus Hnnea, in which only the young form retains a comparatively low spire. In Diaphora this principle is carried to an extreme, some species having shells with greatly produced spires composed of as many as twenty whorls. Indeed, so long does the spire become that the animal can no longer occupy the whole of it, and secretes a new internal wall cutting off part of the upper whorls, which may become decollated, as in D. telescopium Mlldff. A more efficient method of reducing the diameter of the shell (but one 226 HUGH WATSON. which probably necessitates a greater change in the structure of the animal) is for the whorls themselves to become laterally compressed. This has taken place to some extent in Strepto- stele and Obeliscella among the Streptaxide, but it is especially characteristic of the Oleacinide. In some members of this family, such as Streptostyla gracilis Pilsbry, the whorls have become so narrow that the form of the shell closely resembles that of the marine carnivorous genus Conus—a striking example of convergence due to the acquisition of similar habits. Thirdly, the position of the shell may be altered, so that the part of the animal in front of the shell becomes lengthened. This is admirably shown in Hedley’s figure of Rhytida lampra (Pfr.),! but it occurs to some extent in most of the carnivorous genera. A variety of this modification 1s found in such forms as Ennea densecostu- lata Mlldff. and Diaphora eutrachela Mlldff., in which the greater part of the shell has moved further back by the elongation of the last whorl towards the mouth. Fourthly, the shell may not only be moved backwards, but it may become reduced and flattened, until it either disappears altogether as in Selenochlamys, or sinks into the skin as in Apera. No projection will then remain to impede the animal’s progress when it follows worms into their burrows, and it is more especially in the vermivorous genera that this extreme modification seems to have occurred. We see, then, that the acquisition of carnivorous habits may lead to a great alteration in the structure, form, position, or size of the shell. The movement of the shell to the hind end of the body, and its gradual degeneration, will greatly modify the animal itself. Many of the organs will be subject to partial detorsion. The lung and respiratory orifice will move back with the shell and mantle, so that the animal will become opisthopneumic ; and it is probable that it is a direct advantage to a carnivorous form to have its respiratory opening near the hind end of the back, because this is the part which is least hkely to be 1 * Proce. Linn. Soc. N.S.W.’ (2nd ser.), 1891, vol. vi, pl. iii, fig. 3. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 227 covered when the greater portion of the body is under the surface of the ground or inside its victim’s shell. ‘The anus will move back with the respiratory opening, and the rectum will thus be directed posteriorly (text-fig. 5, cf. cand pD, p. 183). With the degeneration of the shell the palhal organs will become modified, and the posterior position of the pulmonary veins may pull round the auricle until it hes posterior to the ventricle, as in Testacella and a few other genera (text- fig. 6, p. 244). The organs which formerly occupied the visceral hump will become reduced in size and pushed forward into the general body-cavity. The pressure of these will retard the backward growth of other organs such as the receptaculum seminis, and it is therefore not surprising to find that the slugs have as a rule shorter receptacular ducts than the snails to which they are probably most nearly allied. The columellar muscle will disappear with the degeneration of the shell, and the various retractors that originally arose from it will become attached to different parts of the skin, which will become thickened to take the place of the shell (see p. 126). And as the anterior part of the body becomes lengthened, the origins of the tentacular and penial retractors will tend to move forwards. ‘Thus the mere fact that animal food is less easily accessible than vegetable food may lead to profound changes in the anatomy of a snail which becomes carnivorous. When the snail has approached its victim it will require to feel exactly where to attack it with its radula, and will, if possible, try to prevent the animal escaping until the teeth are fixed in its flesh. Accordingly we find that nearly all the carnivorous genera have a pair of special feelers at each side of the mouth beneath the lower tentacles. In Huglandina these feelers are very long and are supposed to be used for prehension, but according to F. M. Woodward! their function in Natalina caffra (er.) is probably purely tactile. Miss Davies* has recently discovered that in Paryphanta com- 1 ‘Proc. Mal. Soc.,’ 1895, vol. i, p. 271. 2 * Proc. Roy. Soc. Victoria,’ 1913, vol. xxv, p. 225. 228 HUGH WATSON. pacta Coe & Hedley and P. atramentaria (Shuttl.) these papille are glandular in structure. Now in Peripatus, an arthropod genus whose habits are not unlike those of the carnivorous snails and slugs, there are two large slime-glands which open on the oral papilla, and their secretion is used to entangle the prey. The carnivorous slug Atopus also possesses two large glands opening one at each side of the mouth; and Simroth,! who first discovered them, has suggested that their function may be similar to that of the slime-glands of Peripatus. We have already seen that the pedal gland is developed to an unusual extent among the carnivorous genera. Now this gland also opens close to the mouth, and I would suggest that its secretion may play some part in over- powering the prey. The form of the pedal gland in Apera has come to resemble in some respects that of Simroth’s glands in Atopus, and nearly twenty years ago André ®* suggested that the secretion of the pedal gland in the Stylommatophora might be used in feeding as well as in locomotion. We may at least say this: that the remarkable development of the pedal gland in the Agnatha, as well as the presence of the labial feelers, may not improbably be due to their predaceous habits. It is unnecessary to attempt to prove that the acquisition of carnivorous habits might lead to a modification of the radula and the surrounding structures: obviously these would be among the first organs to be affected. It will be sufficient, therefore, if I point out the chief ways in which the radula has become transformed. In the first place, it has increased in size, and especially in length, and become capable of far greater protrusion than in the herbivorous forms. Secondly, the individual teeth have become much larger, especially those occupying an analogous position in the radula to that held by the canine teeth in the jaw of the Mammalia; and to make room for these, the central teeth and those at the extreme edges of the radula have become vestigial or have 1 *Naturwiss. Wochenschr., 1901, vol. xvii, p. 122. 2 «Revue Suisse de Zoologie,’ vol. ii, p. 532. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 229 entirely disappeared. Thirdly, the main cusps of the teeth have become far longer and more sharply pointed, so as to penetrate the victim’s skin, and the secondary cusps have completely gone, excepting in some of the more primi- tive genera of the Oleacinide. And as all the teeth have become simplified in this manner, the differentiation between laterals and marginals has been lost, though it has been re- acquired in most of the species of A pera and Natalina, and to some extent inGuestieria. Fourthly, the bases have also become lengthened in order to prevent the teeth from being bent outwards. Fifthly, the rows of teeth have become sharply angled in the centre, owing to the way in which the radula is forcibly drawn back over the anterior edge of the odontophoral support into the narrow radula-sac. The muscles of the odontophore have become correspond- ingly enlarged. Inthe more primitive forms the hind end of the radula-sac still projects from between these muscles ; but in those that are more completely adapted to a carnivorous existence the odontophoral muscles have entirely surrounded the radula-sac, and form a very large cylindrical structure. This growth of the odontophore leaves little room for the development of other bulky organsin the anterior part of the body, especially as it is an advantage to a carnivorous form not to have a very large head and neck. I therefore suggest that this is the reason why the genital ducts of carnivorous snails and slugs are without large accessory organs. There would be no room for the development of a dart-sac, for example, in a snail with a very large odontophore. The growth of the odontophore would also tend to increase the size of the body-whorl of the shell. If this were laterally flattened as in the Oleacinida, it is evident that when the snail withdrew into its shell, the hind end of the odontophore would come to press against the outer side of the shell in the neighbourhood of the posterior end of the mantle-cavity. The result of this would be that the posterior limit of the cavity would be pushed forwards in the middle, and with it the heart and the lower end of the kidney. The upper end 230 HUGH WATSON. of the kidney, being near the suture of the shell, would remain in its original position ; and accordingly we find that in the Oleacinide the kidney is‘obliquely lengthened in a very characteristic manner. The cerebral, buccal, and ventral ganglia would be pushed further apart by the growth of the buccal mass and odonto- phore, and thus we find that in many of the carnivorous forms the cerebro-buccal, cerebro-pedal, and cerebro-pleural con- nectives are unusually long (see pp. 141, 146). The extrinsic buccal retractors would become strongly developed at the same time as the intrinsic muscles of the odontophore. So long as they sprang from the columellar muscle, they would undergo little change beyond an increase in their thickness, although I have found that their anterior ends tend to split up mto a number of separate strands in the Rhytidide. But in those forms in which the shell is degenerate and the skin is taking its place, we frequently find that the retractors of the odontophore arise from a large area of the integument; as, for example, in Apera gibbons, Testacella haliotidea Drap., and the ‘l'rigonochla- ming (see pp. 172, 173). A jaw becomes superfluous in a snail or slug in which the ‘adula is protruded far beyond it, and Simroth! has shown that it would be a positive disadvantage in a vermivorous form, because if it bit off pieces from its prey, the latter would escape. Accordingly the jaw is absent or extremely degenerate in nearly all carnivorous genera excepting Plutonia, in which Simroth states that the sharp edge is covered by a softer downward growth. Asin other carnivorous animals, the digestive region of the alimentary canal becomes reduced in size. ‘This is most apparent in the slugs, for in these the pressure of the organs which formerly occupied the visceral hump tends to diminish anything in the body-cavity that is unnecessarily large. In Atopus and Apera the true stomach has almost completely disappeared, the crop passing straight into the intestine ; and 1 Op. cit. p. 118. THE CARNIVOROUS SLUGS OF SOUTH AFRICA, Dom in these forms, and possibly in others also, digestion takes place partially within the lobes of the liver. In nearly all the carnivorous genera the length of the intestine and rectum is more or less reduced. Finally, Simroth considers certain pecuhar connections between the genital ducts of a few of the carnivorous genera to be modifications to facilitate self-fertilisation, and he attri- butes this to the somewhat isolated life which a vermivorous slug is likely to lead. But I have not found these modifica- tions in the forms which I have examined, and it remains to be proved whether they are at all general among the Agnatha. Such are the changes which are likely to take place in the organisation of a snail when it acquires carnivorous habits. We are now in a position to discuss whether Apera, Testa- cella, and other predaceous forms, are really closely related to one another, or whether their resemblance is merely due to the fact that they have all become carnivorous. Tur PHyLogeny or APERA AND OTHER NAKED CARNIVOROUS Sues. I think that Simroth is the only author who has suggested that the genus Apera has been evolved directly from an herbivorous slug. As already mentioned, he put forward the theory that Apera might have been derived from the Janellide. But Plate’ has shown that the slugs belonging to this family differ from most of the Stylommatophora, not only in being without lower tentacles, but in other important characters, such as the respiratory tissue, which takes the form of numerous fine tubes radiating from the mantle-cavity and recalling the trachez of arthropods. Now A pera possesses none of these features. Moreover, it would be difficult to imagine a radula more unlikely to develop carnivorous characters than that of Janella, with its ' * Zool. Jahrb., 1898, vol. xi, pp. 193-280, pls. xii-xvii. 932, HUGH WATSON. myriads of minute multicuspid teeth. I have therefore no hesitation in saying that Apera is in no way allied to the Janellide. Nor do I know of any other herbivorous family from which Apera is at all likely to have been derived. And the fact that the carnivorous characters are so highly developed in A pera is against the theory that the genus has been directly evolved from any herbivorous form. It seems certain that the ancestors of Apera must have been carnivorous for a very long time; and it is not lkely that these ancestors would all die out without leaving any descendants excepting this single genus. ‘The question is whether we can find any carnivorous genus of slugs or snails resembling Apera in characters which are not hkely to have been developed independently through the common acquisition of predaceous habits. Collinge has already pointed out how improbable is the theory of P. and F. Sarasin that Apera is closely allied to Atopus.! This genus and the other members of the Rathouisiide differ widely from Apera and every other carnivorous form in a number of important characters, such as the wide separation of the male and female openings, the presence of Simroth’s glands, the structure of the foot, the very large mantle, and the structure of the liver; while they only resemble them in the radula, the absence of a jaw, and a few other poits connected with their carnivorous habits. I agree with Simroth in regarding the Rathouisiide as being more nearly related to the Veronicellidz than to any monotrematous carnivorous family ; indeed, I have little doubt that, with the exception of the Veronicellidz and the Onchidiide, no family of the Stylommatophora is less closely related to Apera than the Rathouisiide. Plutonia, a carnivorous slug found in the Azores, is perhaps less unhke Apera than is Atopus; but it differs from it in the mantle, the laterally compressed form of the body, the presence of a jaw, the absence of a penial retractor, ' For references, see pp. 111-113. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. Zoe the presence of a secondary ureter, and other characters. ! On the whole it is extremely improbable that Apera is related to Plutonia. It is possible that the latter genus may be allied to the Trigonochlamine, but Simroth has given weighty reasons in favour of its having been evolved from the species of the Vitrininz inhabiting the shores of the Atlantic, and not from any carnivorous group. The only other carnivorous forms without external shells which are known to science are the genera from ‘T'rans- caucasia placed in the T'rigonochlamine. One of these— Selenochlamys—bears a very striking resemblance to Apera, as will be seen from Simroth’s figures.” Almost the only external differences between Selenochlamys and Apera are that in the former genus the visible mantle is a little larger, the foot is more distinctly tripartite, and there is a median dorsal keel extending from the mantle to the hind end of the animal. ‘This last difference is the most conspicuous, and yet it is not greater than the difference between Apera sexangula and A. gibbonsi, dimidia, or purcell1. And when we turn to Simroth’s description and figures of the internal anatomy of Selenochlamys, we at once notice the similarity between the radial buccal retractors in this genus and those of Apera gibbonsi and A. parva. A closer inspection, however, reveals many differences. ‘The nerve gangha of Selenochlamysareall separate; the pedal gland is narrow, straight, and almost entirely embedded in the foot ; the vagina is extremely short; the mght tentacular retractor does not cross the pemis. Moreover there appears to be no shell, so that the morphology of the mantle must be quite different to that of Apera. Indeed, I think that there can be little doubt that we have in the superficial resemblance between Apera and Selenochlamys a remarkable instance of con- vergence due to the acquisition of similar carnivorous habits ; and that the only affinities of Selenochlamys are with the ' Simroth, H.,‘ Nova Acta Acad. Caes. Leop.-Carol. Germ. Nat. Cur.,’ 1891, vol. lvi, pp. 223-229. 2 * Festschrift Leuckarts,’ 1892, pl. vi. 234 HUGH WATSON. other members of the Trigonochlamine, which Simroth has shown to have probably been derived from the Parma- celline. This view is confirmed by the geographical distribution of the genera. We may now consider the carnivorous genera which still retain an external shell, for it seems certain that the shell of Apera was originally external. The New Zealand genus Schizo glossa resembles Apera in many ways, but it differs in its reproductive system, for in Schizoglossa the male organs are much reduced, and there is no receptaculum seminis.! These features alone render it improbable that Apera has been derived from Schizo- glossa. According to Murdoch ? the reproductive organs of Paryphanta busbyi (Gray) bear a considerable resem- blance to those of Schizoglossa, and I think that there can be little doubt that the latter genus has been evolved in New Zealand from that section of the Rhytidide to which Paryphanta busbyi belongs. Strebelia possesses a receptaculum seminis, but in this genus the penial retractor is attached to the vas deferens instead of to the penis itself. And, judging from Strebel’s figures, the salivary glands are united below the crop, and not above it as in Apera.® Moreover, it does not seem likely that a purely South African genus should have been evolved from a slug which is only found in Mexico. There remain the European carnivorous slugs, Daude- bardia and Testacella. Of these Daudebardia differs widely from Apera in its pedal gland, nervous system, reproductive organs, excretory system, etc. There can be no doubt that A pera has not been evolved from Daudebardia. On the other hand, Testacella has many points in common 1 Hedley, C., ‘Proc. Linn. Soc. N.S.W.’ (2nd ser.), 1893, vol. vii, p. 390, pl. ix, fig. 4, pl. x, fig. 9; Collinge, W. E., ‘Ann. Mag. Nat. Hist.’ (7th ser.), 1901, vol. vii, p. 72, pl. ii, fig. 30. 2 «Trans. N. Z. Inst.,’ 1903, vol. xxxv, pp. 260, 261, pl. xxvii, figs. 4, 5. 3 «Beitrag. z. Kenntn. d. Fauna Mexikan. L.-u. Siisswasser-Conchy- lien,’ 1878, vol. iii, pls. i, ii. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 935 with Apera; indeed, if we take into consideration both the internal and external characters, I do not think that any other genus resembles A pera so closely as does Testacella. The pedal gland of Testacella lies free in the body-cavity as it does in Apera. Both genera are opisthopneumic, and in both the primary ureter opens at the posterior end of the mantle-cavity. The reproductive organs of the two genera are on the whole very similar, and in both the right tentacular retractor crosses the penis. But if we compare the genera more closely we find that there are several small differences which it is not easy to explain away. In the first place, the auricle in Testacella has come to he directly behind the ventricle, whereas in Apera the heart has not been rotated so far. Yet Testacella cannot be derived from Apera, because it still retains an external shell. Secondly, the foot of Testacella contains numerous dermal mucous glands, while that of Apera has none. Thirdly, the pedal gland of Testacella has no terminal vesicle. Fourthly, the left parietal ganglion has not become fused with the abdominal ganglion in Testacella, but all the visceral ganglia remain separate. Fifthly, the arrangement of the odontophoral muscles, and the structure of the odontophoral support, and even the blood-supply of the odontophore, show surprising differences in the two genera, as we have already seen. And, lastly, there is the difference in the geographical distribution of the two genera. On the other hand, the resemblances are found to be largely of a negative character, if we except those which might be due to the common acquisition of vermivorous habits by both forms. Therefore it will be well to look further before we assume that Apera is related to Testacella. It is true that there are no other carnivorous genera in which the shell has become reduced, but it is possible that the nearest living allies of Apera may have quite large shells, for when once the presence of a shell is found to be a disadvantage, its degeneration probably takes place rather rapidly. Pilsbry has suggested that Apera may possibly be allied VOL. 3, PART 2. 17 =" »>O=> 236 HUGH WATSON. to the Rhytididez, and I am inclined to believe that Pilsbry is right. Beutler has examined the histology of the skin of Paryphanta, and has found that the foot is without dermal mucous glands, as it isin Apera. The pedal gland in the Rhytididz not only les freely in the body-cavity, but it is often somewhat contorted, and, as in Apera, it ends in a vesicle containing a broad fold. In Paryphanta, Natalina, and Rhytida capillacea (Fer.), the visceral ganglia are all separate, as in Testacella, but in Rhytida inequalis (Pfr.) Fischer has shown that the left parietal ganglion is united with the abdominal ganglion exactly as in Apera. I have already shown that the arrangement of the odontophoral muscles in at least one member of the Rhytidide is very similar to that found in Apera, and, judging from Beutler’s account, the structure of the odonto- phoral support is also similar. The reproductive organs of Rhytida and Paryphanta are on the whole very like those of Apera, and in these genera the right tentacular retractor crosses the penis. The members of the Rhytidide also resemble Apera in having no secondary ureter. Indeed, almost the only differences that I know of between Rhytida and Apera are those which would be likely to be brought about by the degeneration of the shell and its further retrogression to the hind end of the animal, and we have seen that this is a modification which is especially lable to occur in carnivorous forms. Another fact in favour of the theory that Apera has been evolved from the Rhytidide is to be found in the geographical distribution of that family, for it occurs not only in the Australian region, but also in South Africa itself. It is true that Apera has almost certainly not been evolved from those members of the Rhytidide which are now found in South Africa: these have retained their separate visceral ganglia, but have become specialised in another direction, as is shown both by their radula and their reproductive system, in which the penis lies to the right of the tentacular retractors. But the presence of these snails in Africa renders it not at all HE CARNIVOROUS SLUGS OF SOUTH AFRICA. 237 improbable that forms more like Rhytida inequalis may also have once extended into that region, and given rise, by the degeneration of the shell, to Apera, just as in New Zealand another branch of the family has probably given rise to Schizoglossa. It is not unlikely that the Rhytidide originated in early Mesozoic times, or perhaps even before the end of the Paleo- Zoic era, in Gondwanaland—that great Southern continent which is supposed to have extended from Australia and New Zealand across the Indian Ocean, through Africa, and even as far as South America.! ‘Beitrag z. Kenntn. d. Fauna Mexikan. L.-u. Siisswasser-Conchylien,’ 1878, vol. ili, pl. xix, fig. 1. * * Journ. de Conchyl.,’ 1853, vol. iv, pp. 16-29, pl. i, fig. 1. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 943 than in most genera of the Oleacinide excepting Stre- belia, but I have already shown that this is what we might expect in a genus with a degenerate shell. The other female organs seem to be identical. The penis often ends in a cecal diverticulum in the Oleacinide, which, however, is quite short in Poiretia, and absent in Huglandina, Salasi- ella, Streptostyla, and Strebelia. In Testacella we frequently find a vestige of this diverticulum, especially in T. haliotidea Drap. In the Oleacinide an epiphallus is usually, though not invariably, inserted between the penis and the attachment of the penial retractor; this is also the case in 'l'estacella haliotidea, and sometimes in T. maugei (Pl. XXIV, fig. 159). Pfeffer has shown that in Euglandina liebmanni (Pfr.) the retractor springs from the apex of a flagellum, exactly as it does in Testacella haliotidea, and the same author has demonstrated that the internal structure of the penis of Huglandina is not unlike that of Testacella.! Indeed, the similarity between Testacella and the Oleacinidzw seems to me to be even closer than the similarity between Apera and the Rhyti- dide. The fact that the heart in Testacella is on the right side of the kidney, with the auricle directly behind the ventricle, affords further evidence in favour of this view, as will be seen from the accompanying diagram. The usual position of the heart in Hyalinia, Rhytida, and similar genera, is shown in text-fig. 6,a—drawn from a specimen of Hyalinia draparnaudi (Beck). In Daudebardia, according to Plate, the heart is turned slightly to the right as shown in diagram B; but it will be seen that the auricle is still in front of the ventricle, and the kidney is still mainly on the nght side of the heart, though the shell no longer extends far beyond these organs. In Apera the heart has rotated further in the same direction and is not protected by the shell (diagram c); the auricle has now come to le further back than the ventricle, and the kidney is behind and to the 1 * Jahrb. d. Deutsch. Mal. Gesell.,’ 1878, vol. v, p. 81. 2AA, HUGH WATSON. TEXT-FIG. 6. of Diagrams illustrating the rotation of the heart in the evolution of carnivorous slugs from snails. A. Hyalinia, Rhytida, ete. B. Daudebardia. c. Apera. D. Euglandina, ete. E. Strebelia. F. Testacella THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 245 left of the heart. ‘Turning now to the Oleacinide, we find that in Huglandina, Streptostyla, etc., the heart already occupies a position similar to that which it holds in Apera, as is evident from diagram p.' This is probably due partly to the pushing forwards of the ventricle owing to the pressure of the hinder portion of the odontophore,? and partly to the shape and consequent orientation of the shell. Diagram 4 shows the position of the heart and kidney in Strebelia, so far as I can judge from Strebel’s figures, and it will be seen that with the reduction in the size of the shell and its retro- gression to the hind end of the animal, the heart has rotated « little further. It is but a short step from this to the con- dition shown in diagram Ff, whick illustrates the position of the heart in Testacella, with the auricle directly behind the ventricle. Sixty years ago Reymond expressed the opinion that Testacella was “a Glandina with a rudimentary shell,” and if we substitute the word “ vestigial” for “ rudi- mentary,’ I believe that Reymond expressed the truth. At the same time I am aware that the most eminent modern authorities have expressed a contrary opinion. Dr. Pilsbry + states emphatically that the relationships of the Huropean carnivorous slugs (including T'estacella) are with the Aulacopoda, and not with the other carnivorous families ; and Simroth entertains the same view. The only facts which Pilsbry brings forward in support of this hypothesis are (1) that the European carnivorous slugs have lateral and pedal grooves, which he states that the other families do not possess, and (2) that the cerebral ganglia are generally united by a rather long commissure in the European slugs, while in the other families the cerebral ganglia are in close contact. Let us examine these points. In the first place, Jateral grooves occur not only in T’esta- ' See also Strebel, op. cit., vol. iii, pl. v, fig. 5. See p. 229. 3 Cf. Naef, A., ‘Ergebn. d. Fortschr. d. Zool.,’ 1911, vol. iii, p. 131, fig. 19. + * Manual of Conch.’ (2nd ser.), 1908, vol. xix, p. vili. 246 HUGH WaTSON. cella, Daudebardia, and the Trigonochlamine, but some- times also in the Oleacinide, Rhytidide, and Aperide. I have already described them in Apera, where they may be very conspicuous (Pl. VII, figs. 11 and 13). Collinge! shows them in his figures of Schizoglossa novoseelan- dica (Pfr.), and they are present on both sides in Natalina quekettiana (M. & P.), though rather ill-defined. In Kuglandina venezuelensis (Preston) I have also found both lateral grooves, but the right groove is more conspicuous than the left. In these genera, however, the right lateral groove ends in the genital opening, whereas in Daude- bardia it passes below the genital opening and unites with the peripodial groove.? Now, Testacella differs from Daudebardia in this respect, and agrees with the other families of carnivorous snails.® In Daudebardia there is a deep peripodial groove, cutting off a broad foot-fringe, which is cleft at the hind end by a short groove probably representing the lost mucous pore.* A peripodial groove is also present in the Aperide and Streptaxide, and although it is often absent or only shehtly developed in the Rhytidide and Oleacinide, Reymond states that it is fairly deep in Poiretia. In these forms, however, there is no trace of a caudal mucous pore. This is also the case in Testacella, which possesses a peripodial groove shallower than that of Daudebardia and nearer the edge of the foot.’ Lastly, Daudebardia resembles the Trigonochlamine, Limacine, and related groups, in having two conspicuous longitudinal grooves on the foot-sole, approximately parallel to each other. As a rule no such grooves are found in the other families of agnathous snails and slugs, but we have seen that occasionally similar grooves occur towards the hind end 1 “Ann. Mag. Nat. Hist.’ (7th ser.), 1901, vol. vii, pl. ii, figs. 26, 27. 2 Plate, L. H., ‘Zool. Jahrb.,’ 1891, vol. iv, pl. xxxii, fig. 1. 3 See de Lacaze-Duthiers, ‘Arch. Zool. Expér.,’ 1887, vol. v, pl. xxx, pelle 4 Wiegmann, F., ‘Mitt. Zool. Samml. Mus. Berlin,’ 1898, vol. i, p. 62. » Compare Plate’s figs. 3 and 11 on pl. xxxii, op. cit. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 247 of the foot in Apera and Natalina; they are, however, very inconstant, and instead of being parallel they diverge for- wards and seldom extend for more than half the length of the animal. Authorities differ with regard to the foot-sole of Testacella: Plate states that longitudinal grooves are present ; Taylor! says that they are absent; Pfeffer® found them only in the anterior two-thirds of the foot-sole ; Simroth ® saw traces of them towards the hind end of the foot in a few individuals only. I have examined the foot-sole in about a dozen examples of Testacella maugei from various localities, and I found that about half the specimens showed no trace of longitudinal grooves, while in the others the grooves were present, but they were much less conspicuous than in Daudebardia, etc., and diverged forwards from the hind end as in Apera. In other words, the grooves on the foot-sole of Testacella, instead of being constant and approximately parallel, as in the other European carnivorous slugs, are divergent and very inconstant. We see, then, that the evidence of the various dermal grooves of Testacella is against Pilsbry’s hypothesis instead of in favour of it. Turning now to the evidence afforded by the length of the cerebral commissure, we find that while in Daudebardia and the Trigonochlaminz the commissure is moderately long, in Testacella, as in the Oleacinide, Streptaxide, Rhytidide, and Aperide, the commissure is quite short, and the cerebral gangla are in consequence close together. This fact was clearly stated by Strebel* thirty-six years ago, and has been abundantly confirmed by subsequent investiga- tions. It is therefore evident that the only features upon which Pilsbry seems to base his theory of the affinities of Testacella are found on examination to support the contrary hypothesis. ' *Monog. L. and F.-W. Mollusca Brit. Isles,’ 1902, vol. ii, p. 2. OV eCib.. ps, 60. 3 * Nova Acta Acad. Ces. Leop.-Carol. Germ. Nat. Cur.,’ 1891, vol. lvi, p. 241. 4 Op. cit.; vol. iii, p. 9. 248 HUGH WATSON. If the views which I have brought forward are correct, all the carnivorous genera of slugs, with the exception of those belonging to the Trigonochlamine and to the Rathouisi- id, have been derived independently from different snails by the parallel degeneration of the shell. This will be made clear from the following table : American Oleacinide —> Strebelia. European Oleacinide > Testacella. Southern Rhytidide —> Schizoglossa. Northern Rhytidide -. European Zonitine —> Daudebardia. Atlantic Vitrinine Transcaucasian Parmacelline Oriental Ditremata — > Rathouisiide. => Jai )ieneRy = Plutonia. > Trigonochlamine. But even if we admit that A pera has probably been evolved from the Rhytidide and Testacellafrom the Oleacinid e, the question of the affinities of these genera 1s not yet settled ; for Pilsbry! states that the Oleacinidw and Rhytidide are closely related to each other, and if this be the case Apera and Testacella might still have been derived from a common predaceous ancestor, instead of their carnivorous characters having been acquired independently. Now it is true that the Oleacinidz and the Rhytidide have many characters in common, but if we except those which would be likely to be caused by their carnivorous habits the remaining features are chiefly such as are possessed by nearly all the more primitive sigmurethrous Stylommato- phora. And there are at least two important differences between the families—their distribution and their shells. The difference in distribution, however, is less important than it appears, for, on the one hand, it is quite conceivable that the Oleacinide may have once inhabited tropical Africa, and, on the other hand, we have already seen that the Rhytididz may occur in tropical America. But the difference in the shells cannot be so easily explainedaway. In the Oleacinide the 1 Op. cit., vol. xix, p. xii. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 249 shell is elongate, with a pointed spire and laterally compressed whorls, and the columella is truncated or folded. In the Rhytidide, on the other hand, the shell is heliciform or depressed, with a very obtuse spire and laterally expanded whorls, and the columella is neither truncated nor folded. I fail to see how it is possible to derive the shell of the Oleacinidee directly from that of the Rhytidide or vice versa. If these families are really related, it can only be indirectly, and we must postulate a large number of inter- vening forms, with shells intermediate in shape. But it might be asked whether we do not find such a series of intermediate forms in the Streptaxidw; and as a matter of fact in this family we have every gradation from heliciform shells such as Artemon and Imperturbatia to cylindrical shells like Ennea, and from these to pointed shells with laterally compressed whorls such as Streptostele and Obeliscella. May it not be that the Streptaxide have been derived from the Rhytidide and the Oleacinidz from the Streptaxidw? In my opinion the Streptaxidx may possibly have been derived from the Rhytidide, for I have recently dissected a form which proves to be in some ways intermediate between the two families. The great majority of the Streptaxidz, however, have come to differ widely from both the Rhytididz and the Oleacinide in their nervous system, their reproductive organs, and even in their radula; and I think that there can be no doubt at all that the Oleacinidz have not been derived from that family. Therefore, if Pilsbry’s view is correct, we must sup- pose that all the forms intermediate between the Rhytididz and the Oleacinidz have died out completely, which does not seem a probable hypothesis. Dr. Simroth has suggested that the Oleacinide may have been derived from the Achatinide. Now I regard the striking resemblance between the shell of the American genus Kuglandina and the African genus Achatina as almost certainly due to convergence; because Euglandina, with its long labial papille and closely aggregated nerve- 250 HUGH WATSON. ganglia, is probably one of the most recently evolved genera of the Oleacinidee. Indeed, its distribution suggests that it may not have arisen until after the final separation of the West Indian Islands from the maimland. I think that most zoologists who have studied the Oleacinidz will agree that the most primitive genera in the family are those with long spires, namely Spiraxis, Pseudosubulina, and Vari- cella. These genera are very unlike Achatina, but they - TEXT-FIG. 7. A. Curvella caloraphe Preston, Brit. E. Africa. B. Varicella nemorensis Ads., Jamaica. Representative teeth from the radula. x 400. are exceedingly similar to the more primitive members of the Achatinidw, if we include the Stenogyrine and Cceliaxine in that family. It is, in fact, almost impossible to say whether some groups of species should be placed in the Stenogyrine or Oleacinide until their radula has been examined. JI believe, therefore, that the Oleacinide have been derived from gnathophorous snails closely allied to the Stenogyrine or even belonging to that subfamily. The chief differences between the internal anatomy of the Stenogyrine and that of the Oleacinide are to be found in the radula. We know, however, that the Stenogyrine THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 2 are sometimes predaceous,! and that their teeth are lable to alter in form and acquire carnivorous characters. This is well shown in text-fig. 7, 4, which represents part of the radula of an African member of the Stenogyrine; and Crosse and Fischer? have figured the radula of an American species— Leptinaria lamellata (Pot. & Mich.) —in which the outer lateral teeth have become similarly lengthened. Moreover, we find that the radula of the more primitive genera of the Oleacinide is not of the specialised carnivorous type found in the higher members of the family and in the Rhytidide. Thus in Varicella the radula is usually quite small, the rows of teeth are nearly straight, the bases of the teeth are broader and shorter than in the usual carnivorous type, and there is a small additional cusp outside the main cusp (text-fig. 7, B). This last character is especially significant, as the small cusp is obviously a vestigial ectocone, such as we find in so many herbivorous genera. Iam indebted to the Rev. Prof. Gwatkin for kindly allowing me to examine the mounted radule of a large number of carnivorous forms, and I find that this additional cusp occurs in all the species of Varicella represented in his collection, namely, V.nemo- rensis Ads., phillipsi Ads., dissimilis Pilsbry (= simi- lis Ads.,) and venusta Ads., although it is extremely minute in the last species. It is therefore very surprising that it has been entirely overlooked by previous observers, who have figured the radule of V. phillipsi and V. nemorensis. Unfortunately I have not had an opportunity of studying the radula of Spiraxis, a genus which may be even more primitive than Varicella; but I have examined that of Pseudosubulina lirifera (Morelet), and find that in this form the second cusp is also present and is larger than in Varicella, although in other ways the radula has become more highly specialised. Strebel has shown, however, that in Pseudosubulina there is a vestigial jaw, the structure of which is not unlike that of the jaw in the Achatinide. 1 Johnson, C. W., * Nautilus,’ 1900, vol. xiii, p. 117. ? * Mission scientifique au Mexique,’ pt. 7, 1877, pl. xxviii, figs. 8-10. VOL. 3, PART 2, 18 2 HUGH WATSON. In my opinion these facts are sufficient to prove that the Oleacinidez have not been evolved from the Rhytidide, but have arisen directly from a gnathophorous family. Perhaps it might be said that the Oleacinide cannot have been evolved from the Achatinidz, because in the latter family the central teeth of the radula are almost invariably much smaller than the laterals, whereas in the more primitive members of the Oleacinide the central teeth are sometimes nearly as large as those on each side of them, not- withstanding that the teeth in the middle of the radula tend to diminish in size among carnivorous genera (cf. figs. 7,A and B). But I do not maintain that the Oleacinide have been evolved from any of the recent genera of the Achatinide. It must be remembered that already in Upper Cretaceous and Eocene times the Oleacinide were represented by forms which can hardly be regarded as primitive; the family, therefore, cannot have arisen very much later than the Jurassic period. Now the small central teeth are not an absolutely constant feature of the Achatinidz even at the present day, and it is not improbable that i Mesozoic times the radula was still of the more generalised type found in the closely allied family Megas piride. Indeed, it is possible that in the Jurassic period the Achatinidz may not yet have definitely separated from the Megaspiride. Callio- nepion may perhaps be regarded as a descendant of a form intermediate between these two families, and in this genus the central teeth are nearly as large as the laterals. Moreover, the penis has a continuation in Callionepion, which reminds us of the similar structure found in Hug- landina liebmanni (Pfr.), and other members of the Oleacinide.! Lastly, the shells of the typical section of Spiraxis in some respects resemble the Megaspiride quite as much as the Stenogyrine, which also suggests that the Oleacinidz may have diverged from the Acha- 1 Pilsbry and Vanatta, ‘Proc. Acad. Nat. Sci. Phila.,’ 1899, pp. 371- 373, pl. xv, figs. 3, 8, and Pilsbry, ‘Man. of Conch.,’ 1904, vol. xvi, p. 178, pla xxxt, igs. 17,6. THE CARNIVOROUS SLUGS OF SOUTH AFRICA. 253 tinidz when this family was scarcely distinct from the Megaspiride. Nevertheless, until further anatomical investigations have been made, it is impossible to say exactly to which group now living the Oleacinidz are probably most nearly related. It is equally difficult to form an opinion as to the place of origin of the family, for both the Achatinide and the Megaspiride have an extremely wide distribution. Perhaps the Oleacinidz may have arisen near the southern or south-eastern shores of that continent which geologists believe to have stretched from Western Hurope to America during a large part of the Mesozoic era. If this were the case, the south-eastern expansion of the area of distribution would be prevented by the extensive “ Mediterranean Sea”’ which then existed ; the gradual growth of the North Atlantic would separate the Huropean from the American forms, and the subsequent incoming of the Glacial Period would restrict the northern distribution of the family. On the other hand, the Oleacinide may possibly have arisen further south, and have entered America by way of the old land-connection which probably extended from Africa to Brazil. The absence of the Oleacinide from the Ethiopian Region is a possible objection to this theory; but perhaps the family may have once extended into that area, and may have there been exterminated by the elongate Streptaxide — Ennea, Streptostele, and their allies—which would probably enter into competition with them, but appear to have arisen just too late to follow the heliciform Streptaxide across to the Neotropical Region. ‘The hmited distribution of the Olea- cinide in South America is, however, another objection to the theory of their southern origin, and the first view that I have given seems to me to be the more probable. Iam there- fore glad to see that Pilsbry now upholds the more northerly origin of the family,’ although seven years ago he thought that the probabilities favoured the hypothesis that the Olea- '* Rep. Princeton Univ. Exped. Patagonia 1896-1899,’ vol. iii, Zoél., 1911, p. 625. 254 HUGH WATSON. cinidez arose in the Brazil-African continent.! , og ei A. g. rubella. 63a yi --- rection u/ ¥ f 0.00.7 72x25. A.sexangula. Fert i ==== 1 od.m, Ss \ ; G . ned 79x100,4.6.rubella. 71x 32. A dimidia Wine > /* dimidia, 6 Apera: nervous syste Vi / Watson del Huth London Etc. Ann. Natal Mus. Vol. III. PI SOWA. Tentacle, x50 87, Mouth, x 60 90, x 118, Odontophoral Support. 89, x 118, Radula-Sac. Figs. 80-86, 88-90, 4. dimidia. Fig. 87, 4. sexangula. 88, Mouth, x 60 Photo. by Watson. Collotype. Apera : Sections of Tentacle and Buccal Mass. 7 4 L ‘ eo rw a a ~ Ann.Natal Mus.,Vol. Ill. O5. 96. A.g.rubella x a. ere a 4 104. A. dimidia « 6. lla. A.. buruupt x 4. Watson, del Apera: odontophore. A.sexangula x Gz. Huth, London. ts imo) ¥ Ann.Natal Mus Vol. Ill. PV, wa re pigsebet tty e est ie we. Lane kk od.s.- ze ---v.m---- sexs steesea st uw Ewa ro roaieUKe cere OO Tee Ay Ve OMe léa. Watson del. Huth, London, Apera dimidia: trans.sections through the odontophore. x55 Ann. Natal Mus. Vol. III. PA OD bi be + As | LY a * Pam tan cit g) \ ¥ \ N Ve : » \Y ( Hy 2 ANAT yt 4 N e : ; \ « J x cm SAN: , 4 AAG WA a . \ YN) \Y | \ t | Lp} Ay v4 Gy: A hy (; ‘ co van \) an HN Hl Y, Hy 121, A. gibbonsi lupata, x 30. \\ Apera : Radule. SS Ann. Natal Mus. Vol. III. 125, A. gibbonsi s.s., X 25. 127, A. dimidia, - - upi, x 126. wait Y 128, A. burn Photo by Watson. Apera: Radule. 1 oe 182.x3¥2.A.dimidia Watson, del. Apera: rll | . Asexangula. 133 x1%.A. burnupi. © } digestive system. Huth, Londen ‘5 P eo Ww ee es i. a ira eo Ms . ; S27 ue. 4 ae so ae - rs Py “te Fav Co ro Jee Ann. Natal Mus. Vol. III. JP, OA Mle Pa as hit 138, x 135, Salivary Duct. ie ee Ey . ia iter, we 4 hy 137, x 135, Intestine. S. ee os 5 oa : 4 139, x 119, Salivary Gland. 1405 x 19S Ventricle. 142, x 194, Papilla 141, x 135, in Penis. Kidney. Photo by Watson. Collotype Apera: Histology. Fig. 135, 4. sexangula ; 136, 137, 140, 141, A. dimidia ; 138, 139, 142, 4. g. rubella. Ann. Natal Mus Vol. Il. Yer.an —__ fog ~ — — - -pents \ ----cpiph. gen.atr-~~~ Ss 14:5 xe A.g.rubella. 14:4 x 2). A.g. Tubella. $ gl-- 150 «1200 Ag rubella Se 151 x 1200 A dimidia Vatson Huth, London Watson, del , ; Apera: reproductive organs 3 Ann Natal Mus.VoLIl. P1. XXIV. v.77 155 x33. hook Beek. oe \ , vas def; 161x215. 160 %4106. 1l6ax5. 5 Watson del. Huth,London. Testacella maugei; Cape Tow7. PARTHENOGENETIC THNDENCY IN MELANOCERA MENTPPE. 269 The Parthenogenetic Tendency in the Moth, Melanocera menippe (Westwood). By Ernest Warren, D.Sc.(Lond.). Crrrain attempts have been made recently at the Natal Museum to cross several species of Saturniid moths for the purpose of investigating the relationships of hybrids to the parents. It was hoped that if healthy hybrids could be raised it might be possible to cross them among themselves and to discover if any definite segregation of characters occurred. The experiments were effectually closed in connection with the species selected owing to the fact that crossing did not readily occur, and no healthy offspring were secured. Never- theless, results of some interest were obtained. It was found that in one of the species, Melanocera menippe (Westwood), there is a slight, but definite, tendency for parthenogenesis. In this species there occurred some indication of an increased parthenogenesis induced by the male of another species. In a second case the male of one species certainly appeared to influence the female of another, but it was not possible to judge whether this was induced partheno- genesis or incipient hybridism. Material.—In the Black Wattle (Acacia mollissima) plantations of Natal the moth Gynanisa maia (Klug) is at some seasons very abundant, and the caterpillar is capable of doing considerable damage to the trees. During last season, October, 1913, to March, 1914, the moth was very common in the environs of Pietermaritzburg, and Mr. H. F. Pentland, manager of the town wattle plantations, very kindly caused about 130 pupe to be collected for me from the bases of the trees. The pupe were obtained in the month of July. 270 ERNEST WARREN. IT am much indebted to Mr. C. B. Hardenberg, Government Entomologist, Natal, for a supply of pupe of Melanocera menippe bred at New Hanover, Natal, and for other specimens; also for kind assistance and co-operation. The larve of menippe are found on Ficus cordata Thb., and Mr. Hardenberg’s caterpillars, obtained from the eggs of several moths, were fed on the same plant. With a view to discover a potential wattle pest Mr. Hardenberg experi- mented with the wattle foliage and found that the older larvee would eat it, but that the newly hatched caterpillars were unable to do so. About fifty newly hatched larvee were supphed with leaves of the wattle; some refused to feed, while others nibbled a little, but all the individuals died in the course of several days. The change of environment due to domestication or to the shghtly different climatic conditions of Pietermaritzburg and New Hanover would appear to have, however, an influence on the larve in this connection, as the following experience indicates. On pairing a male and female menippe, obtained from the pupe bred by Mr. Hardenberg, fertile eggs were laid, and the larvee on hatching were mostly placed on the leaves of Ficus cordata. A few (some seven in fact) were placed on the leaves of the Black Wattle; but from Hardenbere’s experi- ment it was confidently expected that they would refuse to feed. On the contrary, however, those placed on the wattle grew much more quickly and appeared healthier than those on Ficus cordata. Of sixty-three larve placed on Ficus forty-four gradually died in the course of three weeks, although they were supplied with both young and tender leaves, and with older leaves, and were carefully attended to. The remaining nineteen scarcely grew at all, and, as it was anticipated that all would die, the food was changed to wattle. In the next few days eight died, but the remaining eleven soon assumed a more healthy aspect and began to grow. At the present time (Nov., 1914) these eleven specimens are in a perfectly healthy condition, but not PARTHENOGENETIC TENDENCY IN MELANOCERA MENIPPE. 27] so large as the individuals which have been fed on wattle from the first. Of the seven specimens placed on the wattle at the time of hatching none have died. The experiment should be repeated, but the observations, as far as they go, show the adaptability of the larva and the ease with which the moth might become a wattle pest. In the following table is recorded the fate of the pupe of the wild specimens of maia and the bred specimens of menippe. | Males Females. | 3 1 1 ; c ie . O ts} i a lee | eee ue | Res | se Species. aa 3 | So 2 BS | 3 | oO i) ee pH | oS yr. | | pe) 4 Ct |} Sg |No.) & | cae Bo" No] & | oda "= Ihe a \\) ISG StS a OP che a) I = bow BH q | ‘a dou as qa A Sas ae a | S hem BHO | ia) Hod | nA | | AB | mot n> A if oom | | } | | | } : ese lvays | = i P| /Maia .| 128/60} 40! (an 3 68 38" 244 6 | . yr | 5 ry |Menippe .| 25/11) 10 | me elton eee a aan Orin v2.0 | | [ae See 1 eel | pee ae ad see bee OR 1 Three individuals greatly deformed. ? One individual fully formed in pupa-case. 3 One individual deformed. * Four individuals fully formed in pupa-case. > Fully formed in pupa-case. In comparing these two series it will be at once obvious that the bred specimens had been shielded from the adverse influences which proved fatal to a large percentage of the wild specimens. Out of 128 wild pupe of maia fifty (39 per cent.) failed to produce moths for one reason or another, while with the twenty-five bred pupz of menippe only one individual failed. On opening the pupz from which imagos did not emerge it was noticed that generally the contents had mostly dried, and it was clear that the development of the moth had not proceeded to any considerable extent. The contents were, as a rule, thoroughly impregnated with a fungus; but it is not known whether the fungus was the cause of the death of the pupa, or whether it appeared later. Quite possibly the cause of death was some bacterial disease. In any case in Natal the pup of Saturnid moths die in great numbers in this way, BNO ERNEST WARREN. and the percentage of deaths is sometimes much larger than in the case of maia cited above. In this connection a short digression may be allowable. Several years ago some forty full-grown, wild, and apparently healthy caterpillars of Nudaurelia belina were collected and placed in cages with earth. Pupation occurred in a normal manner, but only four moths were ultimately obtained. The majority of the pupe dried up without any appreciable development of the moth. The question as to the actual cause of death requires investigation, but the present point of interest to note is the great natural mortality which occurs, and therefore the absolute necessity for great reproductive power in the species. An analogous case was observed with the so-called Army Worm, which is the caterpillar of a Noctuid, Caradrina exigua. The caterpillar feeds on grasses, but itis moderately omnivorous, and will consume vegetables and other plants. In the middle of May, 1914, a plague of these caterpillars occurred in certain areas of Pietermaritzburg and its environs. From a piece of veld about half a mile from the town 140 caterpillars were collected from the grasses. These were mostly full-grown and were on the point of pupating. They were placed in large wooden boxes with a layer of soil at the bottom, and were supplied with food in case some of the caterpillars had not quite finished feeding. Care was taken that the food supplied was dry, as the bad effect of wet food is well known. Ninety-two of the caterpillars died without pupating, and became covered with a whitish mould; the remainder burrowed very superficially into the ground and duly pupated. Of these only nineteen emerged as moths, nine had been stung by a large dipteron parasite, and the remainder (twenty) of the pup dried up. Thus in this case only about 15 per cent. of the full-grown larvee produced moths. The only other available Saturniid moth in addition to the specimens of wild maia and bred menippe for use in these breeding experiments were four wild male specimens of Nudaurelia belina caught in Pietermaritzburg. PARTHENOGENETIC TENDENCY IN MELANOCERA MENIPPE. 