The Frfsh-Vater Ijollusca of the Tanganyika Territory

THE FRFSH-VATER IJOLLUSCA OF THE TANGANYIKA TERRITORY AND ZANZIBAR PROTECTORATE, AND OTHER CONTRIBUTIONS TO THE SYSTEATICS BIONO1ICS AND DISTRIBUTION OF FRESH-WATER AND TERRESTRIAL L1OLLUSCA.

THESIS

Submitted for the DEGREE of D.Sc. in The University of Edinburgh by WALTER ALAN MOZLEY, Ph.D.(Edin.)

April 1940. CONTENTS

List of Publications by Walter Alan tIozley.

Papers: 1. The Freshwater Mollusca of the Tanganyika Territory and Zanzibar Protectorate, and their Relation to Human Schistosomlasis. Trans. Roy. Soc. Edinburgh, LIX,(No.26), 1939- A Biological Study of the Sub-Arctic Uollusca. Proc. Amer. Philos. Soc., 78, 1, 1937- The Fresh-Water Mollusca of Sub-Arctic Canada. Canadian Jour. Research, D,16 2 1938- 1 .

PUBLICATIONS by WALTER ALAN MOZLEY

1925 Segmeritina crassilabris Walker in Manitoba. Canadian Field Naturalist, 39, p.85 1926 Molluscs from the Manitoba-Ontario Boundary. Nautilus, 39, pp.121-128. Preliminary List of the Mollusca of Jasper Park, Alberta. Nautilus, 40, pp.53-56. Some Molluscs from Western Canada. Nautilus, 40 0 pp.56-63. Molluscs from the Lake Brereton District, Manitoba. Canadian Field Naturalist, 41 0 pp.59-61. 1927 Some Plancton Organisms from Lake Brereton, Manitoba. Canadian Field Naturalist, 41, pp.84-85.

1928 The Variation of Lymnaea traskii Tryon In Pond and Lake Habitats. American Naturalist, 42, pp.286-288. Post Glacial Fossil Mollusca from a Delta Deposit near Winnipeg, Manitoba. Geological Magazine, 65, pp.267-270.

New Records of Western Canadian Mollusca. Nautilus, 42, pp.13-18. 104 Note on Some Fresh Water Mollusca inhabiting Temporary Ponds in Western Canada. Nautilus, 42, pp.19-20. 1930 Further Records of Western Canadian Mollusca. Nautilus, 43, pp.79-85. Reports of the Jasper Park Lakes Investigations 1925-26. The Mollusca of Jasper Park. Transactions Royal Society Edinburgh, 56, pp. 647-669. 1931 Fresh Water Molluscs in Some Jasper Park Lakes. Canadian Alpine Journal, 1931, pp.148-158. 1932 A Biological Study of a Temporary Pond in Western Canada. American Naturalist,, 66, pp.235- 249. A New Interglacial Pulmonate Mollusc from the Province of Saskatchewan. American Midland Naturalist, 13, pp.236-240. 2 .

Notes on Western Canadian Mollusca - Planorbis campanulatus wisconsinensis Winslow. Nautilus, 46, pp.59-62. 1933 The Local and Geographical Distribution of Some Rocky Mountain Mollusca. Proceedings Malacological Society London, 20, pp.214-221. A New Variety of Lymnaea from the Rocky Mountain Region of Canada. Proceedings Malacological Society London, 20, pp.241-243. 1934 The Discovery of Aearithinula Say in Central Siberia. Nature, 133, Molluscs from the Provinces of Saskatchewan and Alberta, Canada. Proceedings L'Lalacological Society London, 21, pp.138-145. x New Fresh-water Molluscs from Northern Asia. Smithsonian Misc. Collections, 92, No.2. Post Glacial Fossil Molluscs from Western Canada. Geological Magazine, 71, pp.370-382. x 1935 The Fresh-water and Terrestrial Moilusca of Northern Asia. Trans. Royal Society, Edinburgh, 58, pp.605-695. The Variation of two Species of Lymnaea. Genecics, 20, pp.452-465. 1936 The Statistical Analysis of the Distribution of Pond Molluscs in Western Canada. American Naturalist, 70, pp.237-244. 1937 The Ponds, Lakes and Streams of the Kirghiz Steppe. Scottish Geographical Magazine, 53, PP-1-10. The Status of the Greenland Lymnaeae. Proceedings Malacological Society London, 22, pp.186-188. The Terrestrial Mollusca of Sub-Arctic Canada. Proceedings Malacological Society London, 22 2 pp.38-382. A Biological Study of the Sub-Arctic Mollusca. Proc. American Phil. Soc. 78, pp.147-189. Frozen Ground in the Sub-Arctic Region and its Biological Significance. Scottish Geographical Magazine, 53, pp.266-270. 3l./ 3.

1938 The Fresh-Water i1o11usca of Sub-Arctic Canada. Canadian Jour. Research, 16, pp.93-138. The Sub-Arctic Region as a Molluscan Habitat. 1ature, 142, pp.1116-1117. 1939 The Quill Lakes Basin, Saskatchewan, Canada, and its Molluscan Fauna. Internat. Rev. Hydrol. u. Hydrobiol., 38, pp.243-249. The Fresh-water Mollusca of the Tanganyika Territory and Zanzibar Protectorate, and their Relation to Human Schistosomjasjs. Trans. Royal Society Edinburgh, 59, pp. 687-744. The Variation of Lyinnaea stagnalis (Linné). Proceedings Lialacological Society London, 23, itg-q,

X Note: Nos. 21 and 23 only were submitted as Thesis for the degree of Ph.D. - TRANSACTIONS

OF THE

ROYAL SOCIETY OF EDINBURGH.

VOL. LIX—PART III—(No. 26). 1938-1939.

IHE FRESH-WATER MOLLUSCA OF THE TANGANYIKA TERRITORY AND ZANZIBAR PROTECTORATE, AND THEIR , RELATION TO HUMAN SCHISTOSOMIASIS.

ALAN MOZLEY, PH.D.(EDIN.), WANDSWORTH SCHOLAR OF THE LONDON SCHOOL OF HYGIENE AND TROPICAId MEDICINE.

[WITH THREE PLATES AND TEN TEXT-FIGURES.]

EDINBURGH: PUBLISHED BY ROBERT GRANT & SON, LTD., 126 PRINCES STREET, AND WILLIAMS AND NORGATE, LTD., 36 GREAT RUSSELL STREET, LONDON, W.C. 1. AND PRINTED BY NEILL & CO., LTD., 212 CAUSEWAYSIDE, EDINBURGH, 9.

MCMXXXIX.

Price Eight Shillings and Sixpence. ( 687 )

XXVI.—The Fresh-water Mollusca of the Tanganyika Territory and Zanzibar Protectorate, and their Relation to Human Schistosomiasis. By Alan Mozley, Ph.D.(Edin.), Wandsworth Scholar of the London School of Hygiene and Tropical Medicine. (With Three Plates and Ten Text-figures.)

(MS. received December 2, 1938. Revised MS. received March 29, 1939. Read March 6, 1939. Issued separately July 6, 1939.)

CONTENTS.

PAGE PAGE

I. INTRODUCTION . . . . . 687 V. MEANS 01? CONTROL . . . . . 717 II. CENTRAL AFRICAN MOLLUSCA IN RELATION TO Drainage; Filling-in; Flushing; Chemical

HUMAN ScmsTosolalAslS . . . 687 Treatment; Removal of Vegetation; Pro- III. SYSTEMATIC ACCOUNT OF THE MOLLUSCAN FAUNA 689 motion of the Growth of Vegetation; Enemies IV. THE DISTRIBUTION OF FRESH-WATER MOLLUSCA of Mollusca; Education.

IN TANGANYIRA AND ZANZIBAR . . 706 VI. REPRESENTATIVE MOLLUSCAN HABITATS . . 727 i. Geographical Conditions which Affect VII. SUMMARY AND CONCLUSIONS . . . 741 Molluscan Distribution . . . 706 ii. Local Conditions which Affect Molluscan VIII. REFERENCES TO LITERATURE . . 743

. . Distribution 707 ix. DESCRIPTION OF PLATES . . 743 iii. Types of Aquatic Environment . 714

I. INTRODUCTION. During a part of the years 1937 and 1938 a biological study of the fresh-water mollusca of the Tanganyika Territory and Zanzibar Protectorate was carried out with the aid of the grant of the Wandsworth Scholarship of the London School of Hygiene and Tropical Medicine. Up to the present little has been known about the molluscs of this part of Africa, although the importance of certain members of this group as the intermediate hosts of human blood flukes (Schistosoma spp.) has been widely recognized. The following account of the constitution and distribution of the molluscan fauna of this region may therefore be of interest and value. The author is indebted to Dr R. R. Scott and to Dr W. L. Webb, Directors of Medical Services in Tanganyika and Zanzibar respectively, and to their subordinates for their active co-operation in the course of this investigation. The interest, taken in the work by Mr C. E. Bobbins, General Manager of the Tanganyika Railways, and by Dr C. Gilman, formerly Chief Engineer of those railways, and now Water Consultant to the Tanganyika Government, has also been of great aid. Major M. Connolly has given valuable assistance in the preparation of the systematic section of this paper. The author is also greatly indebted to Professor B. T. Leiper, F.R.S., for his constant interest in this work, and to Mr W. A. McDonald for valuable suggestions and aid in many ways. A grant of £50 towards the cost of publication has been given by The London School of Hygiene and Tropical Medicine.

H. CENTRAL AFRICAN MOLLUSCA IN RELATION TO HUMAN SCHIsTosoMIAsIs. At least two species of blood flukes parasitic in man are known to occur commonly in Central Africa. Schistosoma hcematobium is found in Zanzibar, Pemba, Tanganyika, and the adjoining territories. Schistosoma mansoni is widely distributed on the African mainland but is not native to Zanzibar and Pemba Islands. Roughly speaking, from 10 to 70 per cent. of the native population of infected districts harbour these parasites. The Director of Medical TRANS. ROY, SOC, EDIN., VOL. LIX, PART III, 1938-3 (NO. 26). 104 688 DR ALAN MOZLEY ON FRESH-WATER MOLLTJSOA OF TANGANYIKA TERRITORY

Services of Zanzibar in his Annual Report for the year 1935 (p. 16) made the following statement regarding schistosomiasis, " The disease treatment figures for this disease, 692 males and 147 females, give no idea of the extreme prevalence of it among the children of Pemba. In many of the villages in low-lying valleys, or in the neighbourhood of streams and swamps in Pemba, passage of a little bloody urine towards the end of micturition is universal among the boys, and those boys without obvious blood in the urine generally show albumen or even eggs on investigation." Further, in the same report (p. 39), the condition in certain localities is described more precisely. "It will be observed that the percentage of pupils infested with this parasite (S. hcematobium) is extremely high (81-25 per cent. in Nwachani and 7749 per cent. in Kengeja schools). The percentages given in these tables for this helminth must, as in the case of the intestinal parasites, be taken as the absolute minimum, because re-examination of the urines of these children which are found to be negative on the first examination may prove positive on subsequent examination." For the most part human schistosome infections in this part of Africa are light as compared with those found in Egypt. The medical importance of these light infections has been emphasized by Blackie (1932). At the present time it is not possible to evaluate the economic importance of human schistosomiasis in this sector of Africa, but it is doubtless very considerable. The local intermediate host of Schistosoma mansoni is the snail, Biomphalaria pfeifferi Krs. Immature specimens of this schistosome were obtained from the liver of a monkey (Gercopithecus sp.) which had been experimentally infected with cercaria obtained from "wild" Biomphalaria pfeifferi collected near Mwanza, T.T. The infection took place on 25th and 26th October 1937, and the monkey died on 17th November 1937, after which it was dissected and the young schistosomes found. The monkey used in this experiment showed no signs of previous infection with bilharzia, and had been kept under helmintho- logically clean conditions for approximately three months before the beginning of the experiment. The local intermediate host of Schistosoma hcematobium is probably the snail, Physopsis globosa (Morelet). Owing to the small number of cercarie available at any one time, direct experimental proof of this was not obtained in the course of this investigation. The occurrence of urinary schistosomiasis in areas in which Physopsis globosa is common, and the rarity of this disease in areas in which Physopsis does not occur, point strongly to this snail as the local intermediate host. Blackie (loc. cit.) working in Southern Rhodesia found that Physopsis globosa was the intermediate host of S. hcernatobium, and Biomphalaria pfeifferi that of S. mansoni. Gordon, Davey, and Peaston (1934) obtained the same results in Sierra Leone. The only other likely snail hosts of human schistosomes in this region are Bulinusforskalii (Ehrb.), and Bulinus tropicus (Krs.). Specimens of Bulinus forskalii were examined from time to time in the course of this investigation, using a method similar to that described by Blackie (loc. cit., p. 29), but no cercarie of the human type were found. Bulinus tropicus is not a common species in this part of Africa. During the expedition of 1936-37 this species was collected in three localities, near Dodoma, T.T., near Sanya, T.T., and in the vicinity of Nar, T.T. B. tropicus does not appear to occur in the Zanzibar Protectorate. It is conceivable that this species may be of importance from the standpoint of schistosomiasis, but the matter requires further investigation. The seasonal distribution of possible human infection may be a matter of some importance. Oercarie which appeared to be of the human type were obtained from "wild" specimens of AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 689

Physopsis globosa in Zanzibar during the months of June and July—that is, duiing the period immediately following the great rains (April and May). From July to November Physopsis becomes increasingly rare in the Zanzibar Protectorate. After the "little rains" (November or December in those parts of the country where they occur at all) Physops'is is again found in large numbers. The examination of approximately 20,000 specimens of Physopsis globosa from the Muyuni Pond, Zanzibar Island, during the month of January 1938, did not bring to light any evidence of the infection of snails with schistosomes. While this suggests that in this locality Fro. 1.- Fio. 2.- there may be only one season of bilharzial infection each year, the A sinistral shell A dextral shell. evidence cannot be regarded as being conclusive. During February and March living specimens of P. globosa are usually difficult to find in the regions near the sea coast, but may be common in some localities in the interior. The occurrence of B'ioniphalaria pfeifferi in the area studied is too scattered, and the number of observations is too small to permit of any conclusions being drawn regarding its seasonal distribution.

III. SYSTEMATIC ACCOUNT OF THE MOLLUSCAN FAUNA. The most important scientific papers which contain records of i 6 fresh-water mollusca from the Tanganyika Territory and Zanzibar 1?io. 3.—The parts of a Gastro- Protectorate are those of E. von Martens (1897), Pilsbry and pod Shell: 1. spire; 2. line -of growth; 3. superior junction; Bequaert (1927), and Haas (1936). In the present account only 4. outer periphery; 5. aperture; those species are included which have been examined by the author 6. parietal wall; 7. umbilicus; personally. The descriptions given under each species refer to the 8, body whorl; 9. suture; 10. apex. East African shells in hand. The measurements of the shell dimensions of gastropods were made in the following manner:— Length of shell—the distance from the apex to the lower edge of the lip of the aperture. Greater diameter—the greatest width of the shell as a whole, measured along a line at right angles to the long axis of the shell and in the same plane as the aperture. Aperture length—the length of the orifice along a line parallel to the axis of the shell. Aperture width—the width of the orifice along a line at right angles to the axis.

Class. GASTROPODA. Family LYMN2EID2E. Genus Lymncea Lamarck, 1799. Lyrnna3a (Radix) caillaudi (Bourguignat). P1. I, A, figs. 13, 14. L'irnnwa caillaudi Bourguignat, Ann. Sci. Nat. Zool., xv, 1883, p. 89. Lirnnaa knyganica Bourguignat, Moll. Afriq. Equat., 1889, p. 158. Limna3a zanzibarica Bourguignat, Moll. Afriq. Equat., 1889, p. 158. Description.—Shell of moderate size, length up to 23 mm., greater diameter up to 13 mm.; elongate-ovate; thin, fragile; surface bright, with a slight tendency towards malleation, with numerous coarse lines of growth, and very slightly impressed spiral lines; whorls about 5, 690 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY rapidly increasing in size, convex, the last showing some tendency towards oblique flattening in some specimens, regularly rounded in others; spire less than one-third of the length of the shell, pointed, in some specimens somewhat resembling that of L. pereger, in others tending more towards that of L. auricularia; suture distinct, moderately impressed; aperture ovate, regularly rounded, with a very slight shoulder at the superior junction in some specimens, outer periphery rounded or slightly flattened; columella gently curving; umbilical chink varying from broad to narrow. This species may be recognized by its size and shape, fragile corneus shell, and dextral aperture. Lymnaa caillaudi is not known to be of any medical importance. Geographical Range.—Africa, Egypt, Abyssinia, Tanganyika, Zanzibar, Southern Rhodesia. Local Distribution.—Tanganyika Territory: Stream crossing Ngare-Nairobi road approximately one mile from Arusha-Dodoma road. Singida, marsh along shore of Lake Singida. Singida-Puma, in pools in torrential streams at Km. 102 and (approx.) 105 Singida-Manyoni railway. Babati, small brook 1 mile north-east of Babati road junction. Large marsh on flood plain of Little Ruaha River at Mapogoro, 8 miles south of Iringa on the Tosamaganga road. Side channel of the Kimala (?) River near the crossing of the Dodoma-Mbeya road, 57 miles north of Mbeya. Small stream 1 mile east of Kimala River. Stagnant pools in bed of stream crossing Dodoma-Mbeya road 75 miles north of Mbeya. Rungwe District, pond near Lake Nyasa in the vicinity of Myaya, also in lagoons on the shore of that lake. Mwanza, two small streams which flow through the town, pond in valley bottom near railway "tunnel." Mwanza-Mantare, pond in bed of intermittent stream near Mwanza-Mantare road. Shinyanga- Tabora road, 22 miles south of Shinyanga, stream. Tinde, rice fields and streams. Nzega, stream. Marsh at Km. 10409 Central Railway. Marsh at Km. 10352 Central Railway, on Nymphcea. Kigoma-Luiche, pond on flood plain of Luiche River near Km. 12376 Central Railway. Tijiji, marshy lagoon behind sandy beach of Lake Tanganyika. Zanzibar: Pond near Mwera bridge, Chwaka road. Remarks.—Pilsbry and Bequaert (1927, p. 108) refer to Lymnca undussumce as, "undoubtedly a form of natalensis Krauss." Later in the same paper (p. 114) those authors refer to L. undussumce as being doubtfully distinct from L. caillaudi (Bgt.).

Family PLANORBID. Genus Planorbis Muller, 1774. Planorbis (Gyraulus) gibbonsi Nelson. P1. I, A, fig. 15. Planorbis gibbonsi Nelson, Quart. Journ. Conch., i, 1878, p. 379. Description.—Shell of small size, greater diameter up to 32 mm., height up to .9 mm.; disc-like, biconcave; relatively thick; surface bright, glossy, crossed by numerous lines of growth which are coarse in some specimens and fine in others, minute spiral impressed lines visible .on the early whorls; whorls about 3k in the largest specimens, gradually increasing in size, sharply rounded above, obliquely flattened, not angulated; sutures deeply impressed; aperture ovate, width less than one-third of the total greater diameter of the shell; very gently rounded above and below, sharply curved on the distal side in the median line, slightly reduced by the preceding whorl. This species may be recognized by its small size. It is easily distinguished from immature individuals of the larger species belonging to this genus by the greater number of whorls, and their height (maximum height of mature P. gibbonsi is 9 mm.). AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 691

Planorbis gibbonsi is not known to be of any medical importance. Geographical Range.—Africa, Tanganyika, Pemba, Zanzibar, Belgian Congo (Lake Edward), south to Cape Province. Local Distribution.—Tanganyika Territory: Tanga-Pongwe, marshy pond near Km. 66 Tanga Railway. Pond north of Arusha-Dodoma road, 31 miles south-west of Arusha. Mbugwe, ponds and native wells near village. Stream crossing Arusha-Dodoma road 3 miles north of Babati. Babati, Bbati Lake. Stagnant pools in bed of stream crossing Dodoma-Mbeya road 75 miles north of Mbeya. Small stream-1 mile east of Kimala (?) River near Dodoma- Mbeya road about 57 miles north of Mbeya. Malongwe (Km. 730 Central Railway) ponds on flood plain of Mhvala River. Mwanza, ponds near aerodrome. Water-filled furrows in shamba at Km. 3655 Tabora-Mwanza railway. Ngwakulwa (5 miles south of Shinyanga), pond. Tinde, rice fields. Itetemia (3 miles south of Tabora on the Lupa road), pond immediately north-east of the Boma-town road junction. Pemba: Wani, stream. Rice field near first bridge on the Kengeja-Tanga road beyond the Kengeja-Mtembile road.

Genus Biomphalaria Preston, 1910. Biomphal aria pfeifferi (Krauss). P1. I, A, figs. 20, 21. Planorbis pfeifferi Krauss, Suclafrik. Molt., 1848, p. 83. Description.—Shell of moderate size, greater diameter up to 121 mm., height up to 4•7 mm.; biconcave, the concavity approximately equal on the two sides; thick; surface crossed by numerous fine lines of growth, fine impressed. spiral lines also present; whorls 4 to 4 21-, tumid, somewhat rapidly increasing in size, the last appearing to increase proportion- ately more rapidly than those preceding, regularly rounded; sutures impressed; aperture obliquely ovate to sub-circular, very slightly angulated near the superior junction. This species may be recognized by its size, general form, and the biconcavity of the shell. Planorbis pfeifferi is known to be the intermediate host of Schistosoma mansoni in Tanganyika. Geographical Range.—Africa, Tanganyika, Rhodesia, Portuguese East Africa, Natal, Transvaal. Local Distribution.—Tanganyika Territory: stream crossing Ngare-Nairobi road approximately one mile from Arusha-Dodoma road, living snails found on 8.VII.37. Sanya, stream crossing Mioshi-Arusha railway at Km. 21-5, living snails found on 23.IX.37. Stream crossing Arusha-Dodoma road 3 miles north of Babati, one large living individual found on 28.IX.37. Kazikazi, pond near railway. Mwanza, small ponds in marsh on the shore of Lake Victoria near Mwanza hospital, living snails found 12.X.37 and 26.11.38. Pond in bed of stream crossing Mwanza-Musoma road at the first (concrete) bridge, about 1 mile from Mwanza railway station, living snails found on 12.X.37. Tinde, living snails found on 28.V.37. Tabora, Nzega Road 65 miles from Tabora (culvert 23) stream, living snails found on 20.V.37. Habitat.—Marshes, ponds, and streams.

Biomphalaria sudanica (v. Martens). P1. I, A, figs. 18, 19. Planorbis sudanicus v. Martens, Malakoz. Blätter, xvii, 1870, p. 35. Description.—Shell of moderate size, greater diameter up to 138 mm., height up to 43 mm.; deeply concave below, only slightly so above; of moderate thickness; surface 692 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY crossed by numerous very fine lines of growth, and by minute spiral impressed lines; whorls about 4, not particularly tumid, regularly increasing in size, the last increasing in proportion to those preceding, regularly rounded, slightly carinated in the median line in one specimen; sutures impressed; aperture ovate, seldom angulated. Tanganyika specimens of this species may be recognized by their size, the very slight concavity of the upper surface of the shells when they are held with the aperture on the right-hand side and facing the observer, and by the form of the last whorl. Biomphalaria sudanica is not known to be of any medical importance. Geographical Range.—Africa. Local Distribution.—Tanganyika Territory: IRungwe District, lagoon on the shore of Lake Nyasa.

Genus Hippeutis J. de Charpentier, 1837. Hippeutis sp. (near junodi Connolly). Description.—Shell of moderate size, greater diameter 42 mm.; subdiscoidal, biconcave; thin; surface bright, glossy, with numerous minute growth lines; whorls about 4, biconvex, rounded; aperture roughly "U" shaped, nearly flat for a short distance near the superior junction, then very gently curving, and later sharply curving, lower lip nearly straight, a small tooth is present on the lower side, of the interior of the body whorl of some specimens but is absent in others. Local Distribution.—Tanganyika Territory: Pond near Km. 344, Tanga Railway (near Moshi). Zanzibar: Muyuni pond. Remarks.—The specimens collected resemble Hippeutis junodi Connolly, but are not identical with that species. The knowledge of the African species of this genus appears to be very scanty, and the series of specimens available for comparison are small. The description of the Tanganyika and Zanzibar form as a new species is therefore undesirable.

Genus Bulinus 0. F. Muller, 1781. Bulinus tropicus (Krauss). P1. I, A, figs. 16, 17. Physa tropica Krauss, Siidafrik. Molt., 1848, p. 84. Description.—Shell of rather small size, length about 9 mm., greater diameter nearly 7 mm.; fabiform; thin; surface with many minute regular lines of growth giving the surface a minutely costulate appearance; whorls about 3, rounded except at the superior junction, where each is sharply but minutely angulate, rapidly increasing in size; spire slightly elevated; sutures very deeply impressed; aperture ovate, periphery continuous, gently rounded except at the superior junction which is sharply rounded; umbilicus broad and deep. This species may be recognized by its size, shape, and umbilicus. Bulinus tropicus may be a species of medical importance in the Tanganyika Territory. Geographical Range.—Africa, Tanganyika, south to the Cape Province. Local Distribution.—Tanganyika Territory: Sanya, stream crossing Moslii-Arusha railway at Km. 21-5. Nar, small stream near the Nar-Singida road, 2 miles south of Nar (near Mt. Hanang). Dodoma, pond 2 miles north-west of Dodoma station (rare). Habitat.—Streams and ponds. AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 693

Bulinus (Pyrgophysa) forskalii (Ehrenberg). Isidora forskalii Ehrenberg, Synib. Phys. Evert., 1831. Description. —S hell small, length commonly about 10 mm. or less, greater diameter about 3•5 mm. (but specimens may be found in which the length is up to 15 mm.); elongate, sub- cylindrical; thin; surface minutely costulate; whorls about 5, obliquely convex; spire very much elevated, more than two-thirds of the length of the shell as a whole, turreted; sutures deeply impressed; aperture ellipsoidal, angulated above, outer periphery gently curving; umbilicus minutely open. This species may be recognized by its size, shape, and turreted spire. Whether or not Bulinus forskalii is of any medical importance has not yet been satis- factorily determined. Geographical Range.—Africa, Egypt south to the Union of South Africa, Mauritius. Local Distribution.—Tanganyika Territory: Tanga-Pongwe, marshy pond near Km. 66 Tanga Railway; small stream draining pond at Km. 66 above. Kisiwani, same district, pond 2 miles north of village. Moshi, pond near Km. 34314 Tanga Railway. Pond near Mbulu-Ngorongoro road, approx. 10 km. north of Mbulu. Mbugwe, ponds and native wells near village. Small stream near Arusha-Dodoma road 40 miles north of Dodoma. Pond near Arusha-Dodoma road 12 miles north of Babati. Dar-es-Salaam, Bagomoyo Pool. Dodoma, pond near "The Tribal Farm" north of Dodoma. Malongwe-Tura, native well near

Km. 721 Central Railway. Kazikazi, native well near station. Tabora, ditch mile2 north- east of railway station. Kategiri, near Tabora, rice field. Itetemia (3 miles south of Tabora on the Lupa road), pond immediately north-east of the Boma-town road junction. Pond near junction of Kazema and Kermernatnite (?) roads. Tabora, Nzega road, ponds situated at the following distances from the point at which the road crosses the railway to the west of the station, Mile 17, 20, 27, 60. Mwanza, pond in marsh near lake shore, rice fields near lake shore, ponds near aerodrome. Mwanza, found in 3 out of 41 ponds near Musoma Road, between 4 and 6 miles from Mwanza. Mantare, found in 9 out of 60 ponds near Mantare. Malampaka, found in 21 out of 40 ponds near Malampaka. Ngwakulwa (5 miles south of Shinyanga), pond. Tinde, rice fields and ponds. Nzega, ponds in valley west of village. Nguruka (Km. 1060 Central Railway) well near station. Wells near Km. 10731 and 10738 Central Railway. Pond near Km. 1080 Central Railway. Kigoma-Luiche, pond on ,flood plain of Luiche River near Km. 12376 Central Railway. Ujiji, marshy lagoon behind sandy beach of Lake Tanganyika. Pemba: Mkanjuni (3 miles north of Chake Chake), streams. Ole, ponds. Zanzibar: Mtojuu pond near Bambi. Pond near Mwera bridge, Chwaka road. Fumba road, ponds near Mile 65 and 75. Habitat.—Marshes, ponds and small streams.