273 To recapitulate, the moths used in these experiments included forty male and thirty-eight female Gynanisa mala, ten male and fourteen female Melanocera menippe and four male Nudaurelia belina. Pairing of the Moths.—The pup were all carefully sexed and buried under soil in muslin cages which were placed in a well ventilated shed. At the advent of the wet season the soil was occasionally damped. The imagos of maia began to emerge first (September 17th, 1914), and before any of those of menippe had appeared about fifty of the maia moths had become decrepid. ‘The last moth to emerge appeared on November 5th, and it was deformed. The menippe began to emerge on October 6th, and the last moth emerged on October 25th. Only fresh moths were used for pairing, and since it fre- quently happened that fresh individuals of opposite sex of the two species were not available at one and the same time the number of possible pairings was greatly reduced. The total number of pairings that could be effected were : (menippe @ x maia <), (menippe @ xX menippe 3d), c. 7 (maia @ x menippe a), ( ( i S maia ? x belina 2), menippe @ x belina 2), leaving 7 unpaired 2 menippe and 29 unpaired 3 Mala. The paired moths were placed in cubical muslin cages of a capacity of about 1+ cubic feet. A small branch of wattle and a glass pot of diluted honey were placed in each cage. It appears that the copulation of Saturnid moths occurs at night, and in no case was it observed in these experiments. From the account to be given below of the development of the egos laid by the females it would appear that copulation occurred in at least four of the pairs menippe ? X maia d, in the pair menippe ? xX menippe d, and in one of the pairs maia ¢ x belina d. Whether it occurred in any of the other pairings is more doubtful. fe} bed Unpaired Maia.—The unpaired maia females laid eggs 274 ERNEST WARREN. freely. These eggs usually began to shrivel in about twenty days. If cut open about fourteen days after laying, the contents were seen to consist of a fluid, greenish-yellow, homogeneous yolk. In about five weeks the yolk had con- tracted and dried, but as a rule it did not appear to decompose or turn black. The number of eggs laid by an unpaired female varied from about 100 to 200, with a mean of about 150. Kges laid by the unpaired moths were cut open daily and examined with a hand-lens magnifying 20 diameters, but no trace of development or segmentation could be detected. The unpaired female died in about eleven days, and the un- paired male in about seven days. Unpaired Menippe.—The behaviour of unpaired me- nippe females was markedly different. There was the greatest reluctance to lay eggs, and one individual died in about twelve days without layingany. The eggs of menippe are of somewhat larger size than those of maia, and the number produced by a female is about 130. The average number laid by unpaired menippe did not exceed thirty- two as against 150 of maja. The unpaired menippe remained phlegmatic and lived about as long as the paired moths. On dying, the abdomens were mostly filled with unlaid eggs. A number of these eges were opened and examined some five weeks after the death of theinsect. The egg-shells were thin and not so hard as those of the laid eggs. Generally the yolk was still fluid and apparently fresh and normal in character, but no trace of development could be detected. The contents of the laid egg were of a bright yellow colour, and in the majority of unfertilized eggs the yolk gradually dried, and the egg-shell became somewhat indented in the course of four or five weeks. The yolk, asa rule, did not appear to decompose in any way. On one occasion artificial fertilisation was attempted. The spermatophores removed from a fresh male maia were rubbed over the eggs removed from a recently emerged female menippe. In about three weeks the eggs began to shrivel and darken, and on opening PARTHENOGENETIC TENDENCY IN MELANOCERA MENIPPE. 279 them the yolk was found to be very fluid, blackish, and decomposed ; thus the condition of the yolk was in marked contrast with that of normally laid unfertilised eggs. The eggs obtained from six unpaired female menippe, after they had been laid for seven weeks, were cut open with scissors and examined with a strong hand-lens. Ina small percentage of the eggs distinct traces of development could be detected. The condition of the egg as seen by a lens may be indicated by numbered stages. Stage 1, when there is no obvious development. Stage 2, when there is some sign of a blastoderm. Stage 3, when there is an obvious embryo and a few hairs may be visible. Stage 4, when the development of the embryo is consider- ably advanced and many hairs are visible. Stage 5, when the embryo actually hatches. Ordinary fertilised eggs of menippe hatch in about three weeks after being laid. In the case of the pairing of menippe @ x menippe 6d, which occurred on October 22nd, the eggs hatched on November 16th. With regard to the eggs laid by unpaired female menip pe it was quite clear that the maximum development possible had occurred at the time of examination, since considerable desic- cation had taken place, and a few of the eggs were quite dry. In the accompanying table the details with respect to the eges of the unpaired menippe are given. | | | Bees laid by six specimens of unpaired menippe. | Per- Development. 5 = | Total. prises | No.1. | No.2. | No.3. | No.4. | No.5. | No.6. | | | | Stage I, | 32) 937 | 24) 25) 20) | “43: | 1st | 93°8 Stage 2. slip salle eel 0) 2 Seu can |e ak Stage 3. il AN iat | 0 0 1 4 271 Stage 4. 0 0 ) 0 0 0 0 ) Stage 5. 0 0 0 0 OF mee) 0 0 Pitotaleere reed ye B00 Geos) | oo a7 ©) 193 9/100 276 ERNEST WARREN. The four embryos in Stage 3 possessed a few unmistakable hairs. Thus in 193 eggs laid by six unpaired females 4°1 per cent. exhibited some sign of development, and 2°1 per cent. reached a further stage in the formation of an embryo. There is, therefore, a slight, but quite definite, partheno- genetic tendency in menippe. We have seen above in the case of maia that the unpaired moths laid eggs freely, and an average number of about 150 was produced. After the eggs had been laid some four weeks about 150 egos from a number of different moths were opened in addition to those examined daily, but no trace of development could be detected. Thus no parthenogenetic tendency was demonstrated in this species. Menippe ? x Maia d.—We now return to the paired moths (see list, p. 273, a), and the accompanying table gives the necessary details. Eggs laid by 5 paired females, menippe @ | mewsl | cams, Ok: |Percent- Percent-| evelop- | 3 vz = > a ad jage (in- age (ex- AMONG || Sas DMN, ment, | . y g | | No.3 | No.5 | No.6 | No.8 | No. 13 ee | arael | | (paired | (paired | (paired | (paired | (paired NESS ical | Oct. 9th).| Oct. 9th). Oct. 13th). Oct.23rd).|Oct. 18th). Stage | 45 | 84 | 128 | 37 | 101 | 395 | 613) 51-9) Stage 2 45 | 20 2 8 68 | 143 | 22:2 | 27-4 |Stage3) 17 27 0 5 10 | 59 | 92) 115 Stage 4| 12 30 | 0 3 UD |) eso | (Lies os Stage5; 0 | 1 0 0 0 Tey alae on | | = | ay = = | | : | | Totals.) 119 | 162 | 130 | 53 | 180 | 644 | 99:9 | 99°9 | | | | | i} On comparing the above two tables it will be seen at once that with the exception of the pair No. 6 there is very great increase in the degree with which development occurred, and also in the number of eggs in which such development was present. This is shown in the accompanying comparative table. PARTHENOGENETIC TENDENCY IN MELANOCERA MENIPPE. 27) | lle 9? menippe. | Total. | Stage 1.) Stage 2.| Stage 3.| Stage 4. Stage5 | | | ae: =. | i : | 2b | Eggs of 6 unpaired moths . 193 | 181 oe aye 0 | 0 | Eggs of 5 paired moths. | 644 | 395 = 143 | 59 | 4.6 il | It may be noticed that (1) the average number of eggs laid by the unpaired menippe was only thirty-two, but by the paired moths it was 129; (2) with unpaired moths the per- centages of eggs reaching the 2nd, 5rd, 4th, and 5th stages of development were 4-1 per cent., 2:1 per cent., O per cent., and O per cent. respectively ; while with the paired moths the percentages were about 22 per cent., 9 per cent., 7 per cent., and 0-1 per cent. respectively. On examining these two series it was believed at first that the eggs of the paired moths had been duly fertilised with the exception of those of No. 6, in which the amount of development exhibited by the eggs was almost negligible. It was expected that the single larva which hatched would exhibit on close examination characters intermediate between those of menippe and maja. The eggs remaining in the abdomen of some of the paired moths were examined five weeks after death. In No. 3 (see table, p. 276) sixteen eggs were found; one close to the external aperture contained an embryo in Stage 3, three were in Stage 2, and twelve showed no sign of development. ‘The twelve eggs were placed anteriorly in the abdomen. In No.5 three anteriorly placed eggs were found, and these showed no sign of development. In No. 15 two eggs were in Stage 2, and thirty-six showed no sign. The eggs which occurred in the middle and front regions of the abdomen were thin-shelled ; and in some cases the yolk was dry, but in an undecomposed condition, It is possible that in No. 3 the eggs which exhibited develop- ment had come into contact with the stimulating semen of the male in the body of the female, although they were not actually laid. 278 ERNEST WARREN. The Larva. case of No. 5, where one caterpillar hatched, the moths were paired on October-9th and the larva hatched on November 16th, thus the development occupied a period of thirty-eight days, while the time taken by a batch of pure-bred eggs was only twenty-five days, a differ- ence of thirteen days. Strictly speaking, about four days should be deducted from the periods of thirty-eight and twenty-five days as representing the time between pairing and the actual laying of the eggs. On hatching, the young larva was singularly lifeless and languid. It was carefully placed on the leaf of Ficus cordata, but during the first day it refused to feed. Sub- sequently it fed to a shght extent, but no appreciable growth occurred, and the creature died on November 26th, living just ten days altogether. The larva passed but very few fecal pellets, perhaps seven or eight altogether, and these did not appear to be normally black, but they were greenish brown and irregularly shaped. It would seem that the digestive functions were not normal. In general aspect the larva was indistinguishable from that of a pure-bred menippe, but on closer examination it was seen that it was not quite normal. Some of the hairs were less black than in the majority of young menippe larve that I have examined; also several of the tubercles bearing the long hairs of the body were yellowish and translucent instead of being black and opaque. The lateral, roundly triangular, black, raised areas at the extreme posterior end of the body were rather smaller than normal, and the one on the right hand side was only biack around the periphery, the central portion being pale. The posterior right thoracic leg was pale brown instead of dark brown or black. The median dorsal raised area over the anus, which bears tubercles and hairs, did not project posteriorly as far as usual, owing to its antero-posterior median axis being exceptionally short. In addition to these peculiarities the penultimate segment of the body and the segment in front were not separated off from one another on the right side of the body, and on this side PARTHENOGENETIC TENDENCY IN MELANOCERA MENIPPE. 279 where there should have been two lateral rows of tubercles there was only one. The outer lateral surface of the fourth right hand false foot was abnormally swollen and not marked off from the body in the usual manner, and the normal pigmented bar on the side of this foot was absent. Thus, this larva was asymmetrical, there beimg some deficiency in pigmentation, especially on the right hand side, and some deformity of the body on the same side. Comparison of Larve. It will be seen, however, that the differences between this larva and a normal larva of menippe do not appear to connect in any way the former with the caterpillar of mata. The larvee of some of the Saturnid moths are wonderfully alike at the first instar, but there are certain well-marked differences between the young caterpillars of maia and menippe. (1) In maia the general ground-colour is yellowish-brown, with a wide hoop-hke blackish band in the middle of the segment. In menippe the general colour is reddish-brown, with sometimes a faintly defined darker transverse band in the middle of the segment. In the present offspring the general coloration was like that of menippe, but it tended to be somewhat paler. (2) In maia the tubercles on the body are bright yellow, in menippe they are black. In the present offspring they were mostly black, but three were pale and two were black at the base and pale above. (3) In maia the hairs are yellow and translucent. In menippe there is some variation in this matter, more usually they are black, but some tend to be translucent and brown particularly on and around the head. In the present offspring a considerable number were translucent and light brown, especially towards the apex of the hair. (4) In maia the median dorsal posterior raised area above the anus carries large tubercles and a number of equidistant small tubercles along the posterior margin ; these latter carry one hair apiece. In menippe the area is relatively larger 280 ERNEST WARREN. and projects further back over the anus, the tubercles are smaller, and the little marginal tubercles are fewer and not regularly arrangedas in maia. In the present offspring the area was not normal in shape, and perhaps on the whole in general outline it was intermediate between that of maia and menippe. The condition of the tubercles and hairs, however, was in no way intermediate; it was rather peculiar, but obviously it was a modification of the menippe type and | had no connection with the maia arrangement. The fact that the arrangement of the tubercles and hairs in no way resembled that in maia renders it extremely probable that the more or less intermediate shape of the area was a quite accidental occurrence, due to the slight deformity of the posterior portion of the body. (5) The lateral raised area at the sides of the last false foot in the present offspring was like that of menippe and not hke that of maia in general outline and arrangement of hairs, but it was rather small. On the right side of the body, as already remarked, the area was yellowish except along its margin which was dark brown. (6) In maia the head is relatively narrower than in menippe. In the present offspring the head resembled that of menippe. (7) In maia the mean lengths of first, second, and third joints of the true walking legs are in the proportions of 100: 84:48, in menippe 100: 94:70; in the present offspring the proportions did not differ appreciably from those of menippe; they were 100: 105: 74, thus they showed no hint of a shortened terminal joint, which is characteristic of maia. (8) In maia the terminal claws of the thoracic legs are brighter in colour and considerably larger than in menippe, while in the present offspring they did not differ in the least from those in menippe. (9) In all other characters that were observed no differences could be detected between the present offspring and the young larvee of menippe. On the whole it may be said that no influence of the male PARTHENOGENETIC TENDENCY IN MELANOCKRA MENIPPE. 281 moth could be definitely detected in the offspring. ‘here was a certain deficiency of pigment, which, however, was doubtless associated with the unhealthy state of the larva, and was not due, as was at first thought, to the influence of the male. In all the more important characters the present larva was so completely menippe and not maia in character that it is extremely doubtful if the offspring can be regarded as a genuine hybrid. Pseudogamy.—We have already seen that the presence of the male maia had an undoubted influence on the female menippe; it caused the female to lay a good supply of eggs, and these exhibited a greatly increased power of development. The only obvious explanation of this rather remarkable result would appear to be that copulation of the moths actually took place; and that the semen from the male exerted a stimulat- ing action on the normal, weak parthenogenetic power without real fertilisation, or fusion of male and female pronuclei, occurring. This has been termed pseudogamy, and it is the explanation offered by Hans Przibram! in the case of the artificial application of the semen of Mantis religiosa to the egos of female Sphodromantis bioculata. Some of the eggs were rendered fertile, but the offspring appeared to be entirely like the mother, and there was no direct influence from the male. In the present example it is to be regretted that there is no cytological evidence to support this view, but careful consider- tion of the facts renders any other supposition still more difficult to accept. If the offspring was a true hybrid the almost complete absence of any of the male characteristics is totally unlike that which is found in the vast majority of the hybrids between distinct species, since these so usually tend to be more or less intermediate between the two parents. A possible explanation with respect to the absence of a re- enforced parthenogenetic tendency in the pair No. 6 (table, ' Przibram, Hans, ‘Experimental Zoologie, Phylogenese, p. 25, 1910. 282 ERNEST WARREN. p. 276) is that no copulation occurred, but that the close proximity of a male nevertheless influenced the female to lay a good supply of eggs. These eggs simply exhibited the normal amount of parthenogenesis inherent in the eges of this species. The parthenogenetic power, like any other character, would be expected to vary toa certain extent from individual to individual; but the laws of probability are against the supposition that the generally much increased parthenogenesis, and the freedom of egg-laying observable in the paired moths arose purely accidentally through the chance selection of such individuals, while none of the unpaired individuals possessed these characters in a high degree. It may be added that the further evidence that will be adduced from some of the other pairings is entirely against such an explanation. A few observations on the remaining pairings that were effected may be made here. Menippe 2? x Menippe d.—The moths (list, p. 273, 8) were paired on October 22nd, 1914, and 121 eggs were laid. The majority of the eggs were hatched on the morning of November 16th, the same day as that on which the partheno- genetic individual emerged. It was a damp, warm morning. The egos of one of the batches laid by the female did not duly hatch. In this batch there were forty-six eggs. These were opened and examined, and only one showed any sign of development. Of the remaining seventy-five eggs there were twelve irregularly scattered through the different batches which did not hatch. On opening them eight showed no sion of development and four contained well-formed dead embryos. The batch of forty-six eggs which did not hatch were pre- sumably laid before copulation, and the partial development which occurred in the case of one egg was due to the normal parthenogenesis. Maia ¢ x Menippe ¢.—In the case of these seven pair- ings (see list, p. 273, c) itis less clear that copulation occurred, and, if it did, the resulting effect was much less marked than in the reciprocal cross of menippe 2 x maia d. Whether PARTHENOGENETIC TENDENCY IN MELANOCERA MENIPPK. 283 paired or not the female of maia lays freely, but more eggs were, on the whole, laid by females placed with males. It was noticed that the eggs of one of the paired moths were very exceptionally slow in shrivelling, and on this account it was thought that possibly copulation had occurred and more or less normal development was taking place. There was very little shrivelling in five weeks after laying, although unfertilised eggs are obviously indented after a period of fourteen to twenty days. After the eggs had been laid seven weeks, shrivelling had become more marked, and seventy-five egos were opened and examined with a hand-lens. In eighteen egos a whitish granular lump or lumps on the periphery of the yolk could be seen. An examination with the microscope showed that these consisted chiefly of dense clusters of needle- shaped crystals. The crystals dissolved in 50 per cent. hydrochloric acid. They doubtless arose by the concentration of the salts in solution in the egg through evaporation, and their formation appears to depend largely on the slowness of evaporation. In rapidly drying eggs the crystals were not found. In 150 eges from several unpaired maia females only one egg was found containing such crystals. The microscopic examination of the contents of the egas which so successfully resisted desiccation revealed no definite sign of segmentation. Without much more investigation it cannot be said whether the formation of such crystals was in any way connected with an unsuccessful attempt at development after copulation. An ego stimulated by contact with the male fluid, even without actual fusion of nuclei, might be able conceivably to resist desiccation much better than an unstimulated ege. The power of resisting desiccation is undoubtedly a vital act not directly explainable in ordinary chemical and physical terms. With respect to the remaining six paired moths, about fifty egos from each—that is, about 300 in all—were examined, and only six exhibited any obvious crystal formation, and even in these it was less marked than in the case above described. These eggs began to shrivel in about twenty days. VOL. 3, PART 2. 20 284 ERNEST WARREN. The formation of such crystals also occurred in eggs laid by a female maia paired with a male belina, as will be seen in the account given below. Maia ? x belina ¢.—In one of the pairings a female maia was placed with a captured, wild, fresh-looking male belina (see list, p. 273, p). There is here more evidence that copulation occurred, and that the semen exerted an influence on the eggs of the female maia. The moths were placed together on October 16th, and the female soon began to lay very freely. Two hundred and one eggs were laid, and these were very slow in shrivelling. On November 10th some of these eggs were examined, and on December 3rd, when it was obvious that hatching would not take place, all the remaining eggs were opened. Altogether forty-eight (i.e. about 24 per cent.) showed with a hand-lens an opacity on one side of the yolk. Examination with the microscope proved that a certain amount of segmentation of the yolk had occurred; large rounded or polygonal cells crowded with yolk-globules were present. On treatment with weak methyl blue, a deeply staming blue spot became defined in each area in the course of a day, and there appears little doubt that this was the nucleus. There was not much difference in the condition of the eggs on November 10th and on December 3rd, except that on the latter date drying was taking place and death of the large yolk-cells was occurring. The nucleus had become less defined and was of inflated size and of irregular shape. Portions of the cells had apparently disintegrated into an opaque, white, finely granular substance. In some of the eggs clumps of needle-shaped crystals were found, and they were quite the same as those mentioned above, and occurred among, but apparently not inside, the yolk-containing cells. In the case of the second pairing, a female maia was placed with a very shabby, wild male belina. The male soon died. Fairly numerous eggs were laid, but these shrivelled ‘rather quickly and there was no sign of development. Doubtless the male was spent and no copulation took place. PARTHENOGENETIC TENDENCY IN MELANOCERA MENIPPR. 285 Menippe ? x belina ¢.—Only a single pair was avail- able (see list, p 273, 8). There was probably no copulation, or, at any rate, there was no effect; the moths died in the normal time, and no eggs were laid. Summary.—T'o summarise the general differences observed in eggs laid by unpaired moths and paired moths, the accom- panying table is instructive. | i ; (i ; : : ; | | as | : Maia 2 x] pete ee IM enippe® aS. Maia®? x |Maia 9 x|belina 6} Moths. Gees | xmaiad (9 indivi-|™enipped|belina ¢ |(oldshabby| auaiay: (6 pairs). |"quals), | (7 Pairs). | (1 pair). Geoae : pcx : : ae oR | [es | | No. of eggs | | | examined . 193 644 | 300 yhoo Ne ePADLOR I al) Percentage | _showing | some de- | | | | velopment 62 38°7 | 0 0 | 24°0 | 0) | | | Gerstaecker! gives a list of species of moths in which parthenogenesis has been observed : Fam. Sphingide — Sphinx, Smerinthus. Fam. Huprepiade — Euprepia. Fam. Saturniide — 8.-fam. Saturniine — Telea. Fam. Bombycide — Gastropacha, Bombyx. Fam. Liparide — Liparis, Orgyia. Fam. Psychide — Psyche. Fam. Tineidz — Solenobia. Melanocera menippe belongs to the family Saturn- iide, and the sub-family Sphingicampine, in which the larva pupates underground and does not form a silken cocoon as in the sub-family Saturniine. It is interesting to note that it is Just in these and in related families that hybridism between different species has been most often observed.” ' Bronn, H. G., ‘ Klass. u. Ord. d. Thier-Reichs, Arthropoda,’ vol. v, p. 166, 1866-79. 2 See Przibram, Hans, ‘ Experimental Zoologie-Phylogenese,’ pp. 41-51, 1910. 286 ERNEST WARREN. The view that a parthenogenetic tendency is favourable to hybridism is also suggested by the fact that hybridism appears to be especially easy in Hehinoderms,' and it is in this group that the parthenogenetic tendency is strongly marked and artificial parthenogenesis can be most easily effected. Pseudogamy, or the stimulation of partheno- gensis by the semen of the male, without real fertilisation, 1s, in the case of insects, a phenomenon based, as far as I am aware, on indirect evidence, and not on actual cytological observation ; although in Hchinoderms some evidence of this nature has been adduced. In the observations above described there is undoubtedly strong evidence that the females which were placed with the male of another species behaved differently than when un- paired, and the eggs were markedly affected. In the case of menippe ? x maia d, the single larva which hatched was so fundamentally similar to a normal menippe larva, that it would appear extremely improbable that it was a hybrid arising from real fertilisation. The only other alternative view which appears admissible is_ that copulation occurred and that a spermatozoon entered the egg and stimulated the normal parthenogenetic power without actual fusion of the male and female pronuclei, or, possibly, the mere contact of the male fluid acted as a_ sufficient stimulus. It has been stated, but on whose authority I have not ascertained, that painting the eggs of silkworms with different chemicals stimulates parthenogenesis, and if such is the case it could be readily understood that contact with the male fluid might have an analogous effect. In one of the pairings, maia ? x belina gd, the male appears to have had an undoubted influence on the eggs, but iain, without cytological evidence it is impossible to judge whether or not there was true fertilisation, since development did not proceed beyond early segmentation. 1 See Shearer, C., Morgan, Walter D., Fuchs, H. M., “On the Experimental Hybridization of Echinoids,” ‘Phil, Trans. Roy. Soc. Lond.,’ vol. eciv, 1914. PARTHENOGENETIC TENDENCY IN MELANOCERA MENIPPE. 287 Further experimentation with cytological observations among insects might be expected to lead to interesting results and to an increase in our knowledge of the inter- relationships of normal fertilisation, parthenogenesis and hybridism. NEW SOUTH AFRICAN ARACHNIDA. 289 New South African Arachnida. By John Hewitt, Director, Albany Museum, Grahamstown. With 9 Text-figures. Con'rENTS. PAGE INTRODUCTION . : F pe . 290 SysTEMATIC ACCOUNT : ; . 290 Orv. ARANEA : ; . 290 Fam. AGELENID Aj 3 . 290 MUIZENBERGIA g.n. . : ; . 290 Muizenbergia abrahami sp.n._. : . 291 Fam. ARGIOPID Ai : ; . 295 ERIGONOPSIS g. 1”. : : : . 299 Erigonopsis littoralis sp. n. ; : . 296 Fam. CTENIZIDA : 3 . 299 Pelmatorycter dreyeri sp. x. ; . 299 Bessia minor Hewitt : . 302 Bessia fossoria Poe. ; . 304 Spiroctenus armatus Hewitt : : . 3085 Acanthodon abrahami (Hewitt) . a0D Acanthodon ochreolum Poe. : . 306 Acanthodon hamiltoni Poe. : ad Acanthodon grandis sp. n. . 308 Idiops pretoriz (Poc.) ; 2 ald Idiops astutus sp. n. : . 310 Hermacha mazcna sp.n. . Pole HERMACHOLA g.n. . : : . 314 Hermachola grahami sp. n. ; : . ol4 Stasimopus steynsburgensis sp. 1. Jae Stasimopus gigas sp.w. . = . 318 Stasimopus minor sp.n. . : : . 320 Stasimopus oculatus Poe. . ; 2 Sak 290 JOHN HEWITY. PAGE Orv. SOLIFUG Ai : : . pias Chelypus hirsti sp. n. ; . 323 Orp. SCORPIONES .. ; ; ; : . 320 Opisthophthalmus pugnax . 325 Var. natalensis var. n. é . 320 INTRODUCTION. THe miscellaneous assembly of new Arachnids described in this paper are all in the Albany Museum, except when otherwise stated. Perhaps the most noteworthy are the small marine spiders taken by the Rev. N. Abraham at Muizenberg, near Cape Town (Muizenbergia abrahami and Hrigo- nopsis littoralis), a discovery of some interest seeing that previously only one marine genus, viz., Desis, had been known from South Africa. Whether their occurrence on our coast has any special zoo-geographical importance it 1s impos- sible to decide at present, as little or nothing is known of the Arachnid fauna of other shores; in any case it seems very probable that quite a new field awaits exploration on the coasts of the southern hemisphere, in worm-tubes and similar habitats between tide-marks. Some of the descriptions are based on material which has been kindly lent to me by the authorities of the British Museum, London, and of the Natal Museum, Pietermaritzburg ; for this courtesy I am especially indebted to Mr. A. 8. Hirst, the Arachnologist of the British Museum, and to Dr. E. Warren, the Director of the Natal Museum. SYSTEMATIC ACCOUNT. Orpen ARANEA. Fam. AGELENIDA#. Sub-fam. AGELENIN#. ’ Group Cryphecee. Genus MUIZENBERGIA gen. nov. This new genus is at once distinguished from any known genus of this group through the position of the tracheal NEW SOUTH AFRICAN ARACHNIDA. 29] stigma, which is situated about midway between the genital opening and the base of the inferior spinners. ‘The chelicera have two teeth on each side of the fang-groove. | According to Mn. Simon’s ! key to the genera, Muizenbergia should be placed near to Cicurina Menge. Muizenbergia abrahami sp. nov. Text-fig. 1, a—G. The types of this species, a single adult male and a some- what shrivelled female, were collected at Muizenberg, Cape Town, in September, 1913, by the Rev. N. Abraham, who presented them to the Albany Museum. Mr. Abraham found them when searching for Desis, in the interspaces of near the calcareous masses built up by marine annelids ; they are, therefore, marine spiders, probably resembling Desis in habit. Colour.—Cheliceree and anterior portion of carapace castaneous ; hinder portion of carapace pale brown, like the lees and sternum; distal segments of legs more deeply coloured than the basal segments ; on the cephalic portion of the carapace there is some slight infuscation, which is only strongly pronounced around the eyes. Abdomen uniformly dull dirty grey. Carapace.—Longer than wide, the cephalic region fairly strongly convex, gradually becoming much flatter in the thoracic region ; these two regions are not sharply separated though their junction is marked on the margin by an obtuse angle. Anteriorly the carapace is squarish, the clypeus being very wide and very low; the distance from an anterior lateral eye to the anterior margin of the carapace is less than the long diameter of the eye. Fovea long and very narrow, Ocular Area.—Nearly three times as broad as long. Anterior row straight, posterior row slightly procurved. Antero-medians smallest, but not very small, a little nearer to the antero-laterals than to each other. Eyes of posterior row subequally spaced. The laterals of each side very close i Simon, Eugene, ‘ Histoire Naturelle des Araignés, tome ii, p. 266, Paris, 1897. Trxt-FIc. 1. Muizenbergia abrahami sp. nov. x 40, A. Carapace of male. B. Tarsus of male palp with bulb in situ. x 75. c. Male palp with bulb of tarsus dislodged. x 55. pb. Process of the patella of the male palp. x 185. §. Posterior half of abdomen of male in ventral view showing spinners and tracheal stigma. x 40. x 25. F. Spinners of female. a. Ventral view of sternum, labium, cox of appendages, and one chelicera of the male. x 40. NEW SOUTH AFRICAN ARACHNIDA. 293 together. Quadrilateral formed by the four median eyes broader behind than in front, and the posterior breadth exceeds its length. Chelicerz.—Elongated, a little dilated in the basal half, projecting obliquely forwards. Fang strongly curved, reach- ing backwards nearly half the length of the chelicera; on each side of the fang-groove there are two teeth, of which the distal one is large, sharply pointed, and triangular, the more proximal one small. Pedipalp of Male.—Both the patella and tibia carry < strongly projecting process. That of the tibia, which is the longer, is a tapering cylindrical process, strongly curved backwards and ending in a point; that of the patella arises from a broad base and is expanded over half of its length, the expanded portion carrying a group of three bristles, distally to which the process suddenly narrows and even- tually ends in a claw-like hook. The tarsus is oval and elongated, but not twice as long as broad; the distal end of the bulb is not far from the apex of the tarsus. Numerous long, plumose, bristly hairs occur on the tarsus. ~ Legs.—All the hairs, bristles, and spines which invest the legs and the palps are finely plumose. None of the tarsi are spined. The first metatarsus and tibia in the female are quite without spines (in male specimen damaged). The second metatarsus has a pair of long spines, near the apex, and the third metatarsus has two pairs of long spines below ; the third and fourth tibize and fourth metatarsus are beset with a number of long and strong spines. The hairs on the legs are not very densely arranged. ‘The paired tarsal claws of the first leg in the male have nine teeth each, the more basal teeth being minute ; the unpaired claw has two teeth, but that of the female has only one. The unpaired claw of the fourth leg of the male has three teeth, that of the female only two. Labium.— About as broad as long; broadest near the base and tapering gradually towards the anterior margin, which is broadly rounded and carries four long bristles; it YO, JOHN HEWITT. reaches about as far as two-thirds of the distance along the maxillee. Sternum.—Only a trifle longer than wide, broadly truncate in front and narrowing to a point posteriorly between the hind coxe, which are completely separated thereby. It is broadest opposite the coxee of the second legs. A bdomen.—Hlongated, covered with soft, slender, plumose hairs, all backwardly directed. On its ventral surface the tracheal stigma is placed far forwards, being situated only very slightly posterior to the mid-point between the genital aperture and the bases of the inferior spinners.! Spinners.—Inferior spinners widely separated, about twice their own width apart. Superior spinners with the apical segment about half the length of the basal one. Colulus absent. The median and superior spinners arise only very shghtly posterior to the inferior spinners, especially in the male, where the whole six are arranged almost in a transverse line; in the female the arrangement is roughly in two transverse lines. Total length.—Adult male 3°5 mn. This genus in its spinner arrangement approaches the group Hahniine and will probably prove to be closely related to some of the little-known genera of that group. The tibial process of the male palp is very like that figured by Simon? for Hahnia oreophila H. Sim. The position of the tracheal stigma is also apparently an extreme condi- tion of a tendency obtaining in various genera of this group. IT have placed Muizenbergia provisionally in the group 1 Since writing the above, I have examined a large female example from St. James, near Cape Town (coll. J. H. Power). This clearly shows that the median tracheal stigma is double, the actual openings being connected by a short transverse furrow. The tufts of hairs shown on E, text-fig. 1, mark the positions of these openings. Further, the lateral eyes of each side are slightly raised on an obliquely disposed prominence. 972 Op. cit, py 2738: NEW SOUTH AFRICAN ARACHNIDA. 295 Cryphccex mainly on account of the dentition of the cheliceree; probably the separation of the two groups is artificial. Whether this genus can be regarded as a direct derivative of the South African terrestrial fauna, or whether, like Desis, its allies are widespread over the shores of the southern hemisphere, cannot be stated at present; it does not seem to be closely related to the various known httoral species recorded by Mr. H. R. Hogg.! Fam. ARGIOPIDA. Sub-fam. Linypuirn a. Genus HRIGONOPSIS ge. 10v. This generic name is proposed for the reception of the species hereafter described, which combines the characters of various members included in Simon’s world-wide group Hrigonex, but does not agree entirely with any one of them, and in some respects approaches the group Linyphiw. The characters of the male palp will probably prove sufficient to differentiate this from any other known genus. Other characters are: carapace of male simple, precisely like that of the female ; no trace of fovea or median sulcus on carapace ; length of quadrilateral formed by the four median eyes shghtly greater than its hind width; posterior median eyes very shghtly nearer to each other than to the posterior laterals ; sternum with a rather broad posterior prolongation, separating the bases of the fourth legs from each other, followed, posteriorly to the coxx, by a small flat quadrangular piece; abdomen without scute or indurated area; female epigyne comparatively simple; female palp without tarsal claws, the maxille not very broad even in the male (more or less intermediate between Hrigone and Linyphia); chelicere with four teeth on the anterior margin of the fang-groove and three on the posterior margin in both sexes; tarsus of first leg ! Hogg, H. R., “ Subantarctic Islands of New Zealand,” Article IX, Wellington, N.Z., 1909; “Some Falkland Island Spiders,” *‘ Proc. Zool. Soc.,’ 1913, p. 37. 296 JOHN HEWITT. very slightly longer than the metatarsus or subequal thereto, and the metatarsus very slightly shorter than the tibia. Erigonopsis littoralis sp. nov. Text-fig. 2, a—H. The types of this species are one adult male and several adult females taken at Muizenberg, near Cape Town, by the Rev. N. Abraham in September, 1914. The habitat is the same as that of Muizenbergia abraham sp. nov. Colour.—Carapace, sternum, and appendages brownish, without strong infuscation except around the eyes: the cara- pace and sternum are lightly infuscated, more especially so about the middle of the carapace. The skin of the abdomen is for the most part infuscated and is covered with black rather bristly hairs; on the hinder half of the abdomen are five pale transverse stripes dorsally, and there is a pair of pale spots about midway between the anterior stripe and the front margin of the abdomen. Carapace.—Narrowed anteriorly, the front margin broadly rounded ; cephalic area fairly strongly convex, gradually pass- ing into the shghtly convex thoracic portion, the boundary between the cephalic and thoracic portions of the carapace being very ill-defined both at the margin and elsewhere. Clypeus vertical but not greatly elevated, the anterior lateral eyes being about three diameters distant from the anterior margin of the carapace. Anterior margins of anterior row of eyes in a distinctly recurved line; posterior row only shghtly procurved ; anterior medians smallest, but not very small, nearer to each other than to the anterior laterals. Along the mid-line of the carapace in its cephalic portion there are four or five weak bristles, and a few stiffish hairs occur on the ocular area, but none on the margin of the carapace. Chelicere.—Fairly stout, directed vertically downwards, without spines at the sides and without basal spot; fang strong and rather short; anterior margin of fang-groove carrying four teeth, of which the middle ones are strongest ; the posterior margin with three small teeth. TExT-FIG. 2. hh ‘ i BS ih pray i { Hf | Erigonopsis littoralis sp. nov. A. Dorsal view of carapace and abdomen. x 30. B. Tarsal claws of first leg. x 375. ©. Barbed hairs on margin of fourth tarsus. x 185. D. Dentition of chelicere. x 100. £. Maxille, labium, and anterior part of sternum. x 90. F. Posterior part of sternum, cox of fourth pair of legs, and epigyne. x 90. G. Distal portion of male palp. x 200. H. Tibia of same showing the apophysis. x 250. 298 JOHN HEWITY. Legs.—Coxe I and IV of equal length. Metatarsus IV longer than the tarsus. Apart from a few long spiniform setee on the femora and tibie, the legs are spineless; there are a few long setiform spines on the distal segments of the palps. The legs are clothed with short stiff sete; most of these are simple, but on the tarsus, and to a less extent on the metatarsus, there are some long stiff sete, each carrying two or three lateral barbules, usually situated on one side of the hair about the middle of its length. On tarsus IV the stout feathered hairs are arranged in longitudinal rows, about ten inarow. The elongated so-called auditory hair on the fourth metatarsus is present. On tibia IV dorsally there are two projecting hairs, the distal one longer. Onychium present. Paired claws of the first tarsus with only five teeth, the distal one longest, the others much smaller and gradually decreasing in size towards the base; the median claw is fairly long and carries two teeth, the basal one of which is very minute and arises from the base of the larger one. On the claws of the fourth tarsus the teeth are still weaker, the inferior claw being drawn out toa lone fine point and carrying only one weak tooth. Sternum.—Broader than long, quite fused with the labium. Labium.—Very broad, carrying four bristles anteriorly ; there is a strongly thickened slightly concave border anteriorly. The maxille have the long axis very obliquely imclined. Palp.—Male palp short, the femur being about twice as long as the patella, and their conjoint length about equal to the distance from the base of the tibia to the end of the conductor of the style of the bulbal organ. The end of the conductor somewhat resembles a scorpion’s vesicle (sting) in shape ; immediately beyond the distal extremity of the tarsus the style is bent at right-angles on itself, its distal portion continuing to the apex as a broad slightly curved lamina. Tibial apophysis not bifurcated and not large. External branch of the tarsus (paracymbium) slightly arched, but no spines nor hairs occur on the concavity. Total length.—Female 2} mm. Male 1? mm. NEW SOUTH AFRICAN ARACHNIDA. 299 Various members of this group of Argiopidee are known to live in marshes and on the sides of streams, but no marine forms have been hitherto recorded. Fam. CTENIZIDA. Pelmatorycter dreyeri sp. nov. Text-fig. 3. The type consists of a single adult male from Bloemfontein, presented to the Albany Museum by Dr. T. F. Dreyer, in August, 1914. The species is related to P. nigriceps! Purcell, from Johannesberg, but differs therefrom in the TEXT-FIG. 3. oy oo hom oo Pelmatorycter dreyeri sp. nov. Eyes of male. x 45. ocular arrangement, in the dentition of the claws of the fourth leg, and in some minor points. Ocular area.—Wider behind than in front, the anterior row of eyes strongly procurved, the laterals much larger than the medians, the long diameter of the former being quite one and two-third times the diameter of the latter. Posterior laterals decidedly longer than the anterior medians, but shorter than the anterior laterals. Chelicere.—With seven teeth on the fang-groove. Pedipalps.—Near the base of the tibia on its inner side Is a spine. Legs.—Tarsus II, without spines; III with four spines on its anterior side, two or three postero-dorsally situated, and one on the posterior side ventrally situated; IV with no spines on its posterior side. Posterior claw of tarsus [V with 1 Purcell, W. F., ‘ Trans. South African Philosophical Society, vol. xi, p. 358. VOL. 3, PART 2. 21 300 JOHN HEWITY. one tooth only in the outer row and four in the inner row; anterior claw with three small ones in the outer row and four large ones in the inner row. Metatarsus I equal in length to the distance from the centre of the fovea to the anterior margin of the carapace. Patella IV without an apical spine inferiorly. The distal group of spines on femur IV, at its upper outer edge, is composed mostly of weak spines; there are only very few short, fairly stout, spines. Abdomen.—With a few, rather strong bristles near the base above. Posterior sternal sigilla.—Not very clearly defined, but apparently rather more than a length apart, and about half a length distant from the sternal margin. Posterior spinners.—Apical segment a trifle shorter than the penultimate segment. Measurements.—Total length 18°5 mm., length of cara- pace 5°5 mm., width of carapace 4-4 mm., length of metatarsus of first leg 3°5 mm. A female specimen taken at Bloemfontein by Dr. Dreyer, a few days later, presumably belongs to the same species. The characters are detailed below. Ocular area.—More than twice as broad as long, the posterior row distinctly wider; anterior row with anterior margins moderately procurved, posterior margins in a very hghtly procurved line, the medians slightly nearer to the laterals than to each other, their distance apart being distinctly greater than a diameter; posterior row with pro- curved anterior margin and recurved posterior margin, the laterals oblique and elongated, much longer than the medians and as long as the anterior laterals, posterior medians nearer to the posterior laterals than to the anterior medians ; distance between the laterals of each side about one-third the length of the posterior laterals. Posterior sternal sigilla.—About half a length apart and almost a length distant from the sternal margin. Chelicerz.—The inner dental series comprises eight teeth. NEW SOUTH AFRICAN ARACHNIDA. 301 Pedipalps.—Coxa with five teeth on the inferior surface situated far forwards. Tibia with four apical spines inferiorly, also four on the anterior surface below, but none on the posterior surface. Tarsus with two spines not far from the apex inferiorly, and a longer one on each side near the base. Legs.—tTarsi and metatarsi of first two pairs of legs scopulate to the base (on the second metatarsus the scopula is absent on the posterior side in its basal half). Tarsus I without spines, excepting one near the apex inferiorly ; Il with one or two spines inferiorly on the posterior side; III with a row of six spines postero-dorsally, of three antero- dorsally, also about seven antero-inferiorly situated; IV stronely aculeate on the anterior surface and with a number of spines on the posterior surface in its distal half. Meta- tarsus I with one or three apical spines below, and two or three along the inferior surface ; II with three apical spines below and two along the inferior surface ; II] numerously spined on both anterior and posterior surfaces superiorly, and with four weak spines on the inferior surface, in addition to those at the apex, which are long and strong; IV with numerous spines inferiorly, mostly situated on the anterior side, superiorly with two spines on the posterior side. Tibia I and II each without distinct spines below except one at the apex; III without spines at the apex inferiorly, a supero- anterior band of seven spines, also seven dorsal spines, and two near the supero-posterior edge distally ; IV with a single apical spine inferiorly, and with one spine or none on the postero-dorsal edge. Patella III covered with short, stout spines on the anterior surface, the dorsal surface with two stout spines near the posterior edge; IV without spinules on its anterior surface, except two or three immediately at the base of the segment. Femur IV with a dense group of short, strong spines anteriorly at the apex and dorsally. 'Tarsal claws of first leg with four teeth in the outer rows, and five or six in the inner rows; anterior claw of fourth leg with only one small distal tooth; the posterior claw with one or two distal teeth and one at the base, also a very minute 302 JOHN HEWI''. second basal tooth. Coxa III with a dense tuft of stiffish setee on the postero-ventral border. Colour.—For the most part castaneous above, chelicere blackish, patelle and more distal segments of first two pairs of legs and of palps pale brown ; carapace pallid in its hinder portion at the sides. Abdomen pale except over the median area above, where it is infuscated. Measurements.—Total length 23°5 mm., length of cara- pace 8°5 mm., breadth of same 5°5 mm. According to Dr. Dreyer, the nest has two long blind side passages leading into the central tube one on each side in the upper half of its length ; the upper part of the tube projects a little above the surface of the ground, and the distal end is folded inwards, thus closing the entrance to the nest. Bessia minor Hewitt. Text-fig. 4. Bessia minor Hewitt, Records, Albany Museum, vol. ii, p. 469, 1913. An adult male of this species was taken at Alicedale by Mr. F. Cruden on March 12th, 1914, and a description of it is here given. Carapace.—aAs long as the metatarsus and tarsus of the first leg; the lateral margins fringed with strong bristles, and the abdomen also is bristly over its median area superiorly. Fovea hghtly procurved. Chelicerew.—Dentition somewhat similar to that of the female, the teeth not arranged in continuous well-defined rows; the larger teeth are roughly arranged in a double series—the inner of which is weaker—in the middle of the group, and a single series at each end. Pedipalp.—Resembling that of a Spiroctenus. The tarsus has numerous short spines distally above ; the tibia has long spines below; at the base of the coxa inferiorly is a patch of cusps. Legs.—Tarsi not spined, all of them scopulate on the sides, the fourth tarsus only weakly so; paired claws with a single spirally curved row of numerous long teeth as in NEW SOUTH AFRICAN ARACHNIDA. 303 Spiroctenus. Metatarsus I longer than the tibia, shghtly scopulate in its apical third, shehtly but distinctly bowed ; II with several scopular hairs near the apex ; III and IV not scopulate. Apical tubercle of tibia I with a fairly long and TEXT-FIG. 4. Bessia minor Hewitt. Sternum, labium. and basal parts of the appendage of the male. x18: stout spur at its apex, and a shghtly shorter and less stout spur on its base externally ; the proximal tubercle also bears a spur and is situated rather less than a quarter of the length of the segment from the apical border. Anterior surface of patella III with a row of three spines superiorly and two longer ones distally below. All the femora are spined above. 304. JOHN HEWITT. Labium.—With about twenty-four cusps on its anterior half. Sternum.—Posterior sigilla elongated, about one and a half diameters apart and about half a diameter distant from the sternal margin. At the margin of the sternum opposite the base of the second leg there is a deep circular depression in which is situated the second sigillum; a similar depression also occurs in the female, but is not so noticeable, as the sete which fringe its margin are not so strong as in the male. Sternal margin fringed with strong bristles, especially pos- teriorly. Posterior spinners.— Apical segment about three- quarters the length of the penultimate segment. Measurements.—Total length 15°2 mm., length of cara- pace 65 mm., breadth of same 4°8 mm., length of first metatarsus 4 mm., of fourth metatarsus 5°8 mm. Bessia fossoria Poe. Bessia fossoria Poc., Ann. Mag. Nat. Hist., vol. vii, 6, p. 320, 1900. I have examined the type, which is very immature, and have now no doubt but that the much larger specimen from Redhouse, referred to in a previous paper,! belongs to this species. The dentition of the paired tarsal claws was incor- rectly described by Pocock, for the inner distal row is repre- sented by two or three minute teeth oneach claw. Moreover the tarsus of the first lez and the distal half of the metatarsus is rather strongly scopulate, the tarsus of the second leg is scopulate at the sides, but the third and fourth tarsi are not scopulate. The teeth on the chelicere constitute a single row, but the line is irregular, not straight. The genus is evidently very closely related to Spiroc- tenus Sim. 1 Hewitt, J., ‘ Records, Albany Museum,’ vol. ii, p. 471, 1913. NEW SOUTH AFRICAN ARACHNIDA. 305 Spiroctenus armatus Hewitt. Spiroctenus armatus Hewitt, Records, Albany Museum, vol. ii, p. 467, 1913. In drawing up the description of this species the following characters were accidentally omitted. Colour.—Carapace and legs reddish brown, palps pale brown; abdomen pale with ill-defined dark markings above ; ventral surfaces pale. Posterior sternal sigilla.—Hlongated, oval, about one and a half diameters apart and about half a diameter distant from the sternal margin. Measurements.—TYotal length 15 mm. ; length of cara- pace 6°4 mm. ; breadth of same 51 mm. ; length of metatarsus of first leg 4-2 mm., of fourth leg 6°2 mm. Acanthodon abrahami (Hewitt). Gorgyrella abrahami Hewitt, Records, Albany Museum, vol. ii, p. 473, 1913. I now prefer to place this species under the genus Acanthodon, although formerly I doubtfully referred it to Gorgyrella. Itis probably closely related to A. ochreo- lum Poc., of which only the male is known. The male of abrahami seems to differ from that of ochreolum in the closer approximation of the frontal eyes and in the less strongly bent first metatarsus, buc until more Jansenville material is available for comparison the specific distinction of abrahami should be held somewhat doubtful. T'wo adult males of abrahami were taken by Mr. F. Cruden at Alicedale on February 15th, 1914, and I here give a des- cription of the sex. Frontal eyes quite separated, about one-third of a diameter, or slightly less, apart; the quadrangle formed by the frontal and anterior median eyes wider behind, the medians large, about three-fifths to three-quarters of a diameter apart ; hind margins of posterior row about in a straight line, the medians two to two and a half diameters apart and a diameter or slightly more distant from the laterals, 306 JOHN HEWI'T. Tarsi of all the legs scopulate below throughout their length, the first tarsus being only quite thinly scopulate, the fourth without sete along its middle. Tarsal claws of first leg with three teeth, of fourth leg with two teeth and a third tooth may or may not be just indicated. First metatarsus slightly bowed, with two spines at the apex inferiorly and three along the postero-inferior edge, but none along the antero-inferior edge. First tibia shorter than the metatarsus, with four spines along the postero-inferior edge, with the usual pair of tubercles near the apex, the more proximal one being distant from the apex less than one-sixth the length of the segment. Band of spinules on anterior surface of patella IV stretching two-fifths to one half of the length of the seg- ment and comprising only about eight to ten spinules ; III has two to four spines along the distal edge anteriorly, and two to nine on the anterior surface, but only one on the dorsal surface, in addition to one or two on the distal edge. Tibia of pedipalp barely one and two-third times as long as deep, the band of spines bordering the excavation complete, in three rows proximally, but elsewhere in a single row, except at the distal end; altogether there are nineteen to twenty-four such spines. Tarsus of pedipalp with one long spine at the apex superiorly and one or two weaker ones, the distal inferior lobe on the posterior side rather strongly projecting. Process of palpal organ comparatively short, flattened, and twisted. Measurements.