Genus Physopsis Krauss, 1848. Physopsis globosa (Morelet). P1. I, A, figs. 1-12. Physa globosa Morelet, Journ. Conchyl., xiv, 1866, p. 162. Description.—Shell of moderate size, length of mature shells from 10 mm. to 17 mm. (greater diameter from 8 mm. to 12 mm.), thin, light to dark horn-coloured, surface dull or bright, with numerous minute lines of growth; whorls nearly 5, convex, rapidly increasing in size, shouldered in some series of specimens, regularly rounded in others, tumid; spire varying 694 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY from acuminate to very nearly flat; sutures deeply impressed; aperture sinistral, elongate- ovate, outer periphery gently rounded shouldered in some specimens, superior margin angulate; umbilicus a narrow chink. - This species varies widely in different localities. It is readily distinguished from all other members of the East African fauna by its size, shape, the thinness of the horn-coloured shell, and the sinistral aperture. Physopsis globosa is known to be the intermediate host of Schistosoma hce'inatobium in Southern Rhodesia and Sierra Leone. Geographical Range.—Africa, Angola, Sierra Leone, Tanganyika, Zanzibar. Local Distribution.—Tanganyika Territory: Tanga, ponds near Pangani Road. Tanga- Pongwe, pond near Km. 8, Tanga Railway, living snails found on 16.IX.37. Pongwe, pond 100 miles north of Km. 16, Tanga Railway, living snails found on 17.IX.37. IPongwe, rice field at Km. 1615 Tanga Railway (empty shells). Tengeni, in quiet places in small stream crossing Tanga Railway at Km. 431, living snails found on 17.IX.37. Km. 48 Tanga Railway (empty shells). Korogwe, pond 200 miles south-west of railway station, living snails found on 17.IX.37, also in ponds in large marsh near Pangani River 2 km. west of station, living snails found on pond near Bagomoyo road 7 miles south of Bagomoyo (empty shells). Stream near Bagomoyo road 7 miles south of Bagomoyo (empty shells). Kidondoni, near Dar-es-Salaam, ponds and rice fields, living snails found on 23.VITI.37. Pond north of Arusha-Dodoma road, 31 miles south-west of Arusha, living snails collected 26.IX.37. Mbugwe, ponds near village, living snails found on 28.IX.37. Mbuga on plateau near Nar (empty shells). Singida, marshy lake 1 mile south-west of Singida railway station, living snails found on 20.VII.37. Singida-Puma, pond 5 km. south of summit on Singida-Manyoni railway, living snails found on 21.VII.37. Dodoma, ponds near golf-course, living snails found on 7.111.37. Iringa, abandoned, meanders on the flood plain of Little Ruaha River below Iringa, living snails found on 1.X.37. Large marsh on flood plain of Little Ruaha River at Mapogoro 8 miles south of Iringa on the Tosamaganga road, living snails found on 7.X.37. Stream crossing East Mufindi road approximately 10 miles from Dodoma-Mbeya road, living snails found on 2.X.37. Rungwe District, pond near Lake Nyasa in the vicinity of Myaya living snails found on 3.X.37. Kazikazi (Km. 663 Central Railway), native well near station from which railway employees draw their water supply, living snails found on 10.X.37. Pond near Km. 4625, Central Railway, pond dry on 4.IX.37. Nzinge (Km. 469 Central Railway), Nzinge mbuga. Bahi (Km. 516 Central Railway) sluggish channel leading into Bahi mbuga 3 miles south of Bahi, living snails found on 7.VIII.37. Pond from which water is drawn for railway locomotives about one kiloffietre west of Tura station (Km. 706 Central Railway) empty shells, also in two ponds in valley below pond (empty shells). Ma!- ongwe (Km. 730, Central Railway), ponds on flood plain of Mhvala River, living snails found on 6.IX.37. Mwanza, living snails found in 7 out of 41 ponds examined near Musoma road between 4 and 6 miles from Mwanza on 7/9.111.38. Mwanza, ponds near aerodrome, living snails found 29.11.37. Fela (Km. 355 Tabora-Mwanza railway) large native well from which the inhabitants of the village of Fela draw their water supply, living snails found on 12.X.37. Mantare, living snails found in 33 out of 60 ponds near Mantare examined on 10/17.111.38. Malampaka, living snails found in 9 out of 40 ponds near Malampaka, examined on 18/23.111.38. Shinyanga, pond in side channel of Muhumbo (?) River, living snails found 15.X.37. Ngwa- kuiwa (5 miles south of Shinyanga), pond, living snails found on 15.X.37. Rice fields near Tinde, living snails found 28.V.37. Ndomo, near Nzega, bathing pool of Ndomo school, living snails found on 23.X.37 and 29.X.37, no snails found in this pool on 23.V.37. Shinyanga- AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 695

Tabora road, pond 23 miles south of Shinyanga. Ponds in stream north of Nzega, living snails found on 23.V.37. Itetemia, near Tabora, living snails found on 21.V.37. Tabora, IRufita, pond (empty shells). Tabora, ponds near Mile 6, Nzega road, living snails found on 19. 17.37. Tabora, ponds near Mile 65, Nzega road, living snails found on 20. 17.37. Marsh at Km. 10409 Central Railway, living snails found on 18.X.37. Kigoma-Luiche, stream near Km. 12399 Central Railway. Pemba: Pond in valley south-south-east of Tangani. Ole, ponds, living snails found on 24.1/1.37 and 24.1/111.37. Manyudi near Mikanjuni, living snails found on 23.17111.37. Kidimni, east of Chake Chake, living snails found on 22.17111.37. Pond at Umwonde village, east of Chonga, living snails found on 22.17111.37. Pond near Kidimni village, east of Chonga, living snails found on 23.17111.37. Pond near Old Misherafu, eastern side of Pemba, living snails found on 23.17111.37. Pond near Mischerefu, living snails found on 23.17111.37. Pond near Manyundu, living snails found on 23.1/111.37. Kipangani, pond, living snails found on 24.1/111.37. Urnwondi, east of Chonga, living snails found on 21.17111.37. Mizingani, living snails found on 14.1/111.37. 2 miles east of Mzwajani, living snails found on 21.17111.37. Middle part of stream south-east of Ngezi, living snails found on 20.17111.37. Ngezi, living snails found on 13.1/111.37. Chonga, living snails found from 24.1/1.37 to 11.1/111.37. Stream between Mizingani and Bagomoyo, living snails found on 16.17111.37. Ponds in stream below Ingezi-Ndambani, living snails found on 15.1/111.37. Pond near Veleni, living snails found on 12.1/111.37. Pond in quiet water in stream west of Ngwachani, living snails found on 12.17111.37. Pond in valley south-south-east of Tangani, living snails found on 25.III.37 and 21.yIII.37. Small stream with gentle current below Tangani pond, living snails found on 21.17111.37. Pond in upper part of valley of the stream south of mile 15 Kengeja-Mkoani road, living snails found on 19.1/111.37. Rice field near first bridge on the Kengeja-Tanga road beyond the Kengeja-Mtembile road, empty shells found on 18.1/111.37. Kengeja-Tanga road, pond near culvert north of Pamgogo, living snails found on 24.1/1.37 and 18.17111.37. Two streams between Kengeja and Mtomkuu (empty shells). Mtomkuu. Streams near Kengeja' school, living snails found on 17.1/111.37. Mfuuni east of Kengeja, •living snails found on 17.1/111.37. Kengeja, upper part of stream near Kengeja school, living snails found on 14.1/111.37. Pond in stream 4 miles north-west of Mkoani, living snails found on 25.1/1.37. Zanzibar: Mbiji pond, living snails found from 10.1/1.37 to 6.1/111.37. Mtojuu pond near Bambi, living snails found 17.1/1.37. Pond at Mile 925 Chwaka road, living snails found on 17.1/1.37. Pond near Mwera bridge, Chwaka road, living snails found on 15.1/11.37. Tomondo, living snails found on 2.1/1.38. Kombeni-Ziwani, pond near Mile 7 Fumba road, living snails found from 7.1/1.37 to 7.XI.37. Mile 3 Macunduchi road, ponds, living snails found from 16.1/1.37 to 5.17111.37 and from 28.17.38 to 21.1/11.38. Rice field near Mile 325 Macunduchi road. Large rice field below anti-malarial drain near Mile 3 1 Macunduchi road, living snails found on 9.VII.37 Chegu, pond, living snails found on 1.1/111.37. Pond on path leading from Ndajani to Chegu, living snails found on 8.1/1.37. Ponds near Ndajani, living snails found on 25.1711.37. Muyuuni pond, living snails found from 25.IV.37 to 1.IX.37 and from 20.17.38 to 19.1/11.38. Mtende, living snails found on 6.1/1.37. Variation.—Physopsis globosa varies widely in different localities. In order to ascertain whether or not the local races are sufficiently distinct to merit their description as subspecies, several series of shells, including the paratypes of this species,- were measured, the ratios of certain dimensions were calculated, and the frequency distribution of certain shell ratios were plotted, using the method adopted in other papers (Mozley, 1935, 1937). The results are given in the accompanying tables, together with a representative series of measurements, and indicate that there is a considerable degree of overlapping in the range of variation of all the TRANS. ROY. SOC. EDIN., VOL. LIX, PART III, 1938-39 (NO. 26). 105

696 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

series measured. It therefore appears to be reas'onable to conclude that although local races of P. globosa do exist, the description of these forms as subspecies is not justified. List of the Series of Shells Measured. Number Mean Series. Locality, of Shells Length Measured, of Shell. Pond near Tangani, Pemba. 100 15'5 mm. Mbiji Pond, Zanzibar. 100 121 mm. Pond on flood plain of Mhvala River, 25 11-8 mm. near Malongwe, T.T. P. Angola (paratypes). 10 146 mm. B.M. (Nat. list.), No. 76.6.8.57. Physopsis globosa. Greater Aperture Aperture Length. Locality. Diameter. Length. Width. mm. mm. mm. mm. 17'8 116 11'9 6'1 Tangani, Pemba. 16'5 11'4 11'5 5'8 154 10'5 10'4 4.9 15'0 11'1 11'3 5'0 14'4 10'1 10'7 5'3 Malongwe, T.T.

13'6 9'8 10'2 5'2 11 13'6 9.4 9•6 46 Mbiji, Zanzibar.

13'6 9.3 9'6 4.5 . Tangani. 13'5 9'1 9'3 3'8 Mbiji. 131 10'0 9'8 5'0 Malongwe. 13'2 8'9 9'6 42 Mbiji. 12'3 8'2 84 38 Tangani. 122 88 9'0 4'1 Mbiji. 12'1 8'3 8'5 3'7 117 8'7 8'8 4.3 Malongwe. 11'7 8'3 8'0 3.9 Mbiji. 11'4 81 7'6 3'9 Malongwe. 11'4 7'8 7'7 3.5 Mbiji. Physopsis globosa. Frequency Distribution of Ratios of Shell Dimensions. Length Greater Diameter' 1'25 1'30 1'35 1'40 1•45 1'50 1'55 160 1'65 to to to to to to to to to 1'29 1'34 1'39 1'44 1'49 1'54 1'59 1'64 1•69

Tangani, Pemba. 1 20 29 31 11 7 0 n=100.

Mbiji, Zanzibar. 4 12 23 25 26 9 n=100.

Malongwe, T.T. 3 8 8 4 2 n = 25.

Angola (paratypes) 3 4 2 %=1O.

AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 697

Length Aperture Length' 120 1•25 130 135 140 145 150 1-55 1•60 to to to to to to to to to 1•24 1•29 134 139 1•44 1•49 154 159 164

Tangani. 7 22 32 28 8 3

Mbiji. 8 6 18 26 25 12 3 1 1

Malongwe. 1 0 3 6 1 12 2

Angola (paratypes) 1 5 2 1 1

III Aperture Length Aperture WidthS •; . L- C' C C r7 09 op 171 099 ,-4 — — — — — 4 , C11 Cq G9 aq Cq C11 C c C11 o o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4.4 4 4. I + 4 4 4 4. .+ .- I- c' 0 0 iO 0 iO Q LO 0 0 10 0 10 0 10 0 I () 0 O 00 C 19 CP 19 74 71C C? C 10 C4 01 01 01 01

Tangani. 1 1 4 1 2 5 14 20 16 16 10 3 3 3 0 1

Mbiji. 1641210839891374501

Malongwe. 1 1 0 1 5 5 5 3 3 1

Angola 1 1 1 1 2 1 1 1 1 (Para- types). Habitat.—T he usual habitat of Physopsis globosa is in ponds and marshes. In order to obtain more precise information regarding the type of habitat most favoured by Physopsis globosa, and at the same time to gain some idea of the proportion of shallow water habitats which are potentially dangerous from th6 standpoint of bilharzia, 141 rice fields, ponds, water holes, etc., were examined. These were situated between Shinyanga and Mwanza, T.T.—that is to say, in the region lying to the south of Lake Victoria. All of the observations which are included in the table given below were made during the month of March 1938. Local Distribution of Physopsis globosa. Series A. Habitats near the Musoma Road, east of Mwanza, T.T. Habitats near Mantare, T.T. Habitats north of Malampaka, T.T. Habitats south and west of Malampaka, T.T.

Explanation.—The relative frequency of occurrence of P. globosa is expressed in the form of a fraction. For example, the value given in the table under Series B for the rice-field habitat, namely fi, means that P. globosa was found in 22 of the 31 rice fields examined near Mantare.

698 DR ALAN MOZLEY ON FRESH-WATER MOLLTJSCA OF TANGANYIKA TERRITORY

A. B. C. D. Series. Mwanza. Mantare. Malampaka. Malampaka. Total. n=41. n=60. n=20. n=20. n=141.

Types of Habitat: 1. 0 1 24 Rice fields. 2 3T 2 1 Ponds in the bed of streams. 4 5 x 4 Ponds situated at some distance from 1 3 o 5 well-defined channels of drainage. 8 TT 43 1 1 0 2 Shallow native wells ("Water holes"). T7 6 1 Small reservoirs constructed by x x natives. ("Dams.") * * 10 Small seepage marshes. X x ft ft Borrow pits. >( x x ft Water-filled furrows in cultivated

plots. 11: x x x ft Railway ditches. X x x * Small streams. ft x x x ft

7 38 0 9 4Q_ Total.oa. i11 Note.—" x "= no habitats of this kind examined. The conclusions which might reasonably be drawn from these observations are as follows: In the region lying to the south of Lake Victoria (between Shinyanga and Mwanza, T.T.) approximately one pool in three is inhabited by Physopsis globosa. Note.— The term "pool" is used in the general sense as including pools fed by springs, rice fields, borrow pits, etc. The habitats which are most commonly inhabited by P. globosa are rice fields. The next most important habitats are ponds in the bed of streams. Native wells do not as a rule appear to be important habitats of P. globosa (but see exceptional instance near Fela, T.T., below). The water supply of the village of Fela, T.T., is drawn during the dry season from a water hole situated a short distance to the north of the railway station. This pool was found to contain Physopsis globosa on several occasions between October 1937 and March 1938. In the course of the work in this area the impression was gained that between Shinyanga and Mwanza (a distance of 182 km. on the railway, and somewhat less in a direct line) there is on the average one habitat of P. globosa per 2 sq. km . It is important to note, however, that by no means all these ponds are likely to be the site of actual bilharzial infection as many of them are infrequently visited by human beings. It is worthy of note that in the other parts of the Territory examined in the course of this investigation Physopsis globosa is as a rule much less common than in the region lying between Shinyanga and Mwanza.

Family ANCYLID2E. Genus Gundlachia Pfeiffer, 1849. Gundlachia burnupi Walker. Gundlachia burnupi Walker, 0cc. Prs. Mus. Zool. Univ. Mich., No. 175, 1926, p. 1. Description.—Shell small, length 29 mm., width 16 mm., height FO mm. in one specimen (1.2 mm., width FO mm., height 06 mm. in another specimen); calyptriform, laterally compressed; surface dull, crossed by numerous coarse lines of growth, and by fine lines radiating from the apex; apex blunt, broadly rounded, situated posterior to the central point, and to the right of the anterior-posterior median line; anterior slope of the apex slightly convex AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 699 when viewed from the side, posterior slope nearly flat in the upper two-thirds, lower one-third of the slope slightly concave. This species may be recognized by its size, cap-like form, radial lines in the apex, and the lateral compression of the shell as a whole. The members of the genus Gtndlachia are not known to be of any medical importance. Geographical Range.—Africa. The limits of the range of this species in Africa are not known to the author. Local Distribution.—Tanganyika Territory: Itetemia (3 miles south of Tabora on the Lupa road), pond immediately north-east of the Boma-town road junction.

Genus Burnupia Walker, 1912. Burnupia kempi (Preston). Text-figs. 4, 5.

Ancylus kempi Preston, Proc. Zool. Soc. Lonci., ii, 1912, p. 190. Description. —S hell of moderate size, length up to 48 mm., width 36 mm., height 1-7 mm., calyptriform, ovate when viewed from above or below, only slightly compressed laterally; surface dull, with moderately fine growth lines and having a pitted appearance on the apex; apex blunt, broadly rounded, slightly overhanging, situated approximately one-third of the length of the shell as a whole from the posterior end and to the right of the anterior-posterior median line, anterior slope regularly convex, posterior slope in large specimens convex in the upper third, very slightly concave in the lower two-thirds. Fm. 4.—Burnupia kempi Walker. Dorsal view x 6. This species may be recognized by its size, roundly cap-like form, and the pitted appearance of the apex in uneroded specimens. The members of the genus Burnupia are not known to be of any /7Z~~ FIG. 5.—Burnupia kempi medical importance. Walker. Lateral view x 6. Geographical Range.—Africa, Uganda, Tanganyika. Local Distribution. —Tanganyika Territory: Large brook 6 miles north of Tukuyu, Rungwe District. Habitat.—On stones in a swiftly running brook.

Family NERITID. Genus Neritina Lamarck, 1816. Neritina knorri Récluz. P1. I, B, figs. 3, 4. Neritina knorri Récluz, Rev. Zool. Soc. Guy., 1841, p. 274. Description.—Shell of moderate size, length up to 176 mm. or slightly larger (greater diameter 21-8 mm., lesser diameter 11.4); semi-fabiform; thick; surface bright, dark brown coloured or black, with numerous small lines of growth, and minute spiral lines; the last whorl rapidly increasing in size and concealing those preceding; aperture operculate, dextral, semi- lunar, outer lip broadly rounded, inner lip very broad, nearly flat, serrate on the inner margin, dark blue with tints of orange, interior of aperture blue, with a band of bright orange-scarlet near the outer lip; imperforate. 700 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

This species may be recognized by its general form, semi-lunar aperture, and blue-coloured interior of the aperture which has a band of orange-scarlet. Neritina knorri is not known to be of any medical importance. Geographical Range.—Africa: the East Coast, Pemba, Zanzibar, Madagascar, Portuguese East Africa. Local Distribution.—Pemba: Stream mid-way between Fufuni and Mtemani. Sipwese, two streams between Sipwese and Nanguji. Habitat.—Streams. Neritina natalensis Reeve. Neritina natalensis Reeve, Conch. Icon., 1855, p1. 16, fig. 75. Description.—Shell of moderate size, length up to 19 mm. (greater diameter 194 mm.), ovoid; thick; surface bright, dark brown or green in colour with small yellow spots, with small lines of growth, and minute spiral lines; whorls about 3, rapidly increasing in size, the last not completely concealing those preceding; spire somewhat elevated; aperture operculate, dextral, sub-ovate, outer lip regularly rounded in the distal part, nearly straight near the point of junction, inner lip broad, white in colour, minutely serrate, interior of aperture white in colour or slightly bluish; imperforate. This species may be recognized by its general form, somewhat elevated spire, and the colour of the inner lip and interior of the aperture. Neritina natalensis is not known to be of any medical importance. Geographical Range.—Africa: Pemba, and Tanganyika south to Natal. Local Distribution.—Pemba: Mtokuu, near Kengeja. Habitat.—Streams.

Family AMPULLARIID. Genus Pila Roding, 1798. Pila gradata (Smith). P1. I, B, fig. 10. Ampullaria gradata, Proc. Zool. Soc. Lond., 1881, p. 289. Description.—Shell of large size, length up to 647 mm. (greater diameter 608 mm.); suborbicular; thick; surface bright, light to dark brown in colour, with numerous fine lines of growth, and minute impressed spiral lines; whorls about 4, rapidly increasing in size, tumid, shouldered at the superior margin; spire short, broadly pointed; sutures deeply impressed; aperture operculate, dextral, shouldered at the superior junction, interior of aperture bluish in colour with white margin, umbilicus broadly open. This species may be recognized by its size, shape, and dextral aperture. Pila gradata is not known to be of any medical importance. Geographical Range.—Africa: the East coast, inland as far as Lake Nyasa. Local Distribution.—Tanganyika Territory: Mkomazi, Tanga Railway. Hungila. Pond 3 miles south of Bagomoyo and 1 mile east of Dar-es-Salaam—Bagomoyo road. Stream 7 miles south of Bagomoyo. Tabora, rice field near Nzega road, 6 miles from Tabora. Mvhala River at Malongwe. Marsh in lagoon behind sandy beach of Lake Tanganyika at Ujiji. Zanzibar: Mwera River, rice field near Mile 325 Macunduchi road. Mbiji. Habitat.—Streams, large ponds, and marshes. AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 701

Genus Lanistes de Montfort, 1810. Lanistes olivaceus (Sowerby). P1. I, B, figs. 5, 9. Paludinct olivacea Sowerby, Gen. Shells, pt. 41, 1834, pl. clxxxiii, f. 3. Description.—Shell of large size, length up to 695 mm. (greater diameter 442 mm.); subconical; surface bright, light brown to olive green in colour, with numerous fine growth lines, and also in some specimens minute slightly impressed spiral lines present on upper whorls; whorls about 6 Y1 , convex, gradually increasing in size, the last increasing relatively more rapidly than those preceding; spire acuminate; sutures moderately impressed; aperture operculate, sinistral, auriform, outer periphery gently rounded, angulated at the superior junction, umbilicus varying from a minute slit to a width of 1-5 mm. This species may be recognized by its size, pointed spire, and sinistral aperture. Lanistes olivaceous is not known to be of any medical importance. Geographical Range.—Africa: Tanganyika, Pemba, Zanzibar, Portuguese East Africa. Local Distribution.—Tanganyika Territory: Tanga, pond near Pangani road. Korogwe. Hungila. Pond near Bagomoyo road, 7 miles south of Bagomoyo. Stream 7 miles south of Bagomoyo. Kidondoni, near Dar-es-Salaam, ponds and rice fields. Ruvu. Mbugwe, ponds near village. Dodoma, pond near "The Tribal Farm" north of Dodoma. Dodoma, pond near Bahi road 10 miles west of Dodoma. Singida-Puma, pond 5 km. south of summit on Singida-Manyoni railway. Nzinge (Km. 469 Central Railway), Nzinge mbuga.'Bahl (Km. 516 Central Railway), pond 1 mile east of village; sluggish stream leading into Bahi mbuga 3 miles south-west of Bahl. Chaya (Km. 676, Central Railway), Lake Chaya. Pond from which water is drawn for railway locomotives, about one kilometre west of Tura station (Km. 706 Central Railway). Malongwe (Km. 730 Central Railway), ponds on flood plain of Mhvale River. Malongwe, stream from south-west joining Mhvala River immediately above Malongwe railway bridge, also in pond above mbuga near headwaters of that stream. Note. —L. olivaceus was not found in the mbuga. Nyahua (Km. 766 Central Railway), large mbuga 4 km. east of Nyahua station. Mantare, rice field. Usinge (Km. 1029 Central Railway), railway ditch, also in ponds. Pond at Km. 10394 Central Railway. Lake from which the railway draws its water supply for locomotives between Luichi and Kazuramimba. Kigoma- Luiche, stream near Km. 12399 Central Railway. Pemba: Mkanjuni (3 miles north of Chake Chake), streams and rice field. East of Mile 2 Wete road. Stream north-east of Mkanyageni. Jone (Ole area), in large stream. Ole, eleven ponds in the vicinity of the village. Chonga. Wani. Pond near Mischerefu. Mtomkuu. Kengeja, swimming pool in stream . mile east of Kengeja, also in stream between that pool and Sipwese. Sipwese, two streams between Sipwese and Nanguji. Kengeja-Tanga road, pond near culvert north of Pamgogo. Shitwi stream near Kengeja-Tanga road crossing. Large stream north of Ng'engingi. One stream between Kengeja and Mtomkuu. Oka, stream. Small stream in valley south-east of Tangani. Pond near Kiwaa. Rice field near first bridge on the Kengeja-Tanga road beyond the Kengeja-Mtembile road, also in stream at above bridge. Zanzibar: Mbiji. Kinyasini. Tomondo. Small stream and rice fields near Mile 675 Chwaka road. Pond near Mwera bridge, Chwaka road. Mwera River. Rice fields and ditches near Mile 2 Fumba road. Ponds near Mile 3 Macunduchi road. Large rice field below anti-malarial drain near Mile 3 Macunduchi road. Rice field near Mile 325 Macunduchi road. Muyuni, pond. Mtende, ponds. Habitat.—Ponds and streams. 702 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

Lanistes ovum Peters. Lani.stes ovum Peters, in Frosehel, Arc/iiv. Naturgesch., xi, 1, 1845, p. 215. Description.—Shell of large size, length in specimens collected up to 466 mm. (greater diameter 452 mm.), broadly conical, thick; surface bright, light to dark brown or green in colour, with very numerous lines of growth and almost imperceptible spiral lines; whorls about 41 (juvenile specimen), convex, tumid, rapidly but regularly increasing in size; spire short, broadly conical, blunt; suture moderately impressed; aperture operculate, sinistral, broadly auriform or subovate; umbilicus broadly open. This species may be recognized by its size, the proportion of length to greater diameter and broadly conical spire. Lanistes ovum is not known to be of any medical importance. Geographical Range.—Africa: Tanganyika, Belgian Congo, south to Transvaal. Local Distribution.—Tanganyika Territory: Malongwe, ponds on flood plain of the Mhvala River. Rungwe District, lagoon on the shore of Lake Nyasa. Habitat.—Ponds. Lanistes farleri Craven.

Lanistesfarleri Craven, Proc. Zool. Soc. Lond., 1880, p 219. Description.—Shell of moderate size, length 226 mm. (greater diameter 194 mm.), subovate to somewhat elongate; moderately thick; surface dark horn-coloured, with minute lines of growth and numerous coarse impressed spiral lines which are readily visible to the naked eye; whorls 3, convex, rapidly increasing in size; spire elevated; broadly conical; suture very deeply impressed; aperture, operculate, sinistral, auriform, outer periphery regularly rounded up to six dark brown spiral bands plainly visible in the interior of some specimens, absent in others; umbilicus minute. This species may be recognized by its general form, minute growth lines and the coarse spiral lines on the whorls. Lanistes farleri is not known to be of any medical importance. Geographical Range.—Africa: Tanganyika. Local Distribution.—Kiwanda, near Muhesa (Km. 40, Tanga Railway), Tanganyika Territory. Habitat.—Small streams. Family VIVIPARID . Genus Viviparus de Montfort. Viviparus constrictus (E. von Martens). Palvdina constricta E. von Martens, Conchol. Mitth., iii, 1, 1886, p. 16. Description.—Shell of moderate size, length 21-6 mm. (greater diameter 143 mm.), subconical; light greenish brown in colour, surface with numerous coarse lines of growth; whorls nearly 5, with from two to four well-marked carime which disappear on the body whorl in some specimens; spire acuminate; suture very deeply impressed; aperture operculate, dextral, subcircular, angulate at the superior junction; umbilicus narrow. This species may be recognized by its carinated whorls, dextral aperture, and narrow umbilicus. TZiviparus constrictus is not known to be of any medical importance. Geographical Range.—Africa: Lake Victoria. Local Distribution.—Lake Victoria near Mwanza, T.T. Habitat.—On lake bottom near shore. AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 703

Viviparus un'icolor (Olivier). Cyclostoma unicolor Olivier, Voyage, Emp. Othoman, ii, 1804, p. 39. Description.—Shell of moderate size, length 205 mm. (greater diameter 157 mm.); subconical; green in colour; surface with numerous minute growth lines and also spiral lines; whorls about 6, smooth, slightly shouldered; spire broadly conical; suture impressed; aperture operculate, dextral, subovate to subcircular, broadly angulate at the superior junction; umbilicus broad. This species may be recognized by its broadly conical spire, smooth whorls, dextral aperture, and broad umbilicus. Viviparus unicolor is not known to be of any medical importance. Geographical Range.—Africa: Egypt, south to Tanganyika (and beyond ?). Local Distribution.—Tanganyika Territory: Mbugwe. Habitat.—Ponds. Family THIARID. Genus Thiara Boding, 1798. Thiara vouamica (Bgt.). P1. I, B, figs. 6, 7, 8. Tiara crenularis Desh., var. vouamica Bgt.; Moll. Afr. Equat., 1889, p. 183. Description.—Shell of moderate size, length up to 307 mm. (greater diameter 156 mm), in some localities specimens about 15 mm. in length are common; elongated, turriculate; thick, dark olive-green in colour; surface with numerous minute lines of growth and also impressed spiral lines, some of which are visible to the naked eye; whorls about 7, angulated, concave above, convex below, the shoulder with numerous sharp spines; spire acuminate, turreted; sutures deeply impressed; aperture dextral operculate, elongate, periphery sharply angulated at the superior junction; imperforate. This species may be recognized by its turriculate spire, the spines on the whorls and dextral aperture. Thiara vouamica is not known to be of any medical importance. Geographical Range.—Africa: Somaliland to Natal. Local Distribution.—Pemba: Chonga, stream. Oka, stream. Wani, stream. Kengeja, swimming pool in stream .- mile east of Kengeja; also in four streams between Kengeja and Sipwese. Shitwi stream, near Kengeja-Tanga road crossing. Zanzibar: Mwera River. Habitat.—Streams. Genus. Cleopatra Troschel, 1857. Cleopatra ferruginea (Lea). P1. I, B, figs. 1, 2. Melania ferruginea Lea, Proc. Zool. Soc. Lond., 1850, p. 182. Description.—Shell of moderate size, length about 26 mm. (greater diameter 12 mm.); thick, surface bright, dark brown or chocolate in colour in mature specimens, horn-coloured with three dark brown spiral bands in young specimens, with minute lines of growth and numerous very small impressed spiral lines; whorls about 7, convex, somewhat more tumid on the lower side than the upper, gradually increasing in size, the increase of the last whorl in slightly greater proportion than those preceding; spire acuminate;, sutures deeply impressed; aperture dextral, operculate, subovate, periphery continuous, angulated at the superior junction, gently rounded elsewhere; umbilicus a narrow slit. TRANS. ROY. SOC. EDIN., VOL. LIX, PART III, 1938-39 (NO. 26). 106 704 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

This species may be recognized by its general form, smooth whorls, and dextral aperture. Cleopatra ferruginea is not known to be of any medical importance. Geographical Range.—Africa: Tanganyika, Zanzibar, south to Cape Province. Local Distribution. —Tanganyika Territory: Bweembwela, stream crossing Tanga Railway at Km. 485. Stream -1 mile west of Mbiji village. Korogwe, Pangani River. Mkomazi. Pond 3 miles south of Bagomoyo and 1 mile east of Dar- es- Salaam-Bagomoyo road. Hungila. Mantare, ponds. Malampaka, ponds. Zanzibar: Mbiji, pond, also in stream 1 mile west of Mbiji village. Rice field near Mile 325 Macunduchi road. Habitat.—Ponds and streams.

Genus Melanoides Olivier, 1804. Melanoides tuberculata (0. F. Muller). P1. I, A, figs. 22, 23. Nerita tubereulata 0. F. Muller, Verm. Terr. Fluv. Hist., ii, 1774, p. 191. Description.—Shell of moderate size, length up to 328 mm. (greater diameter 120 mm.); elongate; moderately thick; horn-coloured or dark brown; surface rough, with numerous coarse longitudinal and spiral lines; whorls about 9 slightly convex, regularly increasing in size; spire elongated, more than half of the length of the shell, sutures impressed; aperture operculate, dextral, auriform, gently rounded except at the superior junction which is angulate; imperforate. This species may be recognized by its elongated spire, rough surface, and dextral aperture. Melanoides tuberculata is not known to be of any medical importance. Geographical Range.—North and East Africa, Madagascar, Asia Minor, southern Asia, Java. Local Distribution.—Tanganyika Territory: Kiwanda; near Bagomoyo. Pemba: Ole. East of Mile 2 Wete road. Upper waters of Chandegi River near Wani. Mtomkuu, near Kengeja. 1 miles north of Mkoani. Habitat. —Streams. Class PELECYPODA.

Family 5PHERIID iE. Genus Sphcerium Scopoli, 1777. Sphcerium naivashcense Preston. Sphcsrium naivasha3nse Preston, Rev. Zool. Afric., i, 1912, p. 328. Description.—Shell of very large size for African members of this genus, length 126 mm., height 103 mm., breadth 62 mm.; ovate; nearly equilateral; thin; umbo rather small in proportion to the size of the shell; outer surface light brown in colour; growth lines very small, numerous minute radial lines present; cardinal and lateral teeth moderately well developed; interior of shell horn coloured or bluish-white. This species may be recognized by its large thin shell, relatively small umbo, and small growth lines. Sphcerium naivashcense is not known to be of any medical importance. Geographical Range.—Africa: Tanganyika, Lake Naivasha. Local Distribution.—Tanganyika Territory. Malongwe, pond in side channel of Mhvala River. Remarks.—Haas (1936) regards S. naivashcense as a subspecies of S. ha'rtmanni (Jickeli). AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 705

Sphcerium nyanzce Smith. Sphwrium nyanzw Smith, Ann. Mag. Nat. Hist., ser. 6, x, 1892, p. 383. Description. —Shell of moderate size for African members of this genus, length 103 mm., height 86 mm., breadth 6•1 mm.; suborbicular, non-equilateral; thick umbo very large in proportion to the size of the shell as a whole; outer surface light to dark brown, with numerous coarse growth lines, cardinal teeth well developed; lateral teeth strongly developed; interior of shell bluish grey in colour. This species may be recognized by its suborbicular and non-equilateral form, relatively large umbo, and coarse growth lines. Sphcerium nyanzce is not known to be of any medical importance. Geographical Range.—Africa: Lake Victoria, Lake Edward, and Lake Lohondo in Ruanda. Local Distribution.—Tanganyika Territory: Lake Victoria, near Mwanza. Remarks.—Haas (1936) regards S. nyanzce as a subspecies of S. capense (Krauss).

Family CYRENID. Genus Corbicula, von Mühlfeld, 1811. Corbicula africana (Krauss). Cyrena africana Krauss, Sudafric. Molt., 1848, p. 8. Description.—Shell of moderate or rather large size, length of largest specimen collected 128 mm., height 11-0 mm., breadth 74 mm.; subtriangular; equilateral; umbo of large size in proportion to that of the shell as a whole; outer surface light brown in colour, growth lines very coarse; cardinal teeth strongly developed; lateral teeth serrate; interior of shell brownish pink with one or more dark brown bands radiating from the umbo. This species may be recognized by its thick, subtriangular shell, and coarse markings. The members of the family Cyrenithe are not known to be of any medical importance. Geographical Range.—Africa: Central Africa, south to the Cape Province. Local Distribution.—Tanganyika Territory: Lake Victoria, near Mwanza.

Family UNIONID. Genus Ccelatura Conrad, 1853. Ccelatura ruellani (Bourguignat). P1. I, A, fig. 24. Unio ruellani Bourguignat, Molt. Eluv. Nyanza Oukréwé, 1883, p. 10. Description.—Shell of small size for members of this family, length 376 mm., subovate; outer surface of valves brown in colour, inner surface white or bluish with shades of pale pink at the posterior end of some specimens; umbo situated about one-third of the length of the shell from the anterior end; hinge line slightly curving; upper periphery of shell forming a straight line in a downward direction at each end in most specimens; distal periphery gently curving; beak sculpture consisting of "W" shaped ridges, approximately in the median line, which disappear towards the distal periphery of the valve, and of series of wrinkles anterior and posterior to the umbo. This species may be recognized by its small size, subovate form, and sculpture. The members of the family Unionithe are not known to be of any medical importance. Geographical Range.—Africa : Lake Victora. Local Distribution.—Tanganyika Territory: Lake Victoria, near Mwanza. 706 DR ALAN MOZLEY ON FRESH-WATER MOLLTJSOA OF TANGANYIKA TERRITORY

Family MUTELID]E. Genus Aspatharia Bourguignat, 1885. Aspatharia wahibergi (Krauss). Indina wahibergi Krauss, ,Sudafr. Moll., 1848, p. 19. Description. —Shell of large size, length up to 115 mm., elongate, slightly swollen (breadth 431 mm. in a specimen 1554 mm. in length); outer surface light brown in colour, interior of valves very pale pinkish white in colour; hinge line gently curving, umbo at about the middle of the anterior half of the shell; "V" shaped notch at the posterior end of the hinge on the interior of the valves. This species may be recognized by its general form, very slightly pinkish shell on the inner side, gently curving hinge line, and "V" shaped notch. The members of the family Mutelithe are not known to be of any medical importance. Geographical Range.—Africa: Nigeria, Tanganyika, south to Natal and Transvaal. Local Distribution.—Tanganyika Territory. Malongwe. Mhvala River. Habitat.—Streams.

IV. THE DISTRIBUTION OF FRESH-WATER MOLLUSCA IN TANGANYIKA AND ZANZIBAR. i. Geographical Conditions which affect Molluscan Distribution. A study of the natural history of the mollusca in East Africa reveals the fact that geographical conditions play a very important part in governing the distribution of these animals. A grasp of the geographical setting is therefore essential to an understanding of the local problems of snail distribution, and hence eventually, an understanding of the problems of snail control. A brief review of such conditions in Tanganyika and the Zanzibar Protectorate is given below. As a matter of convenience, this geographical account is divided into four sections, which deal respectively with earth structure, climate, vegetation, and soil. It must surely be obvious to the reader that the conditions described are not to be regarded as isolated individual "environmental factors," but that acting through long periods of time they influence living organisms collectively.

EARTH STRUCTURE. Apart from marine invasions and recessions in Zanzibar, Pemba, and the coastal belt of the mainland, the earth's surface in this part of the world, taken as a whole, has been less disturbed than that in the northern part of the northern hemisphere since Palaeozoic or early Mesozoic times. That is to say, there has been no widespread glaciation. The ancient peneplain has been profoundly modified, however, by differential crustal movements and intense differential weathering by wind, rain, and sun. Consequently there has been a complicated series of local changes which, although of a strikingly different nature from those of the Sub- Arctic (Mozley, 1934), exert an equally powerful influence upon the flora and fauna of the region. CLIMATE. The climate of the area under consideration varies between wide limits from year to year, and within a single year from one season to another. In general it is characterized by an intense radiation, a relatively high temperature, and an alternation of drought and plentiful moisture. These conditions are of importance as far as fresh-water mollusca are concerned, in limiting the number of species which are able to exist within the area, and also in governing AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 707 the occurrence and relative abundance of these forms within restricted localities. The effects of climate upon the mollusca are manifested chiefly through the agency of high evaporation during the dry seasons, and of great floods during the wet seasons. It is difficult to assess the relative biological importance of (a) changes in climate in times past; (b) climatic differences over broad geographical areas; and (c) microclimate. It is by no means certain that (b) and (c) above are necessarily coincident and reciprocal. In fact it is certain that within (b) the microclimates vary over a wide range. This may be regarded as a clue to the diversity of distribution met with among the molluscs and other living organisms in many districts. It may be well to point out that the conception of microclimate is a new one, in which the collection of observational data is still in its infancy. Never- theless, sufficient information is on hand to establish the validity of this concept beyond serious doubt. VEGETATION. The vegetation of a given region has a considerable effect upon the aquatic fauna. In by far the greater number of instances, however, the relationship exists between certain associations of fresh-water animals and general types of vegetative cover, rather than between individual species or genera of aquatic animals and those of plants. In the inland regions of Tanganyika the vegetation, except in certain restricted areas, has a distinctly xerophytic character. In addition to this, the physical form of the plant associations is frequently such as to promote the rapid drainage or evaporation of surface water. Thus the forests are usually open formations with widely scattered trees and little shrubby undergrowth, while on the grasslands the plants are usually found growing in compact groups which are separated by bare soil. Much of the run off in the interior of Tanganyika is collected in shallow, grass-covered inundation plains, from which evaporation, aggravated by the prevailing winds, is very rapid. These conditions tend to accentuate the prevailing semi-arid condition of much of the country, and thus to restrict the distribution of the mollusca. In Zanzibar, Pemba, and restricted areas on the African mainland, the growth of vegetation is much more luxuriant than in the interior. The rainfall of these regions is much greater than that of the other parts of the territory covered in the course of this work, so that water conservation by the vegetation in general is possibly of relatively little importance in the present connection. However, the presence of a dense growth of grasses and sedges in damp places, such as valley bottoms, is of considerable importance from the standpoint of schistosomiasis, since such situations, as a direct result of the vegetation are frequently the habitat of dangerous snails. (Text-fig. 7.) SOILS. The discussion of geographical conditions in Tanganyika and Zanzibar must not be concluded without some mention of soils. Unfortunately the knowledge on this subject is not sufficient to allow of any effective treatment. Nevertheless, a relationship between soils and aquatic fauna must exist, and will surely be taken into consideration by later workers. An interesting series of problems is presented by the relation of the character of the soils in a given drainage basin to the.biological productivity of the aquatic habitats.

ii. Local Conditions which affect Molluscan Distribution. In considering this topic it is important to bear in mind the fact that the geographical - conditions mentioned in the preceding paragraphs are liable to be considerably modified in 708 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

different districts. For example, although the overhead climate of large tracts of land in this area appears to present no differences which are of importance, there is no doubt that diverse localities within such areas are subject in widely different degrees to the action of flood and drought. It is therefore essential to regard the conditions met with in any one habitat as forming not only a closely knit complex, but an unique complex. As a matter of convenience, in order to secure emphasis and conciseness, these matters are discussed here under diverse headings, such as flood, drought, and chemical conditions.

FLOOD. Several of the fresh-water molluscs which inhabit Zanzibar and Tanganyika are forms which find conditions of life most suitable in situations which are little disturbed by waves or stream current. The species of snails which are of importance from the standpoint of human schistosomiasis (Physopsis globosa and B'iomphalaria pfeifferi) belong to this group. Great floods are a disturbing and often inimical element in the habitats of these animals, since under such conditions, snail food is carried away and the molluscs themselves in many instances are destroyed. During the course of these investigations rainfall up to 100 mm. (3.95 in.) was met with in a twenty-four hour period, and there are partly authenticated reports of greater precipitation. Such heavy rainfall is followed by severe flooding, in which many potential molluscan habitats are destroyed by erosion, and snails are carried downstream. It is not surprising therefore that many of the molluscs of Zanzibar and Tanganyika have a restricted local range, being found only in situations which are not subject to extreme flood conditions.

DROUGHT. The fresh-water animals of this region have not only to endure conditions of excessive moisture, but also, at different seasons in each year, those associated with an extreme deficiency of moisture. In many parts of the Tanganyika Territory very nearly the whole of the annual precipitation occurs during the period from November to April. The months of June, July, August and September are almost invariably rainless, and those of May and October usually so. Periods of drought also occur along the coast and in Zanzibar and Pemba. There are numerous species of molluscs in diverse parts of the world, including the African Physopsis globosa, and Bulinus forskal'ii, which are capable of surviving severe desiccation. On the other hand, flooding is not necessarily inimical to other members of this group of animals. An alternation of flood and drought, however, each on such a grand scale as is met with in East Africa, has a very marked effect, and considerably limits the distribution of many molluscs. Provided that a given species can survive any degree of desiccation, the duration of exposure to drying is a matter of importance. In certain other parts of the world, an annual active period of about two months is sufficient to permit of the maintenance of a fauna of mollusca in a pond. In Zanzibar and Tanganyika, a longer period, probably a period of nearly four months in each year, is required for the survival of Physopsis globosa. In any event it is noteworthy that only a small number of individuals of P. globosa in a pond appear to survive drying. Casual observations, which remain to be confirmed, give the impression that the individuals of Physopsis globosa found in Zanzibar (which island has a relatively moist climate) are somewhat less likely to survive desiccation than are those individuals of the same species which are found in the dry interior of Tanganyika. AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 709

GEOMORPHOLOGICAL CONDITIONS. The degree of exposure of a given molluscan habitat to flood and drought obviously depends to some extent upon geomorphological conditions in the neighbourhood. Observa- tions made in such diverse areas as Pemba Island and the Lake Province of Tanganyika indicate that an elementary knowledge of such conditions may be a considerable aid in the practical work of finding the habitats of dangerous snails. The following examples, which are typical of the conditions met with in much of the interior of Tanganyika, support this view. (a) The Fauna of Ponds in the Lake Province, Tanganyika. The area in the vicinity of Malampaka, T.T., in common with much of the Lake Province, is one of relict topography. To the north of Malampaka village there is a remnant ridge, remnant hills (" inselberge ") lying near it. As a result of this the landscape is hilly, drainage is rapid, ponds on high ground are isolated from one another, and those in the valley bottoms are subject to severe scouring during heavy rain. The result of an examination of 20 ponds in this vicinity is given below. It is worthy of note that Physopsis globosa was not found in any of these ponds.

Snail Fauna of Ponds lying to the North of Malampaka, March 1938. Physopsis Bulinus globosa. forskalii. Spring well, half-way down gentle slope. 0 x Shallow well on high ground. 0 0 Seepage marsh. 0 0 Pond, 10 m. below summit of ridge. 0 0 Pond, 15 m. below summit of ridge. 0 0 Pond on mid-slope (30 m. below summit of ridge). 0 x Pond, near No. 6. 0 x

Pond, near No. 7. . 0 x Pond on high ground. 0 0 Pond on high ground. 0 x Spring pond at foot of remnant hill. 0 x Spring pond at foot of remnant hill. 0 x Spring pond at foot of remnant hill. 0 0 Spring pond at foot of remnant hill. 0 0 Rice field. 0 x Rice field. 0 x Shallow well. 0 x Shallow well. 0 x Pond. 0 x Shallow well near summit of ridge. 0 x x =present. 0 = absent.

The country lying to the south and west of Malampaka is in a more advanced stage of peneplanation, so that its general aspect is more flat or gently rolling, with the remnant hills situated far apart. The drainage is therefore relatively slow, there is a greater possibility of communication between ponds, and much less tendency towards scouring after heavy rain than in the area lying to the north of the village. Twenty ponds in this neighbourhood were examined. Physopsis globosa was collected in nine of them (see next page). These observations may be taken as indicating a definite tendency for Physopsis globosa to occur more commonly in ponds situated in flat country than in ponds in hilly areas. Additional observations in other localities indicate that Physopsis globosa is occasionally to be found in the immediate vicinity of remnant hills, but in such areas this species almost 710 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

Snail Fauna of Ponds lying to the South and West of Malampaka, March 1938. Physopsis Bulinus ylobosa. forskalii. Pond in bed of gently inclined intermittent stream. x 0 Borrow pit. 0 x Pond at foot of remnant hill. x 0 Pond. x 0 Pond. 0 x Rice field. 0 0 Pond. 0 x Pond. x 0 Pond in bottom of broad valley. x 0 Pond in bottom of broad valley. x 0 Pond in bottom of broad valley, but smaller and shallower than No. 10. 0 x Pond on higher ground. 0 x Pond on remnant ridge. 0 0 Pond on remnant ridge. 0 0 Pond. x 0 Rice field half-way down gentle slope. x x Pond on gentle slope slightly above valley bottom. x x Pond. 0 x Seepage marsh below spring. 0 0 Pond. 0 0 x =present. 0 = absent. invariably occurs in ponds situated in valley bottoms which as a result of particular local conditions happen to be very gently inclined in a longitudinal direction. The observations made in Pemba (text-fig. 7) confirm this view regarding the usual habitat of P. globosa. (b) Ponds at the Foot of Escarpments. The East African landscape is dominated by the results of tectonic movements on a grand scale. Uplift, warping and dislocation have profoundly modified the ancient peneplain, and as a result the surface presents a complicated series of earth forms which in certain instances, such as in the neighbourhood of rifts and tilt blocks, are of some importance in the present connection. From the coast of the Indian Ocean the country rises in a series of sharp steps, some of which are of gigantic proportions, to a central plateau. This structure is of considerable significance in connection with the study of the distribution and control of schistosomiasis. The distribution of aquatic habitats in the neighbourhood of escarpments is indicated in a diagrammatic way in text-fig. 6, in which a ground-plan and cross-section through two escarpments are given. The usual, and as far as the observations in the Tanganyika Territory are concerned the sole, habitat of dangerous snails in a locality of this sort is in ponds, small lakes, and stagnant streams at the foot of the escarpments. The minor streams on the surface of the tilt blocks have too small a drainage basin to contain water for a sufficiently long period to harbour aquatic snails, while the larger streams, such as those which have cut a course across the line of faulting, lack the quiet-water conditions which are necessary for the existence of Physopsis globosa and Planorbis pfeifferi in this region.

LIGHT AND FooD. There appears to be a considerable possibility that light and food are interrelated conditions in governing the occurrence of fresh-water molluscs in East Africa. Exposure to sunlight (either direct or diffuse) for some period of time is required in order that any given volume of water should become a nutrient medium for fresh-water molluscs. Therefore, small ponds and water-holes which are fed by springs are in many instances barren AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 711 of molluscs. Larger pools, such as the small stagnant reservoirs ("dams" with a diameter of 10 m. or more), which are constructed by the natives in many parts of Tanganyika and in which a given volume of water remains for a period of months are in many instances

j TOREHTIAL ST Q.EAJA

ESCARPMENT ST.EAM ESCARPMENT MUCA OR LAKE

1 S(Pfi- STEWN

TILT BLOCK

FiG. 6.—Diagrammatic representation of the distribution of surface waters on tilt blocks on the edge of the Central Plateau, Tanganyika. favourable habitats for snails. In such places Bulinus forskalii and Physopsis globosa are the most common species. Likewise, pools situated in dark caverns in the coral rock of Zanzibar Island are not favourable habitats for fresh-water mollusca. Such caverns therefore constitute a safe source of water supply for domestic purposes, or alternatively safe bathing places, as far as the possibility of infection with bilharzia is concerned. It is a common (but apparently not universal) belief among the natives of the island that such caverns are the resort of evil spirits, so that these places are not visited by all of the inhabitants. Plans for tapping the subsurface waters of the eastern part of Zanzibar Island, where caverns are most numerous, are said to be under consideration by the government authorities. The utilization of such waters is to be highly recommended. Pulmonate molluscs are known to feed upon a wide variety of animal and vegetable matter, such as: Mud and possibly human "excreta" (=faces). (Physops'is globosa, Gordon, Davey, and Peaston, 1934.) Water plants, mud, "decaying leaves," and a decaying cow's horn. (Planorbis pfeifferi, Gordon, Davey, and Peaston, loc. cit.) Living sticklebacks. (Lymncea stagnalis, Ullyett, 1886.) Sand, newspaper, a dead fresh-water mussel, a dragonfly and Pears' soap. (Anon ater, Kew, 1889.) TRANS. ROY, SOC. EDIN., VOL. LIX, PART III, 1938-39 (NO. 26). 107 712 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

During the present investigations the following materials have been observed to serve as food for snails Decaying vegetable matter (principally leaves and stems of aquatic and semi- aquatic plants). (Physopsis globosa, Lanistes olivaceus.) ? Algae growing upon stones, etc. (Nearly all species collected in Tanganyika and the Zanzibar Protectorate, including Physop.sis globosa and Biomphalaria pfeifferi.) Decaying leaves of trees. (Physopsis globosa, Planorbis gibbonsi.) Beef suet. (Lanistes olivaceus.) From the above account it will be fairly obvious that these animals are able to make use of a variety of food materials, so that should a certain type of food be absent from a given habitat, or be exhausted, many kinds of snails are able to find an alternative food supply. A definite instance of this with Planorbis pfeifferi is recorded by Gordon, Davey, and Peaston (1934, p. 348). Under the conditions met with in many habitats in Tanganyika and the Zanzibar Protectorate there frequently appears to be a greater quantity of food available (at least at some times of the year) than can be eaten by the number of snails present. In many localities therefore the food supply does not appear to be an important condition in limiting the occurrence of snails. There is one special condition in this connection which does appear to be of some importance in controlling the abundance of snails, namely, the amount of light which penetrates beyond the surface in a still-water habitat. Physopsis globosa appears to be somewhat less abundant in ponds in which the surface of the water is completely covered with the leaves of water-lilies (Nymphcea spp.) than in those in which although there are some water-lilies there is also some open water. Furthermore, P. globosa is found in greater numbers in those parts of a single pond in which there is some open water than in those in which the surface is almost completely covered by the leaves of water-lilies, the fact that the undersides of the leaves are themselves a common habitat of the snails notwithstanding. Several observations made in the Muyuni pond, Zanzibar Island, on 16th June 1937 indicate that P. globosa is from four to nine times as abundant in relatively open shallow water as in similar and adjacent places in which the surface of the water is covered by water-lily leaves. This is in accordance with observations made by the present author in the northern part of the northern hemisphere, fresh-water molluscs in that region being found in only very small numbers in habitats in which the surface of the water was covered with a dense growth of Lemna. The explanation of this condition possibly lies in the relative scarcity of food in shaded habitats due to a less abundant growth of algae in the shade.

CHEMICAL CONDITIONS. Fresh-water mollusca in general are to be found living under a wide variety of chemical conditions. As a rule each species has a fairly wide range of toleration with regard to the chemical conditions usually met with in nature. Other environmental conditions therefore more frequently exert an influence upon the distribution of molluscs. This appears to be true of Physopsis globosa. The presence of mineral salts in excess, however, such as is reported from certain lakes in central Tanganyika by Orr and Grantham (1931), has a distinctly inimical effect upon these molluscs. In the course of the present investigation no fresh-water mollusca have been found in a saline lake. The results of the analysis of natural waters collected in Zanzibar is given below. It is worthy of note that the quantity of inorganic solids present in solution in samples Nos. 7 and 8 is very small (12.0 parts per million in each instance). Since food, in the form of CHEMICAL ANALYSIS OF WATER COLLECTED ON ZANZIBAR ISLAND. Collected by A. M. Analysis by the Government Chemist, Zanzibar. Z Results expressed as parts per million.

Nitrogen in Loss on ______Oxygen ab- Solids in ignition sorbed from Appearance, -I Description solution (organic (a) (b) (a) (d) Potassium Alkalinity Chlorides Molluscan Fauna. No. Date. etc., of Permanganate (as CaCO3). (as NaCl). of sample. dried at matter Free and Albuminoid sample. Nitrites, in 4 hours 105° C. and some Saline Ammonia. CO,). Ammonia. at 800 F.

0 x x Physopsis globosa 14/5/37 Pond at Nuyuni X x 004 020 x x 35 1 Laaistes olivaceus Hippeutis sp. 0 nil 140 400 180 2 8/2/38 ,, Dull, brownish 1630 960 016 0'78 nil tint, moderate sediment H nil 75 280 100 3 12/4/38 ,, Dull, slight 105.0 560 001 052 nil td brownish tint, moderate sediment

Pond at Mbiji 1950 600 009 040 x x 46 x x Physopsis globosa 4 11/5/37 Lartistes olivaceus H P/la sp. Cleopatra ferruginea H 0 450 006 036 nil nil 643 1330 110 5 4/5/38 11 Opalescent, sligh 1750 yellow tint

x 11.1 X x Physopsis globosa 6 14/5/37 Pond at Fumba X 003 050 x Road, Mile 72 Z

None 7 4/5/38 Pond, Mile 8, Slightly dull, 39.0 270 0016 01 nil nil 1-37 120 150 Marie - Louise yellow tint Road H 0 x x None 0 8 7/6/38 Pond, 1 mile Slight flocculent 546 420 x - x x x x south of Police sediment lines, Zanzibar

x =not determined.