—Total length 12°5 mm., length of cara- pace 4°5 mm., breadth of same 3°5 mm., length of tibia of first leg 2°75 mm. Acanthodon ochreolum Poc. Acanthodon ochreolum Poc., Ann. Mag. Nat. Hist., 7, x, p. 9, 1902. This species is probably the same as either Gorgyrella schreineri Purcell, or G@. abrahami Hewitt, for it has three pairs of sternal sigilla. The following notes on the type specimen will supplement Pocock’s rather incomplete description. NEW SOUTH AFRICAN ARACHNIDA. 307 Frontal eyes quite separate from each other, about half a diameter or a trifle more apart. The quadrangle formed by the frontal and anterior median eyes is appreciably wider behind. Hind margins of posterior row of eyes in a pro- curved line. Posterior median eyes three to three and a halt diameters apart. Tarsus of palp with one rather long stout spine at the apex dorsally, and several much weaker ones. Pocock’s figure of the tibia of the palp is quite inaccurate in respect to the spinulation of the excavation ; the spinules are not arranged in a single row throughout, and are much more numerous than represented in the figure; distally they occupy two or three rows, and proximally the spinules are considerably longer than elsewhere. All the tarsi are scopulate to the base, the fourth being without spiniform sete mesially below. Tarsal claws without a comb of teeth, the first having only one large tooth and the fourth one or. two. The first tibia has four spines along its outer inferior edge, but none otherwise situated, apart from the two large tubercles. The inner dental series of the chelicerze comprises five large teeth. The first pair of sternal sigilla is near to the sternal margin. Acanthodon hamiltoni Poe. Acanthodon hamiltoni Poc., Ann. Mag. Nat. Hist., 7, x, p. 320, 1902. The type of this species is without doubt very juvenile. The inner dental series comprises five strong teeth and a few small ones basally situated. The cox of the legs are without spinules, and along the posterior edge of the third coxa inferiorly is a strip of moderately fine hairs. The second tibia has a row of eight spines on its anterior side. The band of short strong spines on the anterior surface of the fourth patella extends two-thirds of the distance along the segment. The frontal eyes are very close together; the ocular tubercle is deeply grooved above. 308 JOHN HEWIY'. Acanthodon grandis sp. nov. Text-fig. 5. This may be the female of A. pectinipalpis Pure. (Annals S. Afr. Mus., vol. ii, p. 87) described from Zululand, but I think it is probably distinct on account of the wider separation of the frontal eyes, those of pectinipalpis being described as very close together. It is related to a species described by me! from Newington, Transvaal, under the name of Ctenolophus transvaalensis, but differs in that the third coxa is clothed inferiorly with stout bristles, not spimules, and in other respects also. TEXT-FIG. 5. Wy) Yiey My (} eye YY pl p CEES) Acanthodon grandis sp. nov. The eyes of the female. x 16. The type consists of a single adult female in the collection of the Natal Museum from Umfolosi, Zululand (F. Toppin coll. 1905). Colour.—Upper surface of appendages and carapace castaneous, the legs with a reddish tinge. Ventral surfaces paler; sternum and coxe of last two pairs of legs yellowish- brown. : Carapace.—aAs long as the tibia, metatarsus and half of the tarsus of the fourth leg, or the patella, tibia and three- fifths of the metatarsus of the first leg. Ocular area only slightly wider than long, its width subequal to the length 1 Hewitt, John, ‘Records, Albany Museum,’ vol. ii, p. 412, 19138. NEW SOUTH AFRICAN ARACHNIDA. 309: of metatarsus I, its length almost two-fifths of the distance from the anterior margin of the carapace to the centre of the fovea. Area formed by the frontal and anterior median eyes slightly wider in front, the medians small, about two diameters apart, the frontals large, about one-third of a diameter apart and situated on a common tubercle which is deeply grooved above. Posterior row about equally spaced, the area formed by the four medians appreciably wider behind ; posterior laterals large, the medians considerably larger than the anterior medians, the hind margins ina slightly procurved line. Legs.—Metatarsus III with three long weak spines below and three stronger ones at the apex inferiorly ; IV with six or seven spines along the lower surface and three at the apex inferiorly. Band of spines on anterior surface of tibia [ reaching to the base; of II also reaching to the base and including about twenty-two spines; IV with two inferior apical spines and one weak one on the lower surface. Band of spines along anterior surface of patella III imeluding about thirty spines, of which about seven are on the distal edge; on the posterior distal edge there are five spines; IV with a band of short stout spines stretching the whole length of the anterior surface. Coxe of legs without spinules below, the third having a dense patch of short coarse bristles or weak spines along its posterior border ventrally. Labium.—With two apical teeth. Chelicere.—Inner row of teeth on fang-groove including five large teeth; the outer group without any large teeth. Sternum.—Three pairs of sigilla, the third pair small and not so conspicuous as the preceding pairs; first pair sub- marginal, second pair their own length distant from the sternal margin. Measurements.—Total length 25mm.; length of carapace 10:2 mm., width of same 9 mm.; length of tibia of first leg 425 mm. This is the largest species of Acanthodon known to me (excluding Gorgyrella namaquensis and G. schreineri, but Gorgyrella is hardly separable from Acanthodon). 310 JOHN HEWITT. Idiops pretorize (Poc.). Acanthodon pretoriew Poc., Ann. Mag. Nat. Hist., vii, 1, p. 319, 1898. To the description given by Mr. Pocock I can add the following notes, after examination of the type male: Cheliceree with seven strong teeth in the inner row and four in the outer row ; process of palpal organ strongly flattened throughout and twisted; the excavation on the tibia of the palp very shallow and not bordered by spines; band of spinules on anterior side of patella IV stretching five-sixths of the length of the segment and including about thirty short spines; patella III has a continuous strip of spinules anteriorly including about thirteen as well as four on the distal edge ; frontal eyes one-quarter to one-third of a diameter apart, the anterior medians very large, the area formed by the frontal and anterior median eyes much broader behind, posterior medians very much smaller ; posterior margins of posterior row of eyes shghtly recurved, the anterior margins shghtly procurved ; the posterior medians about three diameters distant from the posterior laterals, which latter are a little elongated but not greatly so. Idiops astutus sp. nov. Text-fig. 6. The type consists of a single adult male taken at Bulawayo, November 15th, 1918, by Mr. G. Arnold, who writes of it “ hunting insects under the electric lights, waiting for those that fall.” This species can at once be distinguished from I. arnoldi Hewitt.) which it somewhat resembles, and which also is believed to occur at Bulawayo, through the total absence of a scopula on the fourth tarsus, whereas arnoldi has a broad scopula on the swollen fourth tarsus. Colour.—Upper surfaces brownish-black, lower surfaces somewhat paler, the sternum and cox of the third and fourth legs, the genital sternite, lung opercula and spinnerets being pale yellowish-brown. 1 «Records, Albany Museum,’ vol. iii, p. 21, 1914. NEW SOUTH AFRICAN ARACHNIDA. Sid Lilli Carapace.—Broad and depressed, its length equal to that of the fourth metatarsus together with two-thirds of the tarsus. Surface coarsely granulated, except on the ocular area and lateral thereto. Ocular area very slightly wider than long, its length slightly less than one-third of the distance from the anterior margin of the carapace to the THXT-FIG. 6. A. Idiops astutus sp.nov. B. 1. arnoldi Hewitt. Distal portions of male palps. a x 14. B Xx 18. centre of the fovea; frontal eyes quite separated, about one and a quarter diameters apart. Area formed by frontal and anterior median eyes very slightly wider in front, its length being only slightly greater than its anterior width; the medians about half a diameter apart and much larger than the frontals. Posterior margins of posterior row in a pro- curved line, the medians about two to two and a half dia- meters apart and about a diameter distant from the laterals, which are elongated but rather small. Posterior medians ’ 254) JOHN HEWITT. about one-third of a diameter distant from the anterior medians. Pedipalp.—Tibia shghtly more than three times as long as deep, the excavation completely bordered by a band of spines. Tarsus dorsally without spines, except at the apex, where there are several; the apical portion strongly projects upwards (see text-fig.6, 4). Process of bulb flattened throughout its length, obliquely truncate at the apex, curved and shghtly twisted. Legs.—Tibia I shorter than the metatarsus, with a pair of tubercles near the distal end inferiorly, the proximal one short and small, the distal one with a large and long black projection ; its under-surface without spines along the median area, about ten or eleven along the posterior edge inferiorly, but none on the anterior edge. Metatarsus I distinctly bowed, with seven or eight short spines inferiorly on the posterior side and three or four on the anterior side (includ- ing two near the apex). Tarsus I with eight or nine spines on each side, weakly scopulate in its distal third, IT and III broadly scopulate from apex to base, IV without trace of scopula and not swollen. Claws of all the tarsi with two basal teeth, the more distal one larger than the other; on tarsus IV the posterior claw of the left leg (missing on mght side) carries three teeth, and such is the case on one of the claws on tarsus II. Band of spinules on anterior side of patella IV extending the whole length of the segment ; patella III also has a band of short spines along its anterior surface, ending in a row of about five or six longer ones along the distal edge. Measurements.—Total length 19 mm., length of cara- pace 8°8 mm., breadth of same 7°8 mm., length of tibia of first leg 5-4 mm., of metatarsus of same 7°4 mm., length of first leg 33°3 mm., of fourth leg 31 mm. Hermacha Mazoena sp. NOV. The type is a single female specimen from Mazce, Mashona- land (J. Darling), in the collection of the British Museum. NEW SOUTH AFRICAN ARACHNIDA. ake Colour.—Almost uniformly pale brown. the abdomen superiorly showing indication of a darker tree-pattern. Carapace.—Sheghtly longer than the fourth metatarsus and tarsus together. Ocular area a trifle more than twice as broad as long. Anterior margins of anterior row of eves ina procurved line; the medians rather small, about three- quarters of a diameter apart. Posterior medians rather small, oval, and widely separated, about three and a half times their long diameter apart, their inner margins about in a line with the outer margins of the anterior medians; pos- terior laterals about as long as the anterior laterals. Fovea straight. Chelicerz.—With seven teeth in the inner row below, the distal tooth of the outer group being opposite the fifth tooth of the inner row. Pedipalps.—Coxe with about sixty teeth arranged in a triangular patch at the base. Labium.—Quite without apical teeth. Legs.—Tarsus I with dense undivided scopula, II with scopula divided by a thin median strip of sete, III and IV not definitely scopulate, but carrying a few scopular hairs at the sides, in IV only distally so. Metatarsus I and II scopulate to the base, III and IV not scopulate; I with two spines at the apex below and two or three along the inferior surface, also one or none on the anterior surface. Tibia I with three weak spines at the apex below and two on the inferior surface, also one on the anterior surface superiorly ; very slightly exceeding the metatarsus in length. Posterior Sternal Sigilla.—Elongated, almost touch- ing the sternal margin. Posterior Spinners.—With the apical and basal seg- ments subequal in length, being about one and three-fifths times the length of the penultimate segment. Anterior Spinners.—Not widely separated, less than half their own length apart (the abdomen is somewhat con- tracted, and in fresh specimens the spinners may perhaps be rather more widely separated); moderately long, being about 314 JOHN HEWITT. four-fifths the length of the basal segment of the posterior spinners. Measurements.—Total length 15°2 mm., length of cara- pace 5'6 mm., breadth of same 4°5 mm., length of tibia of first leg 2°5 mm. No species of this genus has previously been recorded from Rhodesia. The described species to which it is most closely allied are probably H. caudata Sim. from Delagoa Bay, and H. bicolor (Poc.) [Brachytheliscus bicolor Poc.| from Durban ; the female of the former is unknown, and the latter species is distinguished from mazcena in the following characters : ocular area, dentition of chelicerze and of labium, and separation of anterior spinners. Genus HerMACHOLA gen. nov. This name is proposed for the reception of a small species which has many of the characters of the genus Hermacha EH. Sim., but differs therefrom principally in the form of the male bulbal organ and in the shape of the fovea as hereafter described. Hermachola grahami sp. nov. Text-fig. 7, a—c. The type is a single adult male example collected at Grahamstown by Mr. F. C. Graham, September 23rd, 1914, who presented it to the Albany Museum. Colour.—Carapace and appendages pale brown, the more distal segments of the first pair of legs with a dull reddish tinge; the hairs which clothe the carapace and appendages are blackish. Abdomen pale dorsally, with a variegated dark pattern. Lower surface pale, but posteriorly just in front of the base of the spmners the abdomen has a dark transverse stripe. Carapace.—Hlongated, about as long as the metatarsus and tarsus of the fourth leg. Fovea transverse at the base, but with a short backwardly directed median extension, which, however, 1s not so deep as the main portion of the fovea. Near the anterior border of the fovea there are two long spini- NEW SOUTH AFRICAN ARACHNIDA. ole form sete mesially situated; otherwise no spiniform sete occur on the carapace except a few on the borders, especially at the postero-lateral corners. Ocular Area.—Not quite twice as broad as long. Ante- rior row with their anterior margins strongly procurved, the laterals much larger than the medians or than the postero- laterals. Posterior laterals subcontiguous with the posterior TEextT-FIG. 7. ~ : SS Hermachola grahami sp. nov. A. Distal portion of male palp. x 55. B. Fovea. x 40. c. Ocular area. X 69. medians and with the anterior laterals. Distance of anterior laterals from the margin of the carapace scarcely as much as half of a diameter of an anterior median. Labium.—With a single small apical tooth. Chelicere. With seven or eight teeth on the inner row below. Anterior surface covered with spiniform setve which are not elongated ; rastellum weak, without spines. Pedipalps.—Coxa with a small patch of about eighteen small teeth at the base. Femur with two or three weak spines VOL. 3, PART 2. 22, 316 JOHN HEWITT. above ; patella without spines. ‘Tibia with numerous elon- gated spiniform setee on its ventral surface on the inner side, more strongly developed in the basal half of the segment. Tarsus without spines. Bulb small, process coiled up spirally like a corkscrew ; there are two complete coils in addition to the expanded and coiled basal portion which invests the bulb and the narrowed distal portion which also has a strong spiral twist; the process reaches backwards as far as the basal fourth of the tibia. Legs.—tTarsi unspined. Metatarsus I almost straight, with three strong apical spines below but no other spines elsewhere, II also with three strong apical spines inferiorly, and in addition with two or three on the inner surface and two on the inferior surface posteriorly, II] and IV numerously spined. Tibia I stout, with three apical spines, none of which are enlarged, also with six spines on the lower surface, of which one distally situated is specially long and stout ; on the anterior surface there is a row of three spines. Tibia II with two or three apical spines inferiorly, three or four on the lower surface, and two or three on the anterior surface supe- riorly ; III with three apical spines below, four on the inferior surface, two on the anterior surface, two dorsally, and two on the posterior surface above, all of them being elongated ; IV spined like III except that the dorsal surface has one or no spines. Patella I without spines, II with one on the anterior surface near the apex, II] with two on the anterior surface and one on the posterior surface, IV with one on the anterior and one on the posterior surface. Femora armed above with long setiform spines. Tarsi I and II scopulate to the base, the scopula being entire, though not dense, and composed of not very fine sete ; III and IV with a few scopular setz, but mostly the setze are long and subspiniform. Metatarsi I and II thinly scopulate from the apex to near the base. Inferior tarsal claw of anterior legs small but easily distinguishable. Sternal sigilla.—Third elongated, shghtly less than its own length distant from the sternal margin; second sub- marginal, first indistinct. 9 NEW SOUTH AFRICAN: ARACHNIDA. Ba WA Measurements.—Total length 8°25 mm., of carapace 3 mm., width of carapace 2:1 mm. Stasimopus steynsburgensis sp. nov. This species is closely related to S. palpiger Poc. from Graaff Reinet and 8. schreineri Purc. from Hanover. It differs from both in having shorter palps and in possessing a scopula on the tarsus of the fourth leg of the male. The type is a single adult male in the collection of the British Museum from Steynsburg, C.C., taken by Miss Leppan. Colour.—Upper surfaces blackish, excepting the tibia and tarsus of the palps and the tarsi and metatarsi of the legs, which are reddish yellow; also the patella and tibia of the fourth legs are brownish; sternum and cox of appendages inferiorly castaneous, abdomen fuscous, the lung opercula and genital sternite yellow, the spinners pale. Carapace.—Subequal in length to the metatarsus of the fourth leg. The three keels well developed anteriorly, but absent posteriorly, none of them approaching the fovea; hairs are absent therefrom and from the ocular area; the sides of the cephalic area are practically smooth except immediately external to the lateral ridges. Anterior margins of anterior row of eyes distinctly procurved; distance between an anterior lateral and anterior median subequal to the diameter of the latter, but less than the long diameter of the former ; posterior lateral almost as long as the anterior lateral, the distance between them about one and three-fifth times the length of the latter; outer lateral margin of anterior lateral about in a line with inner posterior margin of posterior lateral. Pedipalps.—Stretched forwards, the apex reaches a point three-fifths of the distance along the metatarsus of the first leg. Patella longer than that of first leg, but much shorter than the tibia of that leg or of the palp. Tibia and tarsus together shghtly shorter than the carapace. Process of palpal organ longer than the bulb, tapering and slightly twisted. Legs.—Tarsus I absent from the specimen; II a little 318 JOHN HEWITT. swollen, scopulate below, anteriorly with six spines, posteriorly with seven or eight spines; III with about fourteen spines on each side, scopulate below; IV numerously spined on both sides and scopulate mesially below. Metatarsus I and II without trace of a scopula, but with no spines mesially below though strong spines occur on each side inferiorly. Tibia I shorter than the metatarsus, with only one or two spines mesially below and with about eight spines on the anterior side and four on the posterior side in addition to those at the apex; III with five to seven short spines on or near the distal edge on both anterior and posterior sides superiorly. Patella I with one spine at the apex inferiorly; [lJ with an anterior band of short spines, about six to eight i number, but no distal patch of spinules superiorly ; IV with a patch of minute spines extending over about two-thirds to three- quarters of the length of the anterior side. Anterior claw of fourth leg with two large teeth basally situated and with no small ones distal thereto; posterior claw with three large basal teeth. Measurements.—Total length 18°5 mm., length of cara- pace 7°8 mm., breadth of carapace 7 mm., length of tibia of palp 4-2 mm., of metatarsus of palp 5°6 mm. Stasimopus gigas sp. nov. The type is a single male example in the collection of the British Museum labelled “ Vredefort Rd.?, Barrett- Hamilton”; the specimen is in dried condition. Colour.—Upper surfaces black, except the tarsi and meta- tarsi of the legs, the tibiz and tarsi of the palps, also the distal portions of the tibiee of the first two pairs of legs, and of the patellee of the palps, all of which are reddish. Carapace.—Subequal in length to the metatarsus of the fourth leg. The three keels are prominent throughout and reach back almost as far as the fovea. Anterior margins of anterior row of eyes in a very slightly procurved line, sub- equally spaced, the distance between the medians about one NEW SOUTH AFRICAN ARACHNIDA. 319 and a half times the diameter of an eye; inner edges of anterior lateral and posterior median practically in the same line; distance between anterior and posterior laterals quite one and a half times that between an anterior lateral and anterior median. Pedipalps.—Patella slightly longer than that of the first leg, shorter than the tibia of that lege or of the palp. Tibia and tarsus together considerably shorter than the carapace. Legs.—All the tarsi scopulate below ; no trace of a scopula on the metatarsi. Tarsus I with a single spine on its anterior side and four on the posterior side (II wanting in the speci- men) ; II] with numerous spines in a continuous band on both anterior and posterior sides; [V with numerous spines on both sides, especially anteriorly. Metatarsus I with strong spines below, both over the median area and at the sides. ‘Tibia I considerably shorter than the metatarsus, with strong spines below and at the sides; III with a few short spines and spinules on or near the distal edge on both anterior and posterior sides superiorly. Patella I with two spines at the apex inferiorly and one on the lower surface; II with an anterior band of spines, but no distal patch of spinules above apart from those belonging to that band; LV with an anterior patch of spinules extending about three-fifths of the length of the anterior side. Paired claws of tarsus of fourth leg with a basal comb of four large teeth, distal to which is a very small tooth, and on the posterior claw one or two minute teeth occur on the proximal side of the comb. Measurements.—Length of carapace 11:2 mm., breadth of same 10 mm., length of tibia of palp measured from above 7 mm., of patella of palp 5°5 mm., of first metatarsus 8°4 mm. of first tibia 7 mm., of first patella 5 mm., of fourth metatarsus 11-5 mm. This species is distinct from S. nigellus Poc. (Ann. Mag. Nat. Hist., 7, x, p. 319), which also seems to have been taken at Vredefort Rd. by the late Capt. Barrett-Hamilton, in the following characters: ocular arrangement, keeling of cara- 320 JOHN HEWITT. pace, spinulation of first metatarsus, third tarsus, third patella, in the dentition of the fourth claws, and lastly im size, this being by far the largest male yet recorded in this genus. Stasimopus minor sp. nov. ‘The type is a single adult male from Bloemfontein collected by Dr. T. F. Dreyer in August, 1914. It was found on the open veld. This may be identical with S. oculatus Poc., which is based on a female example, or may belong to another species, for Dr. Dreyer has taken females of two species at Bloem- fontein. It is probably closely related to S. nigellus Poe. taken at Vredefort Rd., but differs therefrom in the wide separation of the anterior median eyes. Colour.—Jet-black above, the distal portions of the legs and of the palps paler; sternum, cox of appendages, genital plate, lung opercula, and spinners pale. | Carapace.—A little longer than tibia of first leg or metatarsus of palp. Almost quite smooth and shining in its anterior half, lightly sculptured (not coarsely rugose) in its posterior half. The three keels depressed, only the median one reaching back to the fovea. Anterior margins of anterior row of eyes in a straight line, the medians much nearer to the laterals than to each other, being about one and a half diameters apart; anterior laterals subequal in length to the anterior medians and separated therefrom by about half the diameter of a median. Ocular area sparsely hairy. Pedipalps.—Pressed forwards, the tip reaches a point three-fifths of the distance along the metatarsus of the first leg; patella considerably longer than that of the first leg, but shorter than the tibia of that leg or of the palp; tibia and tarsus together subequal to the carapace in length. Legs.—All the tarsi scopulate below, but no trace of a scopula on the metatarsi. ‘Tarsus I with two or three spines on the anterior side, two or none on the posterior side; II with two anteriorly and two or three posteriorly ; II with NEW SOUTH AFRICAN ARACHNIDA. SAL one anteriorly and one or three posteriorly ; IV with seven anteriorly and one posteriorly. Metatarsus I without strong spines on the mesial area below. Tibia I subequal in length to the metatarsus, with three spines below besides those at the apex, but none at the sides excepting one near the apex anteriorly ; tibia III with about five short spines on or near the distal edge on both anterior and posterior sides superiorly. Patella I without spines below; III with a strip of six to nine weak spines on its anterior side, but no distal patch of spinules above. Patella [V with an anterior patch of spinules extending about three-fifths of the length of the anterior side. Paired tarsal claws of fourth leg each armed with two large basal teeth, with four smaller ones more distally situated on the posterior claw, but only three on the anterior claw, which are weaker than those on the posterior claw. Measurements.—Total length 85 mm., length of cara- pace 4mm., breadth of same 3:3 mm., length of palp 12 mm., of first leg 15mm. Stasimopus oculatus Poe. Stasimopus oculatus Poc., Proc. Zool. Soc., 1897, p. 728, pl. xli, fig. 2. It has been pointed out to me by Dr. T. F. Dreyer that at least two species of Stasimopus occur at Bloemfontein, and Dr. Dreyer has kindly sent to me adult female material of the same. S. oculatus, the larger species, can be distinguished through the presence of a cluster of spines at the apex of the third metatarsus inferiorly and the presence of stout spinules at the apex of the tibia of the palp superiorly, whereas the smaller species has neither of these characters; there is also a difference in coloration, the smaller species having the abdomen fairly uniformly infuscated above, whilst oculatus has the lateral surface of the abdomen and most of the dorsal surface comparatively pale but with a dark mesial blotch anteriorly and some infuscation over the posterior part of the dorsal surface. 322 JOHN HEWITT. Female specimens which seem to be identical with oculatus are known to me from Reddersburg (Dr. Broom), Ladybrand (Dr. Dreyer), and Kimberley (Bro. J. H. Power). Female specimens which are identical with or very closely related to the smaller Bloemfontein species are known to me from Vredefort Rd. (Capt. Barrett-Hamilton), Valsch River near Kroonstad (Prof. H. H. W. Pearson), and Winburg (Miss 8. Brown). A third form characterised by the possession of short stout spinules at the apex of the tibia of the palp but no spines at the apex of the third metatarsus occurs at Kroonstad (Dr. ie Dreyer). A fourth form from Jagersfontein (M. Francis) has a group of stout spines at the apex of the third metatarsus, but no stout spinules at the apex of the tibia of the palp. Lastly, we have a good series of female specimens from Modder River (Bro. J. H. Power) which seems referable to two species, all the small and immature examples agreeing with oculatus, and all the larger ones being similar to the smaller Bloemfontein species in structure, but very much larger than that form; in this case it seems just possible that the small specimens from Modder River are merely the young of the larger ones.! Unfortunately we possess no evidence which can serve to connect any of these forms with the species based on male material. Only four male Stasimopi are known from the Free State, viz. the type of S. nigellus Poc. from Vredefort Rd., the type and a co-type of 8S. gigas sp. nov. also from Vredefort Rd., and the type of S. minor sp. nov. from Bloemfontein. According to Dr. Dreyer, the nests of the two Bloemfontein species are very distinct; the female of oculatus makes a D-shaped lid, whilst the other species has a more circular lid; the tube of the nest in oculatus is very thickly lined 1 Some of the forms just mentioned will be described by me as new species in ‘ Records Albany Museum,’ vol. iii, pt. il. NEW SOUTH AFRICAN ARACHNIDA. S25 by felt-hke silk, forming a much thicker wall than that found in the other species. In a large specimen the longest diameter of the lid of oculatus reaches 50 mm., the greatest outside diameter of the tube at the top being 60 mm. Order SOLIFUG A. Chelypus hirsti sp. nov. Text-figs. 8, 9. TEXT-FIG. 8. Chelypus hirsti sp. nov. Specimen on left is minus the left chelicera, and is shown in ventral view; specimen on right is in dorsal view. x 3. The types consist of four male examples in the Albany Museum collection, two of which came from Rietfontein, Gordonia, presented by Mr. H. Drew, and two are indefinitely located as North-West Gordonia, presented by Mr. C. A. Anderson. Female unknown. This species is closely related to C. barberi! Purcell, but differs therefrom as follows. 1 Pureell, W. F., ‘Annals S. Afr. Mus.,’ vol. ii, p. 224, 1901. 324 JOHN HEWITT. Flagellum.—Not bifid at the apex, but tapering to a point. Near its base the shaft has a ventral, compressed, and keel- like extension along the greater part of the length of its pro- current portion. Chelicerz.—Upper finger furnished internally with numerous short stout spines, occupying an elongated obliquely arranged area extending from the neighbourhood of the large basal tubercle on the inner side of the fang to a point near the basal enlargement of the flagellum. Outer row of teeth in the upper Jaw comprising six teeth, the fourth one being TEXT-FIG. 9. Chelypus hirsti sp. nov. Left chelicera of male viewed from inner side. x 4. minute. Upper surface of cheliceree covered with small granules, a few of which are spinuliform. Pedipalps.—The upper surface of the tibia is only finely granulated, and in its basal third is almost free of granules ; there is some fine granulation and one or two coarse granules at the base of the metatarsus above. Colour.—Cheliceree without dark marks at the base above. ‘Cephalothorax infuscated over its anterior half. Total Length.—27 mm. This species is named after Mr. A. S. Hirst, who has made important contributions to our knowledge of Arachnida and has described various 8. African species. NEW SOUTH AFRICAN ARACHNIDA. 320 Order SCORPIONES. Opisthophthalmus pugnax Thorell, var. natalensis Uar. 10V. The types of this variety are two adult specimens, male and female, from Estcourt, Natal, collected by Mrs. EH. J. Turner and by Mr. Guy Marshall respectively, who presented them to the British Museum. ‘The Natal Museum has a female example, a trifle larger than the type, from Mooi River, Natal, collected by Mr. C. James. The variety agrees with the typical form of pugnax, as known to me through material from Pretoria, in colour and general structure, but differs as follows. No stridulatory lamelle on the chelicere. Median eyes more posteriorly situated, their distance from the hind end of the carapace being less than one-third of the total length of the carapace (more than one-third in typical pugnax). Hands much less coarsely granulated, that of the male on its inner portion superiorly being covered with numerous quite small and isolated round granules, the finger keel with one or two coarse pits in its course, but practically continuous throughout, the more external secondary keel quite obsolete and the inner one, for the greater portion of its length, only represented by infuscated scarcely enlarged granules, the outer surface of the hand with only one keel; that of the female without coarse granulation on the inner part of its upper surface, except quite near the finger, but covered with a much flattened meshwork of more or less coalesced ridges and granules in which the secondary keels are quite absent, the whole surface appearing much smoother than in the typical form, the finger keel well defined, continuous in the distal half, more or less broken in the basal half (in the Mooi River example it is continuous almost throughout). Hand of male slightly narrower in proportion to the length of the hand back than in pugnax, and the fingers a little longer. 326 JOHN HEWITT. The granules on the mesial portion of the last abdominal sternite of the male more numerous and not quite so large as in the male of pugnax proper (about ten or twelve rows can be roughly counted between the anterior and _ posterior margins); in the female this area is covered with low granules and irregularly shaped ridges, which occur throughont the whole length of the segment, more or less obscuring the keels (these are quite obsolete in the Estcourt specimen, but present, though indistinct, im the Mooi River specimen). Infero-median keels of first caudal segment in the male com- posed of about eight to ten coarse granules, and the space between them also includes about ten coarse granules, but is not obliterated thereby nor are the crests rendered indistinct to any great extent; im the female these keels are pitted and more or less coarsely granular (in the Mooi River specimen the keels are not definitely broken up into coarse granules), with a few granules between the keels. The abdominal tergites in the female are all granular throughout, coarsely so in the posterior portions of each tergite. Pectinal teeth: Thirteen to fourteen male, eleven to twelve female (thirteen to fourteen in the Mooi River specimen). Measurements.—lotal length, male 78 mm., female 89 mm.; length of hand, male 17°3 mm., female 18°38 mm.; of movable finger, male 11°75 mm., female 12°3 mm.; of hand back, male 6 mm., female 7mm.; breadth of hand, male 8 mm., female 10°5 mm.; length of carapace, male 11°7 mm., female 13 mm.; distance of median eyes from posterior margin, male 3°75 mm., female 3°7 mm.; width of first caudal segment, male 575 mm.; of fifth caudal segment, male 4°25 mm.; length of same measured along the side 10 mm.; breadth of vesicle, male 4°25 mm.; length of same 10 mm. In respect to the shape of the hand and the granulation of the last abdominal sternite and of the infero-median keels of the first caudal segment in the male, this form is intermediate between typical pugnax and the form described by me from Keilands under the name of O.latimanus Koch var. NEW SOUTH AFRICAN ARACHNIDA. eo) 27 keilandsi.! It resembles that form also in the posterior position of the median eyes; it differs in that the tail of the male is much stouter in natalensis. The fifth caudal segment in an adult male of keilandsi measures 10°5 mm. in length and 3°75 mm. in width. Dr. Purcell” points out in his important monograph on the genus Opisthophthalmus that the presence or absence of stridulatory lamellee on the chelicerz is usually of no specific value. However, I am inclined to regard the character of some importance in the various forms which range around and between pugnax and latimanus. Such lamelle are present in all our examples of latimanus, 132 specimens of all sizes, from Grahamstown, Highlands, Brakkloof, Fort Beaufort, Jansenville, Schurfteberg, Alicedale, Glen Lynden, Mimosa, Redhouse, Alexandria, and Queenstown; they are even present asa row of bristles on the newly hatched young. They are present in latimanus var. austeroides mihi and in every specimen of our series of typical pugnax; they are completely absent in keilandsi. It is not improbable that intermediates between keilandsi and natalensis will be found, in which case it will no longer be possible to maintain pugnax as a species distinct from latimanus. With the discovery of new local forms in the genus Opisthophthalmus the differences between the so- called species are gradually breaking down and the genus is seen to be composed of numerous local forms which more or less completely grade into each other. 1 ‘Records, Albany Museum,’ vol. iii, p. 7, 1914. * «Annals §. Afr. Museum,’ vol. i, p. 134, 1899. My 3} Fae eho ss ee My air. want Csi eh yas “ ay airs . | , vu | * Sys PL? 4¢ OBSERVATIONS ON SOME SOUTH AFRICAN TERMITES. 329 Observations on some South African Termites. By Claude Fuller, Division of Entomology, Department of Agriculture, Pretoria. With Plates XXV—XXXV and 16 Text-figures. CoNnTENTS. PAGE Ll. IntRODUCTION F ; ; , 5 sil) Il. Tort ABUNDANCE OF TERMITES : J wool III. Tot BEHAVIOUR OF WINGED TERMITES ; 5 83h; Hodotermes transvaalensis sp. n. : . 335 Termes natalensis Haviland u E 5 Bio Termes latericius Haviland : . 5 at!) Termes vulgaris Haviland ; : 34.0 Termes incertus Hagen . . d41 Eutermes bilobatus (Haviland) aad Hater mes sp. 346 IV. OBSERVATIONS ON THE NESTING HABITS AND GENERAL Economy OF CERTAIN SPECIES . ; . 349 Hodotermes transvaalensis sp. n. : . 349 Hodotermes viator (Latreille) : : . 306 Calotermes durbanensis Haviland : . 398 Rhinotermes putorius Sjostedt . ; . 360 Termes waterbergi sp. n. : é . 361 Termes natalensis Haviland ; : . 364 Termes badius Haviiand By) Termes latericius Haviland : : . 385 Termes vulgaris Haviland : : mous! Termes incertus Hagen . : . 393 Eutermes parvus (Haviland) : . 396 Eutermes bilobatus (Haviland) . : . 398 Eutermes trinervius (Rambur) . : . 402 330 CLAUDE FULLER. PAGE V. CLASSIFICATION : : : ; . 413 (1) Introduction p : : s . Ads (2) Systematic Account : : : . 423 Hodotermes ; : ‘ . 423 Hodotermes transvaalensis sp. n. . . 425 Hodotermes pretoriensis sp. n. ; . A429 Hodotermes karrooensis sp. n. : . 4353 Hodotermes mossambicus (Hagen) . . 438 Hodotermes havilandi Sharp : . 441 Hodotermes warreni sp. x. . . 444 Hodotermes braini sp.n. . ; . 446 Hodotermes viator (Latreille) ; . 448 Calotermes durbanensis Haviland . . 451 Rhinotermes putorius Sjéstedt : . 453 Termes ; : : . 456 Termes swazle sp. n. 3 5 . 462 Termes waterbergi sp.n. . : . 466 Termes natalensis Haviland : . 470 Termes badius Haviland . : . 475 Termes latericius Haviland ; . A479 Termes vulgaris Haviland . : . 481 Termes incertus Hagen : : . 483 EKutermes parvus (Haviland) . 485 Eutermes bilobatus (Haviland) : . 487 Hutermes hastatus (Haviland) ; . 489 Eutermes trinervius (Rambur) : . 491 VI. APPENDIX—GLOSSARY ‘ : . 494 VII. REFERENCES ; 3 : ; . 496 VIII. EXPLANATION OF PLATES . . : . 497 I. INTRODUCTION. THESE notes do not pretend to be more than fragmentary records which have been brought together for publication at the express wish of my friend, Dr. Ernest Warren. They include observations which have been made or collected only during the past two or three years; for most of them the writer is personally responsible; but, for some, which are duly acknowledged in the text, he is indebted to his colleagues and other friends. Unfortunately, owing to the difficulties OBSERVATIONS ON SOME SOUTH AFRICAN TERMITES. 33] experienced in determining the material collected during the period mentioned, it has not been possible to include all the observations made; these it is hoped to deal with at some future date. African termites having been described somewhat inade- quately, as a whole, and for the most part in foreign languages, the opportunity has been taken to redescribe the soldier caste of some species discussed, and to furnish some comment upon the ganeral characteristics of each. As far as possible those speculations, which the study of termite-nature excites in every observer, have been avoided ; here and there, however, the temptation to speculate has been too great to overcome. For the sake of completeness, some observations made by others are repeated; these are chiefly derived from the Haviland notes (1) or Dr. Sjéstedt’s monograph (2), and the context indicates whether or not they have been confirmed. The opportunity is also taken to correct some misapprehen- sions the author was under in writing his notes, ‘‘ White Ants in Natal” (3). 2 Il. THE ABUNDANCE OF TERMITES. Owing to the fact that a number of South African species are to a great extent subterranean in their habits, little idea exists as to the great abundance of termitesin general. ‘This abundance may not be widespread, but it may be said of many parts of the country that the soil teems with termites. In some parts many species are found whose subways interlace without anastomosing, and all find sustenance in the same pastures. Where, in other parts, kinds are fewer, at least one sort tends to be particularly abundant. Again, upon the rock-strewn hills and mountain-sides a number of species, making their nests under stones, find a congenial environ- ment. As only a few points in the Union have been at all explored VOL. 3, PART 2. jan CLAUDE FULLER. for termites, and none very thoroughly, these generalities call for some verification. In the south-west Cape, where termites are regarded as rare and where, as at Cape Town, white-ant attacks upon wooden structures are practically, if not wholly, unknown, Hodotermes viator exists abundantly, and a brief search in the neighbourhood of Stellenbosch, made for the writer by Mr. C. P. van der Merwe, revealed this and three other species. Inthe Great Karroo Hodotermes karrooensis is almost ubiquitous and, in restricted areas, Hutermes trinervius also abounds; builders of large mounds, other than the last- mentioned insect, are seemingly non-existent, but there is little evidence on this point, and nothing is known of subter- ranean and rock-nesting forms. Over the whole of the central area of the Orange Free State, the abundance of EHutermes trinervius is phenomenal ; other forms are reported from various parts. In Natal and the Transvaal it is no exaggeration to say that the soil is riddled from end to end of the country with termite tunnellings, through which an inconceivable host of insects constantly passes to and fro. In two small areas at Pietermaritzburg (784 square yards each), the writer reported (3) finding fourteen and sixteen nests respectively, the homes of six different species. This, however, was an observation based solely upon surface indi- cations. Latterly, a favourable opportunity presented itself to observe subterranean conditions at Pretoria in the case of extensive excavations which were being made for the founda- tions of certain large Government buildings. These works, covering two to three acres, were visited almost daily, and many of the observations detailed later were made during their progress. It was found that the soil was inhabited by eight species ; to a depth of 4 to 5 ft. it was riddled with their galleries, and not a cubic yard existed which did not contain one or more cavities belonging to one species or another. OBSERVATIONS ON SOME SOUTH AFRICAN TERMITES. 333 Ill. THE BEHAVIOUR OF WINGED TERMITES. It may be taken as an axiom that every colony of the genus Termes is originally founded by the unaided efforts of a pair of winged insects. There is no acceptable evidence to the contrary; and the oft-quoted presumption that all adult termites are helpless and dependent upon adoption by a section of wandering workers is based upon the merest conjecture. Again, the aerial flight is often regarded as having for its chief objective the prevention of inbreeding ; that it may fail im this connection is indeed very obvious, although it must be admitted that the possibility of insects from different colonies mating is sometimes present. The real object of the production of sexual individuals in such vast numbers and their annual exodus seems only to be for the purpose of perpetuating the species ; the countless swarm of emigrants which emerge, most of them to meet with speedy death, has its parallel in the phenomenal number of eggs or of young produced by many organisms. In the case of two species many individuals have been seen to mate and at once to seek shelter in the earth within 6 to 8 ft. of the spot from which they had just emerged. Again, with these two species, and others, it has been noticed on occasion that an aerial flight takes place from one nest and not from others of a lke kind adjoining it, and that insects mating far afield were of the same colony. At times, vast numbers of this species or that, or several kinds together, take to the wing simultaneously over a great area of country; it is on such occasions as these that sexes of different parentage meet. At other times vast numbers emerge in restricted localities only, or from one nest only. The flying powers of winged termites do not appear great, and the majority do not fly far. At the same time, even in the case of those displaying the smallest powers of endurance, some few at least far outdistance the majority. Clumsily as they seem to fly many are well able to direct their movements 334 CLAUDE FULLER. and rapidly avoid obstacles placed suddenly in their way. As might be expected, those which take to the wing after dark are readily attracted by light. It seems almost certain that this phase in the history of a termite is a succession of events which follow upon one another in regular order, and, if the chain is broken at any point, it remains so. This may be illustrated by several specific cases. The night-flyng Eutermes trinervius, if attracted by illu- minating a white sheet, ahghts, and both sexes dealate after meeting. If some are captured at once before meeting and placed in a tube, they do not dealate. Some, so captured, have been kept over four days alive in _ confinement, and whilst a few dealated, or to be more correct lost their wings, no attempt at pairmg was made. ‘The same holds good for the dusk-flymg Termes incertus, but to a modified extent only. Further, with these two species and three others (undetermined) no attempt is made at burrowing if the sexes are kept separate. In the case of two males confined together it was noticed that one would frequently follow the other as if it were a female. The stimulation induced them to burrow together, but they soon desisted from doing so, although stimulated thereto several times by one another; ultimately they died on the surface of the soil. Again, a number of pairs of T. natalensis were captured one evening and placed in a small box; inthe morning it was found that the couples had lost all regard for one another; they were allowed to wander over a table, but none again mated voluntarily. When, however, the sexes were sorted out and the male placed in the position it had been in when captured, the sequence of events was re- established. In a further instance six pairs of Hodoter- mes transvaalensis were collected (December 50th, 1913) from burrows which they were making at the time, the couples were then placed in jars of soil, but were found to have lost the instinct to start burrowing again. They wandered aimlessly about, the males taking no notice of the females. When, however, artificial burrows were made, and the pairs placed in them side by side, the environment reacted upon them at once OBSERVATIONS ON SOME SOUTH AFRICAN 'TERMLTES. 335 and they burrowed to the bottom of the Jar. In 1914, similar results attended the mating and re-mating of this species. The act of pairing appears to differ in the details of its manifestation with each species, as the following accounts concerning several species will show. Hodotermes transvaalensis sp. n.—This species was noticed upon the wing after meghtfall in Pretoria on December 30th, 1918, and was again observed at 8 p.m. on December 7th, 1914. The actual meeting of the sexes was not seen. Many gathered about the electric street lamps, and beneath these some were seen already TExr-=riG. i, paired off with the females lead- ing, whilst other individuals, male and female, scurried rapid- ly about, as if in search of one another. As soon as a mated female began to dig, the male took his place by her side and seconded her efforts. Both in- sects are very expertexcavators, Section of initial burrow and cell made by a pair of H. trans- S it construct a burrow. Much of vaalensis. x i. and when once started soon the loosened soil is scattered by the feet, the action resembling a hen scratching. Pebbles of no mean magnitudeare loosened, carried outin the jaws,and placed well to one side of the pit’s mouth. The larger and grosser female displays just as much activity in this work as does the hither and less obese male. Their strength and their aptitude is best illustrated by the fact that they succeeded in burying themselves in the stones and grit of a macadamised roadway. If interfered with the insects become most excited, and pairs placed on soil in captivity fled around like cockroaches, taking no notice of one another. When, however, artificial burrows were made, and the insects were dropped head first into them, they became normal at once, and just as imstan- taneously began burrowing. aa0 CLAUDE FULLER. Actual burrowing in open land has not been observed, but under the uniform conditions of a box of soil and observa- tion cells, it was found that the insects did not close in their burrow behind them; instead they placed the loose particles regularly around the aperture so as to form a small per- forated mound, which persisted for several weeks (text-fig. 1). Some females laid eggs within the first week, others only after the lapse of a fortnight. On May 25th, 1915, one pair with young were still alive. Termes natalensis Haviland.— The mating of this species has only been observed on one occasion (Pretoria, October 28th, 1913, 6-7 p.m.), when it took place in conjunc- tion with a flight of T. incertus. The latter was, however, quite local, whilst the winged natalensis had come from far afield—probably a mile away. The insects flew high with well-sustained flight, and many passed overhead. The females, as appears usual, alighted first; selecting free and ligh- standing grass spears. Upon gaining a foothold (fifteen to twenty inches from the ground) they at once reversed their position and stood head downwards. Then all four wings were half opened and the apex of the abdomen became visibly swollen. Some males were flying low over the her- bage when this act took place, and in a very short space of time the females were discovered by them. The male alights, as a rule, directly on the dorsum of the female, flying to her with much precision. He rapidly aligns himself, with wings closely folded and head upward, and combs across and across from one cercus to the other the dilated apex of the abdomen of the female. During this the female sits quite still, but presently she dealates and gives evidence of restless- ness; when her movements sufficiently stimulate the male, he dealates and crawls from her back. The female then moves forwards down the stem, the male close behind, with mouth- organs always closely brushing the anal plates of the female. There is no doubt in connection with this mating that, whilst the male is attracted to the female by the sense of OBSERVATIONS ON SOME SOUTH AFRICAN TERMITES. 337 smell, as is the case with IT’. vulgaris, the directness with which he alights indicates further that, within a circum- scribed radius, the female is visible to him; the play of the rays of the setting sun upon her half-spread wings render- ing her quite a conspicuous object. On this occasion a number of dealated pairs were collected and confined over-night in a small cardboard box. Several TEXT-FIG. 2. The mating of Termes natalensis. A. Two views of female in calling attitude on grass-stem. B. Two views of association of 3 sexes. xX 2. of these were successfully re-mated the following morning, three pairs burrowing into the soil contained in large glass jars. All burrowed to the bottom, to a depth of between four and five inches; and, unable to proceed further downwards, contented themselves by making a more or less globular cell about an inch and a quarter in horizontal diameter and three-quarters of an inch high. Owing to this fortunate circumstance it was possible, after removing the dark shield placed around the jar, to obtain some small view of the actions of the insects. The first pair soon died; the individual 338 CLAUDE FULLER. which survived for a time buried its mate. On the four- teenth day one of the second pair died, and its mate was watched whilst it carefully encoffined the corpse in earthern paste. The third pair remained alive, displaying little if any anxiety when exposed to the light. On the sixteenth day (November 13th) a small bunch of yellowish eggs was noticed adhering to the glass side of the cell. On December 29th the first egg hatched, but whether one of those first laid (these appeared to become somewhat desiccated) could not be determined. '‘I'wo days later fourteen young termites were counted, and many eggs were still unhatched. Before they hatched, the eggs were the object of frequent attention by both parents, and they often mspected them, feeling them over with their palpi. ‘The young insects received constant attention from their parents, and apart from being fed by them they seemed to be constantly groomed. On the ninety- eighth day (February 2nd, 1914) some of the workers appeared to be adult, the heads showing yellow chitin. By this time the cell was a crowded mass of life and some eggs were also seen; at the same time no enlargement of the cell had been made and no galleries driven from it. Naturally, neither male nor female had taken any kind of nourishment, but they did not appear any the worse for their long fast nor emaciated by the feeding and raising of their young. This unfortunately is the last note made upon the colony. ‘The writer left headquarters the next day, and during his absence the soil was kept too moist, and the whole colony succumbed and rotted before his return. The burrowing of the pairs of this species resembles that of the smaller kinds ; but, in common with other larger sorts, it is capable of greater exertions, and in beginning the operations the insects can be seen scratching the loose particles with their feet and throwing them behind them, as do most burrowing animals. When a pebble is met with, it is not avoided but picked up in the jaws and carried out and deposited where it cannot roll back into the excava- tion. OBSERVATIONS ON SOME SOUTH AFRICAN TERMITES. 