I co 714 DR ALAN MOZLEY ON FRESH-WATER MOLLTJSCA OF TANGANYIKA TERRITORY

plant material, was present in moderate abundance in both of these ponds, it is conceivable that the absence of fresh-water molluscs was due to the small quantity of inorganic salts, and possibly those of calcium in particular, which is present in the water. The drainage basins of the ponds from which samples Nos. 7 and 8 were obtained lie in areas of leached soils in which there are no outcrops of coral. Should this explanation be true it may apply to other natural waters in Zanzibar which appear to be barren of mollusca. The fresh-water mollusca of Tanganyika and the Zanzibar Protectorate are extremely sensitive to the presence of certain organic and inorganic poisons. This subject is discussed in Section V. CHEMICAL ANALYSIS OF ZANZIBAR TOWN WATER. ("Zanzibar tap water.") Sample collected May 2, 1938. Analysis by the Government Chemist, Zanzibar. Results expressed as parts per million.

Total solids in solution, dried at 105° C. . . 350

Alkalinity as CaCO 3 . . . . . . 296

Chlorides (as NaCl) . . . . . . 21-

Nitrogen in free and saline ammonia . . . 01 in albuminoid ammonia . . . . . 012

in nitrates . . . . . . . . . 32

in nitrites . . . . . . nil

Oxygen absorbed from potassium permanganate in 4 hours at 80° C. . . 031

TEMPERATURE. The temperatures commonly met with in nature in this part of Africa do not appear to have any marked direct effect upon the distribution of mollusca. The indirect effects of the prevailing relatively high temperature, however, are of some significance. For example, the occurrence of warm weather during the period of minimum rainfall results in many ponds becoming dry, and in this manner the length of the annual period during which it is possible for fresh-water molluscs to be active is greatly reduced. This limits the local distribution of many species. It is interesting to observe that the effect of high temperature in this sector of Africa is thus very much the same as that of extremely low temperatures upon the fauna of ponds in the northern part of the northern hemisphere (Mozley, 1932).

iii. Types of Aquatic Environment. In considering the distribution of aquatic animals in this part of the world, it is worth while, simply as a matter of convenience, to describe the places in which these animals live under four headings, namely, ponds, the African equivalent of the marshy bogs of the northern hemisphere which are known as mabuga, streams, and lakes. It should be understood, however, that no genetic relationship is thereby implied, for although there may be some succession in these habitats its course is not yet known, and on this ancient surface succession must surely be very slow. In any event, it is probable that there is a considerable amount of convergence in the development of aquatic habitats in general—that is to say, similar habitats may arise in different ways. A linear arrangement in the classification of habitats, such as is sometimes to be found in "ecological" papers, would therefore be misleading. AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 715

PONDS. Small and shallow bodies of water are of common occurrence throughout a great part of Tanganyika and the Zanzibar Protectorate. The habitats here designated as "ponds" include pools fed by springs, shallow native wells, rice fields, and borrow pits, as well as undisturbed "natural" pools. Life conditions in habitats of this kind vary between very wide limits, but there is little variety in the molluscan fauna. One reason for the sameness of the fauna may be that conditions of life in habitats of this kind are so severe, and approach so closely the limits of tolerance of most molluscs, that there is. little opportunity for the adaptation of additional species and varieties. The most common species of mollusca found in ponds in this region are Bulinusforskalii, Physopsis globosa, Lani.stes olivaceus. In especially favourable habitats Gundlachia burnupi and Hippeutis sp. may also be found. There are a few ponds in Tanganyika in which B'iomphalaria pfeifferi occurs in great abundance. Each of the three common species noted above is frequently to be found as the sole molluscan inhabitant of a pond, less frequently but not uncommonly any two of them may occur in a single pond; while more rarely all three of these forms may be collected in a single body of water. In one pond near Kazi Kazi, T.T., it was observed that Physopsis globosa (5.IX.37) was succeeded at a later season in the year by Bulinus forskalii (27.11.38). Ponds which form in borrow pits resulting from road construction in many instances become fertile habitats of dangerous snails. One of the first steps towards the control of bilharzia in Tanganyika should be the adoption by the road-making authorities of a broad and shallow type of borrow pit in road construction. The origin of many of the ponds in the interior of Tanganyika is of some interest. In the regions of relict topography the weathering of the granite on the remnant ridges and hills results in sand (from the quartz) and clay (from the feldspar) being carried down to lower levels. The clay is separated from the sand by sheet flooding, is carried on to the peneplain, and is there deposited in small depressions in which it forms an impermeable layer, and thus results in the formation of ponds. MABUGA. The East African equivalent of the bogs, swamps, and (in part) marshes of the northern hemisphere appears to be the mbuga (plural, mabuga). These are shallow, badly drained depressions covered with clay alluvium, from which evaporation is very rapid. Many mabuga are dry for the greater part of the year. They appear to have their origin in the filling-in of a lake or pond, or in the dislocation of the primitive drainage by tectonic movements. (Text-fig. 6.) In general, the molluscan fauna of situations of this kind is a poor one. Many mabuga contain no fresh-water molluscs whatever. If animals of this kind occur at all there is a strong probability of their being dangerous, firstly because Physopsis globosa is the species which is most likely to occur in such places, and secondly because habitats of that snail in or around mabuga are frequently used by natives as bathing places, or as sources of water supply for domestic purposes. STREAMS. The streams of Tanganyika and the Zanzibar Protectorate have a relatively varied and abundant fauna of mollusca. The most common species are the following: Thiara vouamica, Melanoides tuberculata, Lymncea caillaudi. It is only under special circumstances that the streams of this region are dangerous habitats from the standpoint of bilharzia. In this

716 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

respect the local situation appears to differ considerably from that described from other parts of Africa (compare with Gordon, Davey, and Peaston, 1934). Over the greater part of the area covered in the course of the present investigation the alternation of flood and drought, both in extreme form, renders the streams unsuitable habitats for dangerous snails. The following examples appear to cover the range of the exceptions to this general rule: — (a) Pemba. Parts of the valleys of certain streams on the island of Pemba are flat-bottomed, and support a rich growth ofsemi-aquatic vegetation. In these marshes Physopsis globosa is

CENTRAL DRY MEADOW MARSH BROOK. FPESU MANGROVE Rl E _SWAMP Flooded AFkr Standing Water ._ Rapid Current Wakr I (Or Spur) l4eovg gain • All 'kar. Little Standing Marsh Water

I I• ILO I

LONG I TU Dl HAL SECTION–

0. POFlLE5 ALONG I AXES INDICATED j b I I I ii I J EN T P RES E NT I I Except I tp LRcL HUMBE.S I OccasionallyI I A B S I NT d d so P S 5 In Ponds

In Small Numbers FIG. 7.—Diagrammatic longitudinal section of a typical stream on the Island of Pemba, showing the distribution of Physopsis globosa.

common. The upper parts of these streams, apart from a few isolated ponds, are too dry for P. globosa, while in those parts of the streams which lie below the marshes the current is too swift to permit of the survival of P. globosa. A somewhat similar situation has been observed in a meandering section of the course of the Little IRuaha River, a short distance to the south of the town of Iringa, T.T. Physopsis globosa and Biomphalaria pfeifferi are common in the marshes which border this stream. Topographical conditions in this locality are in a general way comparable to those in the Pemba streams mentioned above, and in the natural course of back-cutting by the stream the marshes will disappear. This process, however, will take several thousands of years. (b) Sanya Area, T.T. At Km. 21-5 on the Moshi-Arusha section of the Tanga Railway line (i.e. between Moshi and Sanya stations) there is a small stream in which Bulinus tropicus and Biomphalaria pfeifferi were collected in large numbers in September 1937. As these snails were found crawling upon smooth rock and stones in a narrow channel (P1. II, fig. 3), and there appeared AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 717 to be no marshy area connected with the stream which might serve as a retreat for them, their presence and survival in the stream was difficult to explain. The upper waters of this stream were examined in July 1938, and a large marshy area was discovered in which Bulinus tropicus and Biomphal aria pfeifferi were abundant. This marsh doubtless acts as a reservoir from which the snails migrate into the lower waters. (c) Mwanza, T. T. The stream near Mwanza in which Biomphalaria pfeifferi was collected in large numbers is mentioned in another part of this paper (see p. 741). The particular conditions which render this stream a suitable habitat for B. pfeifferi while other streams in this and the adjoining districts are known to be unsuitable habitats is not understood. It is worthy of note, however, that these molluscs were found in large numbers only in the immediate vicinity of a bridge, so that the disturbance of natural conditions may have been a matter of some importance. It is possible that the significant condition is the protection from scouring afforded by the spillway and apron of the bridge.

LAKES. The lakes of the region covered in this report fall into three classes, namely: shallow, pond-like lakes; large fresh-water lakes; and saline lakes. Shallow pond-like lakes are rare. The relative maturity of the landscape may account for this. Certain pond-like lakes, such as those near Bambi, Zanzibar Island, and Ole, Pemba, are of importance since they contain Physopsis globosa. Many such lakes in Tanganyika contain no molluscs whatever. The large fresh-water lakes which have been examined, namely, Lakes Victoria, Tangan- yika, and Nyasa, have a fauna of mollusca which differs from that of the surrounding country. In many respects each of these lakes, particularly Tanganyika and Nyasa, appears to be a distinct and unique microcosm. No evidence has been found in the course of this work of the occurrence of schistosomes in the molluscs which are peculiar to these lakes. Dangerous snails may occur in the waters of these lakes, but they are uncommon. Only one such locality is known to the author, namely, a protected bay of Lake Victoria near Mwanza, in which a few specimens of Physopsis globosa were collected. However, ponds and marshes along the margins of the lakes mentioned, which are not properly speaking part of the lake, are of great importance from the standpoint of schistosomiasis. (See under Kigoma, Mwanza, and Rungwe in Section VI of this paper.)

V. MEANS OF CONTROL. The most effective means of controlling human schistosomiasis are those which are directed towards the extermination of the intermediate hosts of the parasites. Physopsis globosa is a common snail in Zanzibar and the Tanganyika Territory, while Biomphalaria pfeifferi is not uncommon in parts of the latter region. These speies, however, are by no means the predominant, or in certain senses the best adapted, fresh-water snails in this part of the world. They might be regarded as being interlopers from other geographical regions in which their relatives are found in greater abundance and variety. The task of exterminating these snails is therefore not as difficult as would be the control of certain of the operculate snails which inhabit this region. Various methods of control are discussed in the following paragraphs. On account of the great diversity of conditions met with in different localities, the application of these control measures t6 any one particular habitat must be regarded as 718 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY being experimental, and infallibility cannot be claimed in advance. The general outlook, however, is extremely promising. It appears that a relatively small expenditure upon anti- bilharzial measures in Zanzibar Island will result in the eradication of this disease within four or five years. A somewhat greater expenditure directed towards this end in Pemba will result in a greatly diminished infection rate, and eventually, perhaps, to the disappearance of this disease from the island. On account of the vast area to be covered in Tanganyika, it is probable that a much longer period will be required to attain a degree of control comparable to that mentioned in connection with Zanzibar and Pemba. Nevertheless, there is no doubt that useful work can be carried out at relatively small annual expense, and without increase in the European staff. Provided that a sustained effort is made, it appears to be not unreasonable to look forward to a decrease in the infection rate in the Territory of at least 50 per cent. within a period of, say, 15 years. Drainage. The most satisfactory way of ridding a pond of snails is to drain it. There are relatively few ponds in the region covered in this report which cannot be drained. Other considerations, such as cost, or the absence of an alternative water supply for the use of the native inhabitants, will prevent the adoption of this measure in some areas. It is sometimes possible to achieve the same end, as far as molluscs are concerned, by draining a pool for a period of three or four months, as by destroying the pond altogether. The best time for such intermittent drainage will differ in diverse parts of the country, and must be determined by experiment. In areas where drainage for a long period is not practicable, complete drying for as short a period as one week will probably have the effect of greatly reducing the numbers of snails, and thus materially reducing the possibility of infection. In some pools two such treatments each year might be sufficient. Drainage may also be usefully applied to meandering streams, such as those on the Island of Pemba, which have marshy borders in which undesirable snails are found. The best plan in such instances is to construct a ditch as far as possible in a straight line through the lower part of the marshy course of the stream. Full advantage will thus be taken of the inclination of the valley bottom, and the marshy area will be gradually drained. Side drains leading into the main ditch are undesirable, since they may themselves become the habitat of snails. In certain instances, where ponds lie in shallow depressions in coral rock, from which they are separated by only a thin layer of relatively impermeable material, it may be possible to disturb the pond bottom by blasting in such a way as to open up underground channels of drainage. This appears to be a possibility in the large pond near Bambi, Zanzibar Island. Measures such as this should obviously be preceded by a careful survey of the level of the underground water in the neighbourhood.

Filling-in. The filling-in of ponds is usually so costly as to be impracticable. In some instances, however, town refuse, dead leaves and cocoanut husks are available in quantity and can be put to good use. Rice husks, a waste product of the milling industry, can also be used in this way. In damp places rice husks are likely to cause disagreeable odours, but these can be avoided by the addition of a top layer of soil or sand. It may occasionally be possible to take advantage of the erosive power of streams during the rainy seasons to transport materials downstream to or near the pond to be filled. A useful adjunct to draining or filling-in is the planting of species of Eucalyptus or Acacia AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 719 on the pond site. These plants absorb much moisture from the soil and tend to hinder the formation of ponds. These trees and shrubs are known to thrive in parts of East Africa.

Flushing. There is some evidence that in the neighbourhood of Mount Kilimanjaro (and possibly Mount Meru) small streams whose beds are moderately steeply inclined may be important sites of infection with schistosomiasis. One such stream, which crosses the Tanga Railway at Km. 21•5 (from Moshi), has been found to contain both Bulinus tropicus and Biomphalaria pfeifferi. The general character of this stream and the inclination of its bed appear to indicate that it is suitable for the adoption of flushing as an anti-molluscan measure. Tipper and siphon sluices have recently been designed for use in the control of malaria which are auto- matic in action and inexpensive to construct (Williamson and Scharff, 1936). It is probable that the periodical vigorous flushing action which follows the setting up of one of these pieces of apparatus in a stream would result in the destruction of the snails. The construction of one such sluice in a stream as an experiment is highly to be recommended.

Chemical Treatment. Organic Poisons.—Owing to the prejudice which many natives have against the addition of (to them) unknown substances to their water supply, it is desirable, if possible, to find some local natural product with which to treat infected pools. With this end in view the action of a number of well-known native plants which contain poisonous organic compounds was tested in the laboratory. In this connection it may be mentioned that while there is a considerable number of poisonous plants in East Africa, very few of them, because of the lack of reliable sources of supply, are likely at present to be of much importance in the control of bilharzia. Experiments upon the Effect of Aqueous Infusions of some Native Plants upon Physopsis globosa. Note.—Approximately 30 g. of roughly chopped plant material were placed in 250 c.c. of Zanzibar tap water (for chemical analysis see p. 714), 10 freshly collected snails were added, and the result noted after certain periods of time. Ps'jadia arabica (Swahili-Mkeneta). Form I. Cold infusion of leaves and stems which remained in the vessel throughout the experiment. Result after 21 hours—all snails alive but motionless. Psiadia arabica. Form I. Warm infusion (at approximately 800 C. for five minutes, and then allowed to cool). Plant material remaining in vessel throughout the experiment. Result after 21 hours—all snails alive, nine motionless, one active. Psiadia arabica (Swahili-Mbebeta). Form II. Cold infusion of leaves and stems as above. Result after 21 hours-9 snails alive but motionless, 1 snail dead? Psiadia arabica. Form II. Warm infusion as above. Result after 21 hours—all snails dead. Psiadia arabica. Form III. (Recognized as being distinct from Forms I and II above, possibly no special name in Swahili.) Cold infusion as above. Result after 21 hours—all snails dead. TRANS. ROY. SOC. EDIN., VOL LIX, PART III, 1938-39 (NO. 26). 108 720 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

Psiadia arabica. Form III. Warm infusion as above. Result after 21 hours—all snails dead. Mundulea suberosa (Swahili-Utupa wa miti). Cold infusion as above. Result after 21 hours—all snails alive but motionless. Mundulea suberosa. Warm infusion as above. Result after 21 hours—all snails alive but motionless. Tephros'ia vogel'ii (Swahili-Utupa wa Mrima). Cold infusion as above. Result after 21 hours—all snails alive, 9 motionless, 1 active. Tephrosia vogelli. Warm infusion as above. Result after 21 hours—all snails alive but motionless. Synadenium carinatum (Swahili-Mvunja Kongwa). Cold infusion as above. Result after 21 hours—all snails dead. Note.—All the snails were motionless on the bottom of the vessel and apparently dead within 15 minutes of the commencement of the experiment. Synadenium carinaturn. Warm infusion as above. Result after 21 hours—all snails dead. Note.—All the snails were motionless on the bottom of the vessel and apparently dead within 3 minutes of the commencement of the experiment. Euphorbia sp. (Swahili-Mkweche). Cold infusion as above. Result after 21 hours—all snails dead. Euphorbia sp. Warm infusion as above. Result after 21 hours—all snails dead. Euphorbia tirucalli (Swahili-Mchakazi). Cold infusion as above. Result after 21 hours—all snails alive but motionless. Euphorbia tirucalli. Warm infusion as above. Result after 21 hours—all snails alive but motionless. Sapindus saponaria (Swahili-Halita). Cold infusion of fleshy part of fruit, not including seed. Result after 21 hours—all snails dead. Sapindus saponaria. Warm infusion of fleshy part of fruit as above, not including seed. Result after 21 hours—all snails dead. Sapindus saponaria. Cold infusion of seed (roughly broken). Result after 16 hours—all snails alive, one snail crawling (and feeding ?) upon seeds.

Members of the family Euphorbiace (see 15 and 16 in the above experiments) are common in most parts of the Tanganyika Territory, and also occur in Zanzibar, but their use in natural waters which are likely to be utilized for domestic purposes or for watering cattle is highly undesirable owing to the poisonous properties of this group of plants for vertebrates, including man (Raymond, 1936). The possibility of the successful use of the fruits of Sapindus saponaria, which are commonly known in this territory as "soap berries," is more promising. The active principle in this fruit is a saponin. Saponins are glucosides which have a well-marked toxic action upon cold-blooded animals, but are very much less poisonous to mammals. There is no authenticated case of fatal poisoning by a saponin in man. Up to one gramme per day is said to have been taken orally without any effect. Introduced subcutaneously a possibly fatal dose is 26 g. A summary of experiments upon the action of "saponin" as obtained from AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 721 chemical manufacturers upon Physopsis globosa collected in the Muyuni pond, Zanzibar, is given below. The solutions used in these experiments were freshly prepared in distilled water, and subsequently diluted with Zanzibar tap water.

Experiments on the Effect of Saponin Solutions upon Physopsis globosa. Note.—Ten freshly collected snails were added to solutions of saponin at room. temperature (about 28° C.) and the results noted after certain periods of time (24 hours in one series of experiments, and 40 hours in another series). The stock solution of saponin was prepared by the Government Chemist, Zanzibar. Concentration of saponin in solution. Result. 1. 0025 %. No apparent effect after 40 hours, all snails active. 2. •005 %. All snails motionless after 40 hours. 3. •005 %. 2 snails apparently dead, and 8 snails alive but motionless after 40 hours. 4. .005 %. All snails alive but motionless after 24 hours. 5. 01 %. All snails motionless and apparently dead after 40 hours. 6. 01 %. All snails in torpid condition or dead after 24 hours. 7. 01 %. All snails apparently dead after 24 hours. 8. •015 %. All snails dead after 24 hours. 9. 015 %. All snails dead after 24 hours. 10. 02 %. All snails dead after 24 hours. 11. Control—snails in pure All snails alive and active after 40 hours. Zanzibar tap water. 12. Control. All snails alive after 24 hours. 13. Control. All snails alive after 24 hours. The fruits of Sapindus saponaria are said to contain about 6 per cent. of saponin. Experi- ments were carried out with the use of these berries for the purpose of poisoning snails in a pond situated near Mile 7 on the Fumba road, Zanzibar Island. The berries were placed in the pond on three occasions, namely, 30th August 1937 (100 lb.); 1st September 1937 (100 lb.); and 5th November 1937 (200 lb.). In each instance many dead Physopsis and Lanistes were found in the pond within two days of the treatment, and the impression was gained that the majority of the snails in the pond had been killed. Nevertheless, within a short time thereafter, a careful search revealed the presence of living individuals of both of the snails mentioned above, and these evidently served to repopulate the pond, for within two months snails were again common. A possible explanation of the survival of the snails is that there were scattered individuals in outlying or secluded parts of the pond into which the fruit extract did not penetrate to any extent, in spite of an attempt to distribute it evenly during treatment. That snails can be killed by the use of soap berries may be taken as proven. Larger quantities of soap berries should be used in future experiments. The outlook for the utilization of soap berries is promising, but over the Tanganyika Territory and Zanzibar Protectorate as a whole this method cannot be regarded as being more than a very useful adjunct to other means of control. In areas where Sapindus saponaria is found in numbers the use of the fruits is strongly to be recommended, and will probably be effective. It is possible to purchase soap berries in the bazaar areas of Zanzibar and Dar-es-Salaam. The prevailing price in 1937-38 was 15 cents or a shilling per kilo. when small quantities were purchased, and less for larger quantities. Sapindus saponaria is a common tree in some places, particularly at Tanga, T.T., and possibly at Bagomoyo, T.T. Trees have also been seen at Mwanza, T.T., and Mtende, Zanzibar Island. At least 200 of these trees occur on government property in the town of Tanga, and from these one ton of soap berries were collected in 722 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

1937. In Zanzibar 100 lb. of soap berries are said to have been collected from a single tree in 1937, so that with careful gathering in Tanga the annual crop should probably be sufficient for all local needs. The fruit of the tree Balanites ctgyptiaca which also contains a saponin is said to be effective in killing snails. Although this tree is found in Tanganyika it is not particularly common, and certainly does not occur in sufficiently large numbers to be considered in connection with large-scale measures designed to control schistosomiasis. Another tree which may prove to be useful in this connection is Swartiza madagascariensis. This is known to occur in Tanganyika, as well as in other parts of Africa. The seed pods of this tree are said to contain 224 per cent. of "saponin." The roots of Derris spp., originally a Malayan fish poison, have recently come into commercial use in the manufacture of insecticides. The active principle in the roots of Derris may be rotenone, deguelin, toxicarol or some other substance (Miller, 1935). Powdered Derris root in aqueous mixture has no injurious effect upon mammals. 01 gramme of rotenone either injected into the stomach of rats or administered orally to these animals has no effect. The effect of aqueous mixtures of powdered Derris root and water upon Physopsis globosa was tested in the laboratory with the results shown below.

Experiments on the Effect of the Powdered Root of Derris elliptica upon Physopsis globosa. Note.—Finely powdered Derris root was obtained from the East African Agricultural Research Station, Amani, T.T. This was said to contain 81 per cent. of rotenone on a dry weight basis, and 75 per cent. of rotenone in the powder as actually used. The stock mixture of Derris was prepared by the Government Chemist, Zanzibar, by adding 30 grammes of the powdered root mentioned above to 200 c.c. of distilled water. Theoretically this should have given a solution containing 1425 per cent. of rotenone, but as it is unlikely that there was complete extraction, the mixture was regarded as containing not more than 1 per cent. of rotenone. This mixture was allowed to stand for 24 hours, and the supernatant fluid in the bottle was drawn off and used in the experiments. The further dilution of the stock mixture thus prepared was made with Zanzibar tap water. Twenty freshly collected snails were placed in vessels, the mixture of the strength indicated, and the result observed after a period of 24 hours.

Strength of Mixture. Result. 1 c.c. of stock mixture in 500 c.c. of water (1 part All snails dead. of rotenone in 50,000 parts of water). 25 c.c. of stock mixture in 500 c.c. of water (1 19 snails dead. 1 snail alive, having crawled out part of rotenone in 200,000 parts of water). of the mixture on to the side of the vessel. 20 c.c. of stock mixture in 500 c.c. of water (1 part 18 snails dead. 2 snails alive. of rotenone in 250,000 parts of water). 15 c.c. of stock mixture in 500 c.c. of water (1 part All snails dead. of rotenone in 333,000 parts of water). 5 •10 c.c. of stock mixture in 500 c.c. of water (1 part 19 snails dead. 1 snail alive but motionless. of rotenone in 500,000 parts of water). 05 c.c. of stock mixture in 500 c.c. of water (1 part 16 snails alive but motionless. 1 snail active of rotenone in 1,000,000 parts of water). in solution. 3 snails alive, having crawled out of the mixture on to the side of the vessel. Control—pure Zanzibar tap water. All snails alive. AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOM1ASIS. 723

From the result of these experiments it appears that Derris root may prove to be very useful in the control of P. globosa. The growth of Derris elliptica in East Africa is all that could be desired, and the particular strain of this species, which is available in Tanganyika, is capable of producing roots of high quality. The propagation and cultivation of Derris is not a difficult matter. The optimum age for harvesting the' roots is when the plants are about 24 months old (Garnett, 1936). The successful cultivation of Derris elliptica in small plantations in the immediate vicinity of ponds which are inhabited by Physopsis globosa therefore appears to be a practical possibility, and abeginning should be made without delay. After such small plantations are once established in a given locality they will be practically self-perpetuat- ing (by stem cuttings). Once or twice each year a sufficient number of Derris roots (the quantity necessary will have to be determined by experiment) can be dug up, broken into small pieces and thrown into the pond. There appears to be little doubt that such a method of control would be effective in small ponds. The cost, utilizing local native labour, would be trifling. The poisonous properties of certain other plants found in East Africa, including Tephrosia vogelii and Mundulea suberosa, has been investigated by Worsley (1934 et seq.). The supply of these plants available in East Africa at the present time does not appear to be sufficient to merit their consideration here, quite apart from the fact that they are not particularly effective as snail poisons (pp. 719-720). Inorganic Poisons.—The effect of solutions of certain inorganic salts upon Physopsis globosa was tested experimentally. The salts used were potassium alum (commercial), copper sulphate (chemically pure), and ammonium sulphate (crude, fertilizer). The experiments with potassium alum gave no clear-cut results. The effects of copper sulphate and ammonium sulphate are shown in the tables below.

Experiments on the Effect of Dilute Solutions of Ammonium Sulphate upon Physopsis globosa. Note.—The conditions of these experiments were similar to those with copper sulphate, see above. The stock solutions were prepared by the Government Chemist, Zanzibar.

Chemical and Result. Concentration. After 12 hours. After 24 hours.

1. Ammonium sulphate, 1 part in 1 snail dead. 4 snails motionless. 10,000. 20 snails active.

Ammonium sulphate, 1 part in 7 snails dead. 18 snails 7 snails dead. 3 snails motion- 5000. active. less. 15 snails active.

Ammonium sulphate, 1 part in 17 snails dead. 8 snails 19 snails dead. 6 snails alive. 2500. alive.

Ammonium sulphate, 1 part in All snails dead. 1000.

Control. All snails alive. 1 snail dead. 24 snails alive. 724 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

Experiments on the Effect of Dilute Solutions of Copper Sulphate upon Physopsis globosa. Note.—Twenty-five snails were placed in vessels containing solutions of copper sulphate and the results were noted after certain periods of time. The stock solution of copper sulphate was prepared by the Government Chemist, Zanzibar. The water used for the dilution of the stock solution was from Muyuni pond, Zanzibar Island (for chemical analysis see p. 713).

Concentration of Result. CuSO4 . After 12 hours. After 24 hours.

1 part in 4,000,000. 10 snails active. 15 snails 9 snails alive (having crawled apparently dead. out of solution). 16 snails dead.

4 snails active. 21 snails 4 snails alive (having crawled out apparently dead. of solution). 21 snails dead.

1 part in 2,000,000. All snails motionless and ap- All snails dead. parently dead.

1 part in 1,000,000. All snails motionless and ap- All snails dead. parently dead.

3 parts in 1,000,000. All snails dead within 5 minutes of being placed in solution.

Control—pond water only. All snails alive. All snails alive.