339 Termes latericius.—The mating of this species was observed in Pretoria at dusk on December 4th, 16th, and 18th, 1914. The females were seen to alight first, taking up an inverted position on pendulous parts of grasses and herbage. Immediately on assuming this attitude they begin to agitate their wings violently, and keep them in motion until a male becomes associated with them. How long a 9 TEXT-FIG. 3. The mating of Termes latericius. Aands. The female in the calling attitude, violently agitating her wings. c. The sudden discarding of the wings when the sexes meet. 2. female may go on agitating the wings without a male finding her cannot be said, but one watched for twenty minutes never stopped, and after dark was still unmated. The males seem to have great difficulty in finding the females; they fly low and incessantly over the tops of the grasses, and again and again approach quite near to a female, and even circle within a few inches around, only to fly off again far afield. When mating was accomplished during twilight it was noticed that the male flew in narrowing circles around the 340 CLAUDE FULLER. female and alighted near to her. Directly the male touches the female the wings of both drop with inexplicable sudden- ness, those of the female seeming to fall whilst still being agitated. After this the female leads the way to the earth, the male following close behind. Here they soon burrow into the soil, and there form a cell. Two pairs which burrowed. into observation-cells on December 16th had both produced fifteen to twenty eggs eight days later. They were trans- ferred to glass capsules, and on January 2nd had laid a further batch of ten to twelve eggs. Alive May 25th, 1915. Termes vulgaris Haviland.—The mating of this species has only come under observation once { Pretoria, December 4th, 1914). The departure of the adults from the nest and their mating occurred between the hours of 8 and 10 a.m., imme- diately after a heavy and continuous downpour of rain which had lasted over three hours. The flight took place from a nest in the banks of a stream, and, but for this fact, the nest could not have been located. The insects emerged from numerous apertures in the bank and in the surface overlying the site of the nest. Later in the day these were closed and covered with little mounds of soil. From the fungus-garden cavities to the apertures the insects travelled along specially excavated galleries, an inch and a half wide and a quarter of an inch high, driven upwards through the soil so as to form an inclined plane. These galleries are exactly similar to those made for the same purpose by T’. latericius, and form another connecting link between these two similar but quite distinct species. Upon leaving the nest the adults radiated out in all direc- tions from it, and mating commenced within the first 100 yards. The full periphery of the flight was not actually determined, but in one direction it extended for more than half a mile. No wind was blowing at the time, and there was no evidence of an exodus from another nest. Along the: radius of flight observed it must be conceded that all the mating was between insects of like parentage. OBSERVATIONS ON SOME SOUTH AFRICAN TERMITES. 9341 The females came to rest first, ahghting here and there upon grass stems—both long and short—and immediately took up a position at the apex of the stem, with head down- wards and wings closely folded. No particular inflation of the abdomen could be noticed, but it was obvious that some: sense other than sight alone guided the male insects. When a female alighted, a male was soon to be noticed fluttering low over the herbage. TEXT-FIG. 4. This flight was seen to be most erratic, although generally round and about the centre represented by the female. Often the male would fly close by and attempt to alight upon an adjacent stem —sometimes even doing so—-but only to fly off, low across the grass tops for ten to twenty feet. Ultimately the male manages to locate the female, and settles be- low her upon the same stem. He at once crawls up beside her and rapidly passes the mouth-parts across the region between the cerci and touching the cerci them- is selves. Almost simultaneously The mating of Termes vul- : ; : : ; i garis. Female in calling with this action both insects dea- attitude. x Z. late and the female begins to crawl downwards, the male following. In this attitude the insects progress until the female finds a site suitable for burrowing; the male never more than just keeps in touch with her. When burrowing commences the male takes his place beside the female, and in loose soil the pair soon become lost to sight. Pairs of this species did not thrive in the observa- tion jars, but they constructed cells, and eggs were laid within afew days. One pair with a few young May 25th, 1915. Termes incertus Hagen.—It would appear that shortly 342 CLAUDE FULLER. before the time arrives for the exodus of the winged indi- viduals of this subterranean species, a number of wide galleries or chambers are excavated within two to three inches of the surface of the soil beneath some bare spot, and in these the winged forms are assembled, ready to depart at the chosen moment. When this arrives several (one to five) perforations are driven upwards through the surface. From each of these first of all adventure 2-300 workers, but never any soldiers ; TEXT-FIG. 5. The flight of Termes incertus. x x. Underground assembly- 1 cavities shown in section. x 3. these workers swarm around the individual openings out of which they have come, spreading out to form a circular mat four to six inches across. In this circumscribed area they keep constantly upon the move, and none ever wander away from the main body. Their egress is immediately followed by that of the winged. Although the workers and the few soldiers that remain within the galleries hasten about among the imagos, per- > those chance instilling into them “the spirit of the hive,’ without pay no attention to them, their function seeming only to be the formation of a barrier—a living sacrifice—against OBSERVATIONS ON SOME SOUTH AFRICAN TERMITES. 343 the ants which gather viciously around. By their action they secure to the winged a safe departure. Overcome with anxiety to seize upon the winged insects, the excited ants take no notice of the worker termites, and it is quite exceptional to see one of these carried off until after all the winged insects have dispersed. The barrier is effective, however, because when an ant steps in among the workers to reach its most prized prey, it is so discomfited by the nips of the little termites upon its feet that it 1s compelled to retire. At Pretoria, both in 1913 and 1914, the imagos of this_ species emerged in great abundance on various dates through- out the month of November,' and always between the hours of 6 and 7 p.m. or during the brief twilight. The act took place almost invariably 10 to 24 hours after a fall of rain, the evenings selected being uniformly mild and quiet. Very many instances came under observation, but in none of these did one occur at any other time of the day, nor when windy, nor when rain was falling. On the other hand an emergence may often be followed within an hour or so by both strong wind and heavy rain. The exodus commences almost suddenly and ceases quite abruptly. When darkness actually falls the insects are no longer on the wing, so that the period involved is never more than 30 to 45 minutes. Numerous cases came under observa- tion in my garden, on the neighbouring kopjes and elsewhere. During one evening in 1913, and again in 1914, over a mile length of street was traversed, and along the whole length this species was emerging from countless places, in every garden, and at short intervals along the sidewalks. Such instances furnish some idea of how extensively this species inhabits the soil of Pretoria. The winged incertus issue forth from the soil very rapidly and jostle and crowd upon one another, in the narrow confines of the apertures, to such an extent that very often they are so densely packed as to form small upright columns from the ' The actual period is the last week in October into the first week in December. Another emergence occurred in March, 1914. 344 CLAUDE FULLER. apex of which those ahead take fight. As the insects all stream away in one direction, carried on some gentle air current, the emergence as a whole resembles a cloud of smoke issuing from a chimney. If one of the holes from which they are emerging is covered by a glass tumbler the action at once ceases and the workers retire. If a series of holes are so covered the effect is the same, but, very rapidly, fresh openings are made near by and the exodus proceeds. The females are the first to alight, and they do so directly on to the soil surface, always selecting a spot—large or small —which is bare of vegetation. On alighting they immediately dealate ; occasionally they then run forward an inch or two, but more usually stand still and, puffing up the apex of the abdomen, elevate it until one half of TEXT-FIG. 6. that region is at mght angles to the rest of the body. Without doubt some perfume is wafted upwards, for males flying low overhead are immediately ae Be Ae eae attracted and drop to the ground within % 15. 6 to 18 inches of the female. Usually only one alghts, but sometimes several do so. The male at once dealates and hurries with, comparatively speaking, remarkable precision towards the female, even should he fall amidst the grass. In one instance where the females were alghting the ground was overrun with ants to such an extent that as each fell it was captured. Over this area males were seen flying low (20 to 30 inches) and, to all appearance, systematically searching the area for the expected scent. This they did until darkness fell. The female does not remain in the calling attitude in- definitely, and the period is seldom more than a minute at the outside. She usually lowers her abdomen after a short while, runs a little way and then elevates it again. Directly she detects the immediate proximity of the male, and is certain that he is about to approach her, she hurries forward, and the male, catching up to her and placing his head upon the end OBSERVATIONS ON SOME SOUTH AFRICAN TERMITES. 345 of her abdomen, hastens along wherever she may lead. Occasionally two or even three males will follow one and the same female, and later the whole party will proceed to burrow into the soil quite amicably. Once the male has taken up his position, the female’s objective is some crack in the soil or some small pebble she can easily pass beneath. This reached she immediately begins to burrow, and the male, coming for- ward beside her, assists in the excavation. If a scratch is made in the soil, so as to surround an advancing couple, the female on reaching it will begin to dig; reacting at once to the stimulus of this suitable but artificial environment. The burrow is begun by hfting out the earth particles and placing them around the uf ‘. net ae TEXT-FIG. spot. Whenit is sufficiently deep to accommodate the full about length of the body three-quarters of an inch— the entrance is closed by a gradual process of building in the earth particles until ed adomeis made. The gallery T. incertus. i > ¢ ¢ @¢ i " A oo an * ae ~ 9 fi x oy oa ; . id i aa % j - a) i - ‘ la i Sw Vapad 2 s aw ; : i : /' 7 - > : ; : 7 > ~~ ye. ~~ , ( 4, i! - “ a Ps : ; > - ® P 7 7 : ai : cm: : : - , | ff - 4‘ ' ; i i ‘ ’ 7 ; 7 - ' Coa. 5 ; - : - . é ; fon ¥, a. Wane : : : ; ; i ‘ F : ; 4 ; ; ; ; i i : i ; ; ; ; ; ath ar - ' ' i4 ' - ‘ 4 : : : i” satin, ¥ y 7 : ° o : ' , ; P t) J a ce 7 =~ @ aa 9 7 4 i * 5 — pay Pa aaa 7 Ry 2 ‘ = — — 7 — — * a = - — a Ann.Natal Mus.Vol. II. HG.K del adnat : Huth, London. Cubaris burnupi sp.n. figs. 1-10. Cubaris natalensis sp.n. figs 1l—20. Cubaris longicauda sp n. figs. 21-381. Ann. Natal Mus. Vol. III. H.G.K. del adnat Hath, London Cubaris trilobata sp.n. figs. 1—9. Philoscia warreni sp.n. figs. 10—20. Philoscia dilectum sp.n. figs. 21—81. SOUTH AFRICAN BAGWORMS. 587 South African Bagworms. EprrorraL Note. THe study of South African bagworms presents many points of great scientific and economic interest; and it is hoped that there may be published in this Journal a series of papers which will treat on the structure, life-histories, and bionomics of the various forms. It should be remembered that “bagworms ” do not con- stitute a natural group of insects; they belong to several families of moths, and they are here dealt with together on account of the habit which the larve have in common of providing for themselves a case or bag in which they live during active life and in which pupation occurs. Certain marked resemblances in structure and instinct may be observed in the various forms belonging to different families ; and investigation is necessary in order to determine whether these are due to phylogenetic relationship, or are examples of convergence arising through like needs imposed by a similar environment. The rapid change in the habits of certain indigenous insects in attacking the exotic Black Wattle is a menace to the rising wattle industry of Natal, and at the present time bagworms are the most serious of the insect pests which infest the trees. The investigation of these insects has been deputed to Mr. C. B. Hardenberg, Government Entomologist ; and his paper 588 EDITORIAL NOTE. is published in this Journal with the approval of the Chief of the Division of Entomology and the consent of the Secre- tary of Agriculture for the Union. The present paper is the first instalment of the series, and it is preceded by systematic descriptions of the imagos by Mr. A. J. T. Janse and Mr. E. Meyrick. SOUTH AFRICAN BAGWORMS. 589 South African Bagworms; Notes on the Psychidse and on the genera Gymnelema and Trichocossus, with descriptions of five new species. By A. J. V. Janse, F.E.S.L. With Plate XLII. CONTENTS. PAGE PSYCHIDA . : : ; : } , . 590 CGCETICIN & : : 5 é : : a oul CLaAnia WIk.. : : : : . 592 Clania moddermanni fests : : ‘ 092 PSYCHINZ ; 595 ACANTHOPSYCHE Heyl. , . 995 Acanthopsyche (Dasaratha) amodd (Heyl. : 595 Acanthopsyche (Heceticoides) tristis sp. n. Sood Acanthopsyche (Metisa) alba sp.n. . : Logs PsycHE Schrank : : : 5 Bee) Psyche (Manatha) minions Chagas. Vie : . 600 Psyche (Manatha) subhyalina sp. n. . : . 601 CHALIINZ : ; 5 : : : . 601 Monpa WIHk. . ; : : : . 602 Monda delicatissima Wik. ; ; < . 603 Monda rogenhoferi Heyl. : : : . 604 COSSID At : : : : : : : . 604 GYMNELEMA Heyl. . : : : ; . 606 Gymnelema vinctus (WIk.) : . 608 Gymnelema stygialis Heyl. . : : . 610 Gymnelema imitata sp. n. : ‘ : Swill 590 Ae JS of SANSE. PAGE COSSID Ai—continued. Gymnelema stibarodes (Meyr.) ; : 5 ul TricHocossus Hmpsn. : , : : . 613 Trichocossus arvensis sp.n. . : . 6138 TINEID {see succeeding paper by Mr. E. Meyrick): ADELID [ditto]. PSYCHIDAL. The Psychide form a group of great phylogenetic and biological interest, as they exhibit many primitive characters which are fairly constant in the whole family, and the larve make little dwellings, in which they live from the time they emerge from the egg until they are mature, while the females remain in this bag and may even deposit their eggs therein. : Most lepidopterists place this family in the neighbourhood of the Cossidez, and the fact that some Cossid-like moths are also case-dwellers, together with the similarity in structure, certainly points to this, although the females of the Cossids are winged, which is never the case in the Psychids. There must also bea close relationship between the Psychids and certain groups of the Tineidz; in fact, some genera are placed by one author in the Psychids and by another in the Tineids. The latter view is, I think, correct, as the only similarity is that the females are wingless and that the larvee are case-dwellers. Considering that wingless females are found in several families not at all related to the Psychidex, and that we find case-dwellers in the genus Melasina for instance (which is certainly a true Tineid), we shall have to rely on structural characters im making a decision. I therefore remove the genera Epichnopteryx and Fumea from the Psychide. In both genera the hind legs have well-developed middle spurs, and these are, as far as I know, always absent, or at the most very rudimentary, in all of the true Psychids. It is true that the tongue is entirely absent in both genera ; but we also find this in several species SOUTH AFRICAN BAGWORMS. 591f of the Tineids ; moreover, the development of the tongue is,. | think, entirely regulated by the mode of life of the moth. By removing those species that have spurs we really cut out the whole of Heylaert’s subfamilies Canephorine and Psychoidine from the Psychide. ‘he characters of the Psychide, as I now understand the family, are as follows : Mare.—Head, thorax, abdomen, femora and tibiz densely hairy ; ocelli large; tongue rudimentary ; antenne strongly pectinated ; labial palpi short, covered with hair; maxillary palpi absent ; hind tibiz without any middle spurs, end spurs: very short if present ; wings covered with hair or hairy scales.! Fore wing: la forked with 1b; le usually becoming co- incident with 1) beyond middle; 5 more or less approximate to 4. Hind wing: le present; vein 8 connected with the cell by a bar or anastomosing with part of upper median or becoming coincident with 7. Frmate. — Wingless; without legs or well-developed antenne. Larve and females case-dwellers. This family is divided by Heylaerts into five subfamilies, of which I retain three, all represented in South Africa The following key to the subfamilies is based on that given by Sir G. Hampson in his ‘ Moths of India,’ vol. p- 290, who closely followed the classification as originated by Heylaerts. 1a. Fore wing with vein 1b sending several branches to inner margin I. Qcericin#. B. Fore wing with only one branch from 1b to inner margin or no branch at all é : 5, 2a. Fore wing with one branch from 1b to inner margin IJ. PsycHinz. B. Fore wing with no branch from 1) to inner margin, veins 1b and le anastomosing III. CHaLiinz. I. Subfamily GACETICINA H. Fore wing with vein le anastomosing with 1b, which emits | The wings are often apparently naked, but in several cases this has been proved to be due to flying. Bred specimens, killed before they could fly, show a certain amount of loose hairy covering. 592, A. J. VT. JANSE. several branches to the inner margin; veinlet in cell of both wings forked. At present only one genus is known from South Africa. Genus Criania WI1k. ‘Clania Wlk., Cat. iv, p. 963, 1855. Type, lewinii, from Australia ; Hampson, ‘ Moths of India,’ vol. i, p. 291. Eumeta Wik., Cat. iv, p. 964, 1855. Type, cramerii, from Ceylon. Cryptothelia Wlk., Cat. iv, p. 970, 1855. Type, consorta, from Ceylon and India. Lansdownia Heyl., Ann. Soc. Ent. Belg. xxv, p. 66, 1881. Type, (?) macleayi L. Guiid., from Central America. Mate.—Antenne bipectinate to tip; fore tibia with a long process ; wings large and broad ; fore wing with vein 4 and 5 stalked; 6 present; 8 and 9 stalked; hind wing with vein 8 sending several branches to the costa. This genus has a very wide range, having representatives in Central America, Africa, Asia, and Australia. Clania moddermanni (Heyl.). Pl. XLIII, fig. 1. Eumeta moddermanni Heyl., Ann. Soc. Ent. Belg., xxii, p. lix, 1888; Kirby, Cat. i, p. 503. Mate.—Head, thorax above and underneath covered with whitish hairs, mixed on the underside with fuscous (XLVI)! hairs ; antenne and branches fuscous; legs covered with fuscous hairs mixed with whitish hairs; a fuscous spot at frons; base of antennze surrounded by pinkish buff (X XIX) hairs; a fuscous line from between the antenne, becoming forked at base of tegule and continued as a single line over the thorax ; tegule edged by a fuscous line transversely till it reaches an oblique lateral dorsal lne of same colour; Ist segments of abdomen with long whitish hairs in middle, and with a tuft of fuscous hairs, the remainder of abdomen on upper side with cinnamon-buff (X XIX) lateral hairs, and those on terminal segment fuscous or fuscous-black ; abdomen on underside with two fuscous stripes. ‘ All figures in brackets after the name of the colours refer to Ridgway’s ‘Color Standards and Nomenclature,’ edition 1912. SOUTH AFRICAN BAGWORMS. 593 Fore wing thinly covered with fuscous hairs and scales ; costa with a fuscous-black edging; upper and lower median, disco-cellular, vein 1b and veins 2 to 9 thickly covered with fuscous-black scales; postmedial area between veins 2 to 4 and 7, 8 and terminal area between 4—5 and 6-7 very thinly scaled so as to represent whitish macule when the moth is placed above a light underground ; inner margin edged with fuscous-black from one-third ; cilia very short, fuscous, mixed with fuscous-black. Hind wing covered evenly with fuscous scales and hairs ; all veins, except la and lc, sharply defined with fuscous- black ; cilia fuscous. Underside of both wings lke upper- side, but without the dark scaling on the veins. With expanded wings 26-30 mm. Has.—Ladysmith (September 10th, 1915, bred by C. B. Hardenberg) ; New Hanover (September 23rd, 1915, bred by ©. B. Hardenberg) ; Pretoria (bred by A. J. 'T. Janse). The description given by Heylaerts is very good, and leaves no doubt about the correctness of the identification. In structure this species differs from the Indian species variegata in the following details : Fore wing with the branches from 1b to inner margin, not from one point, but at intervals from one-fourth to two-thirds of the length of this vein, and the branches do not reach the inner margin ; stalk of 4 and 5 over one-third, nearly half; 7 remote from stalk of 8, 9; hind wing with le very faint; 4,5 on a stalk of nearly one-third ; 7 not connected by a bar to 8; a very oblique bar from half upper median to 8 ; some- times a second short angled erect bar from upper median at three-fourths to the junction of first bar and vein 8; veinlets from 8 to costa very faint; process of fore tibiz slightly shorter than tibie. SPECIES AUCTORUM: Kumeta cervina Druce, Proc. Zool. Soc. Lond., p. 684, 1887. The following description is given by Druce of this species: “Male like E. cramerii, but the primaries redder in colour, 594, ee 1 ANSE with a marginal row of fine whitish spots, and a dark brown elongated patch at the end of cell. Secondaries smoky brown, with three whitish dots at the apex. Head, thorax and abdomen brown. Antenne brownish black, tegule with white tips ; a white spot at the base of the primaries ; on the under- side the white marginal spots are more distinct than above. Expanse 1? in. Hab.—West Africa; Gambia; Hast Africa; Delagoa Bay. It is impossible to make out from this description if the species is a true Humeta or not, so I leave it provisionally in this genus. Mr. Junod states in his paper (‘ Bull. de la Soc. Ent. Belg.,’ vol. xxvii, 1899, p. 244) that Humeta cervina is the same as E. moddermanni. He seems to make this statement on the authority of Dr. Heylaerts, but after comparing the descrip- tions carefully, 1 must come to the conclusion that this is very unlikely. Of course, the description is very short, but even as it is, I do not think it applies in many respects to K. mod- dermanni. The figures Junod gives (pl. iv, fig. 2a, b) of his specimen and the bag undoubtedly reter to EK. modder- mann. SPECIES OMITTED : Kumeta junodi Heyl., Ann. Soc. Ent. Belg., vol 34 po Cxxx. This species has to be placed in the genus Acanthopsyche. Kumeta zelleri Heyl., Ann. Soc. Ent. Belg., vol. 38, p- xevii; Kirby, Cat. 1, p. 503. I do not think that this species can be maintained, as the description is given of the larva and bag only. The deserip- tion of the larva is much too general to make the identification certain. The larva and bag do not appear to be sufficiently characteristic. I think that the description of the imago only can be considered as valid and not that of eggs, larve, or pupe, still less of the structure made by the larva to live in. SOUTH AFRICAN BAGWORMS. 595 II. Subfamily PSYCHIN 2. In this subfamily the veinlets in the cells may be forked or single; the process of the fore tibiz is absent or present ; the fore wing has always one single branch from 1b to the inner margin. This subfamily is divided into two genera: (1) Acantho- psyche with a long thin process on the fore tibiz and (2) Psyche without such a process. Heylaerts has divided most genera ofthe Psychide into a number of subgenera and these have been adopted also by Sir George Hampson in his “Moths of India.’ It is not quite clear to me why some of these groups do not receive a generic position ; some of them differ more from each other than some genera do in other families. Genus AcanrHopsYcHE Heyl. Acanthopsyche Heyl., Ann. Soc. Ent. Belg., p. 66, 1881. Type, inquinata Led., from Europe; Hmpsn., ‘Moths of India,’ vol. i, p. 293. Ma.e.—Branches of antenne getting shorter from middle, but they are still present at the tip; fore tibiz witha long process. Subgenus Dasaratha Moore. Dasaratha Moore, P.Z.S.L., p. 396, 1888. Type, himalayana Moore, from Punjab; Hmpsn., ‘ Moths of India,’ vol. i, p. 294. The following description of this subgenus is given by Hampson : Mare.—Fore wing with veins 1b and le anastomosing ; vein 6 present; hind wing with vein 6 absent; a bar between veins 7 and 8. ‘The veinlet of the cell of both wings forked. Acanthopsyche (Dasaratha?) junodi Heyl. (Wattle Bagworm). Pl]. XLITI, fig. 2. Eumeta junodi Heyl., Ann. Soc. Ent. Belg., vol. 34, p. exxx, 1890. Mate.—Antenne and their branches, hairs on legs, head, thorax and abdomen on upper- and underside fuscous. Both 596 Aces. ‘0. TANS: wings glass-like and without any colour ; the costz narrowly edged with fuscous-black ; fore wing from base to medial area and along inner margin as far as postmedial area very thinly covered with long fuscous-black hairs; hind wing with still longer hairs of same colour on inner marginal area as far as 1c; cilia of both wings consist of a few ill-developed fusecous scales. Underside as above, but with less hairs. Exp., 24-26 mm. Has.—Probably found all over Natal at altitudes below 5000 ft., and may be expected all over South Africa where wattles are grown. Its original food-plant was doubtless one of the indigenous Acacias. The subgeneric position of this species is rather uncertain, and it may be necessary to form a new subgenus, as it does not fit well in any known to me. It has to be taken out of the genus Kumeta for certain, as vein 1b has only one spur in the fore wing and that spur is very short and faint, never reaching the inner margin in any of the specimens I[ have seen. The long process of the fore tibiz is present, though rather hidden in the long hairs. The venation of the fore wing is the same as in Dasaratha longicauda Warr except for the spur from 16 which comes in junodi from near end of fork of la and 1b and in the other species from near end of fork 16 and le. In the hind wing all the veins are free in junodi, though 4 and 5 come from a point; 6 is present, the missing vein is distinctly vein 7; 8 is free from base, almost parallel to upper median till half length of cell, then it anastomoses with the upper median and vein 7 for its whole length; Ic is very faint ; frenulum very long, nearly half length of costa. The hair-scales of junodi are very loose and come off at once during flight, but even in perfect specimens the hairy covering is very thin. Subgenus Giceticoides Hinpsn. Oiketicoides Heyl., Ann. Soc. Ent. Belg., p. 66, 1881. Type inqui- nata Led., from Europe. (ceticoides Hmpsn., ‘Moths of India,’ vol. i, p. 293. SOUTH AFRICAN BAGWORMS. 597 Hampson gives the following description of this subgenus = Matre.—Fore wing with vein le anastomosing with 1b ; vein 6 absent in both wings. Hind wing with a bar between vein 7 and 8. This subgenus is divided into three sections : I. Veins 4 and 5 of both wings stalked. Fore wing with 7 from cell. Il. Veins 4 and 5 of both wings from cell. Fore wing with 7 stalked with 8, 9. III. Veins 4 and 5 of fore wing from cell. Fore wing with vein 7 from below angle of cell. Acanthopsyche (Hceticoides) tristis sp. n. (Thatched Bagworm). Pl. XLIII, fig. 3. Mate.—Shaft of antenne and hairs on head, thorax, abdomen and legs cinnamon-buff (X XIX), these hairs are mixed with those of a fuscous colour, especially at the last three seg- ments of the abdomen ; branches of antennz fuscous ; wings moderately covered with fuscous hairy scales, leaving the terminal half of fore wing and the terminal third of hind wing slightly more transparent; in fore wing the veins 2-9 are rather darker in this area; costa of fore wing and termen of both wings fuscous-black. Underside like above. Exp., 20 mm. in 1 male type and 1 male cotype. Has.—New Hanover, Natal (bred by C. B. Hardenberg,. March Ist, 1915). This species comes in Section IIT of the subgenus. The bar of the hind wing is, however, not from vein 7 to 8, but from the upper median at two-thirds to vein 8. It is quite well possible that this bar is actually vein 7, which would then become concurrent with vein 8, and if this is the correct view, vein 6 would be present. The veinlets in both cells. are simple; the branch of 1b in fore wing is distinct, but hardly reaches the inner margin; the stalk of 8,9 is only one-fourth and 7 is so far away from the stalk that it might as well be vein 6. The fork of 1b is very long, nearly half the length of vein. Vein 1b of hind wing is forked and 598 A. sds > JANSE, le is very faint; 4 and 5 come from a point, and the disco- cellular is very oblique. Shape of fore wing rather broad and short, costa arched, apex rounded, termen rather erect, nearly straight, tornus somewhat acute. Hind wing rather large, rounded; costa, termen, mner margin, apex and tornus well rounded. The process of fore tibia thin and shghtly longer than the tibia. Subgenus Metisa W/k. Metisa Wik., Cat. iv, p. 957, 1855. Type, plana WIk., from India; Hmpsn., ‘ Moths of India,’ vol. i, p. 296. Babula Moore, 1.A.8.B., lix, pt. 2, p. 262, 1890. The following description of this subgenus is given by Sir G. F. Hampson in his ‘ Moths of India’ : “Mate.—Fore wing with veims 1b and le anastomosing ; vein 6 present. Hind wing with vein 6 present ; 8 comcident with 7! to the end of cell.” Acanthopsyche (Metisa) alba sp. n. Pl. XLIII, fig. 4. Mate.—Antenne fuscous-black ; head and thorax covered with long white hairs; abdomen and legs thinly covered with white and fuscous hairs; wings transparent, without any colour and without any scales or hairs? except along the immer margin of the fore wings, where there are some short whitish hairs, and the same area of the hind wings on which the hairs are longer and white; cilia represented by some short scales only, in both wings fuscous as far as 1b and beyond this vein more whitish; a fuscous edging along the costa in both wings. Fore wing long and rather narrow ; costa nearly straight ; apex and tornus rounded ; termen oblique, shghtly rounded between veins 3-6, and shghtly hollowed out at veins 2-3 ; ' According to the figure given, Hampson means here the upper median and not vein 7 itself, which really begins from the upper angle. 2 It is quite possible that a specimen just emerging has some loose scales and more hairs on the wings, but the bred specimen I have seen also shows no sign of them. SOUTH AFRICAN BAGWORMS. 599 fork of 16 fully two-thirds; 4 and 5 stalked for half their length; disco-cellular very oblique; 7 some distance apart from stalk of 8,9; stalk of 8, 9 nearly half of 9; cell narrow, long, and with two faint veinlets. Hind wing subtriangular; costa much arched, termen slightly hollowed out at le to 2; inner margin arched; apex and tornus well rounded; le very faint; 4 and 5 on a stalk of half and from lower angle; 8 free from base, then anastomosing with upper median just before half of that vein till near vein 7; cell rather broad and long; a rather well-defined veinlet in middle of cell from near base till below half disco-cellular ; a fainter veinlet below it from before half the distance of stalk of 4, 5 and vein 6, where the disco- cellular is angled. Exp., 29 mm. in type; 25 mm. in Durban specimen. Has.—Type, Nkwalem, Zululand (Janse, caught at leht, January 10th, 1916) ; also another specimen (cotype) from Durban (bred by E. HE. Platt, November 4th, 1914). Mr. Platt sends me the following information about this species: ‘I found the bags numerous beyond Sydenham in the month of May, and I have noted that it feeds on Wattle or other Acacia. I remember this moth was dead in the breeding-cage when I found it, and concluded it had beaten its scales off. I did not keep the case.” Genus Psycue Schrank. Psyche Schrank, Fauna Boiea, ii, 2 Abth.. p. 87, 1802; Hampson, ‘Moths of India,’ vol. i, p. 297. The following description is given by Sir G. Hampson : “ Mate.—Antenne bipectinated to tip, the branches short or long. Fore tibiae with no spine. Fore wing with 10 to 12 veins; vein 1h and le anastomosing or separate. Hind wing with 7 to 8 veins.” The specimens which I place in this genus have all lost their fore legs, so that I cannot be certain about the generic VOL. 3, PART 3. 4.0 600 AT. Tei JANSE. position, but all the other characters taken together leave very little doubt that they must be placed in this genus. Subgenus Manatha Moore. Manatha Moore, A.M.N.H. (4), xx, p. 346, 1877. Type, albipes, from Ceylon ; Hampson, ‘ Moths of India,’ vol. i, p. 298. The following description is given by Hampson : “ Mate.—Fore wing with vein 1b and lec anastomosing; 4, 5 and 8, 9 stalked; vein 6 from near upper angle of cell. Hind wing with vein 6 present; a bar between vein 7 and 8,” The three specimens which I take to be M. ethiops Hmpsn., bred by Mr. Hardenberg from a long thin thorn- like bag, vary a little in their venation. One specimen (No. a) has 4 and 5 of fore wing on a short stalk ; another (No. b) has these veins springing from one point; the third specimen (No. c) has these veins far apart. Specimen b has the left hind wing with vein 5 and 6 on a short stalk. In none of the three specimens is vein 8 of hind wing like it is in the typical Manatha; there is no bar, but vein 8 is free and more or less parallel to the upper median for half upper median, then it anastomoses with that vein and with vein 7 for about one-fourth of 7, then it goes obliquely to the costa. Psyche (Manatha) ethiops Hmpsn. Pl. XLIII, fig. 5. Manatha ethiops Hmpsn., A.M.N.H. (8), vol. vi, p. 116, 1910. I have not been able to compare the three specimens | have with the type specimens (from the Cape Colony), but they correspond sufficiently with the short description to leave little doubt about the correctness of the identification. My specimens, however, show no white scaling on the body ; but they are badly rubbed on head, thorax, and abdomen. The wings are thinly covered with buffy-brown (XL) hairs. Has.—Three male specimens, bred by C. B. Hardenberg, from bags collected in Westfalia (Zoutpansberg District). SOUTH AFRICAN BAGWORMS. 601 They emerged in Pretoria on April 20th, in 1910. In collection Janse. Psyche (Manatha) subhyalina sp.n. Pl. XLIII, fig. 6. Mate.—Fore wing with branch of 16 very long and oblique ; 4 and 5 on a stalk of one-sixth ; 6 from a little below middle of disco-cellular ; 7 slightly stalked with stalk of 8,9, or from a point and from upper angle; 8 and 9 on a stalk of nearly half of 9; 10 and 11 free and parallel to 9; a single veinlet in cell from below 6. Hind wing with 4 and 5 on a stalk of one-fifth and from lower angle; 6 equally far from 5 and 7; 7 from upper angle ; 8 connected with a bar to upper median at four-fifths ; disco-cellular very oblique. Head, thorax, and abdomen above and underneath rather thinly covered with avellaneous (XL) hairs; antenne and the long branches blackish-brown (XLV); wings subhyaline, thinly covered with avellaneous hairs, mixed with olive- brown (XL) hairs, so as to give a cupreous gloss in certain hght. Fore wing with the costa more thickly covered with hairs, so as to become fuscous-black (XLVI); veins fuscous (XLV). Hind wing like fore wing, but costa not dark and inner- marginal area glossy cimnamon-buff (X XIX). Cilia glossy cinnamon-buff, with a darker sub-basal shade. Under-side as above, but more thinly covered with hair. Exp., 15-1775 mm. Has.—T ype from Nkwaleni, Zululand, January 10th, 1916 ; three cotypes, one from the same locality, January 11th, 1916; two from Umkomaas, January 18th and 24th, 1914 (Janse, all caught at light). III. Subfamily CHALIIN A. This subfamily is characterised by Hampson (‘ Moths of India,’ vol. i, p. 300) as follows: Fore wing with vein 1b and ¢ anastomosing, and not sending 602 A. J. T. JANSE. any branches to inner margin. Veinlet in cell of both wings: single or forked. Hind tibize with no spurs. The genus Monda, which I propose to place temporarily in this subfamily, has some characters which make its posi- tion rather uncertain, and may at some future time give rise to the creation of a new subfamily. The running of vein le in the fore wing is quite different from any Psychid known to me. Instead of originating free and then anastomosing with 1b or remaining free, it 1s stalked with 1b for nearly half the leneth of this vei, which is short, and ends in the inner margin: le then runs parallel to lower median, and ends. above the tornus. No spurs or veinlets are given off from these veins. Genus Monpa WHk. Monda WIk., Cat. xxxii, p. 406, 1865. Type, delicatissima WIk..,. from South Africa. Proboscis absent ; palpi very rudimentary; antennz short, less than one-third of costa, bipectinate, branches as long as one-third of length of shaft, suddenly becoming shorter at tip, closely ciliated, and each branch ending in one long bristle ; fore, mid, and hind legs short, slender, without any processes or spurs, and very sparsely covered with rather long hair. Fore wing triangular; costa slightly hollowed out; apex well rounded ; termen erect, somewhat undulating at the veins; tornus well rounded; inner margin nearly straight ; cell long, broad at median part of wing; disco- cellular very oblique. Vein 1) simple at base, short, ending at inner margin at two-thirds; le stalked with 1b for nearly half the length of 1), then curved upwards and ending above tornus; lower median curved upwards; 2, 3 and 4 far apart and at equal distance ; 5 from 4 as far as halt distance 3 to 4 and continued in cell as a veinlet; 6 from middle of distance vein 5 to stalk of 8, 9; 7 absent ; 8 and 9 on a stalk of three- fourths, originating from upper angle; 10 absent; upper median well curved downwards from base to two-thirds; IL SOUTH AFRICAN BAGWORMS. 603 from upper median at over half of wing length, short and anastomosing with 12 at three-fourths of vein 12. Hind wing subtriangular, nearly as broad as long; costa shghtly arched; apex well rounded; termen rounded and ‘somewhat undulating at the veins; tornus well rounded; imner margin nearly straight ; frenulum thin, about one-fourth of costa in length; cell only as long as half the wing and rather narrow. Vein lw rather long and curved; 1b long, shghtly curved; le faint, nearly straight and ending in termen near vein 2; 2 from two-thirds upper median ; 3 from nearer to lower angle than from vein 2; 4 from lower angle; 5 from near middle of disco-cellularand continued as a veinlet in the cell; 6 absent; 7 from upper angle; upper median connected at half to vein 8 by a short bar, which is oblique and directed basally. Four South African species, of which I have seen only two, belong to this genus. Monda delicatissima Wilk. Pl. XLIII, fig. 7. Monda delicatissima WIk., Cat. xxxii, p. 407, 1865. Mate.—Antenne white, ringed with fuscous; branches white with fuscous pecten ; head, thorax and abdomen black and sparsely covered with whitish hairs. Wings whitish transparent, sparsely covered with rather short white hairs ; apical area rather thickly covered with fuscous hairs, from beyond half along costa, from before disco-cellular and till near vein 3; fuscous scaling between vein 3 and 4 rather thin; two rounded fuscous spots a little distance away from disco-cellular and lower median, one between vein 2 and 3 and one between vein 3 and 4; cilia consisting of rather long white hairs sparingly distributed. | FrmaLe.—W ingless. Exp., 17 mm. I bred this interesting little moth from small bags supphed to me by Mr. E. E. Platt, Durban. It feedson Desmonodium incanum DC. 604 Ay: 1 TANSE. Has.—Durban ; Singerton, Barberton District, in June (kK. Munro). Monda rogenhoferi Heyl. Monda rogenhoferi Heyl., C. R. Soc. Ent. Belg., vol. 34, p. elxxxii, 1891; Kirby, Cat., p. 518; Junod, Bull. de la Soc. Neuch. des Sciences Nat., vol. xxvii, p. 250, pl. iv, fig. 5, 1899. Has.—Mozambique, Delagoa Bay ; Durban, Natal. A specimen bred from a bag collected by the author in the Stella Bush emerged at New Hanover on April 24th, 1916. In collection Janse. SPECIES AUCTORUM : Monda major Heyl.,C. R.Soc. Ent. Belg., vol. 34, p. clxxxu, 1891; Kirby, Cat., p. 518, Junod, |. c., p. 249, pl. iv, fig. 6a, b. Has.— Mozambique, Delagoa Bay. Monda heylertsi Junod, Bull. de la Soc. Neuch. des Sciences Nat., vol. xxvii, p. 250, pl. iv, figs. 7a, b, c, 1899. Has.— Delagoa Bay. GVENERA AND SPECIES OMITTED: Fumea trimenii Heyl. (Tineide). Hpichnopteryx transvalica Hmpsn. (Tineide). These species were placed by their respective authors in the Fumeine as a subfamily of the Psychide, but as they have well-developed middle spurs on the hind tibize I exclude them trom the true Psychids. COSSID Ai. It is with some hesitation that I keep the genera Gymne- lema and Trichocossus in the family Cosside, but at the same time I do not feel at hberty to follow the other alter- native, namely, to place them in the Tineidz, though it is true that the resemblance to the latter family is very great. Heylaerts placed his genus Gymnelema in the Tineidz between the genera Melasina Boisd. and Diplodoma Zell. (both true Tineids),in which genera most species, if not all, live in portable cases. SOUTH. AFRICAN BAGWORMS. 605 In the Cossidz the larve live as a rule in the stems of trees and smaller plants, making galleries in which they also pupate. No true Cossids are, as far as I know, free feeders, ” that have some Cossid characters is certainly a point of biological and phylogenetic interest. and to find “ bagworm-moths The only characters that throw the genera Gymnelema and Trichocossus out of the Cosside are the presence of well-developed spurs on the hind tibizw, and the under- development of vein lc of the fore wing. The venation of both wings resembles otherwise, perhaps, more the typical Cossid wing than the Melasina wing. If we follow Meyrick’s classification as given in his ‘ British Lepidoptera,’ we come to the conclusion that the two genera in question have to come in the Tineina-group, and not in the Psychina-group (in which the Cossids come), as the middle spurs of the hind legs are present and well developed. On the other hand, in the Tineina vein 5 of fore wing is as a rule less close to 4 than is the case in Gymnelema and Trichocossus. Also, m these two genera the terminal jomt of the palpi is very short, and the whole palpus is covered with long spreading hairs, like a Psychid-palpus, while in the Tineids the terminal joint is usually rather lone and the palpus is covered with dense scales. This is the case even in the allied genus Melasina, but in Diplodoma the scales are rather loose. Sir George Hampson, in describing the genus Tricho- cossus (A.M.N.H. ser. 8, vol. vi, p. 134, 1910), states: “This eenus may be placed in the 'Tineidw when the limits of the families are better defined.” Of the genus Gymnelema (loc. cit.) he writes: “The genus belongs to the undefined borderland between the Cosside and ''ineide; the distinc- tion will probably be proved to be that the former has vein le of the fore wing fully developed and reaching the termen or becoming coincident with 1b; the latter slight and not reaching the termen.” If I knew that placing the genera in question in the 606 A. J. T. JANSE. Tineidz would satisfy Mr. H. Meyrick, I should prefer to do so, as they are more related, I think, to the Tineide than to the Cosside. It may be, however, that at a future date a separate family will have to be formed for Gymnelema, Trichocossus and their allies, as the two genera known to me have certainly several characters peculiar to them as a group. Pending Mr. Meyrick’s view I keep the genera in the family in which they are placed by Sir George Hampson. The characters of this group (family ?) are as follows: In male and female, head, thorax, abdomen, femora and tibiz densely hairy ; antennz pectinate in male, simple, serrate and ciliate in female; maxillary palpi and ocelliabsent; labial palpi short, covered with long spreading hairs; fore tibize without a process; mid tibiz with two well-developed end spurs ; hind tibie with the mid and end spurs well developed ; wings covered with scales. Fore wing with la forming a long fork with 1b, which is more or less curved upwards and comes rather close to 2; le present for at least half length of wing, then becoming obsolescent; cell long and rather narrow, with a simple or forked veinlet; 2 very near to lower angle; 5 rather close to 4; 5 to 12 present. Hind wing le present; 2 to 7 free, veins 2 to 8 present ; cell with a forked veinlet ; 8 free, more or less parallel to costa. The larvee are case-dwellers. Genus Gymnetema Heyl. PI. XLIII, figs. 8, 9, 10. Gymnelema Heyl., Ann. Soc. Ent. Belg., vol. 35, p. ceclxxv, 1891. Type, rougemonti Heyl.; Hmpsn., A.M.N.H., ser. 8, vol. vi, p. 134. The following description is made from G. vinectus H7/k.: Proboscis absent; palpi hardly reaching frons, ascending, three-jointed, third joint very short, acute; antenne about half length of costa, bipectinate in male ; pecten about three times thickness of shaft, gradually getting shorter to tip and SOUTH AFRICAN BAGWORMS. 607 on one side densely ciliated; antennze of female shorter than half of costa, serrate and ciliated; fore tibiz without a process ; mid tibize with end spurs, inner spur slightly shorter than outer; hind tibiz with very long mid and moderate hind spurs; tarsz with long smooth scales. Fore wing of male about twice as long as broad; costa and inner margin nearly parallel; costa gently arched ; apex well rounded ; termen oblique, shghtly arched ; tornus rounded; inner margin first straight, then well curved towards base. Fore wing in female longer, nearly three times width of wing. Vein la forms a fork with 1b of over one-third of 1b; Lb much curved towards upper median at two-thirds; le faint and only present for about half, never reaching termen ; lower median slightly curved upwards; cell nearly two-thirds of wing; 2 from beyond seven-eighths of lower median ; 3 from lower angle; 4 and 2 from nearly the same distance of lower angle ; 5 from nearer 4 than 4 from 5 and curved at middle ; 6 from two-thirds disco-cellular, which is a little curved between 5 and 6; 7 from near upper angle or from areole ; areole formed by 8 or by stalk of 7, 8 and stalk of 9,10; 9 and 10 from areole, either stalked or free; 11 from near and before middle of upper median ; 12 parallel to [1 ; fork of veinlet in cell very long, from half to one-third (in G. stygialis the upper part of the fork is absent). Hind wing triangular, rather longer in female than in male ; costa arched at middle; apex rounded ; termen gently hollowed at between 2 to 6; tornus well rounded; inner margin slightly hollowed out; vein le rather faint; cell half of wing; 2 from two-thirds of lower median; 4 from lower angle; 3 from middle of 2 to 4; 5 from one-third, 6 from two-thirds of disco-cellular; 7 from upper angle; upper median curved beyond middle; 8 somewhat parallel to costa. The type of this genus is unknown to me and comes from Delagoa Bay. 608 A. J. T. JANSE. Gymnelema vinctus (IW/k.) (Crossed-stick Bagworin). Pa ties 10: Cossus vinetus Wlk., Cat. xxxii, p. 583, 1865. Trypanus vinctus (WIk.), Kirby, Cat. i, p. 863. Cossus incanescens Butl., A.M.N.H., December, 1875, p. 402. Gymnelema vinctus (Wik.), Hmpsn., A.M.N.H. (8), vol. vi, p. 134. Mare.—Head and thorax on upper- and underside covered with white and whitish hairs mixed with some fuscous (XLVI) hairs; terminal edge of tegule with fuscous hairs forming a transverse line over the thorax, which becomes broader in middle; shaft of antenneze white, with short (3 times shaft) branches of acream (XVI) colour; legs covered with fuscous hairs mixed with white; abdomen at base whitish mixed with fuscous, terminal two-thirds of abdomen fuscous, hairs on terminal segment whitish, underside with whitish and hair-brown (XLVI) hairs, except last two seg- ments, which have white hairs, lateral hairs on underside fuscous and fuscous-black. Fore wing white; a broad band of drab (XLVI) scales from below upper median to inner margin between sub-basal and antemedial lines; another oblique band of the same colour, its inner edge from before subterminal at costa to near outer edge of first band at inner margin, its outer edge from subterminal line at costa to postmedial line at inner margin; the two bands become more or less confluent at below lower median, so as to form a semi-circular light patch between that vein and the inner margin; fuscous-black patches on costa at sub-basal, antemedial, and postmedial regions; a number of irregular, broken transverse lines of fuscous and fuscous-black scales, especially in medial and terminal regions; a fuscous terminal line; cilia of fuscous and white scales with a basal line. Hind wing drab-grey (XLVI); a number of long fuscous hairs on inner marginal area; terminal line and cilia as on upper wing. SOUTH AFRICAN BAGWORMS. 609 Underside : both wings covered with drab (XLVI) scales; terminal line very pronounced ; costal and cellular regions of fore wing with fuscous scales. Frmate.—Fore wing more elongate; broad fasciz darker, as they are irrorated with fuscous; costal markings and strigule less sharply defined and of a fuscous colour; strigule very diffused at beyond postmedial line; hind wing hghter; underside as in male except that the costal markings of fore wing are more pronounced, especially at apical regions ; head, thorax, abdomen and legs covered more uniformly with hair-brown (XLVI) and fuscous hairs. Exp., male 25 mm., female 34 mm. Has.—New Hanover, Natal (bred by C. B. Hardenberg in January and February from the “ Crossed-stick ” Bagworm). The specimens of the following localities were all caught at heht: Durban (Green, Cooke); Sarnia in January and February (Wilhamson, Janse). Maritzburg in January (Janse); Umkomaas in January (Janse). The venation 7 to 10 of fore wing varies a little in this species. | have nine males and four females that have 9 and 10 on a stalk from the areole, and three males that have these veins free from the areole. In one male (No. 988) the left wing is normal and 9 and 10 on a stalk from the areole, while the right wing has 7 and 8 ona short stalk from the upper angle, 9 and 10 on a long stalk from the upper median, and no areole is formed at all. Though I have not been able to get my specimens compared with the type, I have httle doubt that my identification is. correct, as the description of Walker as well as that of Butler fit my specimens well enough if one allows for more or less pronounced marking. My specimens agree best with Butler’s description, whose species, according to Hampson, is a synonym of Walker’s. vinetus. None of these three writers mention that the Jarva of this species makes a bag. 610 A. J. '. JANSE. Gymnelema stygialis Heyl. Pl. XLIII, fig. 8. ‘Gymnelema stygialis Heyl., Ann. Soc. Ent. Belg., vol. 35. p. ceclxxv, 1891; Hmpsn., A.M.N.H. (8), vol. vi, p. 134. Frmate.—Head, thorax, abdomen and legs covered with tuscous-black (XLVI) hairs on upper- and underside. Fore wing dusky purplish-black (LII]), regularly irrorated with ‘groups of large white scales, each group consisting of three to six or more scales; the groups are arranged in such a manner as to form transverse interrupted lines which are rather close together ; cilia fuscous-black, darker at base and thus forming a well-defined dark line parallel to termen. Hind wing rather thinly covered with dusky purplish- ‘black scales; cilia as in fore wing. Underside of both wings fuscous-black ; on costa of fore wing about six groups of white scales from one-third of costa to near apex, distance between the spots increased from base towards apex. Ovipositor long, antimony-yellow (XV), surrounded by long fuscous-black hairs. Exp., female 27°5 min. Has.—Bred from a bag found by the author at Waterval Onder, ‘l'ransvaal, emerged on January 9th, 1911. The following description is given by Sir G. F. Hampson (loc. cit.) of the male, which is unknown to me, “Wines shorter and comparatively broader; especially the hind wing; fore wing with vein 9, 10 coincident, the white irroration slight, cilia pure white except at tornus ; hind wing with some white in cilia towards apex. Exp. 18 mm. Has.—White river, Johannesburg and Natal.” Hampson’s females were up to 30 mm. expansion. Gymnelema imitata sp. n. Pl. XLIITI, fig. 9. Mate.—Head, thorax, abdomen and tibiw covered with tilleul-buff (XL) and drab (XLVI) very long hairs; fore tarsi white, ringed with fuscous; mid and hind tarsi and SOUTH. AFRICAN BAGWORMS. 