,, 3 snails dead. 22 snails alive 4 snails dead. 21 snails alive.

The results of the experiments with the use of copper sulphate are similar to those of Khalil (1930), who observed the effect of this salt upon Bulinus contortus and Planorbis boissyi. It appears that copper sulphate might well be used in small ponds for the purpose of killing snails. The toxic effect of the solutions of crude ammonium sulphate may be due to the effect of the salt itself, or to certain impurities such as cresols, which are probably present in most samples of this fertilizer. Sisal Waste. The growing of sisal (Agave sisalana) and the preparation of the fibre of this plant for export is an important industry in Tanganyika. There appeared to be a possibility that the waste products of the process of the decortication of the sisal leaves might be of use in the control of undesirable snails. The effect of the waste upon Physopsis globosa was therefore tested in the laboratory. The results of this experiment are given below.

Experiments on the Effect of Dilute Solutions of Sisal Waste upon Physopsis globosa. Note.—Ten freshly collected snails were placed in mixtures of the liquid portion of sisal waste (obtained from a factory near Dar-es-Salaam) with water. The results were noted after certain periods of time. AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 725

Strength of Mixture. Result.

1 part of sisal waste in 5 parts of water, All snails dead after 3 hours.

1 part of sisal waste in 10 parts of water All snails dead after 24 hours.

1 part of sisal waste in 40 parts of water. 5 snails dead after 24 hours. 5 snails alive but motionless.

1 part of sisal waste in 80 parts of water 2 snails dead after 24 hours. 8 snails alive but motionless.

Control—water only. All snails alive and active. The results of these experiments indicate that sisal waste may be of use in killing snails. Owing to the bulk, and hence the high cost of transporting this substance, it is likely that its usefulness will be confined to the immediate neighbourhood of sisal estates. The chemical composition of sisal waste is not known to the author of this paper. Oxalic acid is said to be present, together with other poisonous substances

Enemies of Mollusca. Fresh-water molluscs are preyed upon by a considerable number of animals belonging to widely diverse groups, but little precise information is available regarding the degree of control likely to be exercised by such predators. The outlook for the control of undesirable snails by biological means is promising, but too little is known about this phase of the subject to allow of.more than tentative conclusions being drawn. Certain possibilities are discussed in the following paragraphs. A considerable number of birds in different parts of the world feed upon mollusca, to a greater or less extent. In East Africa conditions are favourable for the use of ducks and geese in the control of undesirable snails. It is not at all unusual to find a few individuals of the Muscovy Duck (Cairina mosehata) roaming about native villages in Zanzibar, Pemba, and Tanganyika. In addition, there may be one or more of the domestic derivatives of the Mallard (Anas platyrhynchos), but these have not been recognized with certainty. These animals eat snails. A wild species believed to be Alopochen cegyptiacus (the Nile Goose) apparently eats fresh-water snails in preference to other food. Ducks from native villages have been observed to eat the local fresh-water snails with avidity. What now remains to be done is to introduce ducks into bilharzial localities in larger numbers. Until this is done further discussion is without point. A common East African fish, Clarias gariepinus (Fam. SiluridEe), (Swahili-Kombali), has been observed on many occasions to eat Physopsis globosa. This fish is able to crush the shells of Physopsis in its jaws with ease. The shells of all but very small immature Lanistes are rejected after unsuccessful attempts to crush them. As a rule, ponds and small lakes which contain Clarias are without Physopsis. In a few instances, however, the fish and snails have been found living in the same pond. Examples of this are to be found at Mbiji, Zanzibar, and one kilometre west of Tura station on the Central Railway, Tanganyika. In these localities it is possible that the presence of an alternative food supply has made it unnecessary for the fishes to feed upon the snails. The possibility of controlling Physopsis by means of the introduction of Clarias deserves thorough investigation. It is possible to collect large numbers of this fish both in Tanganyika and Zanzibar early in the dry season when they are trapped by the drying of temporary ponds and small temporary lakes. Care should 726 DR ALAN MOZLEY ON FRESH-WATER MOLLTJSCA OF TANGANYIKA TERRITORY be taken to see that as large a number of fishes as possible is introduced into a pond in order that a shortage of food will force them to eat snails. Leeches (Hirudinea) and certain water-beetles (Dytiscithe) are known to feed upon fresh- water snails in the northern hemisphere. These animals have been found in the habitats of Physopsis globosa in Zanzibar and Tanganyika, but nothing is known about their feeding habits in that region. The possibility of either leeches or water-beetles exercising an effective control over the occurrence of snails appears to the author of this paper to be somewhat remote. There are numerous parasites of fresh-water molluscs, including insects (Mathias and Boulle, 1933, Jutting, 1938), trematodes, nematodes (Ohitwood and Chitwood, 1937) and protozoa. None of these are known at present to exercise any decisive influence upon the occurrence of their hosts under usual circumstances. This field remains to be explored further. Removal of Vegetation. Since fresh-water molluscs in general, and Physopsis globosa in particular, frequently occur in places where there is a luxuriant growth of aquatic and semi-aquatic plants, there might appear to be some possibility that the removal of the vegetation would result in a diminution of the snail population. Observations made in Tanganyika and the Zanzibar Protectorate during the course of this invetigation indicate that while this hypothesis may be correct in certain instances, it cannot be regarded as a general rule. Measures of this sort have the greatest chance of success in ponds which are fed by springs. In such instances the development of water into a nutrient medium may be so greatly hindered by the removal of the vegetation that snails are unable to exist in such a pond. This appears to be the condition in a spring-fed pond at Ndomo School near Nzega, T.T. This pond is kept absolutely free of vegetable matter, living or dead, by the school authorities and no snails have been found in it, whereas in an immediately adjacent pool in which there is an abundant growth of aquatic plants Physopsis globosa occurs in large numbers. It is questionable whether such a high degree of cleanliness as in the pool at Ndomo can be achieved and maintained in many ponds in Tanganyika without almost constant supervision by European officials. A part of the central area of the Muyuni pond, Zanzibar Island, is occupied by Rottbcellia exaltata and Pannicum maximum, among which Physopsis globosa can sometimes be found in moderate numbers. A rough estimate of their abundance in this habitat in the month of June would be fifty individuals per square metre. A small section of this part of the pond was cleared of these plants in April 1937. The plot was re-examined on 16th June 1937 and 141 individuals of Physopsis globosa were collected in an area of 1-25 square metres (depth of water 4 metres). This is probably the result generally to be expected when a stagnant pond in which the water is derived largely from surface drainage is cleared of vegetation. It is to be noted that under these conditions the water has had ample time to become a nutrient medium, by which is meant that algEe and decaying vegetable matter are present in considerable quantity. Promotion of the Growth of Vegetation. In the course of field work over a wide area in Tanganyika it was noted that marshes in which there is such a dense growth of grasses, and shrubs that direct sunlight does not reach the water, are seldom inhabited by fresh-water molluscs of any sort, and (as far as these observations go) never by Physopsis globosa or Biomphalaria pfeifferi. This suggests that in some areas where other means of control are not practicable, the promotion of the growth of AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SOHISTOSOMIASIS. 727 a dense "jungle" of shrubby vegetation may be an effective anti-bilharzial measure. This will have to be tested by experiment in the field. Snails are not found in large numbers in habitats in which the surface of the water is completely covered by the leaves of plants. The reason for this is very possibly a shortage of food in such situations as a consequence of the lack of light. The leaves of water-lilies (Nymphcea) appear to be moderately effective in bringing about this condition, but they have the disadvantage that they are themselves inhabited by snails in some numbers. Species of Lemna are fairly efficient in this respect, but the occurrence of this plant in East Africa is sporadic. A more promising plant for the purpose of reducing the numbers of snails in this manner is Pistia strat'iotes. This grows well in East African waters, and has been found forming a complete covering over the surface of ponds. No snails have been found in habitats in which such a condition exists. Education. To obtain the acquiescence and co-operation of the native inhabitants is an essential preliminary to the carrying out of the control measures described in the previous paragraphs. The interest of the average native in bilharzia is easily aroused, and there appears to be a considerable possibility that the inhabitants of many districts can be persuaded to make use of only such pools and streams as are known to be safe. It would be a grave error, however, to regard educational methods designed to prevent infection among natives as being an effective substitute for the actual destruction of the parasite and its intermediate host. Under the conditions existing in East Africa it is unlikely that educational methods alone will have a long lasting effect, and there is considerable likelihood that proposals for their adoption may at times serve as a pretext for delaying other action indefinitely. The members of the European and Indian communities in the area in which this investigation was carried out do not appear to be particularly well informed regarding bilharzia, and some educational work might well be undertaken among them.

VI. REPRESENTATIVE MOLLUSCAN HABITATS. The habitats of fresh-water mollusca which are described below are believed to constitute a series which is fairly representative of those which are to be found throughout the territory covered in the course of this investigation. In addition to this, many of the habitats are of interest individually as being important local sites of schistosome infection. It has therefore been thought worth while to make some comments regarding the application of control methods to these particular habitats.

A. Zanzibar Island. Muyuni, P1. II, fig. 2. The village of Muyuni is situated near the Macunduchi road, approximately 50 km. (31 miles) to the south of the town of Zanzibar. Infection with Schistosoma hcematobiurn is very common among the natives in this vicinity. Immediately adjoining the southern and western sides of the village is a large marshy pond. This forms the common washing and bathing place of the villagers. It is from this pond also that much of the water used for domestic purposes is drawn. The greatest length of the pond site is approximately •5 km., and the greatest width is about 3 km. It is only during and immediately following the season of heavy rain that the TRANS. ROY. SOC. EDIN., VOL. LIX, PART III, 1938-39 (NO. 26). 109 728 DR ALAN MOZLEY 014 FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

whole of this area is submerged. During the dry seasons the pond is roughly circular in shape, with a diameter of about 150 m. • This includes an area of relatively deep water (2 to 3 m.) which is somewhat less than 100 m. in diameter and an encircling belt of shallow water (depth 25 cm. to 1 m.) over parts of which there is a dense growth of plants, chiefly Rottbllia exalt ata

FIG. 8.—Zanzibar. Fm. 9.—Pemba. Pies. 8, 9.-1. Weti; 2. Ziwani; 3. Ole; 4. Chake Chake; 5. Chonga; 6. Nwachani; 7. Mtombile; S. Kengeja; 9. Fufuni; 10. Mkoani; 11. Mkokotoni; 12. Pwani Mchangani; 13. Chaani; 14. Mbiji; 15. Mangapwani; 16. Bambi; 17. Dunga; 18. Mwera bridge; 19. Zanzibar Town; 20. Tomondo: 21. Kombeni-Ziwani; 22. Chwaka; 23. Chegu; 24. Bungi; 25. Muyuni; 26. Kizumkazi; 27. Macunduchi; 28. Mtende; 29. Scale 5 miles; 30. Scale 5 miles. and Pannicum maximum. Surrounding this area of relatively undisturbed pond is cultivated ground which is flooded during the rainy seasons. Three species of fresh-water molluscs have beeen collected in the Muyuni pond, namely, Lanistes olivaceus, Physopsis globosa, and Hippeutis sp. The Lanistes is common throughout the year. Physopsis is to be found in great abundance for several months after the great rains, and again after the small rains. Hippeutis is rare. Under certain circumstances, the alternation of wet and dry conditions, such as occurs in temporary ponds in different parts of the world, is known to be eminently suitable for some fresh-water molluscs. This is particularly true of the pond at Muyuni. Two months after the cultivated area surrounding the pond is flooded, it is the habitat of several millions of snails. Food is present in abundance, conditions of temperature and depth of water are favourable, and enemies, such as certain fishes, are absent. Hence both Lanistes olivaceus and Physopsis globosa are present in large numbers. Furthermore, the conditions are also very favourable for Schistosoma hcematobium, since in addition to the intermediate host being present in large AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 729 numbers, the pond is habitually used by several hundred natives for one purpose or another. It frequently happens that the natives work in the submerged rice fields for several hours at a time, and it is a favourite resort for children, so that there is ample opportunity for the transference of the parasite. It is not surprising therefore that the incidence of schistomiasis among the local population is high. In considering any prospective control measures in this locality it is essential that the necessity of preserving the local culture of rice around the pond be borne in mind. It is upon this crop that the natives depend for a considerable part of their food supply. The most practical way of attacking this problem appears to be to regulate the amount of water in the pond in such a manner as to cause the death of the snails, or at least to result in a reduction intheir numbers, while at the same time retaining aquatic conditions for a sufficiently long period to ensure the development of a good crop of rice. The valley of the intermittent stream which drains the Muyuni pond at times of high water is fairly steeply inclined for the first kilometre beyond the pond outlet. There would be no serious difficulty in draining the pond by means of a ditch dug across the rim of the depression. Following this it would be a simple matter to construct a removable dam across the drainage ditch. This would make it possible to flood the whole of the pond area for a period of up to two months from the time of the beginning of the great rains, and thus to ensure the full development of the rice. Up to that time there appears to be little danger of infection with bilharzia. The drying of the pond would very much reduce the number of snails. In order to accelerate the drying of the pond bottom it might be worth while to have it ploughed twice a year, during the dry seasons. Such action would probably be justified on agricultural grounds alone. A further measure, which might be expected to result in the complete extermination of the dangerous snails, would be to uproot the marsh plants which encircle the centre of the pond, and to have them thrown into the deeper parts of the depression, together with cocoanut husks and other local debris, so that in the course of a few years the deeper part of the depression would befilled. The execution of a plan such as is outlined above would result, over a period of years, in the disappearance of schistosomiasis from this district, while at the same time the rice culture would not be adversely affected. Mtende. Mtende, a village situated near the southern end of Zanzibar Island, is believed to be an important centre of schistosomiasis. There are several ponds in this neighbourhood which are inhabited by Physopsis globosa. It may or may not be feasible to drain these ponds by means of channels above ground. An alternative method of drainage would be to pierce the impermeable layer on the bottom of the ponds by blasting, and thus to accelerate their drainage through underground channels. Several of the ponds become dry during the season of drought each year, and at that time crops are grown on the pond-bottoms. Ploughing at that time of year, which would be justified in any event upon purely agricultural grounds, might possibly be found to be efficacious in reducing the numbers of snails by exposing them to more intense desiccation, and also in bringing them within the range of attention of carnivorous beetles. While the measures indicated above are being experimented with it is desirable that some means of giving the inhabitants of this locality immediate protection from infection 730 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

should be undertaken. It would appear that a practical method would be to treat one of the ponds with chemicals and then to persuade the villagers to bathe and wash only in that pond. This possibility has been mentioned to the local mudir and chief, and they consider that it would be possible to gain the co-operation of the villagers in this matter.

Mbiji. The village of Mbiji is situated 27 km. (17 miles) north of Zanzibar town. About one-half kilometre from the village there is a well-defined depression, the centre of which is occupied by an expanse of open water about 200 m. in diameter. The depth of the water in this area during the month of August is between one and two metres. Surrounding the open water is a belt of reeds which varies in width in different places from 20 to 40 m. During the great rains and for some time thereafter this marsh is flooded, but during the dry seasons the outer edges of it are dry. The level of the water in the pond is controlled in a general way by a well-defined intermittent outlet channel which functions only when the pond is full. Three species of fresh-water molluscs are found in the Mbiji pond in great abundance, namely, Physopsis globosa, Lanistes olivaceus and Cleopatra ferruginea. The Lanistes is found commonly in all parts of the pond. Cleopatra and Physopsis are most abundant in the marshy area. A medical examination of the native children attending the school at Donge, which is situated a few kilometres distant from Mbiji, revealed that a considerable number of them were infected with Schistosoma hcematobium. As far as possible all such children were questioned with regard to the places in which they bathed. In every instance where positive information was obtained the infected child had either bathed in the Mbiji pond, or had been there to fetch water. A number of the uninfected pupils stated that they had never visited the Mbiji pond. Many of the natives of this district, including the majority of the uninfected school children, bathe in a stream to the north and east of Donge village. No Physopsis were found in this stream, and the general character of the habitat renders it highly improbable that any occur there. Several other streams in this neighbourhood were examined but no Physopsis were found. It may therefore be taken that the Mbiji pond is the principal, if not the only source of infection with bilharzia in this district. The control measures required in this area are fairly simple and inexpensive. The Mbiji pond lies in a shallow saucer-shaped depression, and since a well-defined drainage channel approaches the rim of this basin it would be a simple matter to deepen the bed of the intermittent outlet stream to a sufficient extent to drain the pond. The reclaimed pond bottom would undoubtedly have some agricultural value; but if it were used for growing rice some supervision would be required to prevent the recurrence of bilharzial infection following too prolonged flooding. Some local diminution of the mosquito nuisance is to be expected following the adoption of the control measures suggested above.

Kombeni-Zjwanj. In the vicinity of the native settlement known as Kombeni-Ziwani, situated about 11 km. (7 miles) to the south of the town of Zanzibar there are 3 ponds which attain a considerable size during the rainy seasons. Two of these, situated respectively near Mile 65 and Mile 75 on the Fumba road, contain water for only two months after the rainy seasons. It appears that this period of possible activity is too short to permit the survival of Physopsis AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 731 globosa. In any event this species is absent from the two points mentioned. Bulinusforskalii is found in both pools in large numbers. Another pond, which is situated approximately mid-way between the two ponds mentioned above, that is to say, near Mile 7, Fumba road, contains water the year round, although it shrinks very much in size during the dry seasons. This pond is circular in shape, the diameter being approximately 200 metres. The maximum depth of the water in the pond at high water is between two and three metres and at low water approximately one metre. A great part of the pond area is covered with a dense growth of rushes. Lanistes olivaceus and Physopsis globosa occur in this pond in large numbers. An examination of the surrounding district indicates that this pool is the sole habitat of Physopsis globosa in this part of Zanzibar Island. Children attending Kombeni School who are infected with Schistosoma hcematobiurn all state that they have bathed in the pond near Mile 7, Fumba road. The only other common bathing place for the children of this neighbourhood is in a cave situated three kilometres south of Kombeni School. No molluscs were found in this cave pool. The extermination of Physopsis globosa in the pond near Mile 7 would evidently free this area from local infection with bilharzia. Experiments with soap berries (the fruit of Sapindus saponaria, Swahili-Halita) as snail destroyers were carried out in this pool. Many of the Physopsis were killed but the species was not exterminated (see p. 721). The most effective way of controlling schistosomiasis in this district appears to be to drain the pond near Mile 7, Fumba road, which was unsuccessfully treated with soap berries. The cost of this has been estimated by the Zanzibar Public Works Department to be shillings (E.A.) 400. The provision of a well to serve as a source of water supply for the local inhabitants is estimated to cost shillings (E.A.) 100. Half a kilometre to the south of the Kombeni pond referred to above, there is another pool in which the molluscan fauna at present consists of only Bulinus forskalii. This is also the condition in a pool situated one kilometre to the north of the Komnei pond. Although these pools are not breeding sites of Physopsis globosa.at present, they appear to be potentially dangerous. The planting of about twenty Eucalyptus trees in each pond, however, should obviate this possibility. Mile 3, Macunduchi road. In the vicinity of Mile 3, on the Macunduchi road, there are two large permanent ponds occupying well-defined depressions which support Physopsis globosa and Lanistes olivaceus in great abundance. This locality has been visited many times in the course of this investigation, and the impression gained is that relatively few natives come into contact with the pond waters. During the drier seasons the shores of these ponds are used for growing vegetables, so that casual visitors are not encouraged. It is conceivable therefore that these habitats play no very important part in the bilharzial infection of the Zanzibari as a whole. Therefore, if the funds available for control purposes are severely limited, these ponds might be omitted from the programme for the time being with greater safety than most of the other Zanzibar ponds mentioned here. Eventually they should be drained. Such expenditure will be recoverable, at least in part, through the increased value of the property. It is possible that, provided the capital is available, the draining of the ponds might be justified upon agricultural grounds alone. 732 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

Mile 2, Fumba road. In the vicinity of Mile 2, Fumba road, there are two shallow pools approximately 100 m. in diameter which contain Physopsis globosa in considerable numbers. These pools are already drained to some extent by an anti-malarial ditch, and as the gradient of this channel in the immediate vicinity of the ponds is very marked, it appears that if it were deepened for a distance of about one hundred metres, at least one of the ponds would be completely drained. Under present conditions Physopsis have been observed to be carried downstream by the current from the pond for a considerable distance, which condition is highly undesirable as it greatly extends the range of possible infection.

Chegu. A large marshy pond near the village of Chegu situated 24 km. (15 miles) east of Zanzibar town may be an important centre of infection with bilharzia. There is no information available as to the actual extent of infection in this neighbourhood. Physopsis globosa occurs in the Chegu pond in large numbers. On account of the rocky nature of much of the surrounding country it may not be easy to drain this pond. An alternative measure would be to provide one or more wells, so situated that the distance from the native huts to them is less than that to the pond. It is probable that under these conditions a minimum of educational work among the villagers would result in a substantial reduction of the possibility of infection with schistosomes. Bambi. In the vicinity of the village of Bambi, which is situated approximately mid-way between Pwani Mchangani and Bungi, there is a large pool which is said to be called the Mtojuu pond. Physopsis globosa is found in this pool in considerable numbers. This pond is evidently visited by a fairly large number of natives, and may therefore be an important source of infection with Schistosomes. The most practicable means of draining this pond may be by blasting through the impermeable layer on the bottom, and thus facilitating the passage of the water underground. The bottom of the pond appears to be composed of a finely divided black mud which is well suited for agricultural purposes.

Other Habitats of Physopsis in Zanzibar. Specimens of Physopsis globosa have been collected in several localities on Zanzibar Island in addition to those mentioned above, but seldom in large numbers of individuals. Moreover, the snails collected in many such habitats are of small size. This may be taken as indicating in a general way that conditions of existence there are not favourable. Physopsis globosa is occasionally to be found in shallow native wells, but on account of presence (in most instances) of alternative bathing places, and the disfavour with which bathing in places which are used as a source of drinking water is regarded by the rural Zanzibari such occurrences have less significance than on the mainland of Africa. In the latter region, owing to the vastly greater number of transients, with whom local social controls carry little or no weight, many wells are likely to be contaminated. It would be a serious error to assume on the basis of the above statements that all the minor habitats of Physopsis globosa in Zanzibar can be ignored with safety. The molluscan fauna of this region, as is true of all other populations, is in a state of continuous change so AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 733 that the conditions which have been observed in the course of this investigation may not continue indefinitely. The following are examples of localities which should be carefully watched: Tomondo; the culvert near Mile 9 on the Chwaka road, and the marshy pond which lies above it; the ponds, streams, and ditches in the vicinity of the Mwera bridge, Chwaka road; the native wells lying to the south of Mile 15, Chwaka road; the native wells situated 2 miles south-east of Ndajani; the large pond near Mile 12, Marie-Louise road; the marshy area between Miles 7 and 8, Macunduchi road; and ponds near Chaani which are used for washing and bathing purposes. B. Pemba Island. Physiographical conditions on the Island of Pemba favour the development of habitats of fresh-water moflusca to a much greater extent than on Zanzibar Island. Physopsis globosa is exceedingly common and widely distributed, as is also infection with Schistosorna hcematobium. The prospect of success for control measures, however, is promising. A longer period of time will probably be required in Pemba to make such measures effective than in Zanzibar, but given a steady effort, and moderate expenditure, good results are assured. The habitats of Physopsis globosa in Pemba are so numerous and at the same time so closely similar to one another that to describe each of them in detail would involve unnecessary repetition. The account given below is therefore a general one, which applies equally well to the majority of Physopsis habitats on the island. In the section of this paper which deals with the local distribution of the mollusca, it was noted that on Pemba Island, Physopsis globosa is found principally in the middle reaches of streams (see P1. III and text-fig. 7). For the most part the extreme headwaters are too dry for these snails, while the lower parts of the streams are subject to too great flushing and scouring to be suitable habitats for this snail. In most instances, therefore, the localities inhabited by these undesirable snails tend to be strictly delimited. The object of the control measures proposed here is to improve drainage at certain crucial points, so that the middle course of these streams will no longer contain standing water. The results of a practical experiment carried on in a valley situated near the village of Chonga, near Chake Chake, indicate that this is possible. In this instance before the beginning of the experiment a considerable area of moist meadow lying in a flat-bottomed valley was found to support Physopsis africana in abundance. The construction of a ditch, 4 feet in width and 2 feet in depth, through the centre of the marshy area gave definite indication of the conversion of the area from low, moist meadow into dry, relatively well-drained meadow in which Physopsis is absent. A marked diminution in the numbers of Physopsis was observed within two months of the time that the ditch was dug. Making allowance for seasonal changes, the numbers of Physopsis present appeared to grow progressively smaller during the whole of the period, subsequent to the construction of the ditch, that the area was under observation (approximately one year). It is difficult to predict whether measures of this nature will invariably result in the complete extermination of Physopsis in the streams concerned, but it appears to be probable that wherever the inclination of the stream-bed is sufficient, the number of individual snails will be reduced to such an extent that those which remain will be of little importance. The general plan of campaign therefore is to convert low, moist meadows which harbour Physopsis into dry meadows which are unsuitable habitats for this snail. The method proposed above for the control of schistosomiasis is expected to have valuable 734 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

incidental results from the standpoint of malarial control and agricultural development. At the present time the utilization of land in Pemba falls far short of the possible maximum. Certain wet meadows which have been neglected, and in which wild vegetation has either been left undisturbed or allowed to re-enter into possession, are the most dangerous habitats as far as bilharzial infection is concerned. The reclamation of such areas would not only destroy a large proportion of the breeding-places of undesirable snails and mosquitoes, but would make available considerable areas of valuable agricultural land. The Medical and Agricultural Departments of the Zanzibar Protectorate strongly favour an increase in cattle-raising on grounds other than schistosomiasis. It appears that an increase in the number of such domestic animals owned by natives in Pemba would have valuable incidental results as far as the control of schistosomiasis is concerned. The most suitable grazing areas for an increased cattle population are in the valley bottoms described above, but before they could be most effectively utilized for this purpose drainage along the lines proposed would be necessary. In certain parts of the iland, notably to the north of the village of Sipwese near Kengeja, such a development has already taken place to some extent, and it is worth noting that no Physopsis have been collected in these well-drained valleys. The cost of these control measures is difficult to estimate, but the impression gained in the course of field work was that an expenditure of shillings (E.A.) 500 upon the improvement of the drainage of each of the 30 most important drainage basins in the southern part of the island would go far towards reducing infection with bilharzia to negligible proportions in the whole of the area from the town of Chake Chake southwards. Should this figure appear at first sight to be too low, it should be borne in mind that the aim of the work suggested is simply to accelerate drainage along existing channels in which the gradient is moderately high, and not to construct any entirely new major channels.