611 spurs cream-buff (XXX) ringed with fuscous. Fore wing Saccardo’s umber (XXIX) irrorated with fuscous; at costa and at base some white scales; a fuscous terminal line; cilia white, drab (XLVI) at base. Hind wing Saccardo’s umber, sprinkled with fuscous scales, darker at inner marginal area, which has also some long tilleul-buff hairs ; .cilia as in fore wing. Underside shghtly lghter than upperside; some drab- coloured spots on costa of fore wing on second half of wing. Exp., type 24°5 mm.; cotypes 22°5 mm. and 26 mm. Has. (caught at ight by the author). Kshowe, Zululand, February 24th, 1916, cotype (bred by C. B. Hardenberg from a bagworm). Barberton January 15th, 1911, type and one cotype ‘I'he female is unknown to me. The bag of this moth resembles in many respects that of the Crossed-stick Bagworm, but the moth is decidedly different. I think this species is closely allied to G. stygialis. The venation of this species is not quite constant in the fore wing. ‘The type and one cotype has 7 and 8 stalked and 9 and 10 stalked for about equal length, originating from the areole not far from each other. The cotype from Barberton has only 9 and 10 stalked, 8 from the areole at place of stalk in other specimens, 7 from half way upper angle and 8. Gymnelema stibarodes (Meyr.). Melasina stibarodes Meyr., Ann. 8.A. Mus., v, p. 378, 1909. I have lately had the opportunity of examining the cotypes of this species closely, and have come to the conclusion that it is better placed in the genus Gymnelema. It is very difficult to define the limits of this genus and those of Melasina, but several small differences taken together con- firm that there is a difference, and this we notice at once when comparing more typical representatives of both genera with each other in a general way, without entering into details. The general build is much stouter in Gymnelema ‘612 A.J. I. JANSE. than in Melasina, and the wings are shorter and broader in the former. In the species stibarodes the build is very much like G. imitata and the bag is almost identical in method of building, and, as Meyrick remarked in his paper on this species, more like one of a Psychid moth than of a Tineid. The palpi of stibarodes are covered entirely with hairs which are spreading, like in the typical Gymnelema. In Melasina all the palpi I have seen are mainly covered with scales which lie close together, though sometimes they have some hairs mixed with the scales, and then also a tendency to spreading and to reduction of length of palpus; never, however, to the extent found in Gymnelema. The neura- tion gives little help, but that 7 and 8 of fore wing are gene- rally stalked, however shortly, points to Gymnelema and not to Melasina; in the female cotype of stibarodes this stalk is as long as we find in Trichocossus arvensis. The lack of consistency in the running of these veins in stibarodes is also found in G. imitata, to which, I think, stibarodes is mostly allied. SPECIES AUCTORUM : Gymnelema rougemonti Heyl., Ann. Soc. Ent. Belg., Vola ao; p- ccclxxy, 189f>° Hmpsn. ADM.N EE (3), voll va, p. 134. Has.—-Delagoa Bay. Gymnelema leucopasta Hmpsn., A.M.N.H. (8), vol. vi, p. 185, 1910. Has.—Cape Colony. Gymnelema pulverulenta Hmpsn., A.M.N.H. (8) vol. vi, p. 135, 1910. Hazs.—White River, Transvaal. Genus TricHocossus Himpsn. Trichocossus Hmpsn., A.M.N.H. (8), vol. vi, p. 133, 1910. Type, albiguttata Hmpsn., from Potchefstroom. The following description is given by Hampson: “Proboscis absent; palpi minute, clothed with very long hair; antennee bipectinate with moderate branches to apex ; SOUTH AFRICAN BAGWORMS. 613 head, thorax and abdomen clothed with long hair; the last with the anal tuft long; tibiz with the spurs long. Fore wing with the costa and inner margin nearly parallel, the apex rounded, the termen evenly curved; veins la and b anastomosing and curved upwards beyond middle ; le slight ; 2 from close to angle of cell; 5 from angle; a forked veinlet in cell: 4, 5,6 from disco-cellulars at intervals; 7, 8 strongly stalked, from apper angle; 9 from angle; 10, 11 from cell. Hind wing with vein 2 from close to angle of cell ; 3 from angle; 4, 5, 6 at intervals from disco-cellulars : 7 from upper anule ; a forked veinlet in cell; 8 free.” The species which I place in this genus agrees with the description in nearly every detail. The stalk of 7, 8 is as long as two-thirds of 7 and from upper angle; vein 9 is, how- ever, a little apart from the stalk; the upper part of the forked veinlet in cell is rather faint. In the hind wings the veins 4, 5, 6 are not always at intervals, as my specimen from New Hanover has vein 5 curved at base towards 4 so as to come with that vein from a point. Trichocossus arvensis sp. 2. (Meadow Bagworm). Pie SES fie Al. Ma.te.—Head, thorax, abdomen, femora and tibiz covered with white hairs; palpi and hairs under the eyes fuscous (XLVI); shaft of antennze natal brown (XL), branches fuscous-black four times shaft at middle. Wings covered with chetura drab (XLVI) scales and with whitish hairs at mer margins; apparently no markings on the two wings either on upper- or on underside ; cilia (as far as present, which is only partly on termen, tornus and inner margin of hind wing) white, with fuscous base. The specimens are in a rather rubbed condition, but the characters are sufficiently distinct to separate them at once from any of the alhed species. Moreover, the bag is also very characteristic and has been found in two different localities. Exp., 19-17°5 mm. (without the cilia, which are absent in the fore wings of both specimens). 614. A. J. T. JANSE. Fremate.—Unknown. Has.—T ype, from Pretoria, January, 1916. Bred by Mr. K. Munro from a bagworm feeding on Vernonia krausel. The apices of the fore wings of this specimen are some- what broken and the scales rubbed off the apical part as far as from half of costa to tornus. Cotype, New Hanover, Natal, January, 1915. Bred by Mr. C. B. Hardenbere. In this specimen the wings are un- broken but more rubbed, and it has also the antennz missing. SPECIES AUCTORUM : Trichocossus albiguttata Hmpsn., A.M:N.H. (8), vol. vi, p. 134, 1910. Has.—Transvaal, Potchefstroom. EXPLANATION OF PLATE XLIIT, Illustrating Mr. A. J. 'I’. Janse’s paper, ‘ South African Bagworms.” Diagrams of fore and hind wings, showing the neuration in the following species of Bagworms. Fig. 1—x 2. Clania modermannia (Heyj.). Fig. 2—x 2. Acanthopsyche junodi (Heyl.). Fig. 3.— x 2. Acanthopsyche tristis sp. n. Fig. 4.— x 2. Aleanthopsyche alba sp.n. Fig. 4a. Variation. Fig. 5.— x 3. Psyche (Manatha) ethiops (Hypson). Fig. 6.— x 2. Psyche (Manatha) subhyalina sp. n. Fig. 7.—x 2. Monda delicatissima WIk. Fig. 8—x 2. Gymnelema stygialis Heyl. Fig. 9.—x 2. Gymnelema imitata sp. n. Fig. 10.—x 2. Gymnelema vinctu's (Wlk.). Figs. 10a ¢ and 10b 2. Variations. Fig. l11—x 2. Trichocossus arvensis sp, n. Ann Natal Mus.Vol II. PL.XLI. Janse del. et hth. 1. Clania moddermanni (Hey!}.). 2. Acanthopsyche junodi (Heyl.). 3. A. tristis sp. n.. 4.A. alba sp.n.. 5. Psyche (Menatha) aethiops (Hmpsn.). 6. Psyche (Manatha) subhyalina sp.n. 7. Monda delicatissima WIk. 8. Gymnelema stygialis Heyl. 9. G. imitata sp. n.. 10. G. vinctus’(W)k.). 11. Trichocossus arvensis sp. n.. SOUTH AFRICAN MICROLEPIDOPTEROUS BAGWORMS. 615 Descriptions of Some New South African Microlepidopterous Bagworms. By E. Meyrick, B.A., F.R.S. ConrENTS. PAGE TINEIDA .. ; : : : ; PI6LS Fumea obscurata sp. i. : : ; ; . 615 Melasina cnaphalodes sp. n. : : ; . 615 Melasina picea sp. n. . : . 616 Melasina tyrophanes sp. n. . : ; . 616 Melasina craterodes sp.n. . : ; : oly, ADELIDA . : 3 : : : : ae Giled Ceromitia xanthocoma sp. n. : : ; . 617 TINEID AB. Fumea obscurata sp. n. Matze.—1l1 mm. Head, palpi, thorax and abdomen dark fuscous, palpi short, loosely haired. Antennal pectinations five. Fore wings elongate, posteriorly shghtly dilated, costa moderately arched, apex obtuse, termen shghtly rounded, oblique: dark fuscous-grey: cilia concolorous. Hind wings dark fuscous; cilia grey, becoming dark fuscous towards base. Has.—New Hanover, in December (Hardenberg), one specimen. Type in collection Janse. Melasina cnaphalodes sp. 1. Matr.—18 mmm. Head, palpi and thorax brownish, palpi short, slender, thorax with moderate crest. Antennal pectina- VOL. 3, PART 3. 41 616 E. MEYRICK. tions four. Abdomen fuscous. Fore wings elongate, moderate, rather dilated posteriorly, costa gently arched, apex rounded, termen rounded, rather oblique; 7 and 8 long stalked ; brownish suffusedly clouded with fuscous and transversely strigulated or reticulated with dark fuscous, forming two or three small spots beneath costa posteriorly: cilia brownish, basal third rather dark fuscous. Hind wings dark fuscous ; cilia fuscous, basal third darker. Has.—Clan Syndicate, in December (Hardenberg), one specimen. Also a female attributed to the same species, but with wings unexpanded and rubbed. Type in collection Janse. Melasina picea sp. n. Mate anp Femate.—Male 19 mm. Female 22 mm. Head, palpi, thorax and abdomen fuscous, palpi short, slender. Antennal pectinations of male five. Fore wings elongate, moderate, rather dilated posteriorly, costa gently arched, apex rounded, termen rather obliquely rounded; all veins separate; fuscous, suffusedly strigulated with dark fuscous, forming three or four small spots on costa posteriorly : cilia fuscous. Hind wing dark fuscous; cilia fuscous, with darker basal line. Has.—Clan Syndicate, in January (Hardenberg), three specimens. Closely allied to cnaphalodes but truly distinct; the different neuration, more obscure colouring, and costal (not subcostal) posterior spots are discriminating characters. Melasina tyrophanes sp. u. Maie.—16-18 mm. Head pale yellowish. Palpi whitish, with dark fuscous hair-scales. Antennal pectinations slender, five. Thorax and abdomen rather dark fuscous. Fore wings elongate, posteriorly somewhat dilated, costa gently arched, apex obtuse, termen obliquely rounded; rather dark fuscous. faintly purplish-tinged: cilia fuscous. Hind wings pale yellowish, base and apex narrowly infuscated, or whole wing SOUTH AFRICAN MICROLEPIDOPTEROUS BAGWORMS. 617 thickly strewn with dark fuscous scales ; cilia pale yellowish, sometimes greyish-tinged, round apex and upper part of termen suffused with fuscous. Under-surface of hind wings pale yellowish, sometimes infuscated. Has.—New Hanover, in August and October (Hardenberg), three specimens. Melasina craterodes sp. n. Frmate.—18 mm. Head, palpi, antenne, thorax and abdo- men dark fuscous; palpi short, slender ; antennz shortly pectinated, clothed with dense rough scales above excep towards apex, apical portion whitish-grey. Fore wings elongate, posteriorly dilated, costa gently arched, apex obtuse, termen obliquely rounded; dark grey, with some black transverse interrupted striz and strigule, costal edge whitish-grey between these ; a straight somewhat oblique blackish fascia before middle; a second beyond middle, becoming much broader on costal half, and enclosing a grey- whitish dot on costa; a blackish subterminal streak, parallel to termen on upper two-thirds: cilia pale brassy-ochreous, mixed with grey, towards tips whitish. Hind wings dark fuscous; cilia brassy-fuscous. Has.—New Hanover, in February (Hardenberg), one specimen. Alhed to primella, but with very different palpi. Larval case 8 mm. long, subcylindrical, narrowed at extremities, composed of silk covered with refuse, strengthened with a series of twigs attached rather spirally round median area. ADELIDA. Ceromitia xanthocoma sp, 1. FemaLte.—ld mm. Head densely haired on crown, wholly ochreous-orange. Labial and maxillary palpi short, whitish. Antenne white, towards base grey. Thorax rather dark purplish-grey. Abdomen dark grey. Fore wings elongate, 618 E. MEYRICK. rather narrow, posteriorly somewhat dilated, costa gently arched, apex obtuse, termen very obliquely rounded; veins 2-10 unusually approximated, 11 widely remote; rather dark purple-grey: cilia grey. Hind wings and cilia rather dark grey. Hazs.—New Hanover (Hardenberg), one specimen. Nearest iolitha, from which it is distinguished by the wholly orange head and unusually approximated veins. SOUTH AFRICAN BAGWORMS. 619 South African Bagworms: their Transformations, Life-history, and Economic Importance. Parr I. By Cc. B. Hardenberg, M.A., Government Entomologist in Charge of Wattle Insect Investigations. With Plates XLIV-XLVI and 6 Text-figures, CONTENTS. PAGE INTRODUCTION : . 620 I. GENERAL SKETCH OF THE iby: HISTORY OF Sener . 621 (1) The bags . : ; é F : . 621 (2) The larvee ; : : : . 627 (3) The pup : : : : : . 629 (4) The imagos ; : ; ; : . 630 (5) Reproduction. : : : . 632 (6) Seasonal history . ‘ : é . 634 II. SysteEMATIC POSITION OF Bacworms : ; . 634 III. Derartep Account OF SPECIES : . 638 1. Acanthopsyche junodi (Heylaerts)—The ‘Wattle Bagworm : : : : . 638 (1) Literature 3 : ; ; : . 639 (2) Systematic position : : : : . 640 (3) Theegg . : : : : ; . 640 (4) The larva : : . 641 (5) Habits, enemies, Seeeae etc., of the ere ; . 653 (6) The cocoon : : : : : . 668 (7) The pupa : ; ; : : . 669 (8) The imago : ‘ ‘ ; : . 675 (9) Habits of the imagos_ 5 ; : . 678 (10) Life-cycle ; : : F : . 683 EXPLANATION OF PLATES . : A : , . 685 620 C. B. HARDENBERG. INTRODUCTION. Durine the past two years the writer has been engaged in a study of the Wattle Bagworm, Acanthopsyche junodi (Heylaerts), fam. Psychide, and other insects injurious to the Black Wattle, Acacia mollissima Wild., for the purpose of finding means for their control and thus minimising their injury to the wattle plantations. It was soon found that, apart from the ordinary Wattle Bagworm, there were several other species, not as common, to be sure, but deserving more than passing notice as being potential wattle pests. A number of these have already firmly established themselves on the wattle as their principal, if not exclusive, food-plant, others are just beginning to take to it, while still another group of bagworms, although not counting the wattle amongst their usual food-plants, will readily eat it when it is offered to them and will pass through their usual transformations on this diet. When we consider that the Wattle Bagworm, above mentioned, some ten or fifteen years ago was in this same stage of transition from its native host-plants, the Thorn bushes, consisting of various species of Acacia, to the cul- tivated Black Wattle, and that it is at present considered by the wattle growers as their worst pest, we have reason to fear that within the near future these other bagworms will also begin to frequent the Black Wattle to a greater extent than now, and that this tree, from an occasional food-plant, may become a preferred one. It is natural, therefore, that we should have taken a great interest in all such species of bagworms which are to be found either on the wattles themselves or on the other vegetation in and around the wattle plantations. Some two dozen various kinds of bagworms have been found thus far. Of these, six have been noticed to subsist almost exclusively on the Black Wattle, five others frequent the wattle occasionally, although this is not their usual food-plant, while the others have thus far not been taken on the wattle itself, and as yet confine their depredations to the other vegetation in and around SOUTH AFRICAN BAGWORMS. 621 the plantations. But even the majority of these will feed on wattle leaves when they are offered to them. The knowledge at present available about South African bagworms appears to be very scanty indeed. While the cases of some half-dozen species have been figured and roughly described, their larvee and adults are for the greater part unknown, and are not yet represented in the collections of our museums and of private collectors of Lepidoptera. The study of the life-histories and transformations of these bag- worms furnishes an as yet practically unexplored and very interesting field of investigation, and, while admitting that our knowledge on these points is still very fragmentary, the information gained thus far is here published for the purpose of stimulating interest in these curious insects, which appear to have been much neglected in the past. Bagworms in other countries.—In other parts of the Old World the bagworms are well represented, and have received considerable attention. From Austraha, Froggatt mentions thirteen species, mostly of the larger and more curious kinds, as having been described up till 1907. From India some fifty species are known, as mentioned in Maxwell- Lefroy’s ‘Indian Insect Life, while Central Hurope also shows the considerable number of thirty-five species. We may thus safely assume that the twenty-four different kinds which up till now have come to our notice represent only a fraction of the total number to be found. ‘Those about to be discussed here are nearly all such as have been taken in Umvoti County, Natal, in connection with the Wattle Insect Investigation, and the search of other localities will doubtless bring many more species to hght. I. GENERAL SHORT OUTLINE OF THE LIFE HISTORY OF BAGWORMS. (1) Tue Bacs. The popular name “ bagworms” has been given to these creatures on account of their peculiar larval habits. The 622 Cc. B. HARDENBERG. caterpillar does not live exposed, but is concealed in a bag of its own manufacture, which it carries about with it. The popular names in other countries, “Basket worms” of the Americans, “ Bicho de Cesto” of the Spanish, “Sacktra- ever” of the Germans, all have reference to this same curious habit; the last-mentioned name being more nearly correct as it actually indicates that the bag is being carried about. But the name which most accurately expresses the condition is that given to these insects by the Zulus, “u-Mahambanen- dhlwana,” meaning a creature which carries its house with it. When the larva hatches from the egg, its first care, before starting to feed, is to construct its ttle house. This is done by gnawing small pieces of leaf or other substance on which it happens to be, and weaving these together with silk spun from its mouth into a collar which is fastened around the neck. ‘To this successive rings are added until the creature is entirely encased in a conical bag, narrowest at its posterior end, and into which it can entirely withdraw when danger threatens. This silken casing is added to as the larva grows; and, as feeding progresses, little bits of the food-plant are attached to the casing. This is done in various ways, charac- teristic for each species. The formation of the first bag may take place before the distribution of the larvee occurs or afterwards, but in any case before the larva starts feeding. The adornment of the tube as a rule proceeds gradually as the caterpillar feeds and grows, but we have reason to think that in some species, such as Fumea obscurata Meyr. and Melasina picea Meyr., for instance, this does not happen until the larva has become fully grown and makes ready for pupation. As these various adornments are in nearly all cases parts of the food-plant, it follows that the same species of bagworm may bear different aspects according to the plant on which the caterpillar has been feeding. But the method of the attachment is always the same for a given species, no matter SOUTH AFRICAN BAGWORMS. 623; what the food-plant, and it is this, and the shape of its silken casing, which gives to each species of bagworm its charac- teristic appearance. In external aspect there is a great variety: some species. are much more careful architects than others. The Wattle Bagworm, Acanthopsyche junodi (Heylaerts) (Pl. XLIV, fig. 1), takes off irregular pieces of whatever part of the food-plant is within its reach and anything which may be attached to it, and fastens it at one end only to the bag, so- that the result is a bag of irregular surface, adorned with leaves, sticks, seed-pods, or even other bags which happened to be hanging on to the leaf on which it was feeding. Another such careless builder is what we have called the “hatched Bagworm,” Acanthopsyche tristis Janse, but here only the blades of grass are used, and as the silken tube: in the case of this species is more cylindrical, the entire bag has a longer and narrower appearance (fig. 2). The “Gum Bagworm,’ Melasina stelitis Meyrick (fig. 3), uses pieces of the broad-leaved foliage on which it feeds, cuts it into more or less square pieces, about a quarter of an inch each way, which it fastens with one side on to its. bag. These are rather closely applied to the surface of the bag, those of the one layer slightly overlapping those of the preceding one, and as a result we have a bag which, although made of the same material and in a similar manner to that of the Wattle Bagworm, has an entirely different aspect. The distinction between the two can be more readily seen than described. The most careless builder of all is Melasina picea Meyr.. (fig. 4), a species which, for want of a better name, we have called the “ Rubbish Bagworm ” from the construction of its bag. It accepts anything which it finds on the ground: remains of beetles, flat egg-cases of spiders, chips of bark, sticks, dried leaves, etc., everything is made use of. These objects are attached to the case in a very irregular fashion, lengthwise, crosswise, or at any angle, so that when the creature is resting on the ground it can hardly be O24 C. B. HARDENBERG. distinguished from the miscellaneous débris with which the ground is covered. The “ Lictor Bagworm,’ Clania moddermanni Heylaerts (fig. 5), uses small sticks which are rounded off at both ends and are attached lengthwise to the bag and fastened along their entire length. These sticks are as long as or longer than the inner casing, and, being put parallel and close together, give the bag a solid, cylindrical appear- ance. ‘The sticks are not all of equal length and generally there are one or two extending beyond the bag for a con- siderable distance. The “Grass Bagworm”’ (fig. 6) (species as yet undeter- mined) takes thin, hollow grass-culms, cut off to approxi- mately equal sizes, and attaches them lengthwise to its bag, giving the appearance of a neat, cylindrical little bundle. ‘hese hollow straws are fastened to the silken case, if not for their entire length, then at least for a considerable distance. Another group of bagworms are in the habit of attaching the various adornments crosswise to their bags. ‘The “Meadow Bagworm,”’ Trichocossus arvensis Janse, whose bag is shown in fig. 7, takes for this purpose small pieces of the blades of grass, which are smoothly apphed to the casing over their entire length, thus following its curvature. The result is a smooth, more or less cylin- drical bag, lined transversely. The “Crossed-stick Bagworm,’ (Wik.), of the Cosside, takes small pieces of grass culms, or little twigs, or the midribs of leaves, which are attached to the casings tangentially at about their middle, the free ends projecting (fig. 8). Hach following stick has its end crossed over that of the preceding one, and the outline of the bag, around the projecting ends of the component sticks, becomes roughly pentagonal, the diameter of the bag, and thus the length of the sides, increasmg toward the upper end, the mouth or neck, of the bag. A similar, but less regular, arrangement of the sticks is a > Gymnelema vinctus SOUTH AFRICAN BAGWORMS. 625 feature of the bag of Gymnelema stygialis Hampson (fig. 9), another Cossid bagworm; while a third species, Acanthopsyche alba Janse, carefully spins a_ silken webbing over the outside of its bag, thus making it appear as if it were an old discarded bag covered with a spider’s web (fig. 10). Still another group of bagworms, also using small sticks, arrange the material neither lengthwise nor crosswise but at a sheht angle, and as these sticks are of increasing lengths towards the upper end of the bag, this covering assumes a spiral aspect. Three species of these ‘ Spiral Bagworms,” all species of Melasina, fam. Tineidz, have come to our notice, evincing various degrees of neatness in the arrange- ment of the material. We have distinguished them as the “Clear Spiral,” (Melasina tyrophanes Meyr. fig. 11), the “Webbed Spiral” (Melasina cnaphalodes Meyr. fig. 12), and the “ Rough Spiral Bagworm” (Melasina craterodes Meyr. fig. 15). Then there is another species, the “Thorn Bagworm ” (Melasina halieutis Meyr. fig. 14), which uses nothing but the fine particles of the leaf or bark on which it feeds. These are arranged into smooth narrow collars, and the bag is made up of a series of these which are accurately jomed together, and, increasing in diameter towards the mouth end, form a bag, shaped like a smooth greyish-brown thorn. Still another, the “Sand Bagworm” (Fumea obscurata Meyr. fig. 15), which lives at the roots of grasses, uses almost nothing but sand-grains in the construction of its bag, and resembles a true caddice-worm in this respect. Lastly, we may mention a group of bagworms which are distinguished from all the previous ones in that they make no silken tube to which to attach various extraneous materials. The caterpillars merely construct a dwelling by glueing pieces of leaf together, leaving a variously shaped cavity in which the larva lives. The “ Flat Bagworm” (Ceromitia xanthocoma Meyr. fig. 16) makes its case from pieces of the blades of grass, which are arranged transversely and are 626 C. B. HARDENBERG. smoothly and accurately joined together at their sides. There is one layer of these on the dorsal side and one on the ventral side of the creature. These layers are only very slightly convex and leave but a narrow slit-like cavity between them. In this space, which is very wide but very low, the little worm les concealed. Asan additional protection the dorsal covering is carried a little farther forward than the ventral, so that, when feeding, the larva does not need to expose itself at all. A very similar arrangement we find in the case of the “ Flat Leaf Bagworm” (fig. 17). Here the case is made of pieces of leaf, rounded off at one end and shghtly concave at the other, the anterior end. ‘Two or three such pieces form each side of the bag which is found attached to the tree trunks. A third species of this group, the “Seed Bagworm ” (Pl. XLV, fio. 1), uses three narrow, pointed pieces of leaf or grass which are joined side to side so as to leave a narrow cavity, shaped like a three-sided pyramid. The entire structure resembles in size, shape and colour a part of a grass-flower which has been blown against the branch or trunk, and is very difficult to detect. All these structures are undoubtedly protective, and made with the object of imitating the surroundings under which the species originally lived. They serve primarily as a pro- tection during the resting stage or pupation period ; for most, if not all, pass the winter in this condition. During the summer the protection offered is, in the case of most of these bagworms, not so great; but this need not be, for at that time insect life is plentiful, and as the bagworm withdraws into its bag at the slightest disturbance, it is doubtful whether an insectivorous bird would take much trouble to secure it while other food is plentiful and near at hand. In the winter the conditions are different, and a greater amount of protec- tion is afforded by the bag of withered leaves or dried sticks which blend with the general colour scheme of the vegetation at that time. In order to judge in how far the protection aimed at is being secured we must look for these various bag- worms in the winter, and amongst their natural surroundings SOUTH AFRICAN BAGWORMS. 627 on their original food-plants ; and our experience has been that such as are conspicuous enough durmg the summer are difficult to locate amongst the dry vegetation in the winter. (2) Tae Larva. The bagworm caterpillars are as a rule of a characteristic appearance, being dull grey to dark brown in colour, smooth, soft-bodied, and having the three thoracic segments provided with dorsal chitinised plates ; these segments being the only parts of the body which are normally exposed when feeding or crawling. Ina few cases the first abdominal seement may also have a dorsal chitinised thickening, but usually the abdomen is soft and membranous. In all cases where we have had occasion to observe the newly hatched larve we found them bearing a remarkable aspect and assuming a characteristic attitude. ‘lhe thorax is very large and massive, while the abdomen is slender and conical, and is carried elevated at almost right angles to the thorax. This pronounced dorso-flexion, so very abnormal in insect larvee, is probably a character arising from the habitual position of the body during the feeding period of the cater- pillar ; for normally, when crawling or feeding, the bagworm is attached to the underside of the leaf or stalk with the bag hanging downward, and the body is thus continually curved dorsad at the juncture of thorax and abdomen. The continuous carrying of the heavy bag would be expected to give rise to a modification in the structure of the body of the caterpillar. We find that the thoracic legs which have to bear the weight have become very well developed; the cox are massive, meeting in the median line, the legs are stout and the claws are strong, so as to enable the insect to retainits hold. And as the bag, which is usually fairly large and roomy, might slip off while hanging down, if the cater- pillar had not been especially adapted to cope with this emergency, we find that the abdominal and anal prolegs, although not being used for locomotion, as in the case of the 628 C(. B. HARDENBERG. exposed feeders, are not rudimentary, as might be supposed, but are well developed and provided with hooks which are strongly curved and sharply pomted so as to hook into the silken lining of the bag. Nevertheless, the legs have become so short as to be entirely useless for locomotion when the creature is removed from its bag. The bagworm caterpillars are very tidy in their habits. When we open one of the bags we find the inside always dry and scrupulously clean. EHjecta and cast skins are removed at once, as the presence of excrement and larval skins would favour the development of fungi and decay with the slight degree of moisture always present in the bag, due to the emanations of the hving occupant. Excessive moisture is their greatest enemy and quickly proves fatal to them. We thus find that the bagworms (at least such as we have had occasion to observe more closely) do not feed during a rain or mist, but they will postpone their meal untilthe fohage has dried. Ifthe weather conditions are such that they are forced to feed during the rainy weather— and the young larve especially cannot stand a fast of long duration—we find that a disease of a bacterial nature quickly appears amongst them with disastrous results. This may be due to two causes: (a) the wet food may not agree with them and cause digestive disorders, as a result of which the evacuations become fluid and the inside of the bag becomes soiled, since they cannot be quickly and completely removed like the solid normal pellets. Under such atmospheric conditions we find as a rule a great number of soiled bags which either are empty, having been abandoned by their inhabitant, or contain a flaccid, diseased larva; or (b) the drops of water carried in by the larva when retreating into its bag after feeding may increase the degree of moisture inside the bag to anextent which affects the larva unfavourably. Whatever the primary cause, we find that under these condi- tions the caterpillars behave abnormally. They turn around in their bag, and such as are able to leave it through the opening at the lower end start making a new bag. While engaged in SOUTH AFRICAN BAGWORMS. 629 this the caterpillar is exposed to its various enemies and the unfavourable conditions of the atmosphere, and thus it makes haste to construct its new dwelling. It is therefore not very fastidious and takes old abandoned bags which it finds on the leaves to supply part of the new covering. The result is a compound bag, made up of from two to five others, of which we find only one inhabited. (3) THe Pups. When the time for pupation approaches, the larva makes preparations to still further secure its retreat and to establish for itself a safe place in which to pass the all-important period of pupation. Generally the creature moves with its bag to a less conspicuous place, or one where its bag blends more perfectly with its surroundings, so as to enjoy additional protection. The clustering of the ‘ Wattle Bagworm” on the stem and in the forks of the branches, where it is more obscured by the surrounding foliage; the retreat of the “Meadow Bagworm” and the ‘ Crossed-stick Bagworm” amongst the exposed roots of the grasses and low plants on the surface of the ground, where it can hardly be distinguished from the normal litter covering these places; the attachment of the “Thatched Bagworm” to the underside of stone ledges or in corners where waste material is allowed to accu- mulate ; the adornment, in the case of the “Rubbish Bag- worm ” with all kinds of débris found on the ground—all these may be classed under such precautionary measures. Secondly the case is made more secure against possible intruders or accidents. The upper end or neck is closed tightly, while usually the lower part of the bag is filled up with a loose irregular webbing of silk which would entangle any of the smaller parasites, spiders or ants which might probe the opening. The lower aperture must be left open so as to allow the moth to escape, or the abdomen of the male to enter for copulation,as the case may be. ‘The bags are more firmly attached, in some cases by a strand of silk spun around the- 630 C. B. HARDENBERG. twig, or the neck of the bag is closely applied to the bark of the tree and firmly fastened all around. In other cases the bag is provided with a fairly long strand of silk so that the bag will swing and offer no resistance when hit by any object. Thus a bird pecking at it would be unable to pierce the covering, the bag retreating before every blow aimed at it. Having thus secured for itself the maximum amount of safety against external enemies, the caterpillar now makes ready for its transformation. In some cases a true cocoon is made inside the bag in the shape of an inner bag of finely woven, very tough, white silk, which is attached at both upper and lower ends to the outer bag, but free in the middle. In other cases no such elaborate preparations are made, but the larva probably contents itself with an additional layer of silk to the inside of the bag. Thus far,the caterpillar has remained in its normal attitude, that is with its head toward the mouth or upper end of the bag. It now turns round so that it faces the lower end (through which the moth emerges) and pupates. The last moult cannot be removed from the bag as it is now closed, also the pupa would be unable to make the necessary move- ments for the expulsion of the exuvie, so that we find the dried skin of the last moult in the bag near its upper end. The length of the pupa stage has in the majority of the bagworm species not been ascertained, but from such as we have had the opportunity to make observations on it appears that the actual pupal period is very short as compared with the length of the resting stage of the caterpillar. (4) Tue Imacos. There is a great diversity amongst the moths of the various bagworms, not only in the type and general habits of the males, but especially in the different degree of degenera- tion of the females. With regard to this we can roughly divide such bagworms as are here under discussion into four groups. In the first group, comprising the Wattle Bagworm, SOUTH AFRICAN BAGWORMS. 631 the fLictor Bagworm, the Thatched Bagworm, and in fact all the bagworms of the true Psychid type, only the males are winged, while the females are vermiform, wingless creatures which never leave their bag during their short period of reproductive activity. In the Wattle Bagworm all vestiges of legs have disappeared in the female, while in the females of the others, the Lictor and the Thatched Bagworms, these legs are rudimentary and are represented by forked mem- branous appendages of the thorax. In the second group, to which belongs the Sand Bag- worm, the females are destitute of wings, but have functional legs. The female leaves the bag, but remains attached to it on the outside near the lower end, and deposits her eggs in that position. In the third group, the Webbed Spiral Bagworm and other species of Melasina, we find the females with rudimentary wings and functional legs. These females are capable of restricted locomotion only, and remain attached to their bags. In the fourth group, including the Crossed-stick Bagworms and other Cossids and also some species of Melasina, the females are fully developed moths, and fly actively hike the males. We find also a great diversity in the habits of flight in connection with the reproductive functions. In those groups where the females are incapable of locomotion the males are very active creatures. The massive thorax, the pointed fore wings with the small hind wings and the short pointed abdomen, indicate that they are strong, rapid flyers. They fly during the daytime, the Wattle Bagworm in the sun- shine only during the brightest part of the day, and the Thatched Bagworm and the Lictor Bagworm in the late afternoon. Thus far we have had no opportunity to observe, under natural field conditions, the males of the second and third group, but in the laboratory they became active in the evening and they are probably crepuscular in their habits. VOL. 3, PART 3. 42 632 ©. B. HARDENBERG. Of the Crossed-stick Bagworm, as a representative of the fourth group, we know that both sexes fly at mght, and are then attracted by the light. Specimens of both sexes have been captured as late as 11 p.m. on the window in front of a lamp. In all the groups the length of life of the male is very short, a few days at the most. The females enjoy a longer period, and may live for several weeks. (5) Repropuction, From our present experiences we are inclined to believe that parthenogenesis does not occur in any of the species of bagworms which we have observed. In all these the males appear to be as plentiful as the females, or are even in the majority, and we have failed in all our experiments to hatch young from eggs laid by females to which males had had no access. In the case of some European species of bagworms a number of parthenogenetic generations are followed by the emergence of individuals of both sexes and mating takes place. We have bred only a few generations of certain bagworms, but all have proved to be of both sexes, and we are probably safe in deciding that these species of bagworms only reproduce in the normal, sexual manner. Kgg-laying.—The females of the first group deposit their egos in the empty pupa-case. The female is merely an animated mass of eggs, lying with its head downward, facing the lower opening of the bag, the eggs are pushed out and gradually fill the upper cavity of the pupa-case until the female is spent. The body of the female shrivels and retreats until at last it drops out of the bag to the ground and dies. The eggs are closely packed together and are mixed with a quantity of hairs from the abdomen of the female. To these hairs are added very fine strands of silk, the origin of which has not yet been ascertained. The fecundity of the females is amazing; in the Wattle Bagworm we have counted a maximum of over 3000 eggs, the SOUTH AFRICAN BAGWORMS. 633 average being about 2300. This enormous amount of eggs is apparently necessary to counteract the great mortality amongst the young at the time of their distribution. In those groups where the females leave their bags, but remain attached to it, the eggs have been observed to be laid on the outside of the bag. It has been claimed for a certain species of bagworm with similar habits that the eggs were deposited inside the bag by the female thrusting her ovi- positor through the neck of the bag, but we have thus far not been able to observe any such procedure on the part of these bagworms ; the eggs were always laid on the outside of the bag or on the bottom of the breeding cage. In the case of the females of the fourth group, which have well-developed wings, it appears that the eggs may be deposited anywhere, as we have found them usually on the bottom of the breeding cage in which the moths were confined. The fecundity of these actively flying females is much less than that of those moths which remain in or on their bags. In the former case such an enormous reproduction is not required, as the female can do her share in the distribution, and the eggs are deposited in places where they are concealed and on or near the food-plant of the young caterpillar. In the case of the confined females the young play an entirely passive réle in their distribution (the wind being the principal agent), and are thus subject to numerous vicissitudes with a more or less remote chance of their being transported to a suitable food-plant, while with the species in which the females are active this distribution has already taken place before the young have hatched, and they at once find their food ready for them. ‘The mortality amongst these is thus apt to be very much less than in the former case, and a great productivity on the part of the female is not therefore so essential. In fact, it would probably be unfavourable, as the heavy mass of egos might impair the flight of the parent, and thus restrict the area of distribution. 634 C. B. HARDENBERG. (6) SeasonaL History. In this respect the various species of bagworm also behave very differently. Of those we have bred thus far, the Wattle Bagworm is the only species in which the adults emerge in mid-winter. Fertilisation takes place at that time, but the incubation for the eggs is long—two months—while the growth of the young is comparatively slow, and so there is only one generation a year. In the case of the Rubbish Bag- worm the conditions are reversed, the moths appearing in mid-summer—January—and here we also have reason to believe that there is only one generation a year. Of the other species the moths are found to emerge at various parts of the season (mostly in mid-summer), and their life-histories have not yet been followed closely during the entire year, so that we do not know their life-cycle accurately. As to the Crossed- stick Bagworm, we have reason to believe that there are two generations in a year, since moths have been found swarming during the early spring and again in February. II. SYSTEMATIC POSITION OF BAGWORMS. The determination of the moths reared from these various bagworms shows that the species all belong to one of the three families: Psychide (the Wattle Bagworm, Acantho- psyche junodi (Heyl.), the Lictor Bagworm, Clania mod- dermanni Heyl., the Thatched Bagworm, Acanthopsyche tristis Janse, and the Meadow Bagworm, Trichocossus arvensis Junse) ; Cossidee (the Crossed-stick Bagworm, and the Webbed-crossed-stick) ; and Tineide (various species of Melasina and Ceromitia). It is interesting to note that the families mentioned are all of the lower and less specialised ones, being placed in the phylogenetic scheme near the Tri- choptera, from which, according to some authors, the Le pi- doptera have developed. The larve of the other members of these families, so far as known, are all concealed feeders, either living in the wood, like the Cossidez, or mining in the SOUTH AFRICAN BAGWORMS. 635 leaves or feeding between leaves which have been drawn together, as in many Tineide. ‘This is an indication that the concealed feeding habit is the more primitive one, and the life-histories of the larvee of these bagworms thus confirm the low position of the families to which the bagworms had been assigned on the streneth of the structural characters of the adults. The habit of building larval cases which are carried about by their inhabitants has been reported from the following families : (1) Perophoride (Lacosomidez).—Example: Pero- phora sanguinolenta, the ‘Hammock Moth” of South- America. (2) Pyralide.—Hxample: Conodomus hockingii Walsingham and species of the genus Nymphula. (3) Psychide.—The larve are case-bearers without exception. Kirby in his Catalogue in 1892 mentions 206 species of true Psychids, and undoubtedly there are many more, (4) Taleporide.—tThe larve of these are also without exception bagworms. (5) Coleop horidz.—Kxample: Amphisbatis incon- gruella Stt. Many species in this family are leaf-miners in their early stages, making a bag in the later instars of their larval life. (6) Gelechiide, subfamily Gicophorine.—Example : Fumea limulus and Borkhausenia flavifrontella Hb. (7) Tineidee, subfamilies Teichobiine and Adeline. —The members of these two subfamilies are without excep- tion case-bearers. (8) Tortricidz.—The case-bearing habit is known only in one species of the family, Exartema latifasciana Hw, which lives on moss in a bag made of the same material. (9) Cosside.—Members of the genera Gymnelema and Trichocossus. In the species to be discussed we have paid much attention to the larval and pupal stages. Descriptions of caterpillars 636 C. B. HARDENBERG. of South African Heterocera are very few, and such as exist are usually based on the last instar and are confined to general statements as to size and colour-pattern. No detailed, careful descriptions are available. This is greatly to be regretted, both from an economic and a scientific point of view. Let us consider the question first from the economic side. It is almost without exception the caterpillar, not the adult insect, which causes the damage, and it is this stage of the insect’s life which is sent to the entomologist together with the complaint about its depredations. In the absence of reliable descriptions (and, owing to the scarcity of preserved larval skins, careful comparison with known specimens of which the adults have been bred is usually impossible), it is necessary to breed the insect to its adult stage before its identity can be ascertained without a doubt. Anyone who has tried to rear such larvee, sent in from a distance, sometimes without the food-plant and in a closed box, realises the comparatively small chance of success. What with the parasites from which such specimens often suffer, the weakly and often diseased condition of the larve, the paucity of material and the artificial conditions under which it has to be reared, not to mention an often compulsory change of food-plant, the specimens either die before pupation, or, if a pupa is formed it is often so lacking in vitality that the moth does not emerge. Determination of the species in question then depends entirely on the examination of the pre-adult stages. With the state of our present knowledge (or rather lack of it) of South African caterpillars, there are thus many chances of wrong determination, even in cases of some of the most common and widely-known pests. As an instance I may mention that a larva, found hibernating under the bark of a pear-tree, was at once declared by one of our most experienced entomologists to be that of the Codlin Moth, while subsequent rearing of the specimens by the writer proved it to be an entirely different species. Also, the moths of two different species may look extremely alike, except to the Micro-Lepi- SOUTH AFRICAN BAGWORMS. 637 doptera expert, while a careful examination of the larve and pupz shows the differences at once. For these reasons alone the careful study and description of the larve and pupz are essential in the discussion of a species ; but there is still another very important reason. When we desire to determine the benefit derived from the presence of certain insectivorous birds or other animals which make, at certain times at least, insects a part of their diet, it becomes necessary to examine the stomachs or ejecta of such birds or quadrupeds, and to scrutinise the chitinous remains of any insect which may be found thereim. It is evident that only the more strongly chitinised parts will escape the digestive and crushing action of stomach or gizzard, and we find in general only head-cases, mouth-parts, or feet, to guide us in our determination. Unless these parts have been carefully figured and described we are at a loss to say with certainty whether a certain injurious species is being preyed upon by the alleged predaceous enemy. From a purely scientific standpoint the study of the imma- ture stages is equally important, as they will show characters which may assist in clearing up doubtful points in the rela- tionships of certain groups, relations which have been based almost exclusively on the characters (wing venation) of the adults. The study of the position of the sete on the cater- pillars has come much to the fore in this connection in recent years, and has proved to be of great importance in establishing such relationships. And here the study of the first instar of the larva is of the greatest importance. ‘lhe first instar of the caterpillar is the most primitive condition, and at this stage it 1s more apt to show its phylogenetic relationships than in the succeeding instars. With the successive moults numerous changes occur which are due to adaptation to the environment, either to afford concealment, or to guard against the attacks of parasites or predaceous enenues. No cases are known to the writer where larve in the first instar have been parasitised, while the smaller size of the newly-hatched larvee, and the usually more or less concealed position of the eggs 638 0. B. HARDENBERG. deposited by the parent moth, make them probably less easily detected, In addition to this, the young larve, being small, find their food close at hand, and there is no necessity for travelling in search of food. Since it is the movement of an object which attracts attention, this certain amount of immo- bility on the part of the first instar larva doubtless aids in its protection. With succeeding moults the necessity for protec- tion, especially in the case of exposed feeders, becomes greater, and therefore most remarkable changes in colour-pattern and armament make their appearance. In the case of the Bom- bycine, Lymantrid, and Lasiocampid moths, there is an enormous development of secondary hair which obscures the position of the primary sete. In many of these we have found that this secondary hair is either absent in the first instar or so feebly developed as to allow the primary setz to be recognised with certainty. In other cases, as in the Saturniide, where the colour-pattern shows a great diversity in the various species, the first instars (when no colour-pattern has yet been developed) of the larve of the different species look remarkably alike. We have thus endeavoured, wherever possible, to obtain the first instars of the caterpillars, and have described them in detail. In a good many cases, however, this has not been possible, since the bagworms as a rule do not mate readily in captivity. This article is based on the results of a couple of seasons’ breeding only, and there will be necessarily many gaps which remain to be filled in by subsequent investigation. III. DETAILED ACCOUNT OF SPECIES. 1. Acanthopsyche junodi (Heylaerts). The Wattle Bagworm—Fam. Psychide. This species is by far the most common bagworm in Natal, and is probably one of the most widely distributed through South Africa. It has been found from near Port Elizabeth (Cape Colony) to the Groot Spelonken in the Zoutpansberg SOUTH AFRICAN BAGWORMS. 639 District of the Northern Transvaal, and its occurrence in South Africa is probably co-extensive with that of its original food-plants, the native thorn-bushes, consisting of various species of Acacia and Mimosa. In Natal it is the Bag- worm par excellence, and it has become notorious during the last decade as one of the worst pests of the Black Wattle plantations. What was merely an “entomological curiosity ” some fifteen years ago has evolved into a pest of prime economic importance within the last few years, seriously threatening the Black Wattle industry in Natal. So extensive became its spread, and so serious its injury to the wattles, that it was realised necessary to make this insect (with other wattle pests) the subject of a special investigation by the Department of Agriculture. Our work on this and the numerous other species of insects injurious to the Black Wattle (Acacia mollissima Weld.) has now been in pro- egress for about three years, and, while studying the Wattle Bagworm primarily from an economic standpoint, many other matters of interest concerning the biology of the species have: been observed, which appear not to have been recorded before. (1) Lireratore. The insect has been known in the wattle plantations for a great many years, and as early as 1889 it was noticed by the then Natal Entomologist, Claude Fuller; but, though wide- spread, it did not appear to cause any appreciable injury and was not considered of economic importance. In 1899 Claude Fuller published the first more or less detailed account of the life-history of this insect im his Annual. Report for that year, where we find three pages devoted to it. In 1909 the same author published an account of this bag- worm and some half-dozen others as Bulletin No. XVI of the Natal Department of Agriculture, in which he devoted some: seven pages to the life-history of the Wattle Bagworm. Owing to the outbreak of 1911-12 in the New Hanover area, another, more detailed investigation of the Wattle Bagworm 640 C. B. HARDENBERG. problem was undertaken by him, the results of which he embodied in an article of some fifty pages, which appeared in the ‘ Agri- cultural Journal of the Union of South Africa’ for June, July, and August, 1913. In this article over forty pages were taken up with general discussion and the economic aspect of the bagworm, while about nine pages were devoted to the life-history of the insect in question. These published accounts of the bagworm, based as they necessarily were on more or less chance and discontinuous observations, do not entirely agree with each other, nor are they fully in accord with our observations during the past seasons. ‘They unavoidably treat the insect almost exclu- sively from the economic point of view, while the scientific aspect has not received the consideration which it so richly ‘deserves. (2) Systematic Posirion. The species belongs to the family of Psychide, sub- family Psychine, and was described by Heylaerts in the “ Ann. Soc. Entomol. Belg.,’ vol. xxxiv, p. cxxx (1890) as Eumeta junodi. Janse in his ‘Notes on the South African Psy chide’ places it in the genus Acantho- psyche, while in Fuller’s articles, mentioned above, it is always referred toas Chalioides junodi, but I do not know on which authority. The species has been described from male specimens only; the female has not been described. (3) Tue Hea. Description.—Ege 0°88 by 0°75 mm., ellipsoidal in shape. Colour a creamy white, darkening with age, becoming grey to dark grey when hatching approaches. Hgg-membrane soft, colourless, smooth, apparently without sculpture. A micropyle could not be distinguished. DEPOSITION. separately, but all in one mass, mixed with hairs from the abdomen. ‘They are contained inside the pupa-case of the The eggs of a female are not deposited SOUTH AFRICAN BAGWORMS. 64] female at the posterior end, filling this part of the case up to the third or even second abdominal segment. The chrysalis containing the egg-mass is encased in a bag of strong white silk (the cocoon or inner bag) which is stretched as a central tube in the bagworm-case, attached to both the anterior and posterior ends. Devetopment.—External changes: As the egg becomes older it changes from the fresh cream colour to a dirty yellowish white, and later to greyish white, ultimately to dark grey with a black spot indicating the head and thorax of the developing larva. The incubation period is from sixty to sixty-five days. (4) THe Larva. First Instar. Length upon emergence 1°9 mm.; ground colour cartridge- buff! (XXX); head jet black; dorsal thoracic shields, plate on eighth abdominal segment and anal plate Van Dyke brown (XXVIII); the third thoracic shield divided by a median stripe of the ground colour. Prespiracular plates of second and third thoracic segments concolorous with dorsal shields, other setiferous plates drab (XLVI). Thoracic legs snuff- brown (XXIX); leg-plates of abdominal and anal prolegs buffy brown (XL); claws of prolegs bistre (X XIX). The larva (text-fig. 1, 4) is more or less cylindrical, very shghtly flattened dorso-ventrally. While in most cater- pillars the head is very large upon hatching and the body tapers from the neck caudad, we find here that the width of the head is surpassed by that of the thorax, which is massive, especially the third thoracic segment, and from the posterior edge of it the abdomen tapers gradually down to the tail, like a cone. The great strength of the third thoracic segment is probably necessitated by the 1 The colours mentioned are those of Ridgway’s ‘Colour Standards and Nomenclature,’ edition 1912; the numbers refer to the plates in this publication. 642 C. B. HARDENBERG. habit of the young larva of carrying its abdomen elevated at nearly right angles, the point of flexion being at the juncture of the third thoracic and first abdominal segments. Owing to the light ground colour of the body, the chitinised seti- ferous plates on the abdominal segments stand out clearly and sharply defined. During the first instar the head and thorax are of a uniform colour without the distinctive pattern which is acquired in the third and fourth instars. Hrap.—Width of headcase 0°45 mm. Adfrontals not separated from parietals (text-fig. 1, 8), front reaching to two-thirds the distance to the vertex. Frontal punctures high up, close together. Frontal sete placed above the punctures, far apart and half-way up the front. Lower adfrontal sete opposite frontal sete, upper adfrontals near apex of front, adfrontal puncture near lower sete. Sete 1 and 3 of parietals rudimentary or absent, 2 below and laterad of upper adfrontals. Antennz.—Second joint short, only slightly longer than the third and fourth combined (text-fig. 1, c). Clypeus.—Not differentiated from the front (8), the sete 1 and 2 as usual near the distal edge, distance between them about half of that between 1 of right and left side. Mandibles.—Short, sub-quadrate, with broad base, the outer edge slightly concave and sub-parallel to the median edge (ec). Cutting edge armed with five teeth, the inner three blunt, the others acute, the second tooth (from the out- side) the largest and with a step on its outer surface. Mandibular sete 1 and 2 close together and situated on basal fourth of the mandible. Labrum.—Outer angles broadly rounded, median notch about one-third of the total height, tip of indentation acute (pv). Seta 1, 2 and 3 in a line directed latero-cephalad, 4 and 5 sub-marginal, 6 not marginal but has moved up towards 1 and is situated about in line with 1 and 5. Maxille.—Cardo very small, triangular; stipes broad, massive, the inner edge where it joins the mentum elongated SOUTH AFRICAN BAGWORMS. 643 TEXT-FIG. 1. Acanthopsyche junodi (Heylaerts). Larva, first instar. A, Side view. xX 25. B. Front view of head, showing position of sete. x 100. c. Antenna. x 200. pb. Labrum. x 360. &. Man- dible, showing step on second tooth. x 200. ¥F. Labium. x 200. «a. Maxilla. x 200. 644 C. B. HARDENBERG. and strongly chitinised; of the exposed, lateral aspect only the middle part chitimised, the basal and apical parts mem- branous (G). The middle part bears two sete. Palpifer broad and short, bearing one seta at its distal edge. Palpus three-jointed, the two basal joints annular, the last in the shape of a cone bearing several small sense papille at the tip. The basal joint has one seta placed more mediad than that on the palpiter. Attached to the median aspect of the first and second palpal joint we find the remains of the maxillary lobes, consisting of two cylindrical processes, each with a sense-cone at its tip and three dagger-shaped sense- hairs. ‘Three minute sense-cones are situated in front of the maxillary lobes. Labium.—Characterised by a very strong development of the chitinous sclerites of the mentum (Ff). Mentum with two short setze near the median line. Labial palpi two-jointed, slender, terminal jomt about half as long as basal joint. Submentum for the greater part membranous and bearing two long sete near its posterior third. THorax.—The thoracic segments, as mentioned, are very massive, and each is provided with a broad dorsal chitinous shield (P]. XLVI, fig. 3). On the prothorax this shield is fused with the prespiracular plate and is pierced by the prothoracic spiracle ; on the meso- and metathorax the pre- spiracular wart is separate from the dorsal shield. This pre- spiracular wart bears two sete. ‘lhe subventral wart or plate is also armed with two sete. The legs are strong, the claws pointed and slightly curved mediad at the tip, with a median lobe at the base. Coxe large, the chitinous coxal plates meeting in the median line. Thoracic plates uni- colourous, no pattern as yet apparent. On the third thoracic segment the dorsal shield on each side does not reach the median line. ABDOMEN.—The abdominal segments are narrow, except the eighth, with several chitinised setiferous plates (Pl. XLVI, figs. 1, 2). On segments 3 to 6 these plates are seven on each side: the dorsal with one seta (1), the subdorsal with one SOUTH AFRICAN BAGWORMS. 645: seta (2), the lateral or supra-spiracular with one seta (3) and. exceptionally with a minute one (8a), in front of this, the sublateral or subspiracular with two sete (4 and 5), the subventral without any setee, the outer leg-plate with a group of two sete (7), and the ventral with one seta (8). The first abdominal bears in addition to the plates mentioned a mid- dorsal plate, which is unarmed. The prolegs on segments 3 to 6 bear nine to eleven. (exceptionally twelve) hooklets (fig. 4), the numbers on several. specimens examined being as follows : 1 2 3 4 Segment. | l =|, | - | ——— — —— || Right. | Left. || Right. | Left. || Right. Left. || Right. | Left. | | | . | ! mn . Abd.3 .|| 10 | 9 || 10 | 10 || 11 | 10 | 10 | 10 nade 9) 99 SN act) ES eres a ee ey | a are ar ae eae). to! Par | a re 1s |) ao) Wes alee Mee cee esalip of om mmeniees \| | | The spiracles are situated just below the supra-spiracular plate and are small. On abd. 7 the plate corresponding to. the outer leg-plate bears only one seta. On abd. 8 the dorsal and subdorsal plates are fused into one which extends to the medial line and there meets the one from the other side. The supra-spiracular and part of the subspiracular (with one of its two sete) are also fused into one and this is pierced by the spiracle, which is larger than those on the other abdominal segments. ‘The position and number of the sete are the same as on the other segments, but here (and also on abd. 9), group 7 consists of one seta only. On abd. 9 the dorsal plate has extended mediad till it meets the one from the other side, while apparently the subdorsal and supra-spiracular plates have fused into one (thus bearing two. 646 C. B. HARDENBERG. sete), while the subventral plate is wanting. The anal plate on abd. 10 contains four sete on each side, three of which are marginal. ‘The anal legs are almost entirely surrounded by a chitinous ring bearing three sete on its lateral aspect, one on the median, and four sete on its caudal aspect. The hooklets on the anal legs are longer and more robust than those on the abdominal legs and are eight to nine in number. ‘The skin behind and between the anal legs is furnished with numerous small sharp-pointed papillae. SECOND INSTAR. Length of larva 6mm. Ground-colour of caterpillar Hesh- colour (XIV). Head, thoracic shields and plates, coxe, thoracic legs and anal plate and leg-plates of anal prolees light seal brown (XXXIV), setiferous plates on 8th and 9th abd. segments chestnut brown (XIV), those on the other segments russet (XV). The thorax has no longer the preponderance in size which was so conspicuous in the first instar; the widest part of the larva is now at the second abdominal segment, from which it tapers gradually, both cephalad and caudad. Hxrap.—Width of head-case 0°83 mm. Head resembles that of the first instar larva, but the suture between frons and clypeus is now indicated. Adfrontals still fused with parietals. Head uniformally coloured; no pattern is yet apparent. THorax.—On that part of the prothoracic shields corre- sponding to the prespiracular wart two more sete have appeared, one dorso-caudad, one dorso-cephalad of the original (Pl. XLVI, fig. 7). The subventral plate has also gained one seta and now bears three. On meso- and metathorax additional sete have appeared on the pre- spiracular (subspiracular) plate, and a minute one just in front of this, while the setee on the subventral plate have increased to three. AppomEeN.—On the abdominal segments the subventral plate SOUTH AFRICAN BAGWORMS. 647 now bears one seta (6) while group 7 has increased to its normal number of three set, instead of two as in the previous instar (fig. 6). On abd. segments 1 and 2 these are divided, so that one of them is implanted on the enlarged mid-ventral plate and the other two on the small wart corresponding with the leg-plate in the proleg bearing segments. ‘The hooks on the abdominal prolegs have doubled and range now from 18 to 20 in number (fig. 6). These are placed in a single loop, not entirely closed, the opening facing medio-caudad. Abd. 7 resembles that of the first instar, except for the appearance of sete 6 (fig. 5). Abd. 8 shows the dorsal plate now divided in two halves, widely separated in the median line, and the development of the seta 6 as on the preceding segments. On abd. 9 the dorsal plates are also widely separated in the median line, the subspiracular plate is split in two, each bearing one seta, while in addition there has appeared on the subventral plate a seta 6. On abd. 10 the arrangement of the sete on the anal plate has remained the same, the anal legs have acquired a posterior leg-plate with four strong sete and on the lateral leg-plate group 7 now consists of three sete as on the other abd. segments. Hooks of anal prolegs doubled, and these are now 16 to 18 in number. THirp INSTAR. Length of larva 9mm. Ground colour of caterpillar pur- plish vinaceous (XXXIX). Head, pattern on prothoracic shield, thoracic shields, thoracic plates, legs, anal plate and leg-plates on anal prolegs, hooks and abdominal prolegs chestnut brown (XIV), setiferous plates on abdominal seg- ments cinnamon brown (XV), ground colour of thoracic shields salmon buff (XIV), body more flattened dorso- ventrally, with the greatest width at the caudal edge of the second abdominal segment. Heap.—Width of head-case 1°'7 mm. Suture between parietals and adfrontals now indicated, especially on the VOL. 3, PART 3. 43 648 - (.-'B. HARDENBERG. upper part of the latter. Relative position of sete the same TEXT-FIG. 2. NYY yy wy" WY WIT Yay Why Wy, hi * Peet a ow asy! Wy ; ne oer “i "wy Wy ly yb, emery My 2 kes ry ’ = ae (| Wy 423 “eye MWS 4 8 = == 3 323 ae 7333 oa hye Gere ere ag. on 3° ” a ieeelinnniieed = Acanthopsyche junodi (Heylaerts), Larva. A. Front view of head of larva in third instar, showing pattern ; set omitted. x 25. 8B. Ventral aspect of head of larva in third instar showing pattern; sete omitted. x 25. c. Dorsal thoracic shields of larva in third instar, showing pattern. x 24. D. Portion of skin around seta 8 on fourth abdominal segment of larva in fourth instar. x 200. §. Antenna of larva in third instar. x 200. F. Armament of skin between and behind anal legs in nearly full- grown larva. x 200. as in previous instars. There is a faint indication of a darker pattern on the head, which, however, appears only on magni- — a SOUTH AFRICAN BAGWORMS. 649 fication, and is as yet not visible to the naked eye (text-fig. 2, A,B). The structure of the head-capsule now shows up as consisting of irregular plaques, giving the appearance of alligator skin. These plaques are especially apparent on the parietals, and are larger and more darkly pigmented on the areas which make up the pattern. On the front they are not distinct, the surface here appears finely pitted. Antenne .—Second joint now considerably enlarged, four times as long as the third and fourth joints combined (text- fig. 2, EB). THoraAx.—The prothoracic shieid is no longer uniformly coloured, but the dark area is now interrupted by several hghter spots, as shown in text-fig. 2, c. There is one mid-dorsal oval spot in the median line, and a broader subdorsal, just above the spiracle, broadening out cephalad to the front edge of the shield. A similar subspiracular hght spot starts below the spiracle and extends dorso-cephalad until it meets the subdorsal. On the front half of the neck-shield these three hght areas meet, and this entire part is thus light-coloured, except for small dark patches around the bases of the sete. The position of the setz corresponds to that during the pre- vious instar. On the mesothoracic shield also we find the first indications of a developing pattern in the shape of a diffused narrow mid-dorsal and a broader sub-dorsal area of lighter colour. The metathoracic shield shows a mere indica- tion of a suffused hght-coloured sub-dorsal patch. On the intersegmental membrane between the second and third tho- racic segments there has appeared a rudimentary spiracle. A pair of ventral plates, fused in the middle line and bearing two sete, have appeared behind the coxe of the third thoracic legs. AspoMEN.—On the first and second abdominal segments a rudimentary seta is found in front of the spiracle. Other- wise they agree in the arrangement of plates and set with the conditions found in the second instar. Abd. 3 to 6 show no differences from the second instar. Prolegs bear on the average twenty hooklets. On Abd. 7 there is no change, 650 ©. B. HARDENBERG. except that group 7 now consists of two sete. Abd. 8 is as in second instar. Abd. 9 also as in second instar, with the addition of a rudimentary seta in front of the single one representing group 7. Abd. 10 no change except a heavier chitinisation. Anal legs with 16 to 17 hooklets. Over the entire body of the larva a great change has occurred in the composition of the skin. While during the first two instars theskin appeared as a structureless membrane, we find this now strengthened and toughened by numerous minute granular plates (text-fig. 2, p). ‘These are somewhat more prominent on the cephalic edges of the segments, and become especially developed at the front edge of the anal segment, where they form a strong band across the dorsal and ventral aspect. Between and posterior to the anal legs the skin is armed with numerous finely pointed teeth directed towards the base of the leg. ‘They are arranged roughly in parallel bands surrounding the medial and caudal part of the leg-plate, and gradually become smaller, and merge into the chitinous tuberosities of the surrounding skin. This structure of the skin becomes more and more pro- nounced with succeeding instars, and reaches its greatest development in the ante-penultimate instar (text-fig. 2, Fr). In the last stage the amount of chitinisation of the skin is considerably reduced. Fourts IwstTar. Length of larva 21 mm. at end of instar, width 6 mm., the place of greatest width at middle of third abdominal segment. General colour of body sorghum brown (XX XIX); on first two abdominal segments Hay’s brown (XXIX). Ground colour of thoracic shield maize yellow (LV), ground colour of head auburn (II), markings on head, thorax, thoracic plates and legs, plates on first abdominal segment, anal plates and legs, seal brown (XXIX), plates on other segments and abdominal legs Hay’s brown (XXIX). Heap.—Width of head-case 3 mm. Adfrontals now dis- tinctly separated from parietals, broad, reaching to two-thirds SOUTH AFRICAN BAGWORMS. 651 the distance between base of clypeus and notch of vertex. Parietals with distinct pattern as shown in text-fig. 2, A. Mandibles with five teeth which are now blunt, the outer two the larger. Labrum strongly chitinised, median notch rounded. Otherwise no change from the conditions in the previous instar. THorax.—There is no change from the conditions in the third instar except a stronger pigmentation and a more pro- nounced pattern on the thoracic shields. Aspomen.—No differences in the arrangement of plates and sete as compared with the previous instar. The small rudi- mentary seta in front of the spiracle (58a?) present on all abdominal segments. Number of hooklets on abdominal pro- legs irregular, varying between 18 and 22, on anal legs from 16 to 18. Firta Instar. The larva reaches a length of 34 mm. Ground colour of body on sides and ventral aspect of abd. 3-6 cinnamon drab (XLVI), on back and ventral aspect of other abdominal seg- ments fuscous black. Setiferous plates on abdominal segments, hooks of prolegs and anal leg, fuscous black (XLVI), anal plate, anal leg-plates, markings on thorax and on head black. Ground colour of head and anterior edges of thoracic shields argus brown (III), ground colour of thoracic shields maize yellow (IV). Spiracles xanthine orange (III). Thoracic legs fuscous black (XLVI) with spots of maize yellow (III) on the coxe and one subventral on the post coxal plates of the third tho- racic segment. Heap.—Width of head-case 4 mm. The pattern on the head has undergone no radical change. T'HoRAX.—On the thoracic segments the pattern has become more broken up, the light ground colour more pronounced, especially on the second and third thoracic segments. On the coxal plates the light spot has also appeared. The structure, position of setze, etc., otherwise agree with those in the fourth instar. 652 C. B.. HARDENBERG. The’ number of hooklets of the prolegs varies. The ar- rangement on three examples, selected at random, was as follows : 1 2 3 Segment. = - | Right. Left. | Right. Left, Right. Left. pee CS ees ese Abd. 3 18 ig "| 2 |) 20 1660) 4s Abd. 4 20 21 24, 20 D0. al ete AES hie een, 20y 8) 20s eed TBE 8 TSA Nese aay Aba 6 - 2 | 20 | 92 20 30: | eg an eeooues Mee ei? ABs ile | nS 9 | | Eee (eed | AGB heer eet oe along ky hye ks AS Sale a 19 | | | ‘ It appears that the males do not pass through the instar corresponding to this fifth imstar of the female. Measure- ments taken of male caterpillars (which remain much smaller than the female) gave a width of the head-case of 3 mm. only, corresponding to that of the fourth instar of the female, and during this fourth instar all caterpillars collected are about of equal size. After the females have moulted they continue growing and feeding, and during this time we find the caterpillars approximately in two sizes, those that will develop into females being the larger. The omission of this instar on the part of the males becomes still more probable in view of the fact that the males pupate fully a month before the females. Srxta Insrar (Uttimate or Prepupan Insvar). Length of female from 24 to 34 mm., average 28 mm., that of the males 22 mm. Ground colour of body on ventral aspect light ochraceous salmon (XV), dorsal aspect orange cin- namon (XXIX). Thoracic and abdominal legs, setiferous plates, ground colour of head and thoracic shields hght buff (XV), SOUTH AFRICAN BAGWORMS., 653 anal plate, anal legplates, dorsal setiferous plates on abd. 8 and 9, as also the pattern on thoracic shields, Verona brown (X XIX), pattern on head cinnamon (XXIX). The structure and arrangement of sete and hooks on prolegs have not changed. The thoracic legs have become comparatively shorter. The entire larva now appears as if re-modelled without growth within the old skin. Width of head-case ot female 3°65 mm., that of male 3 mm. Here we find the remarkable occurrence that after moulting the larva is actually smaller than in the preceding instar and during this instar no food is taken, no growth takes place, and there appears to be an actual shrinkage. As the larva does not expose itself to the outside air and light, it does not acquire its full coloration as in the previour instars, but remains in the pale state characteristic of a newly moulted larva, The actual length of this stage has not been ascertained, for if the larva be disturbed and the bag opened for examination, it immediately proceeds to repair the damage. As the tax on the spinning glands is already fairly heavy, having to spin the cocoon, the extra effort of repairing the bag from day to day (and frequent examination would be necessary) would so weaken the larva that it would probably succumb before pupating, especially when we consider that no food is taken to replace the wasted energy. (5) Hasirs of THE LaRva. The habits of the larva of Acanthopsyche junodii (in common with many other Psychidz) are abnormal in many ways, and different from those of the great majority of cater- pillars. ‘hese peculiarities in behaviour are the result mainly of two factors. The first of these, and the principal one, acting during the earlier part of the life of the insect, is that the distribution of the species is not allotted to the parent moth, but that this responsibility has fallen on the young caterpillar. The second factor is the necessity on the part of the defenceless larva for concealment and protection against predaceous and parasitic enemies. 654 C. B: HARDENBERG. A. Instincts developed in connection with the distribution of the species.—Apart from the ordinary considerations which render a distribution of the individuals advantageous to the maintenance of the species, a dispersal is doubly necessary in the case of the Psychide on account of the enormous productivity of the female (see later). In order to facilitate this distribution over a wide area, there have been developed on the part of the larva, during its first mstar, three special instincts, in contrast with most caterpillars, namely : (a) A decided positive phototropism. (b) An extensive use of the spinning glands. (c) A meandering instinct. (a) Phototropism.—When the young caterpillars emerge from the eggs they find themselves confined in the empty chrysalis-case of the female which after depositing the eges has dropped to the ground. They crawl out of the pupa- case through the lower opening of the bag and move toward the place of the greatest light. When bred in a room, they gathered on that side of the breeding cage which faced the window; and when brought to the table for removal to breeding-cages they all crawled towards the window, which was soon covered with an immense number of young bag- worms that had escaped through the small apertures in the gauze with which the cages were covered. On hatching under normal conditions in the plantations, this instinct prompts them to go to the outside edge of the block or to the tops of the trees. Where the trees have been cut down during the winter for barking, and the branches, with the bags attached to them, are left lying on the ground, preparatory to the subsequent burning when the weather becomes favourable, we see the upper layers of this brushwood and the standing stumps covered with bagworms, all as a result of this photo- tropic instinct. This instinct remains, though not in so pronounced a manner, throughout the feeding period of the caterpillar. As a result of this we see that the tips of the trees are first defohated SOUTH AFRICAN BAGWORMS. 655: and also the outer branches of such trees as form the out- side edge of a block. It also causes the larva to feed on the: young tips of the foliage, as long as any leaves are being produced. Toward the end of the feeding period we find a reversal of this instinct ; the larva now prefers a more con- cealed spot in which to pass its pupal stage. (b) The extensive use of the spinning glands.— While in the majority of caterpillars the spinning glands are not brought into full use until the time for pupation approaches, when the manufacture of a cocoon becomes necessary, we find here that one of the first acts of the young larva is the spinning of a long silken thread. By means of this it descends from the parent bag, and, attaching the thread to a twig or branch, a sheet of webbing is made on which the larva travels up and down. ‘This webbing’ is extremely heht and is disturbed by the shehtest air-current.. This same webbing is produced when the larva has reached the exposed parts of the trees or brushwood, and the area where the larvee have been hatching by the thousands looks as if covered with a layer of spider-web. In later instars the spinning @lands are almost exclusively used in the construction of the bag, and only when the food-supply runs short does the vaterpillar again construct long threads of silk by means of which it descends, and either reaches the ground and crawls. to another tree, or is swayed by the wind to the branch of a neighbouring tree to which it then attaches itself. (c) The meandering instinct.—As a third additional provision to safeguard the distribution of the species we have on the part of the young larva the peculiar instinct that it will not feed during the first few days of its existence. It appears that its desire for a certain amount of travelling has to be satisfied before it will settle down to feed and construct its dwelling. This increases the difficulty of rearing the specimens in captivity from the ege, as the majority of the minute caterpillars leave the fresh youny wattle twigs supplied to them and start travelling about. As will be shown later, the conditions in a plantation after a heavy 656 ©. B. HARDENBERG. infestation are unfavourable to a continued existence of the species therein, and this travelling instinct to a certain extent prevents the larvee from settling in the same spot where their pregenitors have been bred. B. Agents in the distribution.—As the distribution of the species is not effected by the flight of the adult, hke in the majority of Lepidoptera, the period in the life of the insect when this all-important process takes place has been transferred from the adult stage to the first instar of the larval life. In this, except for offering opportunities to facilitate this distribution, the larva plays an entirely passive réle, and is dependent on external agencies. These agents fall under three heads, each responsible for a distinctive type of infestation of a new area. (a) Wind.—By far the principal agent, favouring distri- bution over wide areas, is the wind. Due to their exposed position on the edges of plantations or onthe tips of the trees, brushwood, stumps, etc., the larvee are easily accessible to any air currents, while their lightness, caused by the long silken thread, allows of their being readily picked up and carried for considerable distances. An analogy of this procedure is found in the well-known example of the so-called “ Balloon Spiders.” part of August and until the beginning of October) the climatic conditions in this country are such that heavy wind- During the hatching period of the egg (the later storms are an almost daily occurrence. ‘The wind rises about nine o’clock in the morning and, usually, dies down towards evening. ‘These wind-storms are often characterised by the presence of eddies, and the young bagworms are either driven before the wind or carried up in the vortex to higher strata. In this way they may be carried for miles, and when the wind dies down and the air becomes calm the bagworms slowly descend and fall literally out of the sky on trees, grass, and other objects indiscriminately. Countless numbers perish by landing in places unsuitable for their existence, but such as happen to alight on one of their many food-plants have a chance to attach themselves and start feeding. The writer SOUTH AFRICAN. BAGWORMS. 657 has several times, while walking in the evening, found a number of bagworms attached to his clothing, which could only have settled in this manner. On another occasion it was found that some four dozen young wattle trees, grown in tins on the laboratory grounds for experimental purposes, had become thoroughly infested overnight with young bagworms ; the nearest plantation, or food-plants, on which these could have hatched being a couple of miles away. In this way we must account for a sudden general or gross infestation of a new plantation or block which had previously shown no signs of bagworm injury. It sometimes happens, but such cases are rare, that the young bagworms are forcibly driven before the wind into another plantation near by. When this happens we find that the first few lines of trees have become grossly infested, the infestation becoming thinner and thinner as we penetrate deeper into the block. Only one such lateral or side infestation has come to the writer’s notice, the manner of infestation being such as to leave no room for any other explanation. Evidences of this will be, of course, more con- clusive when the wind has struck the block athwart the rows of trees; if the wind passes through the plantation in the direction of the rows the bagworms may be carried inward for a considerable distance, or even the entire length of the rows, and the distinctive type of such an infestation will be lost. The heavy wind-storms, which are so prevalent during the late winter and early spring, may even carry the young bag- worms after they have made their bag. During the first few days after the bag is made the young bagworm feeds on the upper surface of the leaf and is rather loosely attached. The writer has seen such small bags dropping out of the sky im the same manner as the naked caterpillars. (b) Birds.—lIt is evident that on account of the exposed position of the young bagworms there are many opportunities for them to attach themselves to the feet and plumage of a bird resting temporarily in the trees of an infested plantation. 658 CO. B. HARDENBERG. These birds flying away will carry the young bagworms with them, and when the bird comes to rest on another tree they can crawl off and establish themselves there. In a block which has already been loaded with bagworms as a result of wind infestation, the arrival of these few extra individuals. will not make any difference, but where this has occurred in a plantation or block free from bagworms the results are rather striking, in that we find in the entire block only a tree or two infested at the tip. This may be distinguished as a true bird infestation, cases of which are by no means rare and account for the presence of bagworms in small numbers. in plantations which, through an exceptionally favourable posi- tion, escape the gross infestation by wind or air currents. Inappreciable at first, such occurrences may give rise to an accumulative infestation in succeeding years (see later). Another agent which could cause a similar type of infesta- tion might be furnished by insects flying through the webbing made by the young bagworms, and carrying some larve along with them. Fuller suggests cockchafer beetles as the most likely to be the carriers in this case. Of course, it would have to be a fairly robust type of flying insect, as the smaller ones would become hopelessly entangled in the mass of web- bing. But, unfortunately for the hypothesis, the cockchafer beetles have not yet made theirappearance during the hatching period of the bagworms, their time of emergence being about a month to six weeks after the hatching of bagworms and their distribution is passed. As a matter of fact, the writer has examined several flying insects for the presence of young bagworms attached to them, but has not been able to find any evidence to support this theory. While not precluding the possibility of such an occurrence, the probabilities are so: shght as to make this possible factor negligible. (c) Mammals.—We meet sometimes with another type of infestation, which excludes the possibility of its having been caused by the wind or by a bird, and which only can have been caused by some other individual carrier. This is the type which | have called a sporadic infestation, where SOUTH AFRICAN BAGWORMS. 659 we find in a young plantation that a single tree or group of trees has become thoroughly infested, the remainder of the plantation or block being free from bagworms. The most plausible explanation therefore is that a hare or buck, passing through the web-covered brushwood, has carried a consider- able number of bagworms with it, and entering the plantation of young trees near by has rested in the grass at the base of one of the trees, and the bagworms have crawled off the carrier and ascended the tree or trees in its immediate vicinity. Such cases we find in blocks of young trees where the canopy is not yet formed and there is a great deal of grass growing, which would afford shelter and a hiding place during the day to such an animal. As the trees are accessible to light all around, the bagworms do not in response to their phototropic instinct ascend the tree to the top, but feed on the ends of the branches, and we find the tree infested with bagworms from top to bottom. Several such centres may be found in one block, showing the spots where the carrying animal had sought a temporary halting or resting place. Both these local bird and sporadic infestations may act as centres for the spread of the insect in a new locality, and in the course of a few years may increase to a more general infestation of the entire block. This is what we have termed an accumulative infestation, as distinguished from the general or gross infestation, caused by wind as the carrying agent. By the wattle growers the two are often confused, because the beginnings of an accumulative infestation are not noticed, and it is only when the block becomes thoroughly infested that attention is drawn to it, and this is then de- scribed as a sudden appearance of the bagworms in enormous numbers. We are generally able, however, to prove the accumulative nature of the infestation by the presence of old bags scattered over a few centres in the plantation. C. Instincts developed in connection with the protection of the larva.—These find their expression in the formation of a bag as a protective covering during the feeding period of the larval life, and the gregarious habit 660 C. B. HARDENBERG. which develops when the larva has stopped feeding and is ready to enter the pre-pupal and pupal stages. The first act of the young bagworm, after its dispersal has been accomplished and its meandering instinct has been satistied, is to construct a bag for concealment. This bag is at first conical, widest at the head end of the larva, and during the first days of the first instar it is carried upwards, while the caterpillar feeds on the upper surface of the leaf. Within a few days it starts feeding on the underside, and the bag hangs downwards, the larva, when moving, crawling along the under surface of the twig or branch. In both these positions the abdomen is deflected dorsad at its juncture with the thorax, the entire-weight being carried by the thoracic legs alone. The abdominal and anal prolegs are not used in locomotion, and are probably only brought into play in so far as they hook into the silk lining of the bag and prevent this from shpping off, or the larva from being dis- lodged by asudden jerk. When a full-grown larva is removed from its bag and put back on the tree it then also only uses its thoracic legs in locomotion, and moves with a pronounced dorso-flexion of the abdomen, and along the underside of the branches. Even when placed on a flat surface it does not crawl in a straight line, using both thoracic and abdominal prolegs, but moves on its thoracic legs only, the abdomen being held bent upwards and its caudal part curved ventrad. Whiie it would not be possible for the abdominal prolegs to grasp a small twig between them, as they are very short and rather widely separated so that their hook-bearmg surfaces could not be sufficiently approximated, it is clear, from its method of locomotion on a horizontal surface, that the larva has lost through disuse the power of co-ordination in respect to these organs. The points of the hooks are directed out- wards, so that, when the planta is pushed out, these hooks are spread and hook into the silk of the surface against which the sole is pressed. The points of these hooks are extremely long and sharp. The bag, as stated, is at first conical and just roomy enough SOUTH AFRICAN BAGWORMS. 661 to enclose the larva. With subsequent growth the bag is enlarged and from conical becomes oval, that is, wide in the centre and narrow at each end. ‘he space inside is now larger than the volume of the larva requires. The manner in which the change in the form of the bag is brought about we have not been able to observe. As the larva feeds, particles of the leaves and twigs are being thrown back and attached to the neck of the bag for a short distance. No. regularity is observed except in so far that all these particles are attached lengthwise, and, as they are fastened for a short distance only, they project tangentially in the direction of the long axis of the bag. During the feeding they are loosely attached. The firmer attachment is probably accomplished during the mght, as in order to do this the larva will have to expose itself for a considerable distance, and it may in some cases even be necessary for it to leave the shelter of its bag. Everything which may be attached to the branch or leaf, such as seed-pods, parts of twigs, even other pieces of insect architecture are indiscriminately made use of, and the result is at times a very curious mixture of ingredients. According to the plant on which the bagworm is feeding, the composition of the bag-covering, and thus its aspect, varies considerably, but the oval shape and the partial longitudinal attachment of the component particles are constant characters. Toward the later instars the differences between the male and the female bags become apparent; the bag of the female is large and plump, while that of the male is shorter and has a longer point, but it is not relatively narrower. These differences are not absolute, since according to variations in food-supply we may find small females and large males, and it is therefore not always possible to judge with certainty from the appearance of the bag the sex of its occupant. Measurements of several hundreds of bags, containing mature: larvee, gave the following results : Male bags: Average length 46°95 mm., width 17°32 mm. Maximum length 56 mm., — width 21 mm. Minimum length 39 mm., — width 15 mm. 662 ©. B. HABDENBERG. Female bags: Average length 53°59 mm., width 18°25 mm. Maximum length 63 mm., width 22 mm. Minimum length 46 mm., width 15 mm. the differences in length being thus more pronounced than those in the width of the bag. D. Feeding habits.—While feeding may take place at any time during the day, it is mostly done during the early part of the morning, as soon as the dew has evaporated off the leaves. Usually no feeding takes place durimg the hottest part of the day, nor durmg a rain or mist, when the foliage is wet. No feeding has been observed during the night, but it is possible that the larva shows some activity during that period in making enlargements to its bag. The inside of the bag is scrupulously clean, ejecta and cast skins being carefully removed, as the presence of these might influence the health of the caterpillar. Food-plants.—The original food of Acanthopsyche junodi was doubtless furnished by the leaves of Acacia horrida, A. caffra, and other species of Acacia which form a great part of the native thorn bush flora. Alice Pegler, in an article in ‘Grocott’s Penny Mail, Grahamstown, December Ist, 1909 (reprinted in ‘ Agricultural Journal of the Union of South Africa,’ vol. vi, p. 215, 1909), writes that A. melanoxylon is occasionally attacked, and mentions as another native food-plant Loranthus dregei. We have found it also on Combretum glomeruliflorum, and undoubtedly continued observations will reveal a great many more native plants which will furnish a means of sub- sistence for the larva of this species. With the clearing of the thorn bush and the planting of the congeneric A. mollissima the bagworm has adapted itself very readily to this tree, and, the food-supply being now prac- tically unlimited, has thrived exceedingly. Besides these, the bagworm has lately become troublesome as a pest to various ornamental trees and fruit trees. Amongst the former we have found it on oak, roses, poplar, Pride of India, Pinus insignis and some of the less oleaceous SOUTH AFRICAN BAGWORMS. 663 Eucalypti, while the privet is occasionally attacked. Amongst the fruit trees may be mentioned: apple, plum, cherry, apricot and guava. Peach, pear, and quince do not seem to be favoured, neither the citrus trees nor vines. We have seen quarter-grown bags on citrus trees in a heavily infested neighbourhood, but they did not reach maturity ; while on peach, under similar circumstances, we have found the young bagworms eating the rind of the young fruit. Character and degree of injury.— While the young bagworms to a certain extent destroy the blossoms of the wattle, the main injury is the destruction of the foliage. Not only are the young leaves eaten off as fast as they appear, and the entire tree defoliated, but even the bark of the young shoots is chewed off when the food becomes scarce. Besides this injury to the tree due to the removal of the foliage, the bagworm often causes another kind of damage during its resting period. When it attaches its bag, pre- paratory for pupation, it fastens it securely to a twig by means of a strong silken strand stretched tightly around it. This strand acts as a ligature, the flow of the sap downward is interfered with and a swelling occurs just above the strand. Ultimately the strand cuts in deeper and deeper, and with a sudden gust of wind the distal part breaks off. The writer has seen the tops of several young trees destroyed in this manner. Where the bagworms gather in clumps on the stem of the tree, generally near a fork, for pupation, they attach them- selves by gnawing off a considerable portion of the bark, the particles of which are interwoven with the silk by which they are attached. The injury caused by the two last procedures may be considerable. Progress of infestation.—At first the caterpillars gather on the tops of the trees, which are defoliated in the early part of the season (November). After this the larve slowly travel downwards, consuming more and more of the fohage, so that by the end of December the majority of the trees have become bare, and new growth is kept down as VOL. 3, PART 3. 44. 664 CG. B. HARDENBERG. it appears. The infested plantations can be recognised at a distance by the reddish-brown colour. This condition remains until the end of the feeding period (end of February and first half of March), after which the tree has a chance to grow again, provided the conditions as to rainfall are favour- able. Generally, however, the rains practically cease by the beginning of March and the trees remain bare until the following spring. This means, apart from other considera- tions, the loss of a full season’s growth; and where thousands of acres are affected, the commercial loss is very considerable. HK. Parasites and enemies.—Four species of Hymeno- pterous parasites have been bred from the caterpillar of A. junodi, and the same number of Tachinide. None of these have as yet been determined. ‘I'he total number of bagworms killed by these various parasities amounts to only about 24 per cent. under the most favourable conditions; so that, from an economic point of view, they are not of great importance. Predaceous enemies.—Amongst these, insectivorous birds take first place. The Yellow Weaver Bird has been observed to be an industrious destroyer of bagworms in the fall, watching for the caterpillar to expose itself when feeding or crawling, when with a quick jerk the insect is pulled out of its bag. Fuller mentions the Little Silver-eyed Bird of the thorns as “ carefully examining the lower entrance of the bag to capture, perchance, an emerging male or waiting female, or even the young caterpillars.’ Alice Pegler in ‘Grocott’s Penny Mail,’ mentioned above, states that Sprews are the chief robbers in the Cape Province, while Mr. T. M. Mackenzie has reported to me that a small green parrot occurring in the plantations will break off a twig to which the bagworms are attached, tear open the bag with its strong beak and eat its contents; but I have not been able to observe this myself. During the resting stage in the winter-time many of the bagworms are devoured by rats. This is especially the case in young and grassy plantations. The bag is gnawed open, SOUTH AFRICAN BAGWORMS. 665 usually from the top, and the pupa extracted. Sometimes the bags are thus attacked im sitw, sometimes they are torn off the branches and collected at the base of the tree, there to be consumed at leisure. The rodents responsible for this destruction belong to two species, Mus concha and Mus zuluensis. F. Diseases of the Wattle Bagworm.—The Wattle Bagworm is subject to various diseases, all of which attack it in the caterpillar stage. The most important of these is: The Bagworm Fungus—Isaria psychide Pole Kvans.—This was first found by T’. M. Mackenzie in his plan- tation at Cramond, Natal, and has since been distributed through many wattle plantations in Umvoti County and else- where. ‘The spores of this fungus, when they are attached to the leaves, are eaten by the bagworm. Arriving in the stomach of the insect, the spores germinate and grow, the fungus threads ramifying through the body of the caterpillar until it becomes a sohd mass of hyphe. The caterpillar dies in about five days after infection, and the body becomes mummified. The fungus continues to increase until it grows through the bag and appears on the outside, where spores are formed as conspicuous white pustules. These spores are dis- tributed by the wind and dropping on the leaves are in turn eaten by other bagworms, which thus become infected. Those bagworms which are attacked early in the season, when they are still very young, die and drop to the ground, and, being still very small, are easily overlooked. The older caterpillars which are infected later in the season may reach the stage when they stop feeding and attach themselves. It is from such specimens that we usually see the fungus grow out of the bags in the fall. Hven of these later infected bag- worms a great number drop to the ground, and we have noticed in some plantations the ground covered with such fungus-killed bagworms. Notwithstanding the deadly nature of the fungus and the progressive infection which takes place during the season, the disease is not sufficient to check the pest to any appreciable 666 | C. B. HARDENBERG. extent, as in every plantation known to me, where the fungus is well established, heavy losses due to bagworm injury still occur. This appears to be mainly due to the circumstance — that when the bagworms begin to attach themselves they do so as a rule fairly low down on the trees. When the spores are being produced from the infected bags they, in order to be effective, must reach the young bagworms on the tops of the trees, where they are feeding at that time. The spores must therefore be carried up, and this probably occurs only under exceptional circumstances. This is presumably the reason why the majority of the bagworms become infected only at a later stage in their larval life when they are feeding on the lower branches of the trees. For these reasons the fungus is not as effective as otherwise it might be. Actual counts of some twenty thousand bagworms, gathered in an infected plantation, showed that of the bagworms which reach the resting stage the maximum of fungus-infested indi- viduals was only 22°5 per cent. of the total number. This was in a plantation where the fungus had been well estab- lished, and which was suffering from a gross infestation of bagworms for the third consecutive season, conditions which would seem to be very favourable to promote the maximum efficiency of the fungus. This percentage is, of course, not high enough for the effective control of the bagworm by this means alone. Other diseases.—Besides the fungus there are many other diseases to which the bagworm is subject. Most of these are as yet of an unknown character, but are probably of an intestinal nature. One of these, which we have re- peatedly noticed in the laboratory, causes constipation of the larva with an untimely death as the ultimate result. The caterpillar generally succumbs at a very early age. Another disease is more of the nature of dysentery, and appears to be induced by the consumption of wet food. It is especially prevalent amongst the young bagworms, and is probably the cause of the death of so many after a prolonged wet spell. We have noticed that the bagworm stops feeding ys SOUTH AFRICAN BAGWORMS. 667 during a rain or mist, and only starts again after the foliage has dried. The older bagworms can withstand starvation for a considerable time, and can thus safely pass through a rain or mist of several days’ duration, but the young cannot starve as long. If there be a prolonged period of continuous rain or mist, so that the foliage has no chance to dry, the young larve are by hunger compelled to eat wet food. ‘This appa- rently causes a disease (dysentery) by which they are killed. Or it may be that this wet diet causes more fluid evacuations, as a result of which the bag becomes soiled inside. The larva then leaves its bag and tries to construct a new dwelling. During this time, when the larve are unprotected against the inclemency of the weather and against enemies, probably a great many of them perish. Such as survive, in their haste to make a new bag, take any material which comes handy, and old bags, deserted by their inhahitants, are often made use of. In this case we find a great percentage of so-called compound bags, made up of several, sometimes as many as four or five, of which only one is occupied. An abundance of such bags in a plantation is a sure sign that the conditions have been at some time unfavourable to the bagworms. In addition to the diseases here enumerated, there are still others of an unknown nature, which attack the older bag- worms. ‘I'he effect is a blackening and ultimate mummifica- tion of the larve, and many of the older bags which were opened showed evidence of this form of disease. In fact, these diseases may destroy more bagworms than the fungus. We have found amongst several thousands of bags, collected in the resting stage, that 30 per cent. of the larvee had been killed off by these diseases. As far as we know, none of these diseases are contagious ; we have found an abundance of perfectly healthy larvee amongst the diseased caterpillars. An undoubtedly contagious disease has been noted during the season of 1913-14. Our attention was first called to it by Mr. H. Holley of Wartburg, who noticed a great number of dead bagworms in a corner of a plot about forty 668 O. B. HARDENBERG. acres in extent. When the writer visited this plantation a few days after being notified, the disease had spread over about one-tenth of the block, starting from the corner where first noticed; and in that affected area I found 95 per cent. of the bagworms dead. A month later this plantation was again visited, this time in company with the Government Plant Pathologist, and we found then the entire block infected, so much so that after a couple of hours’ search over the entire forty acres we could find only two live bagworms. Later on, the same, or a very similar, disease was found in the plantation of Mr. EK. T. Hill of Seven Oaks, and evidences of its presence have since been found at several other plantations. The efforts to isolate and artificially propagate this disease have thus far not been successful. All these diseases undoubtedly depend for their virulence to a great extent on climatic conditions, and their efficiency as a natural check to the bagworm therefore varies from year to year. They constitute a factor which is as yet beyond human control and thus must remain unsatisfactory. We can artificially promote the spread of some of these diseases, but cannot control their action. (6) THE Cocoon. Toward the end of the feeding period, just before the larva undergoes its last moult, it prepares for the pupation by putting an extra inner lining to the bag. ‘his inner lining is very thin and closely applied to the original silken covering. It serves to cover the irregularities of the imner surface ; and the inside of the bag, which with increasing age has become somewhat discoloured, is now clean, white, smooth, and somewhat shiny. This inner lining made, the larva moults into the last instar. After this no more food is taken, and the caterpillar remains light in colour. Soon after its moult it begins to spin its cocoon. ‘This consists of a SOUTH AFRICAN BAGWORMS. 669 cylinder of tough white silk, attached to the upper and the lower neck of the bag and free in the centre, extending as a hollow column through the centre of the cavity of the bag, and just wide enough to accommodate the pupa to be formed. The caterpillar then turns round and lies with its head facing the lower opening. The change into the chrysalis now takes place, and the last larval skin, much broken and _ coin- pressed, is pushed to the very top of the cocoon. The males pupate about a month earlier than the females. (7) Tue Popa. Male Pupa.