C. Tanganyika Territory. Dar-es-Salaam. There are numerous bodies of stagnant water within a radius of 10 km. of the centre of Dar-es-Salaam. Few of them are inhabited by Physopsis globosa. A large pond, known as the Bagamoyo Pool, is situated close to the native quarter of the town. The general appearance of this pond, and the rice fields which surround it, is such as to lead an observer to suspect that a varied and moderately abundant fauna of moflusca might be found there. Careful examination of this pond has resulted in very meagre collections. The only species found is Bulinus forskalii. A few individuals of this species have been found in one particular part of a single rice field adjoining the pond and nowhere else. Physopsis globosa has been collected in several ponds in the area known as Kidondoni which lies 3 miles in a direct line north-north-west of Dar-es-Salaam Post Office. It is probable that a large proportion of the cases of schistosomiasis which are treated in the Dar-es-Salaam hospital each year have had their origin from infection in this area. It is possible to treat all of the Kidondoni pools in one or other of the ways mentioned in Section V of the present account. Tanga. The number of probable sources of bilharzial infection in the vicinity of the town of Tanga are limited. Ponds and streams are numerous in this neighbourhood, but few of them contain Physopsis globosa. The most important local habitat of this snail is a series of ponds situated near the Pangani road, approximately 4 km. from Tanga railway station. These pools are a AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 735 source of water supply for domestic purposes for a certain number of natives, some bathing is done in them, and they are also used for watering cattle. The possibility of draining these ponds should be fully investigated. Should it be found impossible to drain these ponds, some means of poisoning the snails in them will have to be resorted to. As a plentiful supply of I I F I 25 r'

I ) I 'I k''u1• F • 24

2 26

1 22 ' i 3J 1Nt 1A3 )i 15 \ c1u 28 \ 16 18 '7 15 Q21

"lb .20

2_9

Fie. 10.-Tanganyika Territory. 1. Pemba Island; 2. Tanga; 3. Zanzibar Island; 4. Dar-es-Salaam; 5. Mafia Island; 6. Indian Ocean; 7. Pangani River; 8. Moshi; 9. Arusha; 10. Tanga Railway; 11. Morogoro; 12. Dodoma; 13. Bubu River; 14. Singida; 15. Central Railway; 16. Ruvu River; 17. Rufiji River; 18. Iringa; 19. Lake Nyasa; 20. Mbeya: 21. Lake Rukwa; 22. Tabora; 23. Manyonga River; 24. Mwanza; 25. Lake Victoria; 26. Malagarasi River; 27. Kigoma; 28. Lake Tanganyika; 29. Scale. 0 100 200 miles soap berries (the fruit of Sapindus saponaria) is available in this locality they might well be used for the purpose of poisoning the snails in this pond. Another locality which may be of some importance is a pond situated a short distance south of Km. 8 on the Tanga-Moshi railway. Physopsis globosa is common in this pond. This pond is situated close to the main road from Korogwe to Tanga, and natives passing on foot have been observed to wash themselves here. During the wet seasons an extensive area in the vicinity of this pond is under water, and the potential infective area may be an extensive one. Further observations are required in this connection. Partial or intermittent drainage, of the pond is feasible. Korogwe. A small pond situated approximately 200 m. south-west of Korogwe railway station is a habitat of Physopsis globosa. Should the district medical authorities consider this pool to be of sufficient importance, it is well adapted for chemical treatment.

TRANS. ROY. SOC. EDIN., VOL. LIX, PART III, 1938-39 (NO. 26). 110 736 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

Physopsis globosa also occurs in other ponds in the marshes which border the Pangani River in the vicinity of Korogwe. The most practical way of dealing with the possibility of infection from this source is to warn the natives not to bathe in these pools. Kilimanjaro and Meru Areas. The country in the immediate vicinity of the Kilimanjaro massif and Mount Meru, in the Northern Province, differs considerably from the remainder of the Tanganyika Territory. The problems of bilharzia control likewise differ. At the present time human schistosomiasis is known to occur in only a few scattered localities in this region. Potentially, however, this district is possibly the most dangerous one in the whole territory. From the point of view of general geography the country immediately surrounding the Kilimanjaro massif and Mount Meru are essentially oases in an arid region. The natural and cultivated vegetation and the human population are dependent upon the rainfall on the upper slopes of the mountains. In order to make more effective use of the available water supply an extensive system of irrigation canals has been developed on the lower slopes, and to some extent on the surrounding plain. At the present time the utilization of the water which includes the irrigation of crops such as bananas in localities having a rainfall of about 2540 mm. (100 in.) per annum, falls far short of that which would be possible under conditions of efficient control. Steps to attain a more effective use of the water are now being undertaken by the Government, and it is likely that within the next decade the area under irrigation will be very greatly extended. In the course of this extension it is likely that many eminently suitable habitats for fresh-water molluscs will be created. At the present time most of the irrigated land is situated on more or less steeply sloping ground, so that the current in the irrigation channels is swift. Such conditions are not favourable to Physopsis globosa in this region. An extension of the cultivated area will probably involve the invasion of the plains which surround the Kilimanjaro massif and Mount Meru. The current in the irrigation channels in this area is bound to be more gentle than in those on the sloping ground, and as a result diverse species of fresh-water molluscs are to be expected to make their appearance in the canals. The intermediate host of both Schistosoma hcematobium and S. mansoni are known to occur in this district, and have been found in quietly flowing water. Therefore unless a careful watch is kept, and preventive measures are taken, serious outbreaks of schistosomiasis are to be expected. The Medical Officer, Arusha, reports that a case of infection with S. hcematobium and S. mansoni in a- European child living in the plains region below Mount Meru came to his notice during 1938. The most effective means of control under the conditions described above, will be to allow the irrigation channels and all connecting waters at low altitudes to dry completely at frequent intervals. It may also be found desirable to treat chemically some of the natural breeding places of the dangerous snails. The installation of a flushing sluice (q.v.) in the stream crossing the Moshi-Arusha railway at Km. 21-5, is strongly to be recommended. Mile 31, Arusha-Babati road. In the vicinity of Mile 31 on the road from Arusha to Babati and Dodoma there is a large marshy pond which is inhabited by Physopsis globosa. As this pond is situated close to a much-travelled road, and for many months in the year is the only available water supply within a considerable radius, the competent authorities might consider the desirability of erecting a notice in the vicinity warning travellers of the danger of infection with schistosomes. Owing to the sparse population of this district it is probably not worth while to attempt to destroy the snails in the pond. AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 737

Mbugwe. The population of Mbugwe, a native village near the Great North Road between Arusha and Dodoma, is relatively heavily infected with schistosomes. The situation of the village, on a nearly plane surface at the foot of an escarpment, is favourable to the development of shallow ponds and marshes. The local fauna of mollusca is a rich one, although, as is true of the surrounding territory, not many species are represented. Every pond and marsh, and nearly every well in the vicinity of the village appears to be either actually or potentially the habitat of Physopsis globosa. That is to say, there are one hundred or more places within easy walking distance of the village where human beings might be exposed to infection. Drainage or other measures on a large scale will therefore be extremely costly. The most practical means of control at the present time appears to be to have one or more relatively deep wells constructed to serve for the domestic water supply, and to treat one of the ponds with chemicals in order that it may be rendered safe for bathing. It appears that the local natives are sufficiently intelligent and co-operative to ensure that a considerable reduction of infection will result from these measures. Morogoro. The source of infection with schistosomes in the vicinity of the town of Morogoro is a large marshy pond situated to the north of the railway line approximately 1 kilometre east of Morogoro station. Physopsis globosa occurs in this pond in large numbers. None were collected in any other place near the town. The pond referred to should be drained without delay. Should the Railway Department raise any objection on the ground that such action would endanger the stability of their permanent way, the Government Water Consultant will probably be able to suggest ways of avoiding this at small expense.

Dodoma. In 1937 the Medical Officer at Dodoma reported that "all of the school children" in Dodoma were infected with Schistosoma hcematobium. The only available source of infection within 10 km. of the town is in two ponds situated near the golf-course and between the railway station and the aerodrome. Physopsis globosa is common in both of these pools. The ponds should be drained and kept absolutely without water for at least four months. It is probable that this treatment would either destroy all the snails or greatly reduce their numbers. If they appear again, an additional period of drying should be tried. In this way it will be possible within a few years' time to discover the longest period in each year during which it is safe to have water in the ponds.

Iringa. The observations made in the course of this investigation indicate that the only likely site of infection near the town of Iringa is in the marshes and marshy pools along the course of the Little Ruaha River. These are situated at some distance from the most thickly populated area. The number of infected persons receiving treatment at the Iringa Hospital each year is reported to be small. It is not feasible to attempt to drain the marshy areas, as quite apart from the expense involved in such an undertaking, it is probable that it would result in a lowering of the water table in the surrounding district. It might be as well to warn all European residents of Iringa of the danger of wading into these marshy areas. 738 .DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

Ru'ngwe District. The greater part of the Rungwe District appears to be free from schistosomes. Owing to the close proximity of high mountains, and the resulting high inclination of much of the land, most of the natural waters of this district are unsuitable habitats for the intermediate hosts of these parasites. In the immediate vicinity of Lake Nyasa, however, bilharzia is very common among the natives. The waters of the lake proper are safe, as are also apparently the lagoons which stand immediately behind the lake beaches. The sources of infection are in ponds which lie slightly further away from the lake shore but adjacent to it. Such ponds are very numerous, and many of them appear to be inhabited by Physopsis globosa. It will probably be found impracticable to drain these ponds. At least a partial solution of this local problem is to be found in the treatment of certain ponds by chemical means to kill the snails. The local inhabitants can then be encouraged to use these and only these ponds for bathing and washing purposes. The natives in this section of the country are somewhat more intelligent than elsewhere, and it is likely that given the right approach a substantial reduction of infection can be achieved at trifling cost. Nzinge and Bahl. An mbuga lying to the south of Nzinge railway station is the habitat of a large number of Physopsis globosa. Provided that the presence of Schistosoma hcematobium in the human population can be proved, and that the matter is regarded as being of sufficient importance to warrant the expenditure involved, the construction of a shallow well near the native huts on the eastern side of the mbuga will probably relieve the situation. The Railway Depart- ment should be informed that water drawn from this source is not fit for human consumption. Several native settlements in the neighbourhood of the village of Bahl appear to be exposed to bilharzial infection. The most probable sources of infection are in ponds in the beds of intermittent streams. Physopsis globosa has been collected in several of these places. As many of these pools are situated at some distance away from native villages, so that the inhabitants are forced to walk for some distance for water, the provision of wells in more convenient places will probably go far towards the solution of the local problem. Should finances in the course of years permit of a moderate expenditure upon the control of schistosomes, the ponds can be treated with chemicals. Tabora. Judging from the published reports of the Tanganyika Medical Department, the most prolific source of infection with schistosomes in the vicinity of the town of Tabora in years past has been a series of pools in the suburb known as Rufita. Several of these ponds have been filled in with town refuse, etc. This should be continued until all ponds are filled. The Public Works Department should be prevailed upon not to dig any more pits in such a manner that other similar pools will be formed. A large Government school for boys is situated on the eastern outskirts of Tabora. The pupils are strictly prohibited from bathing in any of the natural ponds in the neighbourhood. It is highly desirable that this regulation should continue to be enforced. Along the Nzega road near Tabora there are several ponds which contain Physopsis globosa in numbers. In certain of these pools native milk vendors are said to wash the bottles in which they carry their product. This is particularly undesirable in view of the fact that it is the custom of such natives in this district to diluth their milk with an equal quantity of water. Since the number of such pools is not very great they might well be treated in one of the ways suggested in another part of this paper. The ponds which require immediate AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 739 attention are situated near the Nzega road at the following distances from the main railway crossing of that road (west of Tabora railway station), Mile 27 (north side of road); Mile 38 (north side of road); Mile 60 (south side of road). The small stream which crosses the Nzega road at Mile 65 (culvert 23) is a potential source of infection. As there are several relatively safe sources of water supply in the vicinity it might be as well to discourage the natives from using the water in the small intermittent stream which runs through this culvert. A native well situated 1 km. west of Mile 105 has been examined on several occasions and appears to be free from undesirable snails at present. A pool situated near the village of Ititemia near Tabora, which is commonly and by preference used for bathing purposes by the local natives, supports a large number of Physops'is globosa. The inhabitants of the village appear to be willing and anxious to co-operate in control measures, so that conditions are favourable for the use of copper sulphate as a mollus- cide. The water of the pond is not used for domestic purposes. N2e9a and Ndomo. The water supply for the European inhabitants of Nzega, situated to the north of Tabora, is - drawn from a well and pond a short distance north of the Boma. As far as schistosomes are concerned this supply appears to be safe. Physops'is globosa has been collected in two ponds near the drift on the main road 1 km. south of Nzega. Arabs returning to the village at sundown have been observed to wash themselves in one of these ponds before praying. It is likely that the cost of a combined draining and filling of these ponds would be a few shillings only. The small stream in the first valley to the north of the Boma was found to contain Physopsis globosa in small numbers. That is to say, it was with some difficulty that one hundred specimens were collected in an afternoon. Numerous other ponds within a radius of 10 km. of Nzega were examined but no Phy.sopsis were found in them. There is a Government boarding school for boys at Ndomo, 8 km. (5 miles) north of Nzega. The pond in which the pupils bathe is inhabited by a very large number of Physopsis globosa. Infection with S. hcmatobium is very common among the boys attending this school. This pond (diameter 10 m.) was treated with soap berries (see above) on two occasions, namely 25th and 26th October 1937. After each of these treatments many dead snails were found, but apparently a sufficient number of individuals survived to repopulate the pond. Since these methods have proved to be ineffective, the pond, which is at the present time used for bathing, should be drained. The cost of this work would approximately be four shillings. This would leave the spring, which is nearby and free from snails, as a source of water supply. The explanation of the difference between the spring pond and the bathing pool, the one containing many Physopsis and the other none, is that the former is absolutely free of rooted aquatic plants, while the latter is nearly filled with them. It is probable that the school at Ndomo will be closed in the near future. This fact should not make any essential difference to the carrying out of the measures suggested, since owing to the presence of water in the midst of a semi-arid region it is certain that there will always be natives present in and about this locality. The introduction of any irrigation in this locality, no matter on how small a scale, should be prohibited. There is no objection to the watering of crops in this vicinity by hand. Shinyanga and Tinde. During the course of this investigation Physopsis globosa has been collected in only one place in the vicinity of the village of Shinyanga. This habitat is a small pond on the flood 740 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY plain of the Manyonga River about 2 km. north of the town. Other ponds containing this snail may be found in this valley, but at the present time there does not appear to be much possibility of infection with schistosomiasis. The area is potentially dangerous, however, and a close watch should be kept upon any new ponds which may be formed either naturally or through the agency of man. Physopsis globosa was collected in a pond used for bathing purposes by native workers in a diamond mine situated some kilometres to the south of Shinyanga. The manager of this mine expressed his willingness to adopt measures to render the pond safe. The use of chemical measures of control in this pond is strongly to be recommended. Physopsis globosa was also collected in a small pond near the village known as Ngwakulwa, which is situated not far from the Shinyanga- Tabora road about 5 miles south of Shinyanga. This pond is also well adapted for the use of chemical treatment. There is a tendency here for the natives to dig small irrigation channels in the vicinity of the pond. This should be discouraged, although there is no objection whatever to the crops being watered profusely by hand. The resident medical officer at Tinde reports that urinary schistosomiasis is very common among the native inhabitants of that village. There are numerous possible sources of infection in this vicinity. All the larger streams in this neighbourhood contain at least a few Physopsis, as do also many of the ponds. In flowing water, however, a harmless snail (Lymncea caillaudi) is predominant, and it is-apparently only in a limited number of pools in the bed of or con- tiguous to the streams that Physopsis occurs in any considerable numbers. In ponds situated at some distance from streams and in roadside ditches this undesirable snail is common. The cost of exterminating Physopsis in this district would be out of all proportion to the benefits to be gained. However, the expenditure of approximately twenty shillings upon improving the drainage of the stream which lies between the Trypanosomiasis Laboratory and Tinde village is strongly to be recommended. The local natives could then be advised to wash and bathe in this stream. The stream, which crosses the Shinyanga road 1 km. to the north of Tinde village, would be an even better site for washing and bathing, but it is doubtful whether the natives could be persuaded to walk this longer distance. Bathing and washing in pools on the roadsides and in stagnant pools should be discouraged, and when finances permit they should be treated chemically. Mwanza. The town of Mwanza is §ituated on the shore of Lake Victoria. The possible sites of bilharzial infection fall into two categories, namely: Habitats in the lake and around the shores. Habitats in streams which flow through the town into the lake. There are two large rock-bound and partly sheltered bays in the immediate vicinity of the town. In one of these bays Physopsis globosa has been collected in small numbers on submerged vegetation growing in from one to three metres of water and within fifty metres of the shore. It is by no means certain, however, that this habitat is of great importance as far as the transference of schistosomes is concerned. Observations made over a period of years will be required to prove that it is of no practical importance. In several places in Mwanza, natives who go to the lake for the purpose of bathing, drawing water, or washing clothes walk through small marshy ponds which lie between the lake beach and the slightly higher ground on which the town lies. Certain of these ponds are inhabited by Biomphalaria pfeifferi, and are important possible sites of infection with S. mansoni. AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 741

It is highly desirable that the shore should be built up to a higher level by the filling in of the ponds. Such work would destroy important breeding places of Anopheline mosquitoes. Camping along the lake shore should be rigidly prohibited. The carting of sand away from the lake shore should not be allowed. One of the streams which flows through the town appears to be a more serious menace than are the lake waters. This is the small river along which the Mwanza-Musoma road runs about 2 km. beyond the town boundary. At certain times of the year, notably in October and November, Biomphalaria pfeifferi is found in this stream in large numbers. A pond in the bed of this stream near the Musoma road bridge on the outskirts of Mwanza was found to contain about 4000 Biomphalaria pfeifferi. Many of the natives from the country, after spending several hours in the bazaar, stop here on their way home to bathe and also deftecate under the bridge. The pool is therefore almost ideally situated for the acquisition and transmission of schistosomes, and immediate action is desirable to prevent the continuance of the infection which undoubtedly takes place here. The bed of the stream for the whole distance through the town below the bridge should be cleared of obstructions, so that no ponds can form. Owing to the rapid erosion of the banks of the stream during the rainy seasons it is probable that a part of the cost of such measures can be regarded as preventive works to protect the bridge and drifts.

Kigoma. The probable sites of infection with schistosomiasis in the vicinity of Kigoma are limited in number. The most important of these are the marshy areas which lie between the sandy beach of Lake Tanganyika and the higher ground on which the village of Ujiji stands. This series of marshes is inhabited by Biomphalaria pfeifferi. Several hundreds of natives pass through the shallower parts of this marsh each day on their way to the lake shore, and are thus exposed to infection. The waters of Lake Tanganyika appear to be perfectly safe in this respect. A slightly elevated pathway about one hundred metres in length could be constructed at slight expense and would practically eliminate the possibility of infection. The only other localities in the neighbourhood of Kigoma which appear to be of any importance in this connection are side channels (?) of the Mgonia River near Km. 1239 on the railway. Physopsis globosa was collected here. There is no evidence to show that any person has been infected at this place.

VII. SUMMARY AND CONCLUSIONS. (I) The observations upon which this account is based were made in Zanzibar, Pemba, and the Tanganyika Territory during a part of the years 1937 and 1938. (II) A considerable number of the native inhabitants in certain localities on Zanzibar Island, and in the greater part of Pemba, are infected with Schistosoma hcematobium. Infection with that species and with S. mansoni is common in many places in Tanganyika. As a rule the infection in any individual in this part of Africa is light in comparison with the condition found in Egypt. Nevertheless schistosomiasis is a disease of considerable importance in this region both from a medical and an economic standpoint. The local intermediate host of Schistosoma mansoni is the snail, Biomphalaria pfeifferi. The local intermediate host of S. hcmatobium is believed to be the snail, Physopsis globosa. (III) 26 species of fresh-water mollusca are known to occur in the territory covered in the course of this investigation. 742 DR ALAN MOZLEY ON FRESH-WATER MOLLUSCA OF TANGANYIKA TERRITORY

(IV, i) Some insight into the general geographical setting is essential to an understanding of the local problems of snail distribution, and hence, eventually, to a mastery of the problems of snail-control. The principal geographical conditions which govern the occurrence of snails are earth structure, climate, .vegetation, and soils. (IV, ii) The principal local conditions which govern the occurrence of snails are flood, drought, local geomorphology, food and light, and the presence of enemies. As a result of these conditions the distribution of the undesirable snails, Physopsis globosa and B'iomphalaria pfeifferi, is strictly limited, that is to say, these species occur in only a small proportion of the ponds, lakes, and streams. (IV, ill) The kinds of habitats in which fresh-water molluscs are found in this region are as follows: ponds, mabuga (for definition see p. 715), lakes, and streams. (V) The most effective means of controlling human schistosomiasis in this region are those which are directed towards the extermination of the snail hosts of the parasites. The undesirable species of snails can be exterminated in many localities at small expense. The most effective methods of destroying these snails appear to be drainage, flushing, and poisoning by means of chemicals. The use of organic poisons obtained from plants which are native to the region, or have already been introduced, has not yet been thoroughly investigated. Work along these lines should be initiated. There is some possibility of control by biological means, particularly by the use of ducks and fishes. The information available at present is insufficient to permit conclusions being drawn as to the degree of control which is to be expected from measures of this kind. The acquiescence and co-operation of the native inhabitants with regard to the control measures which are to be instituted is an essential preliminary to all such work. (VI) The most important foci of infection with schistosomiasis on Zanzibar Island are in ponds situated near Muyuni, Mtende, Mbiji, and Kombeni-Ziwani. There are also several other possible foci which are mentioned in the text, and which should be kept under observation. The principal possible foci of infection in Pemba are in marshes situated in the middle course of streams, immediately above the present limit of back-cutting, and to a less extent in ponds situated in the upper parts of streams. In Tanganyika a variety of different habitats are likely to be dangerous from the standpoint of schistosomes. Over the Territory as a whole, however, only a small proportion of the ponds, lakes, and streams are likely to be foci of infection. Suggestions are made regarding means of exterminating the undesirable snails in each of the localities dealt with in the text. Should action be taken along the lines suggested in this report, it is probable that human schistosomiasis will disappear from Zanzibar Island within a period of five years or less. In Pemba and the Tanganyika Territory a very considerable reduction in the infection rate is to be expected after measures have been taken for a somewhat longer period. Eventually the parasites can be exterminated thoroughout this part of East Africa. AND ZANZIBAR PROTECTORATE, THEIR RELATION TO HUMAN SCHISTOSOMIASIS. 743

VIII. REFERENCES TO LITERATURE.

BLACKIE, W. K., 1932. "A Helminthological Survey of Southern Rhodesia," Mem. Lond. Sch. trop. Med., No. 5. CuITwooD, B. G., and CllrrwooD, M. B., 1937. "Snails as Hosts and Carriers of Nematodes and Nematomorpha," Nautilus, vol. i, pp. 130-135. GARNETT, C. B., 1936. "Derris Root," B. Afr. Agric. Journ., vol. ii, pp. 111-113. GORDON, R. M., DAVEY, T. H., and PEASTON, H., 1934. "The Transmission of Human Bilbarziasis in Sierra Leone, with an Account of the Life-Cycle of the Schistosomes concerned, S. mansoni and S. hcematobium," Ann. trop. Med. Parasit., vol. xxviii, pp. 323-414. HAAS, F., 1936. "Binnen-Mollusken aus Inner-Afrika," Abh. senckenb. naturf. Ges., No. 431, pp. 1-156. JUTTING, T. VAN BENTHEM, 1938. "A Freshwater Pulmonate (Physa fontinalis (L.)) inhabited by the Larva of a non-biting Midge (Tendipes (Parachironomus) varus Gtgh.)," Arch. Hydrobiol. Plankt., vol. xxxii, pp. 693-699. KEw, H. W., 1889. "Notes made in 1888 upon Anon ater and some Other Slugs," Naturalist, Lond., pp. 103-107. KHALIL, M., 1930. "The Control of Bilbarziasis in Egypt," Rep. publ. Hith. Labs. Cairo, No. 6, 2nd ed. MARTENS, E. VON, 1897. "Beschalte Weichthiere Deutsch- Ost-Afrikas, " In: STUHLMAN, F., "Deutsch-Ost-Africa," part 4, pp. 1-308. MATHIAS, P., and BOULLE, L., 1933. "Sur une larve de Chironomide (Diptere) parasite d'une Mollusque," C.R. Acad. Sci. Paris, vol. cxcvi, pp. 1744-1796. MILLER, N. C. F., 1935. "The Toxic Value of Dernis spp.," Str. Settlem. Fed. Malay Sts., Dept. Agric. Sci. Studies, No. 16. MOZLEY, ALAN, 1932. "A Biological Study of a Temporary Pond in Western Canada," Amer. Nat., vol. lxvi, pp. 235-249. --, 1934. "Post-Glacial Fossil Mollusca in Western Canada," Geol. May., Lond., vol. lxxi, pp. 370-382. -, 1935. "The Variation of Two Species of Lymncece," Genetics, vol. xx, pp. 452-465. --, 1937. "The Status of the Greenland Lymncea," Proc. malacol. Soc. Lond., vol. xxii, pp. 186-188. ORR, D., and GRANTHAM, D. R., 1931. "Some Salt Lakes of the Northern Rift Zone," Tanganyika Territory, Geol. Surv. Dept., short paper No. 8. PILSBRY, H. A., and BEQUAERT, J., 1927. "The Aquatic Molluscs of the Belgian Congo, with a Geographical and Ecological Account of Congo Malacology," Bull. Amer. Mus. nat. Hist., vol. liii, pp. 69-602. RAYMOND, W. D., 1936. "The Poisonous Properties of some Local Species of Buphorbia," B. Afr. med. Journ., vol. xii, pp. 369-374. 1JLLYETT, H., 1886. "Lymncea stagnalis," Sci. Gossip, vol. xxii, p. 214. WILLIAMSON, K. B., and SCHARFF, J. W., 1936. "Anti-Larval Sluicing," Malay med. Journ., vol. xi, pp. 124-150. WORLSEY, R. R. Le G., 1934, et seq. "The Insecticidal Properties of some East African Plants. Parts I, II, III and IV," Ann. appl. Biol., vol. xxi, pp. 649-669, vol. xxiii, pp. 311-328, vol. xxiv, pp. 651-658, pp. 659-664.

IX. DESCRIPTION OF PLATES.

PLATE I, A.

Figs. 1-6. Physopsis globosa (Morelet). Pond near Tangani, Pemba. Figs. 7-9. Physopsis globosa (Morelet). Pond at Ndomo School near Nzega, Tanganyika Territory. Figs. 10-12. Physopsis globosa (Morelet). Mbiji, Zanzibar. Fig. 13. Lymnwa caillaudi (Bgt.). Stream crossing Ngare-Nairobi road, near Moshi, Tanganyika Territory. Fig. 14. Lymncea caillaudi (Bgt.). Lagoon near Lake Nyasa, Rungwe District, Tanganyika Territory. Fig. 15. Planorbis gibbonsi Nelson. Wani, Pemba. Figs. 16-17. Bulinus tropicus (Krauss). Stream near Nar, Tanganyika Territory. Figs. 18-19. Biomphalania sudanica Martens. Lagoon near Lake Nyasa, Rungwe District, Tanganyika Territory. Figs. 20-21. Biomphalaria pfeifferi (Krauss). Stream near Mwanza, Tanganyika Territory. Figs. 22-23. Melanoides tuberculata (0. F. Muller). Chonga, Pemba. Fig. 24. Ca3latura nuellani (Bgt.). Lake Victoria, near Mwanza, Tanganyika Territory. All figures are natural size. TRANS. ROY. SOC. EDIN., VOL. LIX, PART III, 1938-39 (NO. 26). 111 744 FRESH-WATER MOLLTJSCA OF TANGANYIKA TERRITORY AND ZANZIBAR PROTECTORATE.

PLATE I, B. Fig. 1. Cleopatra ferruginea Lea. Mbiji, Zanzibar. Fig. 2. Cleopatra ferruginea Lea. Kiwanda, Tanganyika Territory. Figs. 3-4. Nerit'ina knorri Récluz. Stream near Sipwese, Pemba. Fig. 5. Lanistes olivaceus (Sowerby). Mbiji, Zanzibar. Specimen with an eroded apex. Figs. 6-8. Thiara vouamica (Bgt.). Chonga, Pemba. Fig. 9. Lanistes olivaceus (Sowerby). Mbiji, Zanzibar. Fig. 10. Pila gradata (Smith). Mbiji, Zanzibar. All figures are natural size. Photographs by W. A. McDonald.

PLATE II. Fig. 1. Pond in the bed of a stream crossing the Dodoma-Mbeya road approximately 75 miles north of Mbeya, Tanganyika Territory. Habitat of Lymna3a caillaudi. Fig. 2. Munyuni pond, Zanzibar. Habitat of Physopsis globosa and Lanistes olivaceus. Fig. 3. Stream crossing the Moski-Arusha railway at km. 215. Habitat of Bulinus tropicus and Biomphalaria pfeifferi. An encrustation of salt leached from the volcanic soils and rocks can be seen on the upper part of the bank. Fig. 4. Rice field, Kidondoni, near Dar-es-Salaam. Habitat of Physopsis globosa. Fig. 5. "Water-hole" at Kimpindu, approximately midway between the Ilindi mbuga and Kitva ya Tembo, 22•5 miles north-west of Dodoma, Tanganyika Territory. This pond is frequently visited by elephants and domestic cattle, so that the bottom is disturbed, and the water is muddy. Possibly as a result of this the pond is barren of mollusca. Fig. 6. Pond near Mwanza-Musoma road, 5 miles east of Mwanza, Lake Province, Tanganyika Territory. Habitat of Physopsis globosa. Fig. 7. Intermittent stream, Tanganyika. The length of the period each year during which this stream contains water is too brief to permit of the existence of Physopsis globosa.

PLATE III. Aerial photograph of a stream situated near Mipoponi, approximately 1 miles north of Mkoani, Pemba. Scale approximately 200 feet to the inch. The upper part of the drainage of the stream is a dry meadow in which no molluscs were found. Below this there is an area of wet meadow in which Physopsis globosa lives. Further downstream the water flows along a well-defined channel in which the current is too swift to permit the existence of P. globosa. Still further downstream the water flows slowly through an area in which rice is planted, and a few individuals of P. globosa are to be found. Below this there is a mangrove swamp (not shown in photograph) from which P. globosa is absent. The course of the stream has been slightly accentuated by the engraver. Photograph by Surveys Department, Zanzibar Government. _3

C, I

Ito AO[ y I U03401d j)tU A10411.IJ iAu iu'j aqj Jo 1?SflhtOW h4 -4d aqL 1 (T

XI'T 'IOI. 1 Vi.. LIN.

Dr A. MOZLEY on "The Fresh-water Mollusea of the Taiigaiivika Territory and Zanzibar Protectorate."—PLATE 11.

- - -

.1-. ..

.. ...

W., • •-:- L&3 fJ,L. • I •lj( 4 4 '' • ,t.

fr .. • kk

- •*•-•.,- .• -•

______ T i ,, ,.. LIX.

Dr A. M0zLEY on Tin' Freh-w,ttr \Teiin'- f tb' Tingativika Territory and Zanzibar Protectorate. !—PLATE III.

P. T4 _

%t' -'•. a %:. I

) , r ,41

*4: 4

; p4i '; •) 1

4 at the following Prices :-

Price to the Price to Vol. I Vol Price to the Price to Public. Fellows. Public. Fellows.

I. £2 2 0 1 '.Cl 11 0 XL. Part 3. £1 6 0 £0 19 6 IV. 0 9 0 0 7 0 Part 4. 1 0 0 0 16 0 V. 0 11 0 0 9 0 XLI. Part 1. 1 1 0 0 15 9 VI. 0 11 6 096 Part 2. 196 120 VII. 0 18 0 0 15 0 . Part 3. 2 5 0 1 13 6 VIII. 0 17 0 0 14 0 I XLII. 2 2 0 •1 11 0 IX. 1 0 0 0 17 0 'JXLIII. 2 2 0 1 11 0 X. 0 19 0 0 16 0 IXLIV. Part 1. 1 18 6 1 9 0 XI. 0 14 6 . 0 12 0 c! Part 2. 1 1 0 0 15 9 XII. 0 14 6 0 12 0 XLV. Part 1. 1 9 0 1 2 0 XIII. 0 18 0 T 0 15 0 Part 2. 1 7 0 1 0 0 XIV. 1 5 0 1 1 0 Part 3. 1 13 9 1 5 3 XV. 1 11 0 160 Part 4. 046 036 XVIII. 2 2 0 1 11 0 XLI. Part 1. 1 1 10 0 16 6 XIX. 3 0 0 2 6 0 Part 2. 1 5 8 0 19 4 XX. Part 1. 0 18 0 0 14 0 Part 3. 1 7 3 1 0 11 Part 4. 0 10 0 ' 0 7 6 General Index to Vols. XXXV.-XLVI. XXI. Part 3. 0 0 7 0 5 3 (1889-1908), with the President's Address XXII. Part 2. 0 10 0 0 7 6 delivered at the opening of the New Rooms of Part 3. 1 5 0 . 1 1 0 the Society, 8th November 1909, etc XXIII. 4 80 310 0 XXIV. 4 3 0 3 9 0 XLVII. Part 1. 0 19 9 0 15 0 XXV. 3 0 0 .. 2 4 6 ,, Part 2. 1 3 0 0 17 4 XXVI. 3 12 0 2 18 0 ,, Part 3. 1 0 10 0 15 6 XXVII. Part 1. 0 16 0 0 12 0 ,, Part 4. 1 7 7 . 1 0 9

I, Part 2. 0 6 0 0 4 6 XLVIII. Part 1. 1 2 9 0 17 2 Part 3. 1 1 0 0 16 0 ,, Part 2. 1 9 6 1 2 5 , Part 4. 1 0 . 0 0 16 0 ,, Part 3. . 1 11 0 1 3 3 XXVIII. Part 1. 1 5 0 1 1 0 ,, Part 4. 0 16. 8 0 12 6 Part 2. 1 5 0 1 1 0 XLIX. Part 1. 0 7 6 0 5 8 Part 3. 0 18 0 0 13 6 ,, Part 2. 1 12 6 1 4 5 XXIX. Part 1. 1 12 0 1 6 0 ,, Part 3. 1 10 0. 1 2 5 Part 2. 0 16 0 0 12 0 ,, Part 4. 1 1 9 0 16 0 XXX. Part 1. 1 12 0 1 6 0 L. Part!. 1 5 9 0 18 9

Part 2. 0 16 0 0 12 0 it Part 2. 1 7 0 1 0 0 Part 3. 0 5 0 0 4 0 ,, Part 3. 1 7 0 1 0 0 Part 4. 0 7 6 0 5 8 ,, Part 4. 1 2 0 0 17 0 XXXI. 4 4 0 3 3 0 LI. Part 1. 1 5 0 1 0 0 XXXII. Part 1. 1 0 0 0 16 0 ,, Part 2. 1 13 0 1 5 0 Part 2. 0 18 0 0 13 6 •,, Part 3. 1 15 0 1 6. 0 Part 3. 210 0 1 17 6 ,, Part 4. 1 11 6 1 3 6 ,, Part 4. 0 5 0 0 4 0 LII. Part 1. 1 7 0 1 0 6 XXXIII. Part 1. 1 1 0 0 16 0 ,, Part 2. 1 10 0 1 2 0 Part 2. 2 2 0 1, 110 ,, Part 3. 2 3 0 1 ii 6 Part 3. 0 12 0 0 9 6 ,, Part 4. 2 13 0 2 0 0 XXXIV. 2 2 0 1 11 0 LIII. Part 1. 2 17 6 2 3 0 XXXV.*Part 1. 2 2 0 1 11 0 ,, Part 2. 1 12 9 1 5 0 Part 2. 1 11 0 1 3 6 ,, Part 3. 2 7 0 1 16 0 Part 3. 2 2 0 1 11 0 L1V. Part 1. 1 5 0 1 0 0 Part 4. 1 1 0 0 16 0 ,, Part 2. 2 3 0 1 11 6 XXXVI. Part 1. 1 1 6 0 16 0 ,, Part 3. 2 10 0 1 17 6 Part 3. 1 0 0 0 16 0 LV. Part 1. 2 3 0 1 11 6 XXXVII. Part 1. 1 14 6 1 5 6 ,, Part 2. 1 17 6 1 8 0 Part 2. 1 1 0 . 0 16 0 ,, Part 3. 2 10 0 1 17 6 Part 3. 0 16 0 0 12 0 LVI. Part 1. 2 3 0 1 11 6 Part 4. 0 7 6 0 5 8 ,, Part 2. 2 3 0 1 11 6 XXXVIII. Parti. 2 0 0 1 10 0 ,, Part 3. 2 10 0 1 17 6 Part 2. 1 5 0 0 19 0 LVII. Part 1, 2 3 6 . 1 12 9 Part 3. 1 10 0 1 3 0 ,, Part 2. 2 4 3 1 13 6 Part 4. 0 7 6 0 5 8 ,, Part 3. 2 3 0 1 12 6 XXXIX. Part 1. 1 10 0 1 3 0 LVIII. Part 1. 2 5 9 1 14 6 Part 2. 0 19 0 0 14 6 ,, Part 2. 2 11 3 1 18 0 Part 3. 2 3 0 1 11 0 ,, Part 3. 2 7 0 1 15 0 Part 4. 0 9 0 0 7 0 LIX. Part 1. 2 0 9 1 10 6 XL. Part 1. 1 5 0 0 19 0 ,, Part 2. 2 1 6 1 11 3 91 Part 2. 1 12 6 1 5 6 * Vol. XXXV, and those which follow, may be had in Numbers, each Number containing a complete Paper. Volumes or parts of volumes not mentioned in the above list are not for the present on sale to the public. Fellows or others who may specially desire to obtain them must apply direct to the Society. Fifty per cent. to be added to the prices of parts in the above list up to and including Vol. XXX. The publications of the Society are on sale through Messrs R. Grant & Son, Ltd., 126 Princes Street, Edinburgh, and Messrs Williams & Norgate, Ltd., 36 TRANSACTIONS OF THE ROYAL SOCIETY OF EDINBURGH.

VOLUME LIX, PART III, 1938-1939.

Studies on Plumage in the Male Brown Leghorn Fowl. By J. P. Gnu, B.Sc., Ph.D., Institute of Animal Genetics, University of Edinburgh. Communicated by Dr A. W. GREENWOOD. (With Five Plates and Five Text-figures.) Price: to Public, 7s. Od.; to Fellows, 5s. 3d. (Issued November 1, 1938.)

The Life-history and Development of Agriolimax agrestis L., the Gray Field Slug. By ROBERT CARRICK, B.Sc., Ph.D., Department of Agricultural and Forest Zoology, University of Edin- burgh. (Now Department of Agriculture, University of Leeds.) Communicated by Dr A. E. CAMERON. (With Thirty-four Text-figures.) Price: to Public, 4s. 6d.; to Fellows, 3s. 3d. (Issued December 8, 1938.)

The Structure and Function of the Alimentary Canal of some Tectibranch Molluscs, with a Note on Excretion. By VERA FRETTER, Ph.D., Department of Zoology, Birkbeck College, University of London. Communicated by Professor H. GRAHAM CANNON, F.R.S. (With Twenty-two Text-figures.) Price: to Public, 6s. Od.; to Fellows, 4s. 6d. (Issued January 31, 1939.)

The Cytology of the Thelytokously Parthenogenetic Saw-Fly Thrinax macula Hi. By Professor A. D. PEACOCK, D.Sc., and ANN R. SANPERSON, B.Sc., Ph.D., Natural History Department, University College (University of St Andrews), Dundee. (With Four Plates.) Price: to Public, 3s. 3d.; to Fellows, 2s. 6d. (Issued February 17, 1939.) Scientific Results of the Voyage of S. Y. "Scotia," 1902-04: The Crinoidea. By Professor C. VANEY, Université de Lyon, and D. DILWYN JOHN, M.Sc., Department of Zoology, British Museum (Natural History). Communicated by Dr A. C. STEPHEN. (With One Plate and Three Text- figures.) Price: to Public, 2s. Od.; to Fellows, is. 6d. (Issued March 13, 1939.)

Differential Fertility in Scotland. Part II. By ENID CHARLES, M.A., Ph.D., Leverhulme Research Fellow. Communicated by Professor LANCELOT HOGBEN, F.R.S. Price: to Public, is. 9d.; to Fellows, is. 3d. (Issued March 27, 1939.) The Fresh-water Mollusca of the Tanganyika Territory and Zanzibar Protectorate, and their Relation to Human Schistosomiasis. By ALAN MOZLEY, Ph.D.(Edin.), Wandsworth Scholar of the London School of Hygiene and Tropical Medicine. (With Three Plates and Ten Text-figures.) Price: to Public, 8s. 6d.; to Fellows, 6s. 3d. (Issued July 6, 1939.)

[For Prices of previous Volumes and Parts see page 3 of Cover.]

PRINTED IN GREAT BRITAIN BY NEILL AND CO.. LTD.. EDINBURGH. A BIOLOGICAL STUDY OF THE SUB-ARCTIC MOLLUSCA ALAN MOZLEY The Johns Hopkins University, Sometime Walter Rathbone Bacon Scholar of the Smithsonian Institution, and Fellow of the National Research Council

ABSTRACT A study of the fresh water and terrestrial mollusca as a representative group of animals of the Sub-Arctic Region, with the object of gaining insight into the history in nature of such a group which has recently migrated into virgin territory. This study has been based on field work by the author in Canada during 1924-31, in northern Asia in 1932 and 1933, and in Fenno- scandia during 1934, supplemented by examination of the collections in many European and American museums. The main results and conclusions are as follows: The molluscan fauna of the Sub-Arctic Region is made up of two hundred and twenty-five species and varieties. One hundred and seventy of these inhabit fresh water, and the remaining fifty-five are terrestrial. These species. are drawn from a wide variety of different phylogenetic groups. It appears that the fauna is made up to a considerable extent of especially hardy representatives of groups which are found in greater variety in the territories lying further to the south. There are in addition a very few families, such as the Lymnaddn, which appear to find conditions in this region specially favorable, and occur in great abundance. The geographical affinities of the Sub-Arctic fauna are as follows: in both Eurasia and North America the Sub-Arctic fauna is closely akin to that which inhabits the temperate zone portions of each continent respectively; in addition to this there are a number of species which are characteristic of and common to all parts of the Sub-Arctic Region. It appears therefore that there have been two series of migrations in this region, viz. (i) from south to north in both Eurasia and North America; and (ii) in a circumboreal direction, by way of an isthmus in the neighborhood of the Behring Strait from North America to Eurasia, and vice versa, and also possibly by way of an isthmus in the neighborhood of Greenland. The distribution of both fresh water and terrestrial species within the Sub-Arctic Region is governed largely by climate and the character of the vegetation. Both of these factors have far-reaching direct and indirect effects. There has been some modification of the original molluscan stock since its migration into this region, but the amount of change is small in comparison with that which has been observed in other parts of the world which have a warmer climate than that of the Sub-Arctic. Some species have undergone little change, while others have been profoundly modified. In certain instances the change appears to have resulted from a rearrangement or segregation of genetic factors which are present in the bulk of the population, rather than a

- change of a fundamental character. Certain species which attain a considerable size in the temperate zone are represented in the Sub-Arctic by dwarf

PROCEEDINGS OF THE AMERICAN PHILOSOPHICAL SOCIETY, VOL. 78, NO. 1, OCTOBER, 1937 147

Reprint Printed in U. S. A.

148 ALAN MOZLEY

varieties. A very few totally distinct species appear to have had their origin in the Sub-Arctic Region. CONTENTS Introduction ...... 148 The Constitution of the Fauna ...... 150 The Distribution of the Mollusca within the Sub-Arctic Region .... 161 The Variation of Some Sub-Arctic Mollusca ...... 177 Conclusions ...... 187

I. INTRODUCTION THE Sub-Arctic Region lies between what are commonly known as the Temperate and Frigid Zones. More specifi- cally it may be defined for the purposes of this account as the territory lying between the fiftieth parallel of north latitude and the Arctic Circle. It thus includes parts of northern Asia, North America, Europe, and Greenland. Obviously this definition cannot be applied too literally, since in certain places near the sea coast, notably in northern Europe, the Temperate Zone extends much further northward than in the interior of the great continental land masses. Some question may arise as to whether this territory possesses any noteworthy features which are peculiar to itself. A consideration of the matter, however, must lead to the conclusion that such a condition does exist. As far as the fauna is concerned, there is a point in the neighborhood of the fiftieth parallel of north latitude at which a relatively rapid change takes place. This may be regarded as the southern boundary of the Sub-Arctic Region. Here within a few degrees of latitude many species of animals which have a wide range in the southern and central parts of their respective continents disappear from the fauna, and their places are taken by other forms, which in many instances have only a limited distribution in the south, or do not occur there at all. This change from a southern and central fauna to a northern one appears to be a more or less direct reaction to the severity of the climate, and in some instances at least, is believed to be brought about by BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 149 • restriction in the types of habitat available. Largely as • result of this condition, the outstanding characteristic of the Sub-Arctic fauna is the predominance of a small number of species of each of the classes of animals which are represented in the fauna. These species appear to be well suited to life under a wide range of conditions within the region; occur in large numbers of individuals; and in different parts of their range may split up into distinct local races. For several years a study has been made of the fresh water and terrestrial mollusca as a representative group of animals in the Sub-Arctic Region, with the object of gaining some insight into the history in nature of such a group which has recently migrated into virgin territory. in this connection it may be mentioned that the greater part of the Sub-Arctic Region has been subjected to severe glaciation within relatively recent times, so that the previously existing fauna was exterminated. The questions to which attention was directed in the course of this investigation were as follows 1. What is the constitution of the Sub-Arctic mollusc fauna? 2. From what geographical and phylogenetic sources has it been derived? What environmental factors govern the distribution of the species within the region To what extent has the original immigrant stock been modified since its migration into the Sub-Arctic Region Studies along these lines were made in Canada during the period 1924-31; in northern Asia in 1932 and 1933; and in Fennoscandia during 1934. In addition to this work in nature, the collections in certain museums were examined, particularly those in, Washington, London, Manchester, Copenhagen, Stockholm, Helsingfors, Leningrad, and Mos- cow. Thanks are due to the curators of these museums, and to many individuals in the several countries in which field work was carried out, for their aid in diverse ways. 150 ALAN MOZLEY The grant of the Walter Rathbone Bacon 'Scholarship of the Smithsonian Institution for the period 1932-34 made it possible to carry out the essential investigations in northern Eurasia, and special acknowledgment is due to t1e officers of that, Institution, and of the United States National Museum for their hearty cooperation. The author is particularly indebted to Doctor Nils Hj. Odhner of the Naturhistoriska Riksmuseet in Stockholm for assistance while studying the Scandinavian fauna; to Professor A. Luther of Helsingfors for help during the stay in Fin- land; to Doctor Chas. H. 0 'Donoghue of the University of Edinburgh for his continued friendly advice and encour- agement; and to Professor E. A. Andrews of the Johns Hopkins University, whose assistance in many ways has made possible the completion of this work.

II. THE CONSTITUTION OF THE FAUNA The total number of species and varieties of non-marine mollusca which are known to inhabit the Sub-Arctic Region is two hundred and thirty-five. One hundred and seventy of these are fresh water forms, and the remaining fifty-five are terrestrial. This number does not include Theodoxus fluviatilis and Clausilia bidentata which barely enter the Sub-Arctic Region. The most striking feature of the composition of the molluscan fauna of this region is that while the species represent a variety of phylogenetic groups, many of the genera (even according to a conservative nomenclature) are represented in any given area by a very few species. It may be that these are protean forms, and that their off- shoots in more southerly latitudes are too specialized to survive in the north. In any event it appears that this fauna is made up to a considerable extent of especially hardy representatives of groups which are found in greater variety in the territories lying to the south. This fact, which implies a severe struggle with environmental conditions, and the elimination of many non-hardy species, is BIOLOGICAL STUDY OF SUB-ARCTIC MOLLTJSCA 151 one of the sources of interest in this fauna. It should be pointed out however, that this does not apply to the fresh water snails of the families Lymnoeidw and Planorbidee which appear to be especially well fitted for life under sub- arctic conditions, and are found in greater variety in this region than elsewhere. The lists of species which are given in this chapter are believed to be reasonably complete. No species of doubtful validity .aincluded. With the exception of members of the family SpheriId, actual specimens of each of the forms listed have been examined by the author.

The Interrelationship of the European, Northern Asiatic and North American Faunas Sixteen species of non-marine rnollusca (10 terrestrial and 6 fresh water) have a circumboreal range, being found in Europe, northern Asia, and North America. The majority of the species which inhabit the northern part of Asia are identical with those found in Europe. There are a few endemic species and varieties in the northern part of Asia (e.g., Lymnra zazurnensis, L. palustris saridalensis and Valvata antiquilina) but most of these are closely related to well known European forms. Apart from the, circumboreal species mentioned above, the molluscan fauna of the far northeastern part of Asia has distinct affinities with that of North America. Examples of this are as follows: The occurrence of Planorbis trivolvis, a common North American snail, as a fossil in the Ohuckchee Pen- insula, Siberia (see Mozley 1935). The occurrence of a species of Ancylus (A. coloradensis) in the Rocky Mountain region of North America. Apart from this the genus Ancylus is known only from the Palarctic Region. A number of species of land snails are said to occur on both sides of the Behring Strait (Dali 1905). The 152 ALAN MOZLEY

series of specimens from this region which have been examined in the course of this investigation have been too small to serve as a basis for a critical opinion. The molluscan fauna of North America is characterized by the presence of numerous species and varieties which do not occur in Europe and northern Asia. There is comparatively little evidence of a direct relationship between the fauna of North America and that of Europe. However, a common European land snail, Helix hortensis, is found in Labrador, Newfoundland, and also on the North American mainland for some distance to the south. It is known from Pleistocene deposits in the State of Maine. As this species is not known to occur in northern Asia, and would hardly have been overlooked if it did occur there, its presence in North America is difficult to account for except on the basis of a former land connection between Europe and eastern North America. The existence of a land bridge connecting North America and Asia in the neighborhood of the Behring Strait on one or more occasions in the past may be regarded as extremely probable.

Catalogue of Sub-Arctic Mollusca I. TERRESTRIAL SPECIES

Class GASTROPODA Order Pulmonata Family Pupi1lid Strobilops affinis Pilsbry, North America Pupilla muscorum (Linné), Europe, northern Africa, Central and Northern Asia, Northern North America Pupilla muscorum lu'ndströmi (Westerlund), Northern Europe and Asia Gastrocopta hoizingeri Sterki, North America BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 153

Family Vertiginid Vertigo ?nodesta (Say), Europe, North America, North- ern Asia Vertigo modesta parietalis (Ancey), North America Vertigo goutdii (Binney), North America S. Vertigo ovata Say, North America Vertigo alpestris Alder, Europe Vertigo aretica Wallenberg, Europe Vertigo lii jeborgi Westerlund, Europe Columella edentula (Draparnaud), Europe, Northern Asia (including Japan), North America

Family VailoniIche Acanthinula harpa Say, Europe, Northern Asia, North America Vallonia costata (Muller), Europe Northern Africa, Northern Asia, North America Vallonia puichelia (Muller), Europe, Northern Africa, Northern Asia, North America Vallonia tenuilabris (Braun), Northern Asia (also known as a fossil from Germany) Vallonia albula Sterki, North America Vallonia parvula Sterki, North America Family Coch1icopida Cochlicopa lubrica (Muller), Europe, Northern Africa, Northern Asia, North America Family Endodontid Gonyodiscus ruderatus (Ferussac), Europe, Northern Asia Gonyodiscus angulosus (Mousson), Northern Asia Gonyodiscus anthonyi (Pilsbry), North America Punctum pygmaum (Draparnaud), Europe, North America Oreo helix stantoni Dali, North America 154 - ALAN MOZLEY Family Arionida Anon sub! uscus (Draparnaud), Europe

Family Zonitithe Zonitoides niticlus (Muller), Europe, Northern Africa, Northern Asia, North America Zonitoides arboreus (Say), North America Zonitoides exigua (Stimp.), North America Retinella radiatula (Alder),. Europe, Northern Asia, North America Retinella binneyana (Morse), North America Euconulus fulvus (Muller), Europe, Northern Asia, North America Euconulus chersinus polygyratus (Pilsbry), North America Vitrina pellucida (Muller), Europe, Northern Asia Vitrina limp'ida Gould, North America Family Limacid Agniolimax lcevis (Muller), Europe, Northern Asia, North America Agniolirnax agrestnis (Linné), Europe

Family Eulotid Eulota fruticum (Muller), Europe, Northern Asia Eulota schrenki (Middendorf), Northeastern Europe Eulota nordenskiöldi (Westerlund), Northern Asia Family Heliciche Helix hortensis Muller, Europe, North America Fruticicola bicallosa (Frivaldsky), Northern Asia Arianta arbustorum (Linné), Europe

Family Succineidoe Succinea pfeiffeni Rossniässler, Europe, Northern Africa, Northern Asia BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 155

Succinea putris (Linné), Europe, Northern Asia Succinea oblonga Draparnaud, Europe Succinea turgida Linné, Northern Asia Succinea altaica Martens, Northern Asia Succinea retusa Lea, North America Succinea ovalis Say, North America Succinea grosvenori Lea, North America Succinea avara Say, North America Succinea avara vermeta Say, North America Succinea hawkinsi Baird, North America Succinea chrysis Westerlund, North America Family Auricu1ida Carychium exile Lea, North America

II. AQUATIC SPECIES

Class GASTROPODA Order Pectinibranchia Family Viviparidee Viviparus fasciatus (Muller), Europe, Northern Asia Campeloma decisum Say, North America Family Paludestrinidc B'ithynia tentaculata (Linné), Europe, Northern Asia, North America Bithynia leachii inflata (Hansen), Europe, Northern Asia Hydrobia ventrosa (Montagu), Europe, Northern Africa, Central and Northern Asia Family Valvatida Valvata piscinalis (Muller), Europe, Asia Minor, Northern Asia - Valvata antiquilina Mozley, Northern Asia S. Valvata aliena Westerlund, Northern Asia 156 ALAN MOZLEY Valvata macrostorna Mörch, Europe, Northern Asia Valvata sib erica Middendorf, Northern Asia Valvata sib erica frigida Westerlund, Europe, North- ern Asia Valvata tricarinata (Say), North America Valvata lewisi Currier, North America Valvata lewisi helicoidea Dali, North America

Family Arnnico1ida Amnicola limosa (Say), North America Amnicola limosa porata (Say), North America Amnicola walkeri Pilsbry, North America Amnicola emarginata (Küster), North America Amnicola winkleyi mozleyi Walker, North America

Order Pulmonata Family Lymnid Lyrnnra stagnalis (Linné), Europe, Northern Asia Lymnra stagnalis jug'ularis Say, North America Lymnva stagnalis sancta3maria3 Walker, North America Lymna3a stagnalis lilliana Baker, North America Lymn63a stagnalis wasatchensis Baker, North America Lymna colurnella casta (Lea), North America Lymncea megasoma (Say), North America Lymnca auricularia (Linné), Europe, Northern Asia Lymnca pereger (Muller), Europe, Northern Asia Lymna3a pereger ovata (Draparnaud), Europe, North- ern Asia Lymna3a zazurnensis Mozley, Northern Asia Lymncea palustris (Muller), Europe, Northern Africa, Northern Asia, North America Lymna3a palustris corvus Gmelin, Europe Lymna palustris kolguevensis Smith, Europe Lymnaa palustris saridalensis Mozley, Northern Asia BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 157 Lyrnna palustris kazakensis Mozley, Northern Asia Lyrnna palustris dravert'i Mozley, Northern Asia Lymiuva palustris bolotensis, Mozley, Northern Asia Lymnwa palustris castorensis Mozley, North America Lymncea vahlii (Möller), Greenland Lyrnna vahlii holbolli Möller, Greenland Lyrnnaa traski (Tryon), North America Lymncl3a lagotis (Schrank), Europe, Northern Asia Lymnera truncatula (Muller), Europe, Northern Asia Lymncea caperata Say, North America Lymncea umbilicata (C. B. Adams), North America Lymnaa parva sterkii Baker, North America Lymna dalli Baker, North America Lymnca modicella (Say), North America Lymncea obrussa decampi (Streng), North America SO. Lymncea exigua (Lea), North America Lymna hedleyi Baker, North America Lymna lanceata (Gould), North America. [ymnca catascopiurn Say, North America Lymnea emarginata (Say), North America Lymnra ema'rginata angulata (Sowerby), North America Lymnea emarginata canadensis (Sowerby), North America Lymna walkeriana (Baker), North America Lyrnna preble'i Dali, North Arrierica Lymna randoiphi Baker, North America Lyrnna3a atkaensis (Dali), North America Lymncea glabra (Muller), Europe, Northern Asia (Interglacial species—Lymnra saskatchewanensis Moziey, North America) Amphipeplea glutinosa (Muller), Europe, Syria, Northern Asia Family Planorbid Planorbis planorbis (Linné), Europe, Northern Africa, Northern Asia 158 ALAN MOZLEY