—Leneth of pupa 12 mm., width 4:25 mm., with the place of greatest width across the humeri or place of insertion of the front wings, depth 4°75 mm. near the caudal edge of the mesothorax. General form slender, with humped back due to the prominence of the mesothorax, and with the last three segments of the abdomen curved ventrad. The entire pupa has a somewhat shrunken appearance, not plump and rounded hike most other pupe. Posterior edges of the abdominal segments very prominent, giving the abdomen a telescoped appearance (text-fig. 3, A, B, C). General impression of sculpture: smooth, subshiny. Colour dark reddish brown, nearly black when mature, ventral aspect of wing-cases and the last two abdominal segments lighter in colour. There is no bloom on the pupa. Head medium-sized, shghtly protruding ventrad, shiny, very finely punctulate (text-fig. 3, pb). Clypeus — short, transversely striate, directed dorso-caudad, thus causing a transverse depression in the pupa at the end of the labrum. Paraclypeal pieces distinct, roughly triangular in outline, flat and transversely striated, not convex and rugose as in many Psychide. Labium long with the labial palpi minute, showing as small rounded lobes at its distal end. Maxillary palpi large, distinct, extending caudad as far as the tip of the labial palpi. TExtT-FIG. 3. cnnenem Aypeus -----pare ypeaT AS Seat wrayer ---70%.pa/p ---/abium -\--- Jaéo/ pole. 770F. ont —e AWQOUU TU end AMIN ETAT ry a) Chypeus ~-----+-paraclypea/= SOUTH AFRICAN BAGWORMS. 671 No glazed eye visible, this being apparently covered by the: broadened basal part of the antenna-cases which extend caudad nearly as far as those of the first legs. On top of the head is a slight median swelling, somewhat rugose, continued in a blunt, low, median ridge on the pro- thorax, the cocoon breaker. The entire dorsal aspect of the prothorax is deeply striated transversely. Near the anterior edge it bears a small setiferous puncture on each side of the median line. Mesothorax massive, slightly striated transversely, median line very slightly raised (text-fig. 3, #). On its dorsal aspect it bears a prominent tubercle on each side of the median line, bearing a short seta. Laterad and latero-cephalad of these there are two other similar but smaller punctures. Anterior wing-cases reaching caudad only as far as the end of the third abdominal segment. Posterior lateral angles of wing-cases rounded, the two ventral angles slightly divergent. Metathorax very narrow on the median line. Anterior angle of insertion of hind wing-cases broadly rounded, obtuse. Posterior wing-cases extending in some pupze beyond the posterior lateral angle of the front wing-cases and half-way along their ventral edge. Posterior lateral corner of these: hind wing-cases rounded and situated at the middle of the second abdominal segment. Near anterior edge of metanotum there is a setiferous tubercle on each side of the median line. All the abdominal segments bear sete corresponding in position to the large primary sete of the caterpillar, and a dark oval non-setiferous spot just mediad of the spiracle. Abdominal segments 3 to 5 have a row of small teeth on the middle of the posterior declivity of the segment, which TEXx?T-FIG. 3.—Acanthopsyche junodi (Heylaerts). Pupa. A. A,a. Male pupa, ventral view. x 25. B. B, b. Male pupa, side view. X 25 c. Male pupa, dorsal aspect. x 2°5. Db. Male pupa, an- terior portion, ventral aspect. x 15. &. Male pupa, anterior portion, dorsal view. x 15. ¥F. Male pupa, caudal end, dorsal view. x 15. G. Male pupa, caudal end, ventral aspect. x 15. Hu. Female pupa, dorsal aspect. x 4. 1. Female pupa, ventral view. x 4. J. Female: pupa, anterior portion, ventral aspect. x 8. 672 Cc. B. HARDENBERG. teeth reach their greatest development on abd. 5 (c, F). They are composed of a narrow base, flattened laterally, the tip of the teeth recurved cephalad and ending in a short seta. On abd. segment 6 these teeth are wanting, and the segment is unarmed. On abd. 7 we find at the cephalic margin a small crescent-shaped plate, which bears on its caudal edge a number (8 to 14) of sharp-pointed teeth, flattened dorso- ventrally and directed caudad (r). Abd. 8, which with the following segments is deflexed ventrad, also bears a toothed plate like abd. 7, but stronger, and with more (12 to 14) teeth directed caudad. The sete are longer than those on the preceding segments. Abd. 9 appears to be unarmed, and bears on its ventral aspect the genital scar (G). Abd. 10 is still more deflected ventrad. Cremaster absent, but instead of this there are two sharp hooks derived from the anal pro- legs, flattened laterally, slightly divergent, with broad base, and with strong, sharply pointed hook directed cephalad. Female Pupa.—Leneth 15 to 20 mm., vermiform, cylin- drical with tapering ends. Abdominal segments 2 to 6 well- marked, strongly segmented, large; thoracic segments and first abdominal segment narrow, contracted, and less distinctly separated. Abdominal segments 7 to 10 smooth, appearing nearly confluent (text-fig. 5, H). General colour reddish brown, the posterior third of the seoments with an olivaceous tint. Head, thoracic segments, and dorsal aspect of abdominal segments | to 7, darker pig- mented. , Head small, deflected ventrad, mouth-parts distinctly visible, clypeus small, triangular: paraclypeals prominent (5). Max- illee small, with large maxillary palpi. Leg-cases small, trans- verse, semi-detached in young pup, becoming more obtect as pupa matures. Thoracic segments with median dorsal ridge (cocoon breaker ?), meso- and metathoracic segments have m addition a slight transverse ridge which bears four setiferous tubercles (H). Abdominal segments 2 to 5 bear on their dorsal aspect a row of small spines, directed cephalad and arising from the middle of the posterior declivity of each SOUTH AFRICAN BAGWORMS. 673 segment; segments 6 to 8 unarmed. All of the segments 2—7 each show a dark puncture dorsad of the spiracle. Abdominal segments 1 to 6 strongly striated transversely on their dorsal aspect; ventral aspect, as also the whole of segment 8, practically smooth. Last segment with a minute hook ventrally situated and directed cephalo-ventrad (1). Duration of the pupa-stage.—Individual records of the length of the pupa stage are not available, owing to the practical difficulty in rearing the adult after repeated dis- turbance of the larva in the prepupal stage or the pupa itself ; but from a comparison between the pupation curve and emer- gence curve in text-fig. 4 we can readily deduce the average duration of the pupal period. These curves are based upon the examination of one thousand bags collected in a certain block every week, weather permitting, from April 9th until September 10th, some twenty thousand bags in all. From these curves we see that the males began to pupate during the first week in April, the number of male pup per thousand bags rapidly increasing, until at the end of the month a maximum was reached, showing that the male larve had completed their pupation. This percentage remained practically stationary, allowing for shght variations in the various lots, until the middle of June, after which. some isolated males began to appear. The maximum emergence of males, shown by the empty pupa-cases, was reached on August 16th, the main emergence starting about the middle of June. Therefore, taking as an average the period from the middle of the pupation curve, April 17th, to the middle of the emergence curve, July 31st, we find for the duration of the pupa stage the average of three months and thirteen days. The females did not start to pupate until about a month after the males, the first’ female pupz having been observed on May 10th; the maximum, showing that practically all had pupated, occurred about the middle of June. Shortly after- wards the first female adults were found. The general emer- gence, if it may be so called, started about the middle of TRXU-HIG, 4, 760 770 730 730 740 Soles So 720 700 770 40 50°60 70 80 90 Number of Days after April 7% JO PHIYO2 S009 O00L Kiana Ui Lalli LZrergece, 770/es Lypavio7?,, 777Q/esS LZrrergence, ferma/es PYOM OP. females Pupation and emergence curves of males and females, Acanthopsyche junodi (Heylaerts), SOUTH AFRICAN BAGWORMS. ae OLS July, and reached a maximum during the second week in August, the average duration of the pupa stage of the female being two months and seven days. We thus see that, although the females keep on feeding for a month longer than the males, yet their pupation period is roughly a month shorter, so that adult males and females appear almost simultaneously. (8) THe Inaco, Emergence of Males.—When the time of emergence approaches, the pupa works its way partly out of the lower end of the bag, until it protrudes far enough to free the wing- cases and the legs. ‘The pupa then splits in the usual manner, and the moth after emergence crawls on to the side of the bag, where it remains motionless to allow the wings to dry and harden. ‘The time from the first extrusion of the pupa until the moth is ready to fly is very short, the entire process being accomplished in about ten minutes. Emergence takes place in the middle of the day, generally between the hours of ll a.m. and 3 p.m. ‘The moths are active and on the wing during the hours of the brightest sunshine. Sunshine is a very important agent in promoting emergence, and in young plantations, or on the edges of the blocks, where the sun has more easy access to the bags, the moths appear earlier than in the centre of old shady plantations. Emergence of Females.—The female does not leave the chrysalis case. When mature, the anterior part of the pupa is pushed off hke a cap, exposing the head and the first two thoracic segments of the female moth. The seam along which this cap is severed is on the ventral side fairly regular ; it follows one of the intersegmental incisions, and is here pro- bably made by a restricted turning motion of the head of the insect, which is armed with two pointed prongs on its ventral aspect (text-fig. 5, c). The ventral slit thus made, a pushing forward of the enclosed moth removes the anterior part of the pupa case, which thus breaks along the edge of one of the TEXT-FIG. 5, A \ Copelatory Onenirg \Ovjposior B (MW L ill \ Ml et ve ih Mii me f eT Wp a iT ail Ay i ME. XY AUN x, i] HU fi 0 Tr a i aS 6 H as a Acanthopsyche junodi (Heylaerts). Imago. A. Female in chrysalis-case, side view. X 4. 8B. Posterior seg- ments of the abdomen of female, side view. x 8. c. Female, head and thorax, ventral aspect. x 16. pb. Female, head and thorax, side view. x 16. E. Female, opened dorsally, showing the mass of egos. X 4. F. Tuft of scales of body-segments of female x 12. a. Single scale from tuft; note curved point, constricted base, and striation of surface. x 36. HH. Antenna of male, show- ing pectinations. x 20. 1. A single ramus of antenna. X 250. SOUTH AFRICAN BAGWORMS. 6077. thoracic segments on the ventral side and tears off irregularly on the dorsal side. Normally this anterior part of the female,. including the head and the first two, or all, thoracic segments, is the only portion of the moth which is exposed during the reproductive period (text-fig. 5, a). Percentage of individuals reaching maturity.— The mortality amongst the young after hatching and during the earher larval instars is beyond computation. Owing to the distribution by external agencies, their finding a suitable food-plant is a matter of chance, and such as find favourable surroundings suffer from many vicissitudes during their first larval instars, amongst which unfavourable weather conditions play a most important role, by favouring the development of diseases of various kinds. Of those which pass this infantile stage and become healthy, feeding bagworms, we have found that toward the end of the feeding period only about one-quarter reach the pupa stage. Our examination of some twenty thousand bags, gathered from week to week in lots of a thousand each, showed that the average number of males which pupated was 250 per thousand, of the females, which have a longer feeding period, 190 per thousand, the deaths being principally caused by parasitism. Of this average of 230 male pupx, only 210 male moths emerged, due to the fact that some twenty of these larve, though parasitised, had been strong enough to pupate, the parasite emerging from the pupa later. Of the females there was. practically no difference between the number of pupz formed and the number of moths emerging. The adults.—The male moth was described originally by Heylaerts in the ‘Ann. Soc. Ent. Belg.,’ vol. xxiv, p. cxxx (1890) under the name Humeta junodi, but the species is now placed by Janse in the genus Acanthopsyche, and the latter author gives a re-description of the male of this: species. The female has, so far as known, not been described, and a description therefore follows : Acanthopsyche junodi (Heylaerts)\—Female.—Wing- less. Length 18 to 22 mm., vermiform, body-wall soft and. 678 Cc. B. HARDENBERG. transparent. Head and part of thoracic segments chitinised, the extent of this chitinous area increasing caudad. Colour of body cartridge buff (XXX), diffused with pale pinkish buff (XXIX), sternum of three thoracic segments dull blue- violet (XXXVI). Head deflected ventrad, conical, bearing two slender, pointed, slightly diverging prongs, arising from the vertex. Ventral aspect of head very convex. Antenne and trophi absent. A pigmented black spot on each side near the base of the prongs. Thorax small, partly chitinised, sterna of first and second, sternum and pleure of third segment membranous (text- fig. 5, c). Thoracic legs rudimentary, represented by a minute papilla, surrounded by a chitinous ring, with larval arrangement of sete. Chitimised parts of thorax smooth, shiny. Abdominal segments large, membranous, surrcunded by a ring of downy hairs, mixed with scales (text-fig. 5, r, G). These rings are incomplete on the first few abdominal segments, being represented on abd. 1 and 2 by a paired ventral and prespiracular tuft, on segments 3, 4 and 5 by a paired ventral, a prespiracular and paired sub-dorsal tufts, and become practically complete on segments 6 and 7. On abd. 8 only the subdorsal tufts remain, while abd. 9 and 10 are devoid of these hairs. Colour of hairs ight buff (XV), of scales yellow ochre (XV), the latter predominatin on the sixth and seventh abdominal segments. New Hanover, Natal, July 1915, bred by C. B. Hardenberg. ‘Type (alcoholic specimen) in collection Hardenberg. oO oD (9) Hasits or THE IMaGos. The males are very active, quick and strong flyers, as 1s indicated by the broad, massive thorax and slender, pointed abdomen. At the moment the wings are sufficiently expanded and hardened, the moth takes flight in search of the female. While normally under ordinary plantation conditions the flight need be but short, as the females are near, the males SOUTH AFRICAN BAGWORMS. 679 are well capable of long-sustained flights. Several cases have come under our observation where females have been visited by males which could only have come from a couple of miles distant. Flying only takes place in the bright sunlight and during the hottest part of the day, roughly between 10 a.m. and 3 p.m., while if the sun be temporarily obscured by a cloud, no males are to be seen. Fuller, in his various discussions on the bagworm, mentioned previously, maintains that the males are very weak flyers and can only fly for a short distance. His observations were based on moths bred in the laboratory, and under those conditions, in the absence of sun and free air, the moths make no attempt to fly and visit the females, even when the latter are confined in the same breeding cage. This strong power of flight, although no longer necessary under the plantation conditions, may well have been advan- tageous, or even essential, under the original environment, namely im the scattered thorn bush, where the species used to breed. Here long distances might have to be covered across the veld before a receptive female was found, and in the absence of warning coloration or any other means of defence, its only safety from birds would he in a quick flight. The transparency of its wings (the few hairy scales with which the wings are covered on emergence are shaken off with the first vibrations) also assists in making the moth very inconspicuous in its flight across the open spaces. Copulation.—The male, upon finding a female bag, alights at the lower end, and with its wings and antenne (text-fig. 5, 4,1) vibrating continuously it endeavours to introduce its abdomen through the neck of the bag. This is usually accomplished in afew minutes, the neck of the bag being partly invaginated in the process. Now with a series of efforts the abdomen is fully stretched so as to reach the female. ‘lhe normal length of the abdomen of the male is about 5 mm., but after full extension it reaches a total length of 20 mm., at which point the intersegmental membranes are stretched to their fullest extent and apparently the limit is reached. But even this VOL. 3, PARI 3. 45 680 C. B. HARDENBERG. would not be sufficient to reach the end of the abdomen of the female (which lies head downward) so as to effect copula- tion in the usual manner, and, as a matter of fact, this does not take place. We have opened the bags at various periods, from five to twenty minutes after the male had alighted, and in no case have we seen the abdomen of the male extending further than just inside the anterior edge of the chrysalis case of the female. We are therefore inclined to believe that copulation takes place in the following manner. The male pushes up its abdomen through the neck of the bag until it reaches the edge of the chrysalis-case of the female. Here the abdomen is inserted for a short distance between the body of the female and the pupa-case, the two prongs on the head of the female probably acting as a guide, and the semen is injected into the cavity of the chrysalis. The spermatozoa travel upwards, assisted probably by movements of the female’s abdomen, until they reach the eggs which are being deposited in the upper end of the chrysalis case, where they permeate the mass and fertilise them. Apart from the apparent physical impossibility of the normal method of copulation, we have the following evidence which tends to support this view of the way in which fertilisa- tion is accomplished : (a) Often some of the eggs have already been laid before copulation takes place. (b) The freshly laid eggs are soft, and frequently polyhedral through mutual pressure ; later, presumably after fertilisation they become rounded and firmer. (c) The eggs are, as a rule, not all fertile; between the developing eggs there are found a number of such as have remained soft and have shrivelled, showing that the sperma- tozoa of the male did not permeate the entire mass. After the “copulation” is completed, the male withdraws from the bag and flies away. Whether it 1s able to fertilise another female we have not been able to ascertain. Although conceivably this may happen under natural conditions, we have never observed plural copulation on the part of the: SOUTH AFRICAN BAGWORMS. 681 male taking place in captivity, nor have we found any case where the female bag was visited successively by more than one male. Nevertheless we have not untrequently seen two males on a female bag simultaneously, each one penetrating the neck of the bag, apparently oblivious of the presence of the other. It frequently happens that the male is not able to extricate itself, and it then dies attached to the bag. In this case the neck of the bag remains obstructed, and the young hatching from the eges cannot escape, and they perish in the bag. The life of the adult male, from data obtained in captivity, is short, about three to four days. While the male during the short period of its life is ex- tremely active, the female during its adult stage remains in its chrysalis-case almost motionless. At most, if the male be tardy in arriving, it will crawl down and expose its head through the neck of the bag, retreating again upon the arrival of the male. Should, however, no male make its appearance, the female will expose itself further and further, until it drops out of the bag and dies. Such females as have dropped out of their bags are not attractive to the males, and invariably die without having been fertilised. The female moth lives for about a month. Parthenogenesis.—While it is claimed that in some species of the Psychidw parthenogenetic reproduction may take place for several generations in succession, we have no evidence which would make us inclined to believe that this occurs in the case of A. junodi. The great number of males would make parthenogenesis unnecessary and in fact unlikely, and in our experiments we have never found fertile eges in cages from which males had been excluded. Keo-laying.—Immediately after copulation, and some- times even before, egg-laying begins. The eges are deposited in the upper part of the chrysalis-case, where they are tightly packed and mixed with the hairs and scales from the abdomen of the female. As the chrysalis becomes filled the body of the female shrinks (Pl. XLV, fig. 13) and retreats until, when 682 C. B. HARDENBERG. all the eggs are laid, the empty female drops out of the bag and dies. Fertility.—The fecundity of the female bagworm is remarkable. Fuller, in his first account of the Wattle Bag- worm in 1899, gave the number of eggs for one female as 1600, while in his later account, in August, 1913, he said: “Some six hundred eggs are laid; the number may be more or less in accord with the size of the female. Actual counts of eggs in 1911 and 1912 ranged from five hundred to seven hundred and fifty per female.” Our counts of several egg-masses during 1913 and 1914 gave the following results: Locality. | Year. Number of eggs in bag. | | | New Hanover | ME Ll } — = Wartburg .| 1913 1544 | 1888 | 2604 2678 | 3098 — = Greytown .| 1913 | 1177 | 1670 | 2197 | — — — Clan Syndi- | | cate . .| 1914 | 438] 678 | .910| 984] 997 | 1023 | 1082 | 1090 | 1101 | 1101 | 1206 | 1211 | 1849 | 1410 | 1416 | 1433 | 1517 | 1564 | 1584 | 1599 | 1605 | | | | | | From this table it is evident that the fertility of the female varies considerably, ranging from a minimum of 438 in a specimen from the Clan Syndicate plantation to a maximum of 3098 obtained from a female in the Wartburg plantations. The Wartburg specimens were collected in a plantation where the infestation was slight, and the bagworms were thus allowed to develop to their full extent. We see that the range here is from 1544 to 3098, with an average of 2580 per female. This is a high average and will only be reached under the most favourable conditions. The Greytown specimens were obtained from a block where a sudden heavy infestation had occurred, resulting in complete defoliation of the trees before the bagworms had become full-grown. As a result a certain amount of starvation had taken place, which, while not being sufficiently severe to SOUTH AFRICAN BAGWORMS. 683 cause the death of the bagworms, checked their development and produced mature insects which were below the average in size. Here the maximum was 2197 and the minimum 1177, with an average of 1661. This agrees closely with Fuller’s first figure (1600) given in 1899, his specimens probably having been obtained from a locality or a tree where just such conditions prevailed. The females of the Clan Syndicate plantations were gathered from a block which had suffered severely from bag- worm attack for two or three successive seasons, each time resulting in a complete defolation, and where the conditions otherwise were very unfavourable for the normal develop- ment of the bagworm. Here we see that with an average of 1209 the number of eggs drops as low as 438, the maximum being 1605. ‘The maximum does not even reach the average of the Wartburg lot. The Clan Syndicate bags can thus no. longer be considered as representing the normal reproductive power of the bagworm, but are interesting as showing how this function depends to a great extent on the size and vigour of the female, determined by the conditions prevailing during the teeding period of the larva. Apparently the bags on the contents of which Fuller based his later conclusions in 1913 were obtained from such an unfavourable area, for the years 1911 and 1912 were successive seasons of unprecedented infestation. Taking into consideration the most favourable, the normal and the most unfavourable localities, we come to a mean average of 1756 eges per female, which we may consider a fair indication of the multiplication of the bagworm over the entire area of its distribution. The actual number of eggs deposited by any one female may greatly excel or fall far below this average, depending on local conditions. (10) Lire-Cyctie. The incubation period of the eggs, judging from the time of copulation to the hatching of the eggs, is about two months, the average being sixty-two days. 684, CG. B. HARDENBERG. As the males are emerging during a period of about six weeks, it follows that the hatching of the young caterpillars also stretches over a similar period, namely from the middle of August until the beginning of October. The life-cycle of the species may be briefly represented by the accompanying diagram (text-fig. 6). The Wattle Bagworm has been discussed at some length, because it is by far the most important from an economic TEXT-FIG. 6. Jan _| Feb, Sa\doldoldolgoléolgoléo SHt+l¢ PO/QO!90/9 0/20/90 ie Ke * ¥ —Larvoe O 2ygae + Lmagines ° Laes Acanthopsyche junodi (Heylaerts). Note.—The above diagram shows the development of the entire generation, not of one particular brood. LOSE TFL DP point of view, and as such has been studied in greater detail than the other Psychidz, and also because its life- history is typical of the family. Of the remaining members of the family, as also of the bagworms of the other families, our data are as yet more incomplete, and therefore these species must be treated more briefly. SOUTH AFRICAN BAGWORMS., 685 EXPLANATION OF PLATES XLIV-XLVI, Illustrating Mr. C. B. Hardenbere’s paper, “South African Bagworms,” Part I. PLATE XLIV. Fie. 1.—Nat. size. Wattle Bagworm, Acanthopsyche junodi (Heyluerts). Fie. 2.—Nat. size. Thatched Bagworm, Acanthopsyche tristis Janse. Fig. 5.—Nat. size. Gum Bagworm, Melasina stelitis Meyr. Fre. 4.—Nat. size. Rubbish Bagworm, Melasina picea Meyr. Fie. 5.—Nat. size. Lictor Bagworm, Clania moddermanni (Heyl.). Fie. 6.—x 3. Grass Bagworm (species undetermined). Fie. 7.—Nat. size. Meadow Bagworm, Trichocossus arvensis Janse. Fie. 8.—Nat. size. Crossed-stick Bagworm, Gymnelema vinetus (Wik.). Fie. 9.—Nat. size. Gymnelema stygialis Hipsn. Fie. 10.—Nat. size. Webbed-crossed-stick Bagworm, Acantho- psyche alba Janse. Fie. 11.—x 3. Clear Spiral Bagworm, Melasina tyrophanes Meyr. Fie. 12—x 5. Webbed Spiral Bagworm, Melasina cnapha- lodes Meyr. Fie 135.—x 35. Rough Spiral Bagworm, Melasina craterodes Meyr. Fie. 14.—Nat. size. Thorn Bagworm, Melasina halieutis Meyr. lie. 15—x 3. Sand Bagworm, Fumea obscurata Meyr. Fie. 16.—x 3. Flat Bagworm, Ceromitia xanthocoma Meyr. Fie. 17.—x : ae Flat Leaf Bagworm (undetermined Tineid). PLATE: XLV. Fie. 1— x 3. Seed Bagworm (undetermined species). Fie. 2—x 3. Gibson’s Bagworm, Manatha wthiops Hmps. Fig. 3.—Nat. size. Lichen Bagworm (undetermined). Fie. 4.—Nat. size. Gravel Bagworm (undetermined). Fig. 5.—Nat. size. Sectional Bagworm, Melasina ci:cophora Meyr. 686 CG. B. HARDENBERG. Fic. 6.—Nat. size. Trunk Bagworm (undetermined). Fie. 7.—x 3. Lesser Lictor Bagworm (undetermined). Fie. 8—x 3. Delicate Bagworm, Monda delicatissima Wl/k. Fie. 9.—x 3. Turret Bagworm, Monda rogenhoferi Heyl. ¢ * ~ Fig. 10—x 3. Single-stick Bagworm. A small hollow twig is. adapted as a dwelling and carried about by the larva. In the figure on the right the stick is opened, showing the pupa contained therein. Fie. 11.— x 3. Single-straw Bagworm. A hollow straw is adapted as a dwelling. In the figure on the right the straw is opened, showing the empty pupa-case. Acanthopsyche junodi (Heylaerts). Fie. 12.—Nat. size. Male bag on the right; female bag on the left, showing inner bag and female imago in chrysalis case. Fie. 13.—Nat.size. Lower figure, female moth nearly spent ; upper figure, chrysalis case filled with eggs; a single egg is shown in centre. Fic. 14.—Nat. size. Male moth with the body of normal length. Fria. 15.—Nat. size. Male moth with the body extended for copulation. Fie. 16.—Nat. size. Female bag with two male moths both trying to copulate. Fia. 17.—Nat. size. Same bag as fig. 16, opened to show the female inside, and the abdomens of the two males entering the neck of the bag. PLATE XLVI Acanthopsyche junodi (Heylaerts). Fie. 1.— Larva, first instar; plan of abdominal segments 10-7, showing the arrangement of the setiferous plates and the sete. The hooks on anal proleg have been drawn relatively too large. Fig. 2.— Larva, first instar; plan of abdominal segments 6-3, showing distinguishing names of the plates. Fic. 3.—Larva, first instar; plan of abdominal segments 2-1 and thoracic segments 3-1. Fic. 4.— Larva, first instar; plate of abdominal proleg of larva, showing the small number of hooks and only two setz (group “ 7”). Fic. 5.— Larva, second instar; plan of abdominal segments 10-7, showing the setiferous plates and the sete. Fic. 6.—Larva, second instar; plan of abdominal segments 6-3. Fic. 7.—Larva, second instar; plan of abdominal segments 2-1, and thoracic segments 3-1. Fic. 8.—Larva, fourth instar; plate of abdominal proleg of larva, showing fully developed number of hooks and the full number of three. setee (group “7”’). XLIV. PGs BaGs OF VARIOUS BaGworMs, Adlard & Son §& West Newman, Ltd, ©. B. Hardenberg ANN. c B. Hardenberg Narat Mus., Vou. IIL. BaGs OF VARIOUS BAGWORMS ACANTHOPSYCHE suNODI (HbyLAnrts) Idlard & Son XLV. Pu. XLVI. Abd DT Ann. Natat Mus., Vor. III. Abdi AbSWE Abo wr. Aba H ' Aba X --aorsa/ __ svbadorsa! _-Jorera/ 9 ws v 8 S205 Aoleral bs : x ve Nee ventral - - “eg plate ee EE a & Sy LS ventra/ 4 3 7h.T Aba Z Abd I 7). AbaIE Aba UE G {lard {Son & West Newman, Ltd., Tmpr SerirerRous PLATES or LARVA. ACANTHOPSYCHE JUNODI (HeyYLArRTS), C. B, Hardenbera, NEW sSOUI'H AFRICAN ARAGHNIDA. 687 Descriptions of New South African Arachnida. By John Hewitt, Director, Albany Museum, Grahamstown. With Plate XLVII and 4 Text-figures. CONTENTS. Orv. SOLIFUGA:: Hemiblossia idioceras sp. n. . : . . 687 Orv. ARANEZ: ' Fam. CTENIZIDA: Stasimopus longipalpis sp. a. : . 689 Stasimopus insculptus Poe. var. aedléhemc is var, n. 690 Stasimopus spinipes sp. n. ‘ : : . 692: Fam. DIPLURIDZ: Lepthercus rattrayi sp. n. : : : . 699 Fam. CLUBIONID : Amaurobioides africanus sp. n. . 704 Orver SOLIFUGAL. Hemiblossia idioceras sp. nov. Text-fig. 1, a—c. THE type of this species is a single adult male example: from Kimberley collected by Bro. J. H. Power (November,. 1915). The species closely resembles H. o’neili Purcel/ in coloration, but has a more complicated flagellum. ‘The fourth tarsus is unsegmented, and carries on each side six. ‘688 JOHN HEWITT. spines, whereas, according to Dr. Purcell’s figure (Ann. 5. Af. Mus. ii, p. 218), there are only five spines in o’neili. I may remark, however, that the spine-armature in specimens from Alicedale which appear to be referable to o’neili does not agree with the above-mentioned figure, but resembles that of idioceras, which has three spines, instead of two, on the distal portion of the tarsus. TExreriG, i, ay im Hemiblossia idioceras, sp. n. A. Flagellum of adult male seen from mesial side. x 44. B. Distal end of flagellum in face view, also portion of mesial surface. x 32. c. Flagellum from outer side. x 44. The description of o’neili contaims no allusion to the presence of some characteristic thickened bristles on the sternites, a secondary sexual character of adult males; such bristles certainly occur in our Alicedale specimens, and are present in the form now described. Flagellum.—The main portion of the flagellum is a thin flat lamina rotatable at the base: on its mesial side, upper and lower out-growths of the lamina approximate to form an elongated flattish pouch which occupies almost the whole length of the flagellum, and has a long and fairly wide shit- NEW SOUTH AFRICAN ARACHNIDA. 689 lke opening along the middle, the edges of which remain separate distally but fuse proximally (text-fig. 1). The distal margin of the lamina is flattened out, and presents a foli- aceous concave extension on its outer side inferiorly—the side adjacent to the chelicera—the sides of this extension gradually converging proximally into a sheht keel which runs for a short distance along the outer surface of the lamina ventrally. The dorsal margin of the flagellum is slightly curved, and has an acutely pointed termination: the ventral margin is almost straight. The distal margins are for the most part fringed with hairs. The stridulatory area of the chelicera presents seven longitudinal ridges. Sternites.—Modified hairs are present on the third post- genital sternite as numerous long stout rods, finely pointed at the apex; there are roughly about twenty on each side, arranged irregularly and rather sparsely. In the neighbour- hood of these hairs is a bright yellow exudate. Numerous hairs of this kind, but more slender, are also found on the first post-genital sternite. Measurements.—lotal length 9 mmn., length of flagellum about 1°2 mim. Orper ARANEA. Fam. C'THNIZIDA. Stasimopus longipalpis sp.nov. Pl. XLVII, fig. 5. This species 1s founded on three adult male examples from Kimberley, collected by Bro. J. H. Power. Colour.—Carapace dark brown or pale brown; legs also brown, the two anterior pairs and the palp being somewhat darker than the hinder two pairs, except on the distal portions of those appendages, which are pale. Carapace.—As long as the tibia of the palp, longer than tibia I. The three keels do not reach back so far as the fovea. The surface more or less lightly sculptured or roughened throughout, being nowhere quite smooth. Anterior 690 JOHN HEWI'T. margin of anterior row of eyes in a straight line; anterior medians slightly nearer to the laterals than to each other. Pedipalps.—Pressed forwards, the tip reaches a point about one-quarter to one-third of the distance along the tarsus of the first leg, or only to the end of metatarsus I; patella con- siderably longer than that of the first leg and a little longer than tibia I. Legs.—All the tarsi scopulate below, but no trace of a scopula on the metatarsi. Tarsus [I without spines or with only a single one on the posterior side, or with one on the posterior side and two on the anterior side; II without spines or with a single one on the anterior side or on both sides; III with 0-1 spine anteriorly and 0-2 posteriorly ; IV with 5-11 spines anteriorly but 0 posteriorly. Meta- tarsus I without strong spines along the mesial area inferiorly or with only 1. Tibia I very slightly shorter than metatarsus I, with about 5-10 spines on the lower surfaces besides those at the apex inferiorly ; tibia III with 1, 2 or 3 spines near the distal edge superiorly on the anterior side, and 3 or 4.on the posterior side superiorly. Patella I without spines below, III with a discontinuous strip of about 5 or 6 short, weak spines on the anterior side, but no distal patch of spinules above ; IV with an anterior patch of spinules extending about three-fifths of the length of the anterior side. Paired tarsal claws of fourth leg with a basal comb of 2 or 3 teeth, or even only | may be well developed. Measurements.— Largest specimen: total length 11:0 mm., length of carapace 5°65 mm., breadth of same 4°9 mm., length of tibia of palp 5°8 inm., of patella of palp 54 mm., of metatarsus I 51 mm., of tibia 1 4°8 mm.; of patella I 3 mm.; of metatarsus IV 5:4 mm. Stasimopus insculptus Poc. var. peddiensis var. ”. Pl. XLVII, figs. 6 and 7, and text-fig. 2. The types of this form are four adult male examples collected at Peddie by Mr. B. Marais during March, 1916. NEW SOUTH AFRICAN ARACHNIDA. 691 They are approximately equal in size, and share the same coloration, being entirely black above except the metatarsi and tarsi of the legs, which are reddish. They are slightly smaller than the type of insculptus, the carapace of which is 4°5 mm. long (not 6 mm, as stated by Pocock! in the original TEXT-FIG. 2. Stasimopus insculptus var. peddiensis var. nov. Palp of adult male. x 16. description) ; in the Peddie specimens the carapace is about 4mm.long. They differ from insculptus as follows: Metatarsus I without spines below or at the sides, apart from those at the apex, or only a single spinule on each side in the distal half, or even three spinules on the posterior side ; tibia I without spinesjinferiorly (insculptus has several strong spines inferiorly) ; tarsus I with O-5 spines on the anterior side, and 0-4 on the posterior side. ! Ann. Mag. Nat. Hist., 7, vii, p. 285. 692 JOHN HEWITT. The tibia of the palp seems short and comparatively stout, being swollen towards the base, and not quite twice as lone as the patella. The spine of the palpal organ is elongated, so much so that if it were straightened out in a line with the axis of the bulb the whole organ would be quite as long as the tibia of the palp. Total length 10 mm., length of tibia of palp 2°35 mm., of first leg 15°3 mm., of fourth leg 16°8 mm. The female of this form closely resembles that of S. insculptus Poc., and was referred to in my key (‘ Albany Museum Records,’ i, p. 84) as connecting together artifex and insculptus. It has neither spies nor spiniform sete at the apex of metatarsus ITI inferiorly, though stiffish setee may be present. Stasimopus spinipes sp. nov. Pl. XLVII, figs 3 and 4. The types of this species are two adult males collected at Kast London durmg the earher part of June, 1916, by Dr. Geo. Rattray, who presented them alone with numerous female examples to the Albany Museum. Colour.—Black throughout, except the two distal segments of each leg which are red, the tibia, tarsus, and bulb of the palp which are brown, the sternum and ventral surfaces of coxe ITI and IV which are castaneous, the lung-opercula and genital sternite which are yellow, and the spinners which are pale. Carapace.—Without hairs on the head region, except several in front of the antero-median eyes. ‘The usual three keels are present, the mesial one being continued backwards to the fovea. The surface is roughened throughout, except in the head-groove laterally and between the three keels dorsally. The eyes of the anterior row are comparatively large and close together, much more so than in schénlandi: the distance between an antero-lateral and antero-median is con- siderably less than the long diameter of the former, and distinctly less than the diameter of the latter. The carapace NEW SOUTH AFRICAN ARACHNIDA. 693° is a little longer than metatarsus IV or than the tibia and tarsus of the palp. Pedipalp.—Pressed forwards the palp extends ar far as the proximal fourth of metatarsus I. The patella is a trifle shorter than patella I. Legs.—All the tarsi are scopulate below, but no trace of a scopula on the metatarsi. Tarsus I with short strong spines on each side, 7-16 anteriorly and 8-21 posteriorly ; I also with rather numerous spines on each side; IV with about 10-15 rather weak spines posteriorly but more numerous and stronger ones anteriorly. Metatarsus I with numerous strong spines im several rows on each side, and one or two may, or may not, occur along the mesial surface inferiorly, the mesial area on the whole being devoid of spines. Tibia I distinctly shorter than metatarsus I, thickly spined below; tibia [1] without a eroup of spinules at the apex dorsally, or such spinules when present are small and weak. The patch of minute spines on the anterior surface of patella [V extends about two-thirds of the length of the segment. Measurements.—Total length 21:0 mm., length of cara- pace 7°5 mm., of patella I 3-5 mm., of tibia I 4°7 mm., of patella of palp 3-0 mm. The palpal measurements of these specimens may point to a relationship with the Grahamstown species S. schonlandi Poc., the ocular characters of which are, however, pro- nouncedly different from those of spinipes, whilst apparently the spinulation of the lower surface of metatarsus I will distinguish the two forms. Probably the species is closely related toS. insculptus Poc. and to 8. qumbu mehr, though the palps of the latter species are relatively shorter than in Kast London specimens. The characters of the female are sharply separated from those of schonlandi, and agree very closely with those of the section including insculptus Poc., kolbei Pure., ken- tanicus Pure., umtaticus Pure. and qumbu mihi (see : Records Albany Mus.” in, p. 84). There is no trace of ‘694 JOHN HEWIT?. ‘spines or spiniform sete at the apex of metatarsus II] inferiorly. At the apex of the tibia of the palp superiorly is a group of spinules. In the larger specimen and ‘several others, patella III has a number of short stout red ‘spinules at the apex superiorly, similar to those on the tibia, but in somewhat smaller examples, though apparently adult, ‘such spinules are quite absent or weak. The antero-lateral eyes are very large, their long diameter being considerably greater than the distance between antero- median and antero-lateral ; the posterior lateral eyes are much smaller than the anterior laterals. The patch of spinules at the apex of tibia I dorsally is rather longer than that at the base of the metatarsus. The carapace and appendages are usually dark castaneous, and the abdomen infuscated above ; sometimes the two distal segments of each leg are paler than the more proximal segments, and more or less tinged with red. Length of carapace 11:0 mm., breadth of carapace 9°3 mm., breadth of ocular area 3:0 mm. The large series of female specimens collected by Dr- Rattray at East London seems to be referable to a single species. They are considerably smaller than specimens from Debe Nek, and from the neighbourhood of King Wilhams- town, which presumably belong to insculptus Poc. Perhaps they are co-specitic with one or both of the two forms described by Dr. Purcell from the Kentani district, viz. S. kolbei and S.kentanicus. ‘The question cannot be finally solved until adult males of these species are available. The adult females from Hast London are smaller than those found at Peddie, but very similar in structural characters ; examples of approximately equal size can be distinguished by the characters of the posterior row of eyes, which are all larger in peddiensis than in spinipes, the disparity in size between the anterior and posterior laterals beimg greater in spinipes. In distinguishing the species of Stasimopus on _ the characters of adult males the most convenient character 1s NEW SOUTH AFRICAN ARAOCHNIDA. 695 that based on the varying length of the palp and of its segments. It should be noted, however, that large and small males of the same species present some appreciable difference in this respect; in small males the palpal sezments tend to become rather more elongated than in large ones of the same species. I give tentatively the following key : PRELIMINARY KEY TO THE SPECIES OF THE GENUS STASIMOPUS, BASED ON STRUCTURAL CHARACTERS OF THE ADULT MALES. I. 1. Pressed forwards the palp reaches a point about + or > of the distance along tibia I, the patella distinctly shorter than patella I: metatarsus I not scopulate, with strong spines on either side but not along the median area inferiorly: tarsus I with 2 or 3 spines on the anterior side and 7-8 on the posterior side, IV with a strip of spines along the anterior side except in the basal 1 or 2: tibia III with a few spinules at the apex above: the three keels of the carapace are depressed, the lateral ones flattened out and transversely ridged, the ridges becoming lost in the reticulation of the sides of the cephalic area. (King Williamstown.) S. inseulptus Poe. a. Similar to the typical insculptus but differing as follows: Metatarsus I without spines or spinules below or at the sides, or only with one or several spinules at the sides. (Peddie.) S. insculptus var. peddiensis var. nov. b. Similar to insculptus but differing as follows : Metatarsus I with strong spines below both over the median area and at the sides and bearing a few scopular hairs near the apex: tibia IIT with a dense patch of short spinules at the apex above: tarsus I with 10-12 spines on both anterior and posterior sides, IV with many spines on the anterior side and about 15 on the posterior side. (Qumbu.) S. qumbu Awtt. 2. Pressed forwards the palp not quite reaching the apex of tibia I, the patella short, half as long as tibia I: VOL, 3, PART 3. 46 696 JOHN HEWITT. metatarsus I scopulate below in the distal fifth or sixth, the under side thickly spined on each side, and also with 1-3 spines along the median line: tarsus I with 1-2 spines on the posterior side but none on the anterior side, III] with 4-6 anterior and 7-9 posterior spines, IV with 2-3 on the posterior side and 14-15 on the anterior side: carapace with 3 low keels, the lateral ones abbreviated behind, the median one continued to the fovea. (Bushman’s Drift, near Ashton.) S. brevipalpis 3. Pressed forwards the palp reaches a point about ? of the distance along tibia I, the patella subequal to patella I in length: metatarsi I and IT scopulate at the apex below, I without strong spines along the midline inferiorly: tarsus I without spines anteriorly but with 1 posteriorly, I1V with 10 spines anteriorly and 2 posteriorly, III with only 1 spme on each side: carapace with the 3 keels obsolete, only the median one being distinguishable posteriorly. (Worcester.) S. erythrognathus II. 1. Pressed forwards the palp reaches a point about + of the distance along metatarsus I, the patella very slightly shorter than patella I: metatarsus I not scopulate inferiorly, with strong spines below both over the median area and at the sides: tarsus I with 3-7 spines anteriorly but 7-8 posteriorly, IV with many spines anteriorly and 8-12 posteriorly: the 3 keels of the carapace are flattened, especially the Purcell. Purcell. median one. (Grahamstown.) S. schonlandi Poe. a. Similar toschénlandi but differing as follows : Metatarsus I without strong spines on the median area inferiorly, or only 1 or 2: distance between anterolateral and antero- median eyes less than the diameter of the latter (much greater in schonlandi) : median keel of carapace distinct from the ocular area to the fovea: distal segments of palp brown (red in schonlandi.) (Kast London.) S. spinipes b. Similar to schonlandi but differing thus: A weak but distinct scopula at apex of meta- tarsi I and II and no spines along the median area of this segment: tarsus I with 1 spine anteriorly and 2 posteriorly: the sp. nov. NEW SOUTH AFRICAN ARACHNIDA. 697 3 carapace keels fairly sharp and carrying long hairs which also oceur on the ocular area. (Bedford; the type females however came from Pearston and Jansenville.) S. astutus Poe. a. Similar to astutus but differing thus: Palp when stretched forwards barely reaching or only slightly surpassing the apex of tibia I: metatarsi I and II with only a few scopular hairs apically below: anterior row of eyes rather more dis- tinctly procurved than in astutus. (Alicedale; the type female from Per- severance, near Port Elizabeth ; the same species was found by me (25/6/1916) at Grahamstown on the drive above the Grey reservoir.) S. patersone Huwtt. III. 1. Pressed forwards the palp reaches a point about 4 of the distance along metatarsus I or a trifle further, the patella very slightly longer than patella I: metatarsus I with a few scopular hairs at the apex inferiorly and without spines along the median area inferiorly: patella III with a single spine anteriorly in the middle and in addition with some weak, short, subspiniform setie, also with several spines on the distal edge, but dorsally there is no distal patch of spinules: tarsus III with 6 or 7 spines on its anterior side situated in the distal half, I with 1-2 anteriorly and 2-3 posteriorly : claws of tarsus IV with a basal comb of about 7 teeth, the more distal ones being short but not sharply separated from the rest of the series : lateral keels of carapace depressed, only distinct along their inner edges and obsolete in the hinder half: anterior margin of anterior row of eyes in a straight line. (Vredefort Rd.) S. nigellus Poe. 2. Pressed forwards the palp reaches a point * of the distance along metatarsus I: the patella considerably longer than patella I but a little shorter than tibia I: metatarsus I not scopulate and without strong spines on the mesial area below: patella III with a strip of 6-9 weak spines anteriorly: tarsus III with 1 anterior spine and 1-3 posterior spines, IV with 7 anterior and | posterior spines. (Bloemfontein.) S. minor Hwtt. 698 JOHN HEWITT. IV. 1. Patella of palp slightly longer than patella I, appreciably shorter than tibia I, the tibia and tarsus together considerably shorter than the cara- pace: tarsus IV with numerous spines on both sides and III also with numerous spines in a continuous band on both sides: metatarsus I not scopulate, inferiorly with strong spines both over the median area and at the sides: the 3 keels of the carapace prominent throughout and reaching back almost as far as the fovea. (Vredefort Rd. ?) S. gigas Hewitt 2. Pressed forwards the palp reaches a point 2? of the distance along metatarsus I, patella longer than patella I but much shorter than tibia I, the tibia and tarsus together slightly shorter than the cara- pace: metatarsus I not scopulate, inferiorly without spines along the median area: tarsus III with about 14 spines on each side, IV numerously spined on both sides: the 3 keels of the carapace well developed anteriorly but absent posteriorly, none of them approaching the fovea. (Steynsburg.) S. steynsburgensis Hewitt V. 1. Pressed forwards the palp reaches almost to the tip of the first leg, the patella being about 13 times as long as patella I and only very slightly shorter than tibia I: tarsus I with 4-5 spines anteriorly and 6-7 posteriorly: metatarsus I not scopulate and without spines along the median line inferiorly or only an odd one proximally. (Pretoria.) S. robertsi Hewitt. 2. Pressed forwards the palp reaches to the end of metatarsus I or slightly beyond, the patella being a little longer than tibia I and 1+ times as long as patella I: tarsus I with 0-2 spines anteriorly and 0-1 spine posteriorly: metatarsus I not scopulate and without spines along the median area inferiorly or with only one. (Kimberley. ) S. longipalpis Hewitt. VI. 1. Differing from all the above-mentioned species in that tarsus IV is not scopulate: pressed forwards the palp reaches to the tip of the first leg, the patella being distinctly longer than tibia I, the tibia being longer than metatarsus IV, which again is subequal to the carapace in length. (Graaff Reinet.) S. palpiger Poc. NEW SOUTH AFRICAN ARACHNIDA. 699 a. Very near to palpiger but apparently differ- ing as follows: Pressed forwards the palp extends beyond the apex of metatarsus I, the patella at least equal to tibia I in length. the tibia as long as metatarsus IV and very distinctly longer than the carapace. (Hanover.) S. schreineri Purcell. Fam. DIPLURIDA. Lepthercus rattrayi sp. nov. Pl. XLVIL, figs. 1 and 2, and text-fig. 3, A—E, ‘he type of this species is an adult male example collected at Kast London by Dr. Geo. Rattray and Master G. Rattray during May, 1916. The females present no conspicuous structural characters which would serve to separate them generically from Hermacha Sim., except perhaps in the weaker scopule of the tarsi, but the adult male differs considerably from that of any known species of Hermacha in the modification of the first leg. The male agrees more closely with that of L. dregei Purcell (Trans. S. Af. Phil. Soc., x1, p. 379, 1902) from the Zuurberg, but differs there- from in the armature of the maxille, in the dentition of the chelicere, and in the modification of the first metatarsus ; the female of dregei is unknown. Colour.—The general colour is dull brown, the surfaces of the carapace, abdomen, and appendages being thickly covered with long silky appressed pale brown hairs ; on the abdomen and appendages these are accompanied by stiffer black hairs more sparsely disposed. The silky hairs on the carapace, cheliceree, and coxe of the legs and palps are paler than those on the femora of the legs, which are almost olivaceous, but the silky hairs of the patelle and more distal segments of the legs are almost as pale as those on the carapace. The posterior spinners are conspicuously banded, the tips of the terminal segments, and the articular membranes between the various segments, being almost white. 700 JOHN HEWITY. Carapace.—The lateral margins converge in front so that the anterior margin is narrow. The radial depressions are obscured by the clothing of hair. The fovea is transverse. Stiffish black hairs fringe the margins in the posterior half, a few occur just in front of the ocular area, and several weaker ones near the fovea. The carapace is as long as the metatarsus and half the tarsus of the fourth leg. Ocular area.—Anterior row of eyes decidedly procurved when seen from above, the antero-laterals being one-half longer than the medians and distant from the anterior margin of the carapace about the length of a median. Posterior row of eyes slightly recurved, the laterals being distinctly smaller and shorter than the anterior laterals, whilst the medians, which are oval, are a trifle longer than the anterior medians. Labium.—Muticous, but maxilla armed with numerous (about fifty) minute cusps at the base in a small patch. Chelicerze.—With nine teeth in the inner row; the outer row is obsolete. Pedipalps.—Superiorly the femur carries a number of curved bristles but no distinct spines; there 1s a weak spine at the apex of the patella superiorly on the inner side ; there is a setiform spine on the inner surface of the tibia, and a number of long stout sete superiorly and inferiorly. The tarsus is without spines or bristles. The spine of the bulbal organ is fairly long and stout, being longer than the bulb; it is strongly twisted. Posteror spinners.—Long and slender, decidedly longer than the sternum and labium together, being approximately two-thirds as long as the abdomen. In side view the three segments are about equal in length, the slender apical segment being only a little longer than the penultimate segment. Legs.—Tarsus I is long and slender, longer than II. All the tarsi are scopulate, the scopula of IV, and to a less extent of III, being divided by a mesial strip of sete ; in III the sete and scopular hairs are not so easily distinguished, the sete being finer than in IV. he scopular hairs at the NEW SOUTH AFRICAN ARACHNIDA. 701 distal end of each tarsus are elongated into a terminal tuft ; a small median tarsal claw is present, this being fairly con- spicuous on tarsus IV. Metatarsus I is shghtly bowed, TEXT-FIG. 3. Lepthercus rattrayi, sp. n. A. Distal segments of the right leg of the first pair, seen from the outer side: adult male. x 13. B. Palp of adult male viewed from the inner side. x 19. c. Palp of adult male viewed from the outer side. x19. bp. Eyes of adult female in dorsal view. x 37. E. Lower surface of abdomen of adult female. x 23. and on its ventral surface presents a conspicuous rounded protuberance ata point nearly two-thirds of the length of the 702 JOHN HEWITT. segment from its base ; this is covered with numerous black spinules. There is a pair of spines at the apex inferiorly, one on the inner surface about halfway along the segment, and one on the lower surface externally just proximal to the protuberance. A distinct scopula is only present in the distal half of the segment, though scattered scopular hairs occur in the basal half also. Tibia [ stouter than the patella, and very much stouter than the metatarsus ; at the apex inferiorly is a strong projecting spur, bearing a slightly curved stout spine at its end; on the lower and lateral surfaces there are also about 6 long but rather weak spines. In the pale articular membrane between tne tibia and metatarsus inferiorly there is a well-defined indurated brown area at the base of the latter ; this also occurs as an isolated element on the second leg, and is more or less distinctly represented on IIT and IV. Patella I with a weak spine near the apex on the inner side. Metatarsus II with 2 spines at the apex inferiorly, also 5 on the lower surface and 2 or 3 on the inner surface ; the scopula is fairly well developed in the distal half of the segment. Tibia IT has 3'spines at the apex inferiorly and 3 on the lower surface, also 2 on the inner surface. The tibize and metatarsi of III and IV bear long, strong spines, the metatarsi being devoid of scopular hairs. The femora superiorly bear long spiniform bristles. Sternum.—The margins are fringed with bristles; the general surface carries brown silky hairs and black stiff hairs. Abdomen.—On the dorsal surface anteriorly there are some long curved bristles. Measurements.—Total length (including spinners and cheliceree) 14°5 mm., length of carapace 4°75 mm., breadth of carapace 3°5 mm., length of metatarsus of fourth leg 3°65 min., length of posterior spinners 3°65 mm. The female greatly resembles the male in general charac- ters. The chelicere have 10 teeth in the inner row, and the outer row is represented by a series of minute denticles, the end of the outer series being opposite to the eighth tooth NEW SOUTH AFRICAN ARACHNIDA. 703: from the distal end of the mainrow. All the tarsi are scopu- late, the scopula of II being divided by a thin median strip of sete, and III and IV by a broader strip. Metatarsus I with 2 spines at the apex inferiorly, and 3 on the lower surface,. II with 2 at the apex and 4 on the lower surface. The anterior spinners are a trifle more than two diameters apart. The terminal segment of the superior spinners is subequal to the basal segment and barely one and a quarter times the length of the middle segment. Labium without apical teeth. Surfaces of body and appendages clothed with pale brown hairs, the abdomen showing no pattern or markings superiorly. The carapace is broader anteriorly than in the male. Measurements.—Total length 17°75 mm., length of carapace 5'4 mm., breadth of carapace 3°8 mm., length of posterior spinners 4 mm., length of tibia of first leg 2-5 mm. This genus probably approaches Brachythele Auss. and Hapalothele Lenz. It does not seem to agree with any of the Diplurine genera recorded from Australia by H. R. Hogg or by W. J. Rainbow, nor with the South American genera dealt with by the late F. O. Pickard-Cambridge. Most of the genera belonging to this section, the Brachy- thelez of Mr. H. R. Hogg, are very imperfectly known, and no author has had sufficient material at his disposal for corre- lating the genera of different zoological regions. The genus Hermacha is also represented at Hast London. The females are larger than those of the species above described. Young specimens of Hermacha can be dis- tinguished from females of L. rattrayi by the presence of one or two cusps on the labium, the weaker spinulation of the first and second metatarsi, and the greater elongation of the terminal seement of the posterior spinners relative to the middle segment. Another related species, Hermachola erahami mihi, which occurs at Grahamstown, has the terminal segment of the posterior spinners considerably: 704 JOHN HEWITY. longer than the middle segment, and is at once distinguished from L. rattrayi by the relative paucity of fine hairs on its surfaces, the dark dorsal pattern of the abdomen being ‘conspicuously displayed through the weak covering of hair ; the characters of the adult male are very different in the two species. Fam. CLUBIONIDA. Amaurobioides africanus sp. nov. Pl. XLVII, fig. 8, and text-fig. 4, a—p. The species described under this name is based on a series of adult male and female specimens collected recently at Kast London by Dr. Geo. Rattray and his son, Master G. Rattray. Specimens were first discovered on the seaward face of the rocks near Bats Cave; their retreats, made of tough silk, lodged in the pits and crevices of the rock surface, were situated near to or just below the average high-water mark, where they were liable to complete submergence at spring tides. Other examples were found on rocks between tide marks aloug the banks of the Buffalo River, and a few at Cove Rock. After visiting their habitats on various occasions, Dr. Rattray is satisfied that many of these retreats are not necessarily submerged at each invasion of the tide, but are often merely drenched by the spray of the waves. On the other hand, he found numerous small retreats occupied by immature specimens amongst the wet seaweed exposed at low water, and these retreats must certainly be submerged every high tide. The species closely resembles the other two recorded members of the genus, viz. A. maritima 0. P. Cambr. (Proce. Zool. Soc., 1883, p. 356, pl. xxxvi, fig. 3), the types of which were sent to Mr. Pickard-Cambridge labelled “ marine spiders,” having been found on rocks in the sea at Allday Bay, Otago, N.Z., and A. piscator Hogg (‘Sub- antarctic Islands of New Zealand,’ Wellington, 1909, article NEW SOUTH AFRICAN ARACHNIDA. 705 1x, p. 162, pl. vu, fig. 4) from the Campbell Islands, where this species also was found on rocks. Some months ago | submitted immature examples from East London to Mr. H. R. Hogg, who very kindly compared them with the types of piscator, afterwards reporting that he was unable to find any essential difference in structure between the South African and New Zealand specimens. Apparently, therefore, the only satisfactory basis of distinction between africanus and piscator is that which may possibly be offered by the external sexual characters. The males of the New Zealand species have not been described in either case, and the epigynal character of the adult female of A. piscator, as figured by Mr. Hogg, appears to be quite distinct from that of africanus, the characters of which are constant in the series of six or seven adult specimens now at my disposal. Colour.—Carapace brown, becoming black-brown on the ocular area, Chelicere black-brown throughout, except in the terminal half of the fang, which is reddish. Lip and maxille reddish-brown. Legs and sternum pale brown. Upper surface of abdomen yellow, with a median dark tree pattern constituted by a series of six dark transverse chevrons, the three anterior ones being connected together by a median stripe which extends considerably in advance of the most anterior chevron. ‘The anterior chevron is much broader than any of the succeeding ones, and the second is broader than the third, though not quite so long. Towards the posterior end of the abdomen, the colour is dark throughout except for faint indications of yellow stripes. The lateral surfaces of the abdomen are dark with three yellow, backwardly directed, oblique stripes. ‘These stripes superiorly pass into the yellow of the dorsal surface, but inferiorly the two posterior ones taper finely and end blindly near the posterior end of the abdomen, whilst the anterior one is quite short and broad, not extending far towards the ventral surface. The ventral surface of the abdomen is pale, the dark coloration of the sides gradually merging therewith. This pattern on the dorsal and lateral surfaces of the 706 JOHN HEWITT. abdomen is very similar to, but not quite identical with, that of the New Zealand species. In those species, according to Mr. Hogg’s recent notes on piscator and judging from the original figure of A. maritima, the dorsal surface presents a yellow pattern on a dark background rather than dark on yellow as in africanus. Carapace.—The ocular characters are not quite as described in piscator. In the hind row, the distance between each pair is distinctly greater than the diameter of the posterior lateral. The area of the anterior laterals is quite twice as great as that of the posterior medians, and is a little larger than that of the posterior laterals. The distance between anterior and posterior laterals is about two-thirds the diameter of the latter. (In both the New Zealand species the lateral eyes seem to be rather more widely separated from each other.) At the sides and in front, the ocular area projects above the general surface of the carapace; posteriorly, how- ever, along its whole width, the ocular area passes quite insensibly into the general surface of the carapace. Apart from the general clothing of short grey hairs, there are a few scattered long black hairs on the carapace ; these latter are most numerous in the ocular region. Viewed under a high power of the microscope, the grey hairs are seen to be finely ciliate. West Newman, Ltd, INDEX OF SOUTH AFRICAN MOLLUSCA. (VOL. IIL.) Synonyms are indicated by italics. New genera, species and varieties described in this volume are distinguished by (g.n.), (sp. n.) and (var.n.) being placed after the names of the authors. PAGE Achatina : 250 Achatina panthera ( Fer.) 138 Achatinidse 249, 250, 251, 252, 254 Actzon 5 138 Agnatha 228, 224, 228, 231, 254, 255 Agnathomorpha 113, 255 Agriolimax agrestis (Lin.) 5 IBS Amphineura 138, 182 Ampullaria 138 | Apera 107, 108, 110, 111, 112, 113, 115, 116, 119, 120, 122, 124, 125, 126, 127, 130, 131, 136, 137, 138, 140, 143, 144, 145, 146, 150, 151, 152, 153, 156, 157, 158, 159, 160, 161, 163, 164, 165, 168, 169, 170, 171, (725 Vise 74s 175s 180, 181, 182, 186, 188, 190, 191, 192, 193, 205, 209, 220, 224, 226, 228, 229, 230, 231, 232, 234, 235, 237, 240, 241, 243, 244, 245, 246, 247, 248, 254 Apera, affinity of Testacelaand 222 characters of 193 digestive system of 155 excretory system of 186 VOL. oO: PAGE | Apera, history of genus 109 nervous system of = 136, 148 phylogeny of 217, 231 vascular system of 181 | Apera burnupi Smith 108, 110, 111, 112, 118, 114, 115, 118, 119, 121, 122, 127, 135, 140, 141, 142, 146, 158, 161, 162, 163, 164, 165, 2 eeli Ose lui lad Oslo ple le, 182, 184, 185, 188, 190, 191, 193, 210, 213, 217, 218, 219, 259, 260, 261 dimidia Watson (sp. n.) 108, 114, 115, 117, 118,119, 120, 121, 124, 125, 126, 127, 128, 129, 130, 152, 133, 134, 135, 137, 138, 141, 147, 148, 149, 151, 158, 156, 157, 160, 161, 162, 164, 165, 172, 177, 179, 180, 182, 184, 185, 186, 188, 189, 190, 192, 193, 204, 209, 213, 218, 219, 259, 260, 261, ATC pL, 132, 152, 262, 263, 264, 265, 266 gibbonsi (Binney) 107, 110, N46, Wy Pist2On Tat 4.8 714 INDEX OF PAGE Apera gibbonsi (cont.)— 123, 126, 127, 128, 132, 133, 134, 135, 187, 140, 141, 142, 146, 156, 157, 158, 159, 160, 162, 164, 166, 169, 170, 172, 173, 174, 175, 176, 177, 178, 179, 181, 182, 184, 186, 189, 190, 192, 193, 194, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 207, 209, 218, 219, 230, 258, 259, 263, 264, 265 gibbonsi (Binney) gracilis Watson (subsp. n.) 107, 159, 184, 199, 259, 261, 264. gibbonsi (Binney) lapata Watson (subsp. n.) 107, 158, 159, 160, 170, 184, 200, 201, 261, 263, 264 lupata var. duplex Watson (var. n.) 159, 201 eibbonsi rubella Watson (subsp. n.) 121, 122, 123, 148, 159, 160, 184, 191, 192, 197, 200, 201, 260, 261, 262, 268, 264, 265, 266 Apera natalensis Clige. 111, 112, 210, 212, 217 gibbonsi Apera parva Watson (sp. n.) 114, 115, 116, 117, 118, 132, 133, 141, 157, 158, 160, 161, 162, 164, 172, 173, 174, 175, 176, 177, 178, 179, 181, 182, 184, 185, 189, 191, 192, 193, 202, 218, 219, 259, 260, 261, 263, 264, 265 purcelli Cllge. 108, 112, 115, Hak, ails, abate), WPAO)S aeal, ALPAr( Psi, Pe) BPE TBs, NB, ey 141, 146, 158, 160, 161, 162,” 164, 168, 176, 177, 180, 182, | 184, 188, 191, 193, 207, 217, 218, 219, 259, 261 sexangula Watson (sp.n.) 108, 114, 115, 116, 117, 118, 119, 108, | SOUTH AFRICAN MOLLUSCA. PAGE Apera sexangula (cont.)— 121, 127, 129, 132, 133, 135, 140, 141, 142, 146, 149, 152, 155, 156, 158, 161, 162, 163, 164, 165, 172, 176, 177, 178, 179, 182, 185, 186, 188, 189, 190, 191, 198, 218, 217, 218, 219, 259, 260, 261, 262, 263, 264, 265, 266 Aperide . 107, 113, 246, 247, 254. Aplysiella 150 Artemon : 249 Aspidobranchia . : 138 Atopus 111, 112, 113, 114, 132, 180, 228, 230, 239 Aulacopoda 245, 254, 255 Auriculidee 150 Cephalopoda 138 | Ceratoconchites schultzei (Simroth) 220 Cerebral ganglia, nerves of 139 Cerithiopsis (Seila) natalensis E. A. Smith (sp. n.) 2,6 Chlamydephorus : ee lelel: eibbonsi Binn. 109, 193 Chlamydophorus 109, 110 gibbonsi Binn. 194 Circinariidee 255 Coeliaxinee : : . 250 Conus : : . 226 Cryptodon murchlandi Sow. . 5 Curvella coloraphe Preston . 250 Cystopelta Daudebardia 111, 124, 125, 181, 134, 136, 140, 148, 149, 165, 181, 190, 234, 288, 239, 240, 243, 244, 246, 247, 248 saulcyi (Bgt.) 174 Daudebardia novoseelandica Hutton Daudebardiine . INDEX OF SOUTH AFRICAN MOLLUSCA. PAGE Delos 190 Diaphora : 225 eutrachela Mild ff. 226 telescopium Mild ff. 225 Diplomphalus . ; . 225 Ditremata 248, 254, 255 Ennea, descriptions of South African : 5 ty) Ennea . 5 34, 249, 253, 255 (Edentulina) 149 arnoldi Stur. Ennea berthxe Melv. § Pons . 29, 68 30, 42, 43, 44,45 | callista M.& P. . 30, 61, 62 Ennea columnella M. § P. f. 29, 30, 57, 58, 60, 638, 82 columnella M. & P. var. vitreola (M. & P.) connollyi Melv. 4 Pons. 30, 69, 82 50, 51, 81 typica . 74%), 59 consobrina Ancey crassidens Pfr. darglensis Melv. §° Pons. f. typica 29, 48, 79, 81 darglensis M. § P. var. illovoensis 29, 31, 49, 81 densecostulata Mildf. elliptica Melv. & Pons. f. typica elliptica M. & P. var. celata Burnup (var. n.) 29, 31,40, 80 elliptica M. & P. manca Burnup (var. n.) 29, 31, 39, 80 farquhari M. & P. f. typica 29, 30, 41, 42, 43, 44, 46, 61, 81 farquhari M. 4° P. var. avena Burnup (var. n.) 29, 31. 46, 81 Ennea farquhari M. g P. var. berthe (M. ¢ P.) 29, 44: var. 226 29, 38, 80 53 | Ennea | | PAGE Ennea hypsoma Melv. & Pons. 30, 31, 43 Ennea inhluzaniensis Burnup (sp. n.) 30, 31, 71, 82 isipingoénsis Stur. f. typica 29, 30, 31, 33, 36, 80 Ennea isipingoénsis Stur. var. 30, 31, Ennea isipingoénsis Stu. var. cylindrica Stur. 33 discrepans Stir. 29, 30, 35, 80 Ennea isipingoénsis Stur. var. simillima Stur. 30, 35 Ennea isipingoénsis Stwr. var. sturanyl Burnwp (var. n.) 29, 36 Ennea juxtidens Melv. 4 Pons. 30, 72, 82 Ennea labyrinthea Melv. §& Pons. 30, 41, 42, 43 Ennea marie Melv. 4 Pons. 29, 30, 61, 62, 63, maritzburgensis M. ¢ P. f.typica 29, 31, 64, 66, melvilli Burnup (sp. n.) 29, 31, 55, 58, 80 microthauma Melv. & Pons. 30, 41, Ennea montana Melv. 4° Pons. 30, 74, 75, 79, mooiensis Burnup (sp. 1.) 29, 31, 62, 82 munita Melv. & Pons. 29, 52, 43 82 obovata Pfr. ; Ennea oppugnans Melv. § Pons. 30, 41, 43 parallela Melv. & Pons. periploca Melv. 4° Pons. 30, 41, 42, Ennea ponsonbyi Burnup (sp.n.) 30, 31, 49, 78, 82 premnodes Stu. 30, 73, 76, 716 PAGE Ennea sylvia Melv. & Pons. 29, 52, 54, 57, 64, 80 tharfieldensis Melv. | 538 Pons. thelodonta Melv. & Pane 29, 50, 51, 81 Ennea vitreola Melv. § Pons. 30, 59 Euglandina 148, 172, 185, 227, 241, 242, 243, 244, 245, 249 corneola (Binn.) 162 liebmanni (P/r.) 243, 252 sowerbyana (Pfr.). 171 truncata (Gimel.) 162 | venezuelensis (Preston) 165, 171, 186, 246 138, 147, 153 | Euthyneura Fissurella 182 Glandina : 109 | Glyphis levicostata /. a Smith (sp. n.) : 2,6 | tenuistriata eSouerea) 3 | Gnathophora 223, 224 | Guestieria 229, 238 Gymnosomata 150 | Haliotis . 182 | Helicinidee 147 | Helix aspersa Miill. 138 | Hyalinia. 238, 239, 240, 243, 244 draparnaudi (Beck) 243 Imperturbatia 249 Janella 110, 231 Janellide 112, 114, 231, 232 Leptinaria lamellata (Pot. 4 Mich.) 251 | Limacide 163 Limacinee 246 INDEX OF SOUTH AFRICAN MOLLUSCA. PAGE | Loripes burnupi E. A. Smith (sp. n.) 5,6 Mangilia shepstonensis E. A. Smith (sp. n.) : 5 ls | Megaspiride 252, 253, 254 Natalina 132, 136, 144, 148, 165, 168, 169; 170, 171; 172; 190, 229, 236, 237, 247 caffra (Fér.) 122, 168, 227 morrumbalensis (M.f P.) 237 permembranacea Preston 2387 quekettiana (M. & P.) 168 trimeni (Melv. § Pons.) 144 Neritidi 147 Obeliscella ; ‘ 226, 249 Oleacina : : 241 | Oleacinide 147, 149, 157, WL, 172, 178, 224, 226, 232, 241, 244, 248, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255 229, 230, Parmacelline 234, 248, 254, 255 Paryphanta 124, 125, 182, 186, 187, 148, 150, 165, 190, 225, 236, 240, 241 atramentaria (Shuttl.) 128, 228 busbyi (Gray) 234. compacta Cow & Hedley. . 227 hochstetteri (P7r.) 134, 139, 144, 146, 154, 163, 165 Pectinibranchia 138, 1538, 180 Phrixolestes 224 Plectrophorus orbignyi Fér. 258 Pleurobranchide 138 Pleurotomaria 150 INDEX OF SOUTH PAGE Plutonia 131, 190, 230, 233, 248 Plutoniin» 254, 255 Poiretia 190, 241, 243 Polyplacophora . 126 Pseudosubulina 250, 251 lirifera (Morelet) 168, 251 Rathouisiide 232, 248, 254 | Rhytida 132, 136, 160, 190, 243, 244 capillacea (Fér.) 148, 192 236 franklandiensis (Forbes) 162 | ineequalis (Pfr.) 148, 149, 236, 237 226 113, 114, 186, 188, 157, 168, 172, 187, 193, 224, lampra (Pfr.) . Rhytidide 234, 236, 238, 243, 246, | 247, 248, 249, 251, 252, 254 Salasiella 237, 241, 248 novoseelandica ( Pfr.) 175, 24 Seila attenuata Hedley . ; 2 Selenochlamys 116,148, 178, 190, 226 Siphonaria 184. Spiraxis . : : 250, 251 Stenogyrine . 250, 251, 252 Strebelia 190, 234, 241, 242, 243, 24.4, 245, 248 122, 189, 148, 157, 190, 226, 237, 238, 246, 24:7, 249, 258, 254, 255 Streptaxide Streptaxis ‘ 132, 225 | funki Pfr. 93 Streptostele 226, 249, 253 Streptostyla 148, 190, 243, 245 gracilis Pilsbry 226 shuttlworthi (Pfr.) 171 Stylommatophora 115, 187, 222, 223, 224, 228, 232, 248 AFRICAN MOLLUSCA. 148, 243 | Schizoglossa 111, 132, 136, 175, 234, | PAGE Succineidze : ‘ . 24) Tayloria . 237 Testacella 109, 110, 111, 113, 114, 116, 118, 123, 124, 125, 126, 132, 134, 136, 137, 148, 149, 160, 164, 165, 171, 172, 181, 185, 187, . 190, 192, 220, 221, 222, 224, 227, 231, 234, 235, 236, 238, 2389, 240, 241, 242, 243, 244, 245, 246, 247, 248, 254 Testacella, phylogeny of 38 Testacella altx-ripe Grat. . 208 asinina de Serres 256, 25 asininum de Serres . 25 aquitanica Grat. . 258 aurigaster Layard . 221, 222, 258 brownians Grat. . 258 bruntoniana de Serres 257 burdigalensis Gassies & Raulin 257 canariensis Grat. 258 deshayesii Mich. 257 Testacella gestroi (Isse?) 174 Testacella haliotidea Drap. 162, 230, 243, 256 haliotidea var. scutu- lum Moqg.-Tand. . 256 haliotoides Lan. . 256 lartetii Dupuy . 257 Testacella maugei Fér. 108, 170, 192, 221, 222, 243, 247, 256, 266 maugei Fér., the oc- currence at Cape Town of 220 Testacella monspessulana Grat. 258 nouleti Bourg. 258 oceanica Grat. 257 occitanixe Grat. . 208 Testacella scutulum Sow. 162, 241 Testacella vagans Hutton 258 718 INDEX OF SOUTH AFRICAN PAGE Testacellide 107, 109, 110, 111, 112, 114, 157, 170, 198, 254, | | | | 255 Tivela_ . : : ; 1 Tivela alucinans Sow. . : 4 Tivela compressa Sow. . yeh Tiwela dolabella Sow. . : 4. | Tivela dunkeri (Réme7) 4, 5,6 | natalensis Dkr. . : 4 | polita Sow. é : 4 | rejecta EH. A. Smith (sp. n.) 4,6 transversa Sow. . : 4 Trigonochlaminze 131, 187, 280, 234, 246, 247, 248, 254. | Trigonochlamys Varicella MOLLUSCA. dissimilis Pilsbry nemorensis Ads. phillipsi Ads. . Varicella similis Ads. Varicella venusta Ads. . Venus dunkeri Romer Veronicella Veronicellide Vitrinine Vivipara Zonitide Zonitine 163, PAGE 190 239, 250, 251 . 251 250, 251 251 251 251 Des teh 121, 126 232, 254. 248, 254, 255 138 239, 240, 255 248, 254, 255 GENERAL INDEX (EXCLUDING MOLLUSCA). (VOL. IL) Synonyms are indicated by italves. New genera, species and varieties des- eribed in this volume are distinguished by (g. n.), (sp. n.) and (var. n.) being placed after the names of the authors. PAGE 548 529, 540, 639 379, 662 Aberia tristis Sond. Acacia : caffra Willd. horrida Willd. melanoxylon R. Br. mollissima Wild. 662 269, 620, 379, 548, 662 | 6389 robusta Beh. 379 | Acalypha punctata Meisn. var. radula 537 Acanthaceous herbs d44 Acanthodon 309 abrahami (/Tewttt) 289, 305 erandis Hewitt (sp. n.) . 289, 308 hamiltoni Poe. 289, 307 ochreolum Poc. 289, 805, 306 pectinipalpis Pure. 308 pretoriz Poc. 310 Acanthopsyche 594, 595, 640 alba Janse (sp.n.) 614, 625, (Metisa) alba Janse (sp.n.) 589, 598 / > PAGE Acanthopsyche junodi (Heyl.) 614, 619, 620, 623, 6384, 63%, 643, 648, 653, 662, 671, 674, 676, 677, 681, 684, 685, 686 (Dasaratha) junodi (Heyl.) 589, 595 tristis Janse (sp. n.) 614, 623, 634, 685 (Heeticoides) tristis Janse (sp. n.) 589, 597 Acerina cernua Lin. 20 Achyranthes aspera Linn. 538 Acidanthera platypetala Baker 540 Adelide . 590, 615, 617, 6385 Adiantwn ethiopicum Lin. 544. capillus-veneris Lin. 544 Agapanthus umbellatus L’Herit. 544 Agelenide 289, 290 Agelenine : : . 290 Agrimonia eupatoria Linn. var. capensis Harv. Oo ILIBRARY |ao 720 GENERAL INDEX. PAGE Agrostis eriantha Hack. . 533 lachnantha Nees. 538, 549 Air, movements of the . 5 byAS) Ajuga ophrydis Burch. . . 549 Albuca affinis W. 4 E. . . 537 pachychlamys Baker . 537 Alchemilla capensis Thb. . 550 woodii O. Kuntze . 539 Alepidea amatymbica E. f° Z. . 536 ciliaris La Roche . 536 concinna Diimmer . 536 jacobszie Diimmer . 536 longifolia E. Mey. var. angusta. . 9036 natalensis Wood 4& Hvans : . 586 setifera N. E. B. . 536 thodei Dimmer . 552 Aloe kraussii Baker. . 5388 minima Baker ; . 538 natalensis W”. 4° E. . 542 Amaurobioides africanus Hewitt (sp. n-) 687, 704, 708, 711 maritima O. P. Cambr. 704, 70 piscator Hogg. on) 704, 705, 707 Amphisbatis incongruella St. 6385 Anaphiloscia 577 Andrea ‘ - 599 Androcymbiumn ealonse Baker 552 Andropogon , : 531, 535 appendiculatus Nees. 533 auctus Stapf. . 533 ceresiieformis Nees. 5383 cymbarius Linn, . 533 filifolius Steud. . 588 hirtus Linn. , bse nardus Linn. var. marginatus 5 Spill nardus Linn. var. prolixus . 549 plurinodis Stapf. . 533 PAGEK Andropogon schirensis Hochst. var.angustifolia 533 schcenanthus Linn. var. versicolor . 533 Aneura fastigiata L.. 2 31) Angrecum : . 546 Anoiganthus brewiions Baker. 550 Anomobryum filiforme Dicks . 559 Anthericum capitatum Baker . 537 Anthistiria : . Dd, 582, 534 imberbis Retz. . 533 Anthoceros natalensis Sim. . 559 punctatus L. . 559 Antholyza paniculata Klatt. 5388, 540 Anthospermum . : 536, 538 Apodytes dimidiata H. M. 545, 548 Aponogeton natalense Oliver . 550 spathaceum Hook. 552 Ara macao (Lin ) ; vag LD militaris (Lin.) : = 29 Arachnida, descriptions of New South African 687, 711 New South African 289 Aranee . . 289, 290, 687, 689 Argiopide ; : 289, 295 Argyrolobium . : . 5388 marginatum Bolus 549 pilosum Harv. . 562 tuberosum FH. & Z. : . 536 Aristea . E : 587, 540 majubensi Baker . 540 montana Baker . . 540 paniculata Baker - 540 Aristida . : 532, 534, 549 angustata Stapf. . 533 barbicollis Trin. & Rupr. : . 533 bipartita Rupr. . 533 junciformis Trin. & Rupr. : . 933 Armadillo (isopod) : . 568 Army worm (noctuid) . . 212, Artemisia afra Jacq. . . 549 GENERAL PAGE Arthrosolen gymnostachys C. A.M. ; é . 538 Arundinaria tesselata Munro 546, 548 Arundinella ecklonii Nees. . 549 Asclepiadacer . : . 586 Asclepias humilis Schlechter . 586 cucullata Schlechter 536 Ascolepis capensis Ridl. 539, 550 Asparagus . 3879, 5388, 544, 546 africanus Lam. . 562 | plumosus Baker — . 562 stellatus Baker . 562 Asplenium adiantum-nigrum Linn. . 557 bipinnatum (Forsk.) 544 cuneatum Lam. 546, 557 | lunulatum Sw. 544 | monanthes Linn. 544: platyneuron (Linn.) 557 premorsum Sw. 546 protensum Schr... 544 theciferum (H.B.K.) 546 Aster. : ; . 538 filifolius Vent 549, 557, 565 perfoliatus Oliv. . . 561 uliginosus W. 4° LE. EY) | Athanasia : : . 538 montana W.¢ FE. . 561 punctata Harv. . 549 thodei Bolus . 561 Athrixia . : . 538 | angustissima DC. 5 tol | elata Sond. : . o61 | fontana MacOwan Do2 gerrardi Harv. . 561 pinifolia N. E. Brown 561 scandicinum (Willd.) 558 | Avenastrum caffrum Stapf. . 5388 turgidulum Stapf. 583 Axonopus semialatus Hook. . 533 Bagworm, clear-spiral . 625, 685 cocoon of wattle 619, 668 Cossid é . 625 INDEX. 21 PAGE Bagworm, crossed-stick 624, 629, 631, 632, 634, 685 diseases of the wattle. 619, 665 duration of the pupa-stage of = . 673 ego of ‘ 619, 640 enemies of wattle. 619 feeding habits of wattle. . 662 flat . : 625, 685 flat-leaf : 626, 685 fungus of . . 665 Gibsons. . 685 e@rass ; 624, 685 eravel A . 685 eum : 623, 685 ima@os of the wattle 619, 680, 675, 678 larva of wattle 619, 627, 641 lesser lictor . 686 lictor 624, 631, 684, 685 life-history of wattle 619, 683 meadow 624, 629, 684, 685 parasites and enemies of wattle . 664 pupa of wattle 619, 669 rough spiral. 625, 685 rubbish 623, 629, 634, 685 sand 625, 631, 685 seed . . 626, 685 sectional. . 685 single-stick 686 single-straw . 686 thatched 623, 629, 631 634, 685 thorn 3 625, 685 trunk ; . 686 turret F . 686 wattle 620, 6238, 629, 631, 634, 638, 665, 682, 685 722 PAGE Bagworm, webbed-crossed- stick 634, 685 webbed-spiral 625, 631 685 Bagworms, account of species of . 619, 638 bags of 619, 621 literature of 619, 6389 microlepidopterous 615 | South African 587, 589 systematic position of . 619, 634, 640 Barbacenia viscosa (Bak.) Pax. 557 Basket worms : 622 Begonia natalensis Hk. 544 Berkheya : 5: evansil Schltr. 56 grandiflora Willd. var. alternifolia . 562 radula =Stobea radula Harv. 562 speciosa =. speciosa DE. . 562 multijuga=8. multi- juga DC. 552 purpurea = 8. pur- purea DC. . 562 Bersama abyssinica Firesen 547 Bessia fossoria Poc, 289, 304. minor Hewitt Black wattle 620, Blechnum attenuatum Sw. 549, australe Linn. inflexum Kunze punctulatum Sw. Bombycide Bombyx Bopusia seabra Presl. : Borkhausenia flavifrontella Hb. Bowkeria gerrardiana Harv. Brachycorythis pubescens Harv. Brachymenium pulechrum Hook. Brachypodium flexum Nees Brachythecium . . 289, 302, : or or Or or or mw or cr mw uno Cacatua alba (P. L. S. Will.) GENERAL INDEX. PAGE Brachythele ; : 703 | Brachytheliseus bicolor Poe. 314 Bromus leptocladus Nees. 533 natalensis Stapf. 5338 Brownleea cerulea Harv. 550, 562 galpini Bolus 540 Brunsvigia cooperi Baker 537 Bryum afro-turbinatum C. M.. 559 bimum Schreb. 556, 558 stellipilum C. M. 555 umbraculum Burch . 556 | Burchellia capensis R. Br. 543 | Buchenreedera glabrifolia NE. Br. 538 lotononoides Scott Ell. 536 sparsiflora W.¢ E. 538 Buchnera dura Bth. — . . 638 Buddleia : . 548 salviefolia Lam. 542, 5438, 544, 545, § Bulbostylis 550, 559 cinnamomea C. B. Clarke 5389 Buphane disticha Herb. 537 Burchellia capensis R. Br. 546 Bush formation d11, 544. >} citrinocristata (Fraser) 22 ducorpsi Jacq. & Pucher. 2 galerita (Lath.) 7, 9, 11, 12, 13, 14, 15, 17; 18, 20, 22, 28, 505, 506, 507, 508, 509, 510 goftini (Finsch) ee egymnopis Sel. . 22 hematuropygia (P. L. S. Miill.) 22 leadbeateri (Vig.) 22 licemetorhyncha (Bp.) 10 moluccensis (Giv.) 22 ophthalniea Sel. 22 parvula (Bp.) . . 22 GENERAL PAGE Cacatua roseicapilla Vieill. 22 sanguinea Gould 22 sulphurea (Gim.) 22 triton (Temm.) 22 Calodendron : 545 capense Thb. 545, 547 545 Calotermes 451 durbanensis Havi- 329, 330, : 451, 453, 497, 508 madagascarensis land Wasm. 453 Calothrix parietina (Naeg.) Thun. var. africana nov. var. 554 Calhonepion : : . 252 Callitriche bolusii Sch. §° Paw. 550 Calypogeia bidentata Nees 558 Campy lopus chlorotrichus C. M. 555 Canephorinz 591 Caradrina exigua 5 a Carex drakensbergensis C. B. Clarke . . 550, 552 dregeana Kth. 539 esenbeckiana Boeckl. 550, petitiana A. Rich 549 phacota Spreng. . 590 spicatopaniculata C. B. Clarke . 550 Carissa arduina Linn. 547 Cassinopsis capensis Sond. 548 Celastrus acuminatus Linn. 546 albatus N. EF. B. . 543 buxifolius Linn. 543, 546 undatus Thb. . 543, 546 Celtis kraussiana Bernh. 545, 547, 548 Cenia hispid Bth. §° Hk. 561 Cerastium capense Sond 536 dregeanum Fenzl. 562 Cephalaria attenuata R. 4 S. . 536 Ceromitia 634. xanthocoma Meyr. (sp. n.) 615, 617, 625, 285 Ceterach cordatum (Thun.) 544, 558 Chaliine 589, 591, 601 | | INDEX. 723 PAGE ' Cheilanthes hirta Sw. 557 multifida Sw. . 562 Chelypus barberi Purcell 323 hirsti Hewitt (sp. n.) 290, 323, 324 Chilianthus arboreus 4. DC. 548 Chironia krebsil Griseb. 539 Chloris virgata Swartz. 533 Chomophytes 556 Cicurina 291 Clania moddermanni ( Heyl.) 589, 592, 614, 624, 634, 685 543, 546 544, 546 511, 553 547, 561 Clausena inequalis Bth, Clematis brachiata Thb,. Cliff vegetation . Cliffortia 538, linearifola E. 546, & Z. 542, 543, 561 natalensis J. M. Wood 543, 561 prostrata Schltr. 543 serrulata Engl. 561 Climate and vegetation 511, 519 Clubionidee 687, 704: Cluytia 543, 546 aftinis Sond. 557 Cocculus villosus DC. 502 Cockatoos, description of hybrid 12, 13, 14 hybridism among. 7 Codlin moth . 686 Coenodomus hockineii Walsing- ham : : . 6395 Coleophoridee 635 Combretum glomeruliflorum Sond. 662 kraussu Hochst. . 122 Commelina africana Lin. 5388, 552 Corotoca akermani Warren (sp. n.) 105 melantho Schiddte 104, 105 phylo Schiidte 105 Corycium magnum Rolfe 540 nigrescens Sond. 540 | Corydalis cracca Schl. 562 724 GENERAL INDEX. PAGE Cosside 589, 590, 604, 605. 624, 631, 634, 685 Cossus incanescens But!. . 608 vinectus WIk. 2 . 608 Crassula . : 535, 538 rubescens Schonl. & Bkr.f. 552 vaginata E. & Z. . 536 Crypheecere ; : 290, 295 Cryptocarya acuminata Schinz. 545, 546, 547 Cryptothelia —. ; . 592 Ctenizidee . 289, 299, 687, 689 Cubaris . : : 567, 568 burnupi Clige. (sp. n.) 567, 572, 583 flavescens Brandt 568 eriseo-albus (Dollfus) . 575 limbatus Brandt 568 longicauda Cilge. (sp.n.) 567, 574, 576, 583 natalensis Cllge. (sp. n.) 567, 573, 583 nigricans Brandt . 568 reticulatus Cllge. (sp.n.) 567, 570, 572, 582 trilobata Cllge. (sp. n.) 567, 575, 584 warreni Cllge. (sp. n.) 567, 569, 582 Cussonia paniculata H. & Z. . 542 spicata Thb. 542, 543, 560 Cyanophycere . 558, 555, 562, 565 Cyanotis nodiflora Kth. . 538 | Cyathea dregei Kze. . . 541 Cyclophorus africanus (Kze.) . 549 Cyenium racemosum Bth. - 537 Cynodon dactylon Pers. = Florida Grass : 390 Cynoglossum micranthum Desf. 538 Cyperus compactus Zam. var. flavissimus . 5 849) fastigiatus Rotth. . 550 flavus Ridl. . . 552 | marginatus Thunb. . 550 Cyperus natalensis Hochst. schlechteri C. B. C. Cyphia Cyrtanthus : : angustifolium Ait. Cyrtomium faleatum (Lin. fil.) Cystopteris fragilis (Lin.) 546, Dasaratha : longicauda Warr. Deilephila euphorbiz (Linn.) . vespertilio (Esper.) . Desis : : Denekia capensis Thb. . Dianthus seaber Thb. Diascia cordata N. E. Brown purpurea N. FE. Brown . rigescens H. M, Dicoma anomala Sond. . anomala Sond. var. circioides Didymodon ; ; Dierama pendula Baker 520, Digitaria diagonalis Stapf. monodactyla Stapf. . tricholenoides Stapf. Dimorphotheca caulescens Harv. nudicaulis DC. var. latifolia | Dioscorea : : 544, Diplodoma : : 604, Diploexochus : Dipluridee 687, Disa cephalotes Reichb. /. chrysostachya Sw. erassicornis Ld. frigida Schltr. oreophila Bolus pulchra Sond. macowani Reichb. f. Disperis cardiophora Harv. fannine Harv. tysoni Bolus venusta Bolus . or w (=P) or Or or or or GO <1 or hw or Ow ww vo Ww ww or or Ol bo 562 54.6 605 569 699 540 540 540 550 540 540 540 540 546 900 550 GENERAL Disperis wealii Reichb. . Doryopteris concolor (L. & F.) 558 Drakensberg, temperature of . 521 vegetation of 511 Dryopteris athamantica (Kze.) 544 bergiana (Schl.) 549, 559 elongata (Sw.) 544 Ecology of the Drakensberg, plant 511 Ehrharta erecta Lam. 538 Ehretia hottentottica Burch. . 379 Eleodendron 543, 545, 547 Eleocharis limosa Schultes 550 Elaphoglossum — spathulatum (Bory) . o49 Elionurus argenteus Nees 538, 549 Euealypta 558, 559 Encephalartos ghellinehii Lehm. 541, 562 Epichnopteryx . : . d90 transvalica Hmpsn. 604 Epilobium flavescens E. M. 550 Equus burchelli (Gray) 20 caballus Lin. P A = P10) Eragrostis brizoides Nees 533 cesia Stapf. 533 chaleantha Trin. 533 chloromelas Steud. . 533 curvula Nees 533 natalensis Hack. 5383 nebulosa Stapf. 533, 549, 550 superba Peyr. . 533 Hrica algida Bolus 582, 562 alopecurus Harv. 538, 552, 562 aspalathifolia Bolus 538 caffra Lin. . 548 caffrorum Bolus 538, 552, 557, | 562 cerinthoides Linn. 5386, 557, 962 cooperi Bolus 548 INDEX. 725 PAGE Erica drakensbergensis Guth. § Bol. : 5388 flanagani Bolus . 562 frigida Bolus , 502, 562 lasiocarpa Guth. § Bol. 588 natalitia Bolus . 538 oatesu R. A. Rolfe 552, 562 schlechteri Bolus 548, 562 thodei Guthr. 4° Bol. 552, 562 woodii Bolus 552, 562 Hrigonee 295 Hrigone . : : . 295 Erigonopsis littoralis Hewitt (g.e. sp.n.) 289, 290, 296, 297 Eriocaulon bauri N. FE. Brown. 550 Eriosema 538 Eriospermum cooperi Baker 537 Euclea : 547 undulata Thb. 537, 542, 548, 546 Eucomis. : : 550 humilis Baker. 538 Eumeta . : . 592 cervina Druce . 593, 594: crameril : 593 jJunodi Heyl. 594, 595, 640, 677 moddermanni Heyl. 594. zelleri Heyl. . 594 Eumorphia sericea W. & E. 552, 559, 561 Euphorbia striata Thb. 537 Euprepia 285 Euprepiadee : 285 Euryops evansii Schi/tr. 561 montanus Schltr. 552 Eutermes 103, 418 bilobatus (Haviland) 329, 346, 398, 400, 401, 402, 404, 412, 487, 502, 504 hastatus (Haviland) 330, 489, 490, 491, parvus (Haviland) 329, 361, 372, 396, 397, 398, 485, 500, 501, 504. 004 726 PAGE Eutermes trinervius (Rambur) 108, 329, 330, 332, 334, 350, 355, 3899, 401, 402, 403, 407, 410, 412, 418, 423, 491, 500, 502 Exartema latifasciana Hw. 635 Exceecaria 560 Faurea saligna Harv. . . O41 Felicia . : : enon amelloides Schltr. . 561 drakensbergensis W. 4 E. 552, 561 levigata (Sond.) O. Hoffm. 561 pinnatifida W. f E. 552 Festuca costata Nees 533 seabra Vahl. . . 533 Ficinia stolonifera Boeckl. 539, 550 | IMU 3 ; . ool capensis Thb. . 548 cordata Thb. ; . 270 Fimbriaria marginata Nees . 556, 558, 559 Fimbristylis 539, 550 Fissidens bryoides Heda. . 558 taxifolius Hedw. 556, 558 Florida grass = Cynodon dac- tylon 390 Fossombronia ioreomanehe! L.& Cae : : - 609 Frullania diptera Nees . 505 Fumea limulus . ; . 635 obscurata Mey. (sp. n.) 615, 625, 685 trimenii Heyl. 604 Fynbosch or Maquis formation 511 Galium rotundifolium Lin. var. hirsutum . , aXe subvillosum Sond. 538 wittenbergense Sd. var. glabrum 539 Galtonia candicans Dene. 540, 562 Gastropacha : . 285 GENERAL INDEX. PAGE Geissorhiza 540 Gelechiide j 635 Geology, topography and _ soil conditions of Natal 511, 512 Geranium incanum Burm. 562 ornithopodum E. & Z. 536, 549 pulchrum Neb. 550 thodei R. Kunth. 562 Gerbera ambigua Sch. Bip. . 562 parva N. BL. Brown piloselloides Cass. Geum capense Thd. 5) Gladiolus saundersii Hk. f. . 5 Glcocapsa sanguinea Kiitz. 558, 554 Gomphostigma scoparioides Turcz. 549 Gorgyrella abrahami Hewitt 305, 306 namaquensis 309 schreineri Purcell . 306 Grewia occidentalis Linn. 548, 546 sutherlandi H. 4& H. = Baakhout ; 542, 5638 Grimmia apocarpa Hedw. . 555 Gunnera perpensa Linn. 550 Guthriea capensis Bolus Gymnelema 589, 604, 605, 606, 611, 612, 685 imitata Janse (sp. n.) 589, 610, 611, 614 leucopasta Hmpsn.. 612 pulverulenta Hmpsn. 612 rougemonti Heyl. . 612 stibarodes (Meyr.). 590 stygialis Heyl. 589, 607, 610, 625, 685 vinctus (W/k.) 589, 606 608, 614, 624, 685 Gymnopentzia pilifera N. E. Brown Gymnosporia buxifolia Tian. Gynanisa maia (Klug.) 561 379 269, 273 Hemanthus hirsutus Baker 557 Habenaria bonatea Reichb. f. GENERAL PAGE Hahnia oreophila 2. Sim. 294. Hahniine ; . 294 Halleria lucida Linn. 543, 545, 548 Hammock moth = Perophora sanguinolenta 635 Hapalothele 703 Haplocarpha scaposa 536 | Harpechloa capensis Kth. 533 | Harveya coccinea Schltr. 537 | speciosa Bernh. 537 Helichrysum 535, 536, 538 adenocarpum DC. 520 | album N. E. Brown 592, 061 alticolum Bolus var. montanum 552 appendiculatum Less. 561 | aureo-nitens Sch. Bip. 552 calocephalum | Schltr. 552 chionospherum | DO. 552, 557 | confertum WN. E. Brown 557 cooperi Harv. 552 cymosum Less. 552 elegantissimum DC. 552 flanagani Bolus 552 fulgidum Willd. | 552, 557 fulgidum Willd. var. monocepha- | lum . o61 fulgidum Willd. var. nanum 597 hypoleucum Harv. 557, 561 lanatum Harv. 561 randii Moore 552, 561 reflexum N.E.Br. 549 retortoides N. E.B. 552, 561 | Heliophila subulata Burch. 552, INDEX. 127 PAGE Helichrysum sessile DC. 561 setigerum Bolus . 552 setosum Harv. 557 splendidum Less. 549 subelomeratum Less. 552 sutherlandi Harv. tenuifolium 549, 557 trilineatum DC. . Hemiblossia idioceras Hewitt (sp. n.) 687, 688 Hermacha 314, 699 bicolor ( Poc.) 314 ecaudata Sim 314 mazcena Hewitt (sp.n.) 289, 312 Hermachola Hewitt (g.n.) 289, 314 erahami Hewitt (sp. un.) 289, 314, 315, 703, 704: Herschellia baurii Kranz. 538 Hesperantha radiata Ker. 540 Heteromorpha arborescens Ch. & Sch. 548, 546 Heteromma : . 538 simplicifolia W. 4° EH. 561 Hibiscus ethiopicus Lin. 537 leiospermus Harv. 537 saxatilis W. J: EB 537 Hippobromus alatus E. & Z. 542, 546 354, 355, 356, 363, 411, 415, 416, 423, 437, 447, 448, 452, 466 451 Hodotermes aurivellii braini Fuller (sp. n.) 380, 446, 503, 504. havilandi Sharp 330, 353, 355, 438, 441, 448, 447, 503, 504 karooensis Fuller (sp. n.) 3830, 332, 425, 429, 433, 503, 504 728 PAGE Hodotermes mossambicus (Hagen) 380, 424 431, 434, 438, 442, 443, 503, 504 pretoriensis Fuller (sp. n.) 3380, 429, 432, 433, 441, 503, 504 transvaalensis Fuller (sp. n.) 329, 330, 334, 335, 349, 354, 355, 388, 425, 430, 432, 438, 434, 436, 488, 439, 442, | 443, 444, 445, 446, 497, 503, 505 | viator (Latreille) . 329, 330, 856, 358, 448, 497, 503, 504 | warreni Fuller (sp. n.) 330, 425, 44.4, 504 Humidity and precipitation in Natal : - O24 | Huttonea grandiflora Schltr. . 540 oreophila Schltr. 550 Hybrid cockatoos 7, 505 fishes 20 mammals 20 moths - 20 Hymenophyllum tunbridgense Sm. . : : 559 Hypericum ethiopicum Thb. 936 lalandiui Choisy 539 Hypnum cupressiforme L. . 555 Hypocheeris radiata Lin. 53 Hypoxis baurii Baker . . 55 rubella Baker . 55 Idiops arnoldi Hewitt 310, 311 astutus Hewitt (sp. n.) 289, 310, 311 pretoriz (Poc.) . 289, 310 Ilex capensis Harv. § Sond. 547, 548 Impatiens capensis, Thb. 544. | Imperata arundinacea Cyr. 949 | GENERAL INDEX. PAGE Indigofera hedyantha E. & Z. 549, 562 hilaris E. & Z. 536 hirsuta Harv. 536 schlechteri Baker f. 549 Ipomcea crassipes var. hirsuta. 537 | Iridomyrmex humilis Mayr. 490, 491 Isaria psychide Pole Evans 665 _ Isopoda of Natal 567 Ixia brevifolia Baker 562 Juncus dregeanus Kunth. 549 effusus Linn. 550 exsertus Buchen. 550 elaucus Ehrh. 550 Kalanchoe thyrsiflora Harv. 538 Kiggelaria dregeana Turcz. Natal Mahogany 545, 547, 548 Kniphofia F . 550 Keeleria cristata Pers. 533, 534, 549 Kyllinga melanosperma Nees . 550 Lacosomide 635 Lansdownia Heyl. 592 Lasiosiphon krausii Meisn. 538 meisnerianus Endl. 5388 Lejeunea flava L. 556, 558, 559 (Ptycholejeunea) striata Nees 556 Lepidozea : ; 558 Lepthercus dregei Purcell 699 rattrayi Hewitt (sp. n.) 687, 699, 701, 703, ill | Lessertia flanagani L. Bolus 552 perennans DC. 549 Leucosidea . 548 scrub 548, 549 sericea EH. gf Z. 548, 546, 560, 5638, 565 Leucospermum gerrardi Stapf. 540 Licmetis nasica (Temm.) 9, 11, 18, 14, 15, 17, 18, 20, 22, 28, 509, 510 Limosella aquatica Linn. 550, longiflora O. Kuntze 550 maior Diels . 550 GENERAL INDEX. | PAGE Linyphia 295 Linyphiine 295 Liparide 285 Liparis . : : 285 Litanthus pusillus Harv. 557 Lobelia preslii A. D. C. 562 vanreenensis O. Kuntze 536 Loranthus dregei E. & Z. 662 Lotononis . 538 wyliei Wood . 562 Luzula africana Drege . 550, 552 Lycopodium enidioides Linn. 546 verticillatum Linn. 546 Lythrum 550 | Macowania glandulosa N. L. Brown A 552 Macromitrium lycopodioides Burch. 555 Madotheca capensis G. . 556 Mahernia malvefolium N. £E.B. 536 Malva parviflora Linn 537 Manatha . 600 | ethiops Hips. 600, 685 Mantis religiosa Linn. . 281 Mariscus congestus C. B. Clarke 550 deciduus C. B. Clarke 550 capensis Schrad. 539 Melanocera menippe ( Westwood) 269, 270, 273, 285 Melasina circophora Meyr. enaphalodes Meyr. 615,625, 685 | craterodes Meyr. 615, 617, 625, 685 | 615, 616, 622, 623, 685 623, 685 picea Meyr. stelitis Meyr. . stibarodes Meyr. 611 tyrophanes Meyr. 625, 685 tyrophanes sp. n. 615, 617 Melasma capensis Hiern 539 secabrum Berg. 539 VOL. 3. 590, 604, 605, 611, 612, | 625, 631, 634 | 685 | 729 PAGE Melianthus ; 549 Melolobium cernuum EF. 4 Z. . 586 decumbens Bth. 5386 Mendelism 14, 15 Mentha aquatica Linn. 550 Mesembryanthemum 538 nubigenum Schltr. . 552 Metalasia muricata Less. 538, 561 Metisa ; : . 598 Microchloa altera Stapf. var. nelsoni . 5338 caffra Nees 533 Migide 710 Mimulus gracilis R. Br. 539 Mnium punctatum L. 556 Mohria caffrorum Desv. 544,557, 558 Monda 589, 602 delicatissima Ji lk. 589, 608, 614, 686 heylertsi Junod 604 major Heyl. 604. rogenhoferi Heyl. 589,604, 686 Monoplius emulator 358 inflatus 358 segillatus 358 Monsonia attenuata Harv. 536 Morea edulis Ker. 562 natalensis Baker . 540 spathacea Ker. . 520, 540 tricuspis Ker. 587 Moths, the pairing of 2738 Muizenbergia Hewitt (g. n.) 289, 290, 291, 294. abrahami Hewitt (sp. n.) 289, 290, 291, 293, 296 Muraltia ecornuta N. E. B. 537 flanagani Bolus 562 saxicola Chodat 537, 557 Mus coucha A. Smith 665 zuluensis 665 Myosotis afropalustris C. H. Wright . 550 Myrica ethiopica Linn. 548, 561 49 ~JI oo (=) GENERAL INDEX. PAGE PAGE Myrsine melanophleos R. Br. _ Osyridocarpus natalensis A. DC, 542 545, 547, 548, 560, 565 Osyris abyssinica Hochst. . 542 umNai=Elewodendron sp. Cee Othonna natalensis Less. . 507 Nasturtium fluviatile E. M. 550 Sudehout ‘ ; - 547 Nemesia flanagani Hiern. . 958 | Oxalis convexula Jacq. . eee floribunda Lehm. . 5386 Oxyanthus ; ; aie foetens Vent. var. lati- folie ; 558 Pachycarpus. é . 5386 melissefolia Bth. 5 pio) | Panax gerrardi Harv. . Bey Ninmbin. 568 Panicum ecklonii Nees . . 533 Nidorella depauperata . . 536 eae ieee Notholwna eckloniana Kze. . 507 HEU ENB 2318 Nudaurelia belina Westw. 272, 273 yeas serratum Spreng. . 538 Nymphula 635 | Papillaria floribunda C.M. . 556 _ Passerina ericoides Lin. . 542 Odina ecatfra Hook. : 5 Bele filiformis Lin. 557, 561 (Keeticine : : 589, 591 | Pavetta caffra Thb. : . 043 (Eceticoides ; : . 596 | Pelargonium aconitifolium (Ecophorine —. : . 685 | E. & Z. 536 Oiketicoides : : . 596 | alchemilloides Oldenlandia amatymbica Willd. 562 Kuntze 588 | bowkeri Harv. . 537 Olea laurifolia Lam. . 545, 548 | dispar Neb. . 936 Olinia eymosa Jhb. 545, 547, 548 | flabellifolium Opisthophthalmus latimanus | Harv. 536 Koch var. | inquinans Ait. . 562 keilandsi. 326 pulverulentum pugnax Tho- Colv. 536 rel. van. schlechteri : 5386 natalensis Pellwa hastata Link. . . d57 Hewitt | viridis (Forsk.) . . Od (var. n.) | Pelmatorycter dreyeri Hewitt 290, 325 (sp. n.) F : 289, 299 Orchidace ; ; . 540 nigriceps Purcell 299 Orgyia. ; : . 285 | Pentanisia variabilis Harv. . 536 Ornithogalum . : . 549 | Pentaschistes natalensis Stapf. 533 flanagani Baker. 552 | Pennisetum sphacelatum Dur. natalense Baker. 588 | & Schinz. . 583 zeyheri Baker . 587 | thunbergil Kunth. Orthosiphon macranthus Gurke 538 549, 550 Orthotrichum . : . 555 Pentzia cooperi Harv. . Bool Orycteropus capensis Smuts . 366 | Perea fluviatilis Rond. . . 20 Osmites . : ; 549, 561 Peripatus : ; . 228 Osteospermum moniliferum Lin. | Perophora sanguinolenta . 6385 538, 542 | Perotis latifolia Ait. . , ae GENERAL PAGE Phalaris arundinacea Lin. 5 88B} minor Retz. 533 Philonotis ’ ‘ [DDG Philoscia 567, 568, 576, 577, 580 dilectum Clilge. (sp. n.) 567, 579 warreni Cllge. (sp. n.) 567, 578, 584 Phygelius equalis Harv. 549 | Phylica paniculata Willd. 561 | Pinus insignis Douwgl. 662 | Piperomia reflexa A. Dieter 557, 558 | Plagiochila corymbulosa Pears 555 javanica N. 4 M. 595 spinulosa (Dicks) Dum. 3 556 | Plant ecology of the Drakens- berg Range 511 | formations and associa- | tions 511, 527 | formations, the succes- | sion and _ inter-rela- | tionships of the olde o625| Planula, development of 91 Planula of Schizotricha sim- plex Warren 83 Platycercus eximius (Shaw) 19 pallidiceps Vig. 19 pennantii Vig. 19 Plectranthus , 588, 544, 558 Plectronia mundtiana Pappe 543, 546, 548 spinosa Klotzsch 548 ventosa Lin. 5438, 546 | Pleurostylia capensis Turez. 545 | Plumularia armata Allman 89 aglaophenoides Bale 90 | buskii Bale 89 | campanula Busk 89 Podaxon carsinomalis 410 pistillaris 410 Podocarpus elongata L’ Herit 045, 547, 548 545, 547 faleata R. Br. INDEX. 731 PAGE Podocarpus thunbergii Hook. . 545, 547, 548 Polygala hispida FE. Vey. 562 hottentota Pres!. 536 rarifolia DC. Bal tenuifolia Link 5387 virgata Thb. var. decora 537 serrulatum 550 Polypodium lanceolatum Lin. 544, 558 lineare Thb. 544 polypodioides Lin. 544: vulgare Lin. 544, 557 Polyplumularia . : > 0) Polystachya F . 546 ottoniana Reichb.f. 557 Polystichum aculeatum Lin. 546 aristatum Forst. . 546 Porcellionides 568 Prepotency, sex . 19 Printzia . : : . 538 laxa N. E. Brown . 561 Protea abyssinica Willd. 540 flanaganii Phillips . 540 hirta Klotzsch. . 540, 548 lanceolata E. M. 540, 548 multibracteata Phillips 540 rouppellize 540, 565 subvestita N. E. Br. . 540 veld 511, 540, 565 Psammotropha . ; . 538 androsacea Fenzl. 562 myriantha Sond. 552 Pseudogamy 281 Psittacidee ‘ 19 Psoralea pinnata Lin. . . 549 Psyche (Manatha) ethiops (Hipsn.) 589, 600, 614 (Manatha) subhyalina Janse (sp.n.) 589, 601, 614 285, 589, 591, 595, 620, 634, 685, 638, 640, 653, 669, 681, 684 Psychide 732 GENERAL PAGE | Psychinee 589, 591, 595, 640 Psychoidine : 591 Pteroxylon utile FE. & Z. = Sneezewood. 545, 547 Pteridium aquilinum Lin. 539 Pteris biaurita Lin. 546 eretica Lin. 546 dentata Forsk. . 546 Pterocelastrus 546, 560 Pterygodium hastatum Bolus. 550 magnum Reich. 550 Pycreus angulatus Nees 550 | ferrugineus C. B. Clarke 550 flavescens Reichb. 550 macranthus C. B. Clarke 550 rehmannianus C. B. Clarke 550 umbrosus Nees £50 Pygeum africanum Hk. f. = Red Stinkwood 545, 547 Pyralidee 6385 Ranunculus baurii MacOwan 550, 559 cooperi Oliv. 550, 559 meyerl Harv. 550 Raphionacme divaricata Harv. 586 Rhamnus prinoides L’Herit. 5438, 546, Rhinotermes putorius (Sjostedt) 329, 380, 415, 453, Rhodobryum syntrichioides (C. M.) Par. Rhynchosia : gibba #. M. Rhyncospora glauca Vahl. Rhus dentata Thb. discolor FE. M. Richardia rehmanni Engl. Riccia fluitans L. Riocreuxia torulosa Dene Rocky scrub formation . 5386, 511, 548 453 503 558 538 536 550 543 54:6 538 559 5044 542 INDEX. PAGE Romulea thodei Schltr. 552 Royena cordata BE. M. 538, 542, 562 hirsuta Lin. 562 lucida Lin. 542,543, Rubus rigidus Smith pinnatus Willd . Rumex ecklonianus Meisn. Rumex woodii N. E. Br. 546, 547 544, 546 544, 546 587, 550 557 Salvia : F 537, 538 stenophylla Burch. 562 Sandersonia aurantiaca Hook. . 562 Saturnia pavonie (Lin.) 20 pyri Borkh. ae FAO) Saturniidee 285, 638 Saturnine : . 285 Satyrium atherstonei Reichb. f. 550 cordifolium Ldl. . 540 cristatum Sond. . 550 longicauda Ldl. . 540 macrophyllum Ldl. . 550 neglectum Schlechter 550 ocellatum Bolus 550 parviflorum Sw. 540, 550 spherocarpun Ld. 550 Seabiosa africana Lin. 5386 columbaria Lin. var. dissecta 536 Schmidelia africana DC. 543, 546,548 monophylla Pres/. 546 Schistostephium crategifoliumn Fenzl. : 561 Schizochilus angustifolium Rolfe 562 gerrardi Bolus 562 Schizoglossum : 5388 Schizostylis coccinea Backh. & Harv. 549 Schizothrix epiphytica Fritsch (sp. n.) 5538, 554 miilleri Naeg. 554 Schizotricha simplex Warrer (sp. n.) 5 hil unifureata — All- man 90, GENERAL PAGE Seilla concinna Baker 549 | kraussii Baker » bbe | natalensis Planch. 5387, 557 | rigidifolia Kth. var. ner- vosa 5387 Scirpus falsus C. B. Cl. . 552 | fluitans Lin. 550 | hystrix Thb. 550 | Scleria woodii C. B. Cl. _ OY} Seolopia ecklonii Szysz. = Red Pear 545, 548 | mundtii Arn. . . 547 Scorpiones 290, 325 | Serub formation 511, 543 Seutia commersoni Brogn. 543, 546 Sebwea marlothi Gilg. . 552 procumbens A. 7. Hill 552 spathulata Steud. 552 thomasil Schinz . 552, 559 | Selaginella depressa A. Braun. 558 kraussiana A.Braun 546 | rupestris Sprengel . 557 Selago : x 857/ flanagani Rolfe 549, 557 | Senecio : 536, 588, 546 barbatus DC. 552 deltoideus Less. 544 drakensbergensis Klatt. 562 evansi N. FE. Br. 562 gramineus Harv. 552 macrocephalus DC. var. hirsutissimus 562 macrovlossus DC. 544 reptans Turcz. . 552 seminivea Wood Evans 552 tanacetoides Sd. 549, 562 tugelensis W. §& H. 552, 562 viscidus N. E. Brown 539 Serpula : > 8B Setaria imberbis R. 4 8. 538, 550 Silene burchellii Off. 536 capensis Ott. 562 Sium thunbereii DC. 550 INDEX. Smerinthus ocellatus (Lin. populi (Lin.) Solenobia Solifugee . Sopubia cana Harv. simplex Hochst. Sparmannia palmata H. MW. Spheralcea pannosa Bolus Sphingicampine Sphingidee Sphinx : : Sphodromantis bioculata Spirachtha Spiroctenus Stachys bolusii Skan. Staphylnids Stasimopus artifex Poc. astutus Poc. brevipalpis Purcell erythrognathus Purcell cigas Hewitt 290, 302, armatus Hewitt 733 PAGE 285 20 20 285 323, 687 539, 550 539 54.2 537 285 285 285 281 108 308, 804 289, 305 562 1038 321, 322, 694, 695 692 697 696 696 g (sp. n.) 289, 318, 322, 698 insculptus Poc. 692, 693, 694, 695 insculptus Poc. var. peddiensis Hewitt (var. n.) 687, 690, 691, 695, 711 kentanicus Purcell kolbei Purcell 694. 694. longipalpis Hewitt (sp. n.) 687, 689, 698, 711 minor Hewitt (sp.n.) 289, 320, nigellus Poe. oculatus Poc. palpiger Poe. paterson Hewitt 399 OL4, 697 319, 320, 322, 697 289, 320, 321 698 697 734. PAGE Stasimopus qumbu Hewitt 695 robertsi Hewitt 698 schonlandi Poc. 696 schreineri Purcell . 699 spinipes Hewitt (sp. n.) 687, 692, 694, 711 steynsburgensis Hewitt (sp. n.) 289, 317, 698 Stenoglottis fimbriata Ld. 544 Stiburus alopecuroides Stap/. 549, 550 Stigonema hormoides B. §° F. 558, 534: informe Kiitz. 588, 554 Stipa dregeana Steud. . . 533 . 5b2 538, 549, 557 511, 548 Stobea multijuga DC. . cinerea Thb. Stream bank vegetation Strelitzia augusta Thd. . 366 Streptocarpus : 544, 559 pusilla Harv. 558 Striga elegans Bth. 537 Strophanthus capensis A. DC. 546 538, 549 var. Sutera : crassicaulis Hiern purpurea c . 902 Sutherlandia frutescens R. Br. 538, 557, 562 Syncolostemon macrophyllus Gurke 538 Syringodea flanagani Bkr. > 552 Syrnium woodfordi (A. Smith) . 8 Taleporide : : . 635 Tambootie grass : . 531 Teichobiine 6385 Telea 285 ‘Tephrosia : : 538 Termes badius Haviland 3829, 380, 362, 363, 364, 375, 377, 378, 379, 381, 384, 418, 419, 421, 422, 423, 425, 456, 457, 475, 500, 502, 504 GENERAL INDEX. PAGE Termes bellicosus Smeathman . 459, 470 goliath Sjost. 470 incertus Hagen 329, 330, 334, 336, 341, 342, 345, 346, 354, 363, 366, 372, 398, 394, 396, 466, 483, 500, 504 latericius Haviland 329, 330, 3389, 340, 362, 3638, 364, 367, 378, 380, 383, 385, 386, 389, 390, 479, 481, 482, 499, 501, 504 monodon Gerstacker 456, 475 natalensis Haviland 329, 330, 334, 336, 337, 362, 3638, 364, 368, 370, 371, 372, 373, 378, 382, 383, 391, 392, 414, 418, 422, 423, 458, 470, 498, 499, 502, 504 swazie Fuller (sp. n.) . 3830, 416, 462, 469, 497, 5038 terricola Sjostedt 456 transvaalensis Sjéstedt 382, 456, 457 491, 502 vulgaris Haviland 329, 330, 3837, 340, 841, 368, 364, 388, 391, 392, 395, 481, 482, 501, 504. waterbergi Fuller (sp.n.) 329, 330, 361, 362, 416, 466, 498, 503 Termites, abundance of 329, 331 behaviour of winged 329, 333 nesting habits of 829, 349 observations on some trinervius Rambur. South African 329 Termitomimus . ; ~ Los entendveniensis Trig. . 104 Termitophilous beetle . . 103 Termitoxenia , ; . 104 GENERAL INDEX. PAGE Themeda forskalii Hack. ) BSR Thesium : : _ sts! Thorn veld 540 Thunbergia steplicifola E.M. 537 Thuidium angustifolium Jaeg. 556 tamariscinum Hedw. 555 Thysanura 5 : . 376 Tineid bagworm ; 685 Tineidee 285, 590, 604, 615, 625, 634, 685 Toddalia lanceolata Lam. = White ironwood = 545, 547, 548 Tortella czspitosa (Schw.) Lempr. 555, Tortricide F : 635 Trachypogon Tose Hack. 583 on Ur or for) Tribulus terrestris Linn. 5 OSH! Trichocossus 604, 605, 606, 612, 635 albiguttata Hmpsn. 614 arvensis Janse (sp n.) 590, 612, 6138, 614, 624, 634, 685 Tricholena rosea Nees . 5 688} setifolia Stapf. 5838, 549 Trichomanes pyxidiferum Linn. 546 Trichoptera : . 684 Trichopterix flavida Stapf. . 538 simplex Hack. . 583 Trifolium africanum Ser. . 039 Trimeria alnifolia Planch. 545, 548 Tristachya leucothrix Trin. . 583 Tritonia lineata Ker. . . 540 kraussil Baker . d40 Tulbaghia leucantha Baker =. 587 Ursinia brevicaulis W. 4 EF. . 538 montana DC, . - 902 Valeriana capensis Thb. 539, 549, 562 ADLAKD AND SON AND WEST NEWMAN, LTD., 739 PAGE Vangueria caffra Sim. . 543 lasiantha Sond. 549, 560 Veld formation é 5, 530 grasses : : . 533 plants, associated 534, 536, 537, 539 Venidium arctotoides Less. . 588 Vernonia 3 : . 538 kraussil Sch. Bip. 614 natalensis Sch. Bip. . 552 Veronica anagallis Linn. . 590 Vitis capensis Thb. 5 . 046 Vlei formation . ; 5 Gplal Vleis : : ; . 549 Wahlenbergia depressa W. f° E. 552 montana DC. . 558 undulata 4. D.. 539 zeyheri E.f Z.. 536 Watsonia densiflora Baker 5 BYE Weisia bewsii Sim ; . | 558 viridula LD. 556, 558 Widdringtonia cupressoides DC. 549 Woodsia burgessiana Gerry, =. 558 Wurmbea kraussii Baker » 552 pusillus Phillips . 552 Xanthoxylon capense H. = Knobwood : . O45 Xyris ; ‘ 550, 559 Xysmalobium ieolaceatene Dene. . . 586 parviflorum Harv. 536 stockenstromense E. Elt. . 9086 Zaluzianskya flanagani Hiern . 552 goseloides Diels. 549, 557 montana Hiern. 552, 557 IMPR., LONDON AND DORKING, i, Aghios ot Wes ' Tae has Cay ees Sie Aw 5 vib te sa alt. a . say Fil em ah dels’ Hie ht PAA Ear a let: te a get ca eee hea ey pe ale, pene Lae np (ae fo hp tee We = am) ; ee age e vibes AW gi et i whi ‘i Meet ate fe pes) dotinayt 4 chi “i gsen oe ee D Raniah a Nd ys ty ats paint ia yume fo, mt) havnt: uy ear seh haat pon S . PN fl Bo ton, a th ike OP aes (i ec) ve fa ae re 4 ie sou 5G a pe: ta part i ee. ‘ ny iad ag ed sere ek Mie aoa fal fut? pela a Be nm ee Picren's acyar . is » an ae Find a ae : we ATUL F, ar - pi, HadT rs 2 mie i Pes VOL. Ill. PART l. SEPTEMBER, 1914 ANNALS NATAL MUSEUM EDITED BY ERNEST WARREN, D.Sc.(Lonp.)., Direcror. PRINTED BY ORDER OF THE TRUSTEES. LOW WO: ADLARD & SON, BARTHOLOMEW CLOSE. Price 7s. 6d. net. . “AN Eas * 7 py Ce ew i- »:; se 2 ee “a * a 4 x ~ ) oe . y ese: tee wa _S = 3 . r. : ess ie hea m a : , o y, ¢ : am / ¥ " Fey - a > £ o% “te “3 ey You CONTENTS. Descriptions of some South African Marine Shells. By Evear A. Smiru, 1.8.0. (With Plate I) A Case of Hybridism among Cockatoos. By Ernest Warren, D.Sc.(Lond.). (With Plate IT) On South African Ennex, with Descriptions of New Species and Varieties. By Henry Crirpen Burnvp. (With Plates III, IV, V, and Text-figs.) On the Development of the Planula in a Certain Species of Plumularian Hydroid. By Ernest Warren, D.Sc. (Lond.). (With Plate VI and 4 Text-figs.) Note on the Occurrence in South Africa of a Termito- philous Beetle of the Genus Corotoca. By Ernest WarRREN, D.Sc.Lond. ADLARD AND SON, IMPR., LONDON AND DORKING. PAGE 1 VOL. III. PART 2. OCTOBER, 1915 ANNALS NATAL MUSEUM EDITED BY ERNEST WARREN, D.Sc.(Lonp.)., Dirucror. PRINTED BY ORDER OF THE TRUSTEES. LONDON: ADLARD & SON, BARTHOLOMEW CLOSE. 1Oto: Price Iss. net. ] y )