Pianorbis corneus (Linné), Europe, Northern Asia Pianorbis antrosus Conrad, North America Pianorbis antrosus sayi (Baker), North America Piano rbis antrosus royalensis Walker, North America Pianorbis trivolvis Say, North America Planorbis trivolvis pilsbryi Baker, North America Planorbis trivoivis koiy'mense Lindholm, Far North- eastern Asia Pianorbis corpuientus Say, North America Pianorbis corpuientus 'muiticostatus (Baker), North America Pianorbis infracarinatum (Baker), North America Piano rbis campanulatus wisconsinensis Winslow, North America Piano rbis campanuiatus rudentis Dali, North America Pianorbis campanulatus davisi Winslow, North America Pianorbis exacuous Say, North America Pianorbis exacuous me gas Dali, North America Pianorbis defiectus Say, North America Pianorbis arcticus Möller, Europe, Northern Asia, North America Pianorbis 'umbiiicateii'us Cockereli, North America Planorbis crista (Linné), Europe, Northern Africa, Northern Asia, North America Pianorbis acronicus Ferussac, Europe, Northern Asia Planorbis borealis Loven, Europe, Northern Asia Pianorbis vortex Linné, Europe Planorbis compressus Michaud, Europe, Northern Asia Pianorbis spirorbis (Linné), Europe, Northern Africa, Northern Asia Planorbis johanseni Mozley, Northern Asia Pianorbis leucostoma Millet, Europe, Northern Asia Pianorbis contortus (Linné), Europe, Northern Asia Planorbis compla'natus (Linné), Europe, Northern Asia BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 159

Planorbula armigera (Say), North America Planorbula crassilabris (Walker), North America Planorbula campestris Dawson, North America Segmenti'na nitida (Miller), Europe, Northern Asia Family Physid Physa fontinalis (Linné), Europe, Northern Asia Physa sartlandinensis Mozley, Northern Asia Physa gyri'na Say, North America Physa gyrina hildrethia'na Lea, North America Physa ancillaria Say, North America Physa integra Haldeman, North America Aplexa hypnorum (Linné), Europe, Northern Asia, North America Family Ancylidoe Ancylus lacustris (Linné), Europe, Northern Africa, Northern Asia Ancylus .coloradensis Henderson, North America Ferrissia parallela(Haldeman), North America Ferrissia rivularis (Say), North America

Class PELECYPODA Order Eulamellibranchia Family Unionida Margaritana mar garitif era Linné, Europe, Northern Asia? (specimens not examined), North America U'nio pictorum Linné, Europe, Northern Asia Quadrula q'uadrula (iRafinesque), North America Amblema costata Rafinesque, North America Fusconaia flava (Rafinesque), North America Strophitus rugosus (Swainson), North America Anodonta cygnea (Linné), Europe, Northern Asia Anodonta anatina (Linné), Europe, Northern Asia Anodonta grandis Say, North America 160 ALAN MOZLEY

Anodonta grandis footiana Lea, North America Anodonta kennicotti Lea, North America Mnodonta beringiana, Northern Asia, North Amer- ica (no adequate series of specimens available) Anodontoides ferussac'ianus (Lea), North America Lasmigona complanata katherince (Lea), North America Lasmigona compressa (Lea), North America Proptera alata megaptera (Rafinesque), North America Actinonaias carinata (Barnes), North America Ligurnia recta latissirna (Rafinesque), North America Lampsilis ventricosa (Barnes), North America Lampsilis siliquoidea rosacea (De Kay), North America Lampsilis superiorensis (Marsh), North America Family Sphwriidoe Sphceriurn corneum (Linné), Europe, Northern Asia Sphcerium sulcatum (Lamarck), North America Sphcerium crassum Sterki, North America Sphcerium solidulurn Prime, North America Sphcerium stamineum (Conrad), North America Sphceriurn emarginaturn (Prime), North America Sphcerium torsum Sterki, North America Sphcerium acuminatum (Prime), North America Sphcerium vermontanurn Prime, 'North America Sphcerium striatinum (Lamarck), North America Sphceriwin occidentale Prime, North America Bphcerium tenue (Prime), North America Sphceriurn notaturn Sterki, North America Musculiurn lacustre (Muller), Europe, Northern Africa, Northern Asia Musculium transversum (Say), North America Musculium truncatum (Linsley), North America ilinsculiurn rosaceum (Prime), North America BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 161

Musculium rycholti (Normand), North America Musculium securis (Prime), North America Pisidium amnicum (Muller), Europe, Northern Africa, Northern Asia Pisidium astartoides Sandberger, Europe, Northern Asia Pisidium casertanum (Poli), Europe, Northern Asia Pisidium nitidum Jenyns, Europe, Northern Asia Pisid'ium sub truncatum Maim, Europe, Northern Asia Pisidium henslowanurn (Sheppard), Europe, Northern Asia Pisidiurn supinurn A. Schmidt, Europe, Northern Asia Pisidium parvulum Clessin, Europe, Northern Asia Pisidiurn steenbuchi (Möller), Greenland, Europe, Northern Asia Pisidiuni liii jeborgii Clessin, Europe, Northern Asia Pisidium compressum Prime, North America P'isidium fallaz Sterki, North America Pisidiuin punctaturn Sterki, North America Pisidium variabile Prime, North America Pisidiu'm va'riabile brevius Sterki, North America Pisidium minusculum Sterki, North America Pisidium adamsi Prime, North America Pisidium neglectum Sterki, North America Pisidiurn scutellaturn Sterki, North America Pisidium roperi Sterki, North America Pisidium subrotundurn Sterki, North America Pisidium splendidulum Sterki, North America Pisidium tenuissimurn Sterki, North America Pisidium rotundatum Prime, North America Pisidiurn ferrugineum Prime, North America

III. THE DISTRIBUTION OF THE MOLLTJSCA WITHIN THE SUB-ARCTIC REGION The range of habitats within a given territory in which an organism can survive is to some extent a key to the 162 ALAN MOZLEY geographical distribution of that organism. Nevertheless such a principle as " Local distribution is the function of geographical distribution," cannot be successfully applied in an arbitrary manner, for apart from the possibility of the presence of physical barriers, it sometimes happens that the habits of an animal differ in diverse parts of its geographical range. Unfortunately the information at present available which relates to the distribution of the mollusca is not sufficient to make it possible to follow up this line of thought. In the following account the local distribution of the land and fresh water molluscs in the Sub-Arctic Region is first discussed from the standpoint of general geographical conditions as they find expression in the landscape types. Then a brief description of the kinds of places in which these animals live is given. Finally the most important environmental factors which govern the occurrence of these animals in any particular habitat are summarized.

The Landscape Types The primary types of landscape met with in the Sub- Arctic Region are the Tundra, Taiga, Forest-Steppe, Steppe, and Steppe-Desert. These may be regarded as representing fundamental natural regions in that they possess distinctive features in the nature of the substratum, relief, climate, soils and vegetation. Well defined examples of each of these landscapes are to be found commonly in the northern part of Asia and in North America, but in Europe, as a result of long-continued human occupation they are often difficult to recognize. The species and genera of plants and animals which characterize each of these great natural regions differ considerably in northern Asia and North America. Some diversity, involving the replacement of one species by another is also to be found in different parts of each of these continents. Nevertheless as far as the form of the landscapes is con- BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 163 cerned, there is a marked similarity, or even identity between northern Asia and North America. A question may arise as to whether the principal agencies which contribute to the genesis and morphology of these landscapes are not to be found outside the living organism, and to a great extent in the climate of the region. A case in point is to be found in the form and distribution of certain terrestrial and aquatic habitats on the Forest-Steppe of Canada and Siberia. On both continents the Forest-Steppe is a border-line territory in which the available moisture is barely sufficient to support tree growth even in the most favorable localities. The chief source of water supply is the snow which accumulates during the winter. During the thaw in spring ponds are formed in the vicinity of the largest snowdrifts which generally occur near groves of trees. It is not clear whether the occurrence of a grove of trees, with its associated terrestrial fauna, in a particular place is due to the presence of a pond during the spring, or whether on the other hand the initial factor is the presence of the trees which act as snow collectors, and so result in the presence of the snowdrift. Since the associated terrestrial and aquatic habitats are inter-dependent, and their development is simultaneous it is conceivable that there is a tendency in the natural economy of the region which governs their occurrence. Whether or not this view is a valid one, and can be extended to other natural regions remains to be seen, but it suggests interesting possibilities. The Tundra or Barren Ground is a belt of treeless land which occupies the extreme northern part of Europe, Asia, and North America. Part of this territory is low-lying and water-logged, but there are also considerable areas of dry sandy and rocky country. As far as the climate is concerned, the out-standing characteristic of this region is the low prevailing temperature, and the shortness of the warm season. In all probability this latter condition is an important. one in limiting the variety of the fauna. The fact that part of the country is water-logged has a similar 164 ALAN MOZLEY effect, particularly since a considerable proportion of the remainder of the territory consists of lichen tundra and bare rock, which are equally unsuitable for many organisms. In the southern part of this region, where groups of semi-prostrate trees are occasionally to be found, there may be a few species of terrestrial mollusca,' as for example, Vertigo sp. near Fort Churchill on the shores of Hudson Bay, but for the most part the land snail fauna of the Tundra is a meager one. The most important aquatic habitats on the Tundra are ponds. Although these may have a considerable surface area they are generally shallow, so that the temperature of their waters during the long days of summer may be relatively high. The fauna of such ponds however, is meager in comparison with that of ponds on the forest-steppe and grassland. The Taiga, or Northern Coniferous Forest, forms a broad belt to the south of the Tundra on all the continental land areas of the Sub-Arctic Region. The rich growth of vegetation, which in some respects dominates the animal life of the region, is made possible by the high temperature and long growing season in comparison with the Tundra, and by the abundance of moisture as compared with the grasslands to the south. The vegetative cover over most of this territory consists of continuous forest, composed in the main of species of Finns, Picea, or Larix. Places in which there is too much standing water to permit of the growth of trees are invaded by shrubs such as Ledum and Vaccinium, together with grasses (Festuca, et al.), and mosses including Sphagnum and Polytrichum. In many instances the occurrence of terrestrial mollusca within this territory is profoundly affected by the character of the vegetation. Pure stands of needle-leaved trees, whether Finns, Picea, or Larix, are not suitable habitats for land snails, and as a rule only scattered individuals of two or three species are to be found. Naturalists working in different parts of this region have been inclined to attribute this condition to the presence of resinous ex- BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 165 tracts of decaying pine needles, etc., but it is possible that the absence of suitable food is an equally important factor. The fact that the presence of a relatively small number of birch Or poplar among the conifers generally results in a considerable increase in the abundance and variety of the molluscan fauna, lends some support to this view. Never- theless it must be admitted that the fallen and decaying logs of birch (Betula spp.) support a much richer fauna than do those of the needle-leaved trees. The characteristics of the aquatic habitats of this region may also be considerably affected by the vegetation. Many of the ponds and lakes have been invaded by Sphagnum, grasses and shrubs to such an extent that they have ceased to be suitable habitats for fresh water molluscs. This is the principal reason for the rarity of temporary ponds in the Taiga. In addition to the purely physical action of the semi-aquatic plants in crowding out the aquatic fauna, it is probable that Sphagnum has an inimical effect upon many animals owing to extracts which arise from it or at least are generally associated with its growth. This explanation of the poverty of fauna in peaty waters has been put forward by a number of writers, and is almost pro- verbial among field collectors. West (1910) has advanced the view that peat extracts play an important part in governing the character of the vegetation in many Scottish lochs. It may be worth noting that when such waters are aerated, as occurs in streams with falls and rapids, the flora and fauna becomes much richer. Many of the larger lakes of the Taiga region occupy more or less deep depressions in the bed rock which thus lies exposed along their shores. Such shores are subject to severe wave action, which renders them unsuitable habitats for the majority of fresh water molluscs. A few species however, have become adapted to life under such conditions, and may be found in certain localities in considerable numbers of individuals. In instances where the slope of the shores of a lake is sufficiently gradual, and conditions of exposure and 166 ALAN MOZLEY bottom are suitable, a number of species of aquatic and semi-aquatic plants may be found forming colonies of some size. For the most part these plants are species of Phrag- mites, Typha, Scirpus, Spa'rganium, Carex, Juncus, or Equise turn. Such plant associations provide food and shelter from the waves and so support a large and varied fauna of mollusca. The Forest-Steppe is a narrow belt of groves and glades which is situated between the forested territory described above and the grassland. In the northern part of the For- est-Steppe the groves become gradually larger until they fuse to form a continuous forest, while along the southern border they become gradually smaller and finally disappear altogether, leaving opengrass1and. In North Amer- ica the most common tree on the Forest-Steppe is the Trembling Aspen (Populus tremuloides), while over the greater part of the northern Asiatic Forest-Steppe the prevailing tree species are Birches (Betula verrucosa var. latifolia, and B. pubescens var. allapica). The Forest- Steppe presents a particular combination of conditions which is distinctly favorable for such species of fresh water and terrestrial mollusca as are able to endure the severe climate which prevails in this northerly situation. Mois- ture is abundant at certain seasons, but a sodden condition of the soil and vegetation is seldom met with. The growth of trees, while sufficient to cast an effective shade, and thus to result in some moisture being retained in the forest litter, is not so dense as to prevent the growth of herbs, especially near the borders of the groves. In addition, the leaves of the trees, when they fall to the ground, form the basis of an excellent food supply for land snails. As regards moisture and food therefore, life conditions are suitable for molluscs, but although the number of individuals may be very considerable, there is no great variety in the species which occur in such habitats. This is a characteristic feature of the Sub-Arctic fauna. Similarly favorable conditions are met with in the aquatic habitats. BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 16.7 Sphagnum bogs are rare, but on the other hand, a very high concentration of mineral salts in solution, such as is of frequent occurrence in the waters of the Steppe and Steppe-Desert, is not commonly found on the Forest- Steppe. The marshes of this region are particularly rich, and in many instances have the most varied fauna of any habitats in this Region. The borders of semi-stagnant meandering streams also support a rich fauna, the species for the most part being the same as those found in marshes. It is in the Forest-Steppe that the temporary ponds attain their greatest development, and gave their'most varied and abundant fauna. This is due largely to the reliability of their water supply as a result of the snow in winter collecting about the groves of trees as mentioned in a previous paragraph. In consequence of this the 'ponds have a longer annual aquatic phase than do those in the other natural regions lying to the south, and are more regular in their occurrence year after year. The intermittent streams on the Forest-Steppe may also support a small fauna of mollusca, as for example Planorbis umbilicatellus in an intermittent stream lying to the north of Birtle, Mani- toba; and Lymncl3a palustris et al. in Paskwegin Brook, Saskatchewan. The Steppe or grassland, known in Canada as the Prairie, is characterized by the absence of trees, and the considerably greater concentration of mineral salts in solution in many of the natural waters as compared with those of the Tundra, Taiga, and Forest-Steppe. Both of these conditions are attributable to the occurrence both periodically and at irregular intervals of a serious deficiency of precipitation, and also to the high evaporating power of the air. Except in saline localities there is generally a luxuriant growth of grasses on the Steppe, and these form a thick and continuous turf which reduces to a minimum the erosive action of the floods which follow the thaw in spring. In valleys, or on hillsides, when the turf is once broken deep ravines may be formed, and these are 168 ALAN MOZLFY barren of life. One of the notable features of the Steppe is a marked restriction in the number of different kinds of habitats which are available for settlement by aquatic and terrestrial animals as compared with the Forest-Steppe. With this is correlated a diminution in the number of species which occur in a given individual habitat. These conditions become more pronounced in a southerly direction in continental Asia, and towards the western interior of North America. The periodical shortage of moisture during the summer, combined with the low winter temperatures have the effect of limiting the number of habitats in which terrestrial mollusca are found. For the most part the only places in which these animals occur on the Steppe are marshes, along the borders of streams, and occasionally under shrubs on the flood plains. In a few places on the Kirghiz Steppe of northern Asia empty and bleached shells of land snails have been found in debris carried down by intermittent streams the drainage basins of which are entirely devoid of trees and shrubs. The habitat of these shells is, not known, but it is possible that during moist seasons 'a few terrestrial molluscs may live on the grasses and herbage. An alternative explanation is that the shells found were fossils washed out of some Post- Pleistocene deposit. The marshes of the Steppe region have not as rich a fauna as have those on the Forest-Steppe. In many instances this is due to the presence of a considerable concentration of mineral salts in the waters, and is also to some extent due to the fluctuations in water level in such lakes which results in the destruction of many of the plants (see Mozley 1935, p. 673). Temporary ponds are of common occurrence on the Steppe, but as the duration of their aquatic phase is not as long as is that of those on the Forest-Steppe their fauna does not include as many different species. Some of the Steppe ponds, especially those which occur in valley bottoms, and old lake beds, are saline, and this has the effect Of still further reducing their fauna. Similar conditions considerably intensified, are BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 169 met with in the intermittent streams. Most of the permanent streams of the grassland regions of both northern Asia and North America draw their water supply from neighboring mountainous areas, so that they cannot be regarded as Steppe streams in the strict sense of the term. The fauna of ponds situated on the flood plains of these streams is generally a rich one, and in all probability these habitats serve as natural reservoirs from which much of the neighboring territory is populated with aquatic animals - during years of unusually abundant moisture. The Steppe-Desert is situated to the south of the Steppe proper. The most characteristic feature of this region is the growth of the grasses in compact clumps with bare spaces intervening. This is the result of the much lower precipitation as compared with the Steppe. The deficiency of moisture has an unfavourable effect upon the aquatic fauna, and the majority of the bodies of standing water are saline. The temporary ponds are an exception to this general rule, their waters almost invariably being fresh (except under the conditions noted above—see under Steppe). The prolonged severe droughts which frequently occur in this region however, result in many of the temporary ponds being without water over a period of two years or more. This results in their fauna being somewhat meager. Marsh development is at a minimum, and so the rich association of species which are found in those situations in other natural regions is generally absent from the Steppe-Desert. The only exceptions to this which have been observed have been along the margins of streams which rise in the neighboring mountainous regions. Most of the other streams in the Steppe-Desert are of the intermittent type, and rarely if ever contain molluscs. The next natural region lying to the south, the Desert, is situated outside the Sub-Arctic Region. 170 ALAN MOZLEY The Molluscan Habitats The habitats of non-marine molluscs in the Sub-Arctic fall into two obvious classes, those of the. land areas, and those of fresh water. There is little overlapping between these two types of station except for the occurrence of a few fresh water species such as Lymna truncatula (in Eurasia) and L. parva sterkii (in North America) on mud flats along the borders of streams, and of members of the terrestrial genus Succinea in similar situations and in marshes. Specific examples of the fauna of representative series of molluscan habitats have been given in a previous paper (Mozley 1935). The terrestrial mollusca of the Sub-Arctic Region are found occupying a somewhat limited range of habitats. With few exceptions the land snails which inhabit this territory are forest dwellers: The high degree of uniformity which is to be found in the fauna of widely separated areas may to some extent be correlated with this uniformity of habitat. While it is true that there is some diversity in .the species of trees which make up the forests, this appears to have but little effect upon the character of the fauna. Pro- vided that food of some sort is available, it appears that the trees are of importance principally as providing shade and thus checking evaporation. The Temperate (as distinct from Sub-Arctic) parts of Europe, Asia, and North America each have a distinct land snail fauna which is made up of a large number of characteristic species. Very few of these invade the northern parts of their respective continents, the fauna of those regions being composed of other species which in most instances either have a circumboreal range, or are represented in Eurasia and North America by closely related and ecologically equivalent forms. It is probable that the widely variable conditions of moisture, and the low temperature in winter combine to exclude the majority of the species which are found in the neighboring regions to the south. The most favorable BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 171 habitats for terrestrial molluscs in the Sub-Arctic Region are moist (but not water-logged) situations in forests of Aspen (Populus tremuloides) in North America, and Birch (Betula spp.) in northern Asia. There is much greater diversity i14 the habitats of the fresh water mollusca in the Sub-Arctic Region than in those of the terrestrial species. In this account the habitats are discussed under several headings, namely: the temporary ponds, permanent ponds and small lakes, large lakes, and streams. It should not be assumed however, that these are absolutely distinct types. There are numerous intermediate examples which are difficult to place definitely in one class or another. As a rule the quiet-water habitats have a richer fauna of mollusca than streams. Temporary ponds which are formed by the melting snow in spring are among the most characteristic aquatic habitats in the Sub-Arctic Region. These ponds have a short period of submersion in the spring, which is followed by progressively drier conditions during the summer and autumn, and by low temperatures during the winter. The period of activity of the aquatic animals which are permanent residents of such pools, and not merely migrants is about two months in'each year. It is obvious therefore that the conditions of life for aquatic animals in these habitats are severe; nevertheless a large number of animals belonging to diverse phylognetic groups are able to main- tain themselves in them. Some species in fact are to be found only in temporary ponds. This is true of certain phyllopod crustacea such as Apus, Lepidurus, and Bra'nch- ipus, and also of certain molluscs (see Mozley 1932). It is probable that the richness in species of the temporary pond fauna in the Sub-Arctic Region is to some extent due to the regularity of their occurrence in the spring of each year. This is the result of their being fed by snow water, and not being dependent upon the occurrence of seasonal rains. Another important factor may be the absence of winter thaws. In this connection it may be worth noting 172 ALAN MOZLEY that Shelford (1919) states that the occurrence of Branch- ipus in the Chicago area of North America is extremely sporadic, and indicates that in all probability this is due to the occasional winter thaws during which these temporary pond crustacea. commence their development, only to be killed by subsequent severe frosts before they have reached maturity. In North America and northern Asia the most common temporary pond molluscs are Lymncea palustris and Aplexa hypno'rum. In the course of field work covering an extensive area in Scandinavia only one typical temporary pond was found (near Strömsund, Järntland), and this contained only Aplexa hypnorum. No temporary ponds were found in Finland, and Professor A. Luther of Helsingfors states that to his knowledge no specimen of Branchipus, or any closely related form (Eubranch- ipus, etc.) has ever been collected in that country. Ponds which contain water permanently, and small lakes, are of particular importance as habitats in the Sub- Arctic Region, since they are of common occurrence, and in most instances present conditions which are favorable for molluscan life. For the most part these are shallow bodies of water of such small area that great waves are not formed on the surface by the wind. Their essential characteristic is the quietness of their waters and the abundance of food. This renders possible a greater abundance of living organisms per volume of water than in any other type of aquatic habitat in this region. Submerged vegetation, including such plants as Potamogeton, Elodea, and Utricularia, is usually common, and provides a site for the growth of microscopic organisms which serve as food, either directly or indirectly, of many of the larger animals. This is also true of the rooted plants which have leaves floating upon the surface, e.g. Nyrnphea, Polygonum, and Zizania. Snails and small fresh water mussels are frequently to be found upon such plants in large numbers. The shores of many ponds and small lakes are occupied by marsh plants such as Typha, Phragmites, Scirpus, and BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 173 Sparganium which generally support a varied molluscan fauna. As has been mentioned in a previous paragraph, the presence of Sphagnum in a body of water results in a much reduced mollusc fauna. On the Steppe and Forest- Steppe the fauna of ponds closely resembles that of marshes. In the Taiga the pond fauna is usually meager owing to the presence of Sphagnum, while on the Tundra of Eurasia a limited number of observations indicate that Lymntha pereger, Piano rbis acronicus, Apiexa hypnorum and Pisidium spp. are the most common molluscs. Ponds on the North American Tundra have been found to contain Lymnaa palustris, Planorbis arcticus, Physa gyrinai and Aplexa hypnorum. Ponds and small lakes in several parts of Greenland are inhabited by Lymna vahlii and L. vahli holboili. Lymna palustris kolguevensis inhabits Kolguev Island. The large lakes of the Sub-Arctic Region fall into two classes, namely those which have outlet streams and those which have not. The molluscan fauna of lakes, of the latter type has been discussed in previous papers (Mozley 1930, 1935). The characteristic features of the lakes of the first group are the possession of a moderate or great depth (5 meters or more), and a considerable expanse of open water which results in the shores being subject to wave action. The most striking example of a body of water of this kind in the Sub-Arctic Region is Lake Baikal (depth 1741 m.). Conditions in this lake however, are so extreme that the great majority of characteristic Sub-Arctic animals are unable to exist there. In many of the lesser, though still large lakes of this region, and particularly in North Amer- ica, three distinct molluscan associations are to be found, namely: (i) The exposed shore association: Lymncea stagnaiis lacustris in Europe; Lymnwa emarginata with L. catascopium, or else either L. stagnaiis sanctcemarie or L. stagnaiis lilliarnv in North America. 174 ALAN MOZLEY

The bottom association: Lyrnnwa pereger in Europe; Amnicola iirnosa, A. waikeri, Vaivata tricarinata, Lampsilis siliquoidea and Anodonta grandis footiana (or Anodonta kennicotti) in North America. The marsh association: Lymncea stagnalis, Planorbis pianorbis, and P. corneus in Europe and northern Asia. Lymnca stagnalis jugularis, Piano rbis tn- voivis and Piano rbis arcticus in North America. The molluscan fauna of streams in the Sub-Arctic Re- gion varies widely with the character of the habitat. Slug- gish meandering streams have a fauna which closely resembles that of marshes. In hilly or mountainous country the high inclination of the stream beds, the presence of scree and silt, and the low temperature may act singly or collectively to limit the variety and abundance of the molluscan fauna (see Mozley 1933). There are numerous streams of large size in the Sub-Arctic Region. Such rivers as the Pitea and Kalix (Sweden), the Kemi (Fin- land), Ob, Irtish and Yenissei (northern Asia), and the upper parts of the Athabasca (North America) have a very meager molluscan fauna in their main channels. It is only in the side waters of most of the great rivers of this region that fresh water molluscs are found in any numbers. It is worthy of note however, that several species of Unionid are found in considerable abundance in the Nelson River and the Red River of the North (not the Arctic Red River). Ponds on the flood plains of large streams are important as habitats for molluscs in some parts of this region (see Moz- ley 1935).

Important Environmental Factors Governing the Distribu- tion of Molluscs within the Sub-Arctic Region Two general classes of environmental factors play a predominant part in governing the distribution of both fresh water and terrestrial molluscs in this territory, namely those relating to the climate of the region and those con- BIOLOGICAL STUDY OF SUB-ARCTIC MOLLTJSCA 175 nected with the vegetation. Not infrequently it is difficult to distinguish between these two classes. It may be well to point out that within a single geographical area there may be significant local differences which play an important part in governing the occurrence of molluscs and other animals. On the Forest-Steppe of both Asia and North America for example, the presence of groves of trees has a profound effect upon the fauna by providing food (dead leaves and twigs), and in hindering evaporation from the surface of the soil. As a result of these conditions a number of species of mollusca are able to live in these situations. In comparison with the territories lying to the south, the presence or absence of lime appears to be of minor importance in influencing the distribution of land snails in any particular part of the Sub-Arctic Region. This may be because a certain minimum of lime, sufficient for the needs of the molluscs, is present almost everywhere in the Region. The critical role in determining the presence or absence of these animals therefore passes to other environmental factors. As far as the freshwater molluscs are concerned, the effect of climate and vegetation upon distribution in the Sub-Arctic Region appears principally in five ways, namely: (i) through the mineral salts and other materials in solution; (ii) through the action of waves in beating upon the shores of bodies of standing water; (iii) through marsh development; (iv) through the force of the current in streams; (v) through desiccation. The action of these factors maybe summarized as follows: (i) Mineral Salts in Solution. The changes in climate which have taken place in the Sub-Arctic Region since the end of the last period of glaciation have resulted in certain of the lakes in this area losing their connection with the sea. Owing to the absence of an outlet and continued evaporation there is subsequently a progressive increase in the concentration of mm- 176 ALAN MOZLEY eral salts in solution in the waters of such a lake. This process culminates in a saturated solution and the deposi- tion of salt crystals around the shores as the lake dwindles in size and finally becomes a salt flat. A high concentration of mineral salts has an unfavorable effect upon the molluscan fauna. This is manifested in two ways, namely, in a direct physiological manner, and through the destruction of the aquatic and semi-aquatic vegetation upon which many molluscs are dependent for food and shelter. Wave Action. The action of waves in breaking upon the shores of lakes plays an important part in determining the suitability of such situations for occupation by molluscs. Most of the gastropods which inhabit the Sub-Arctic Region live in shallow water, and apparently find it difficult to exist in large bodies of water where the shore line and neighboring bottom are frequently disturbed by wave action. For the most part any stragglers which find their way into situations of this sort are either broken to pieces or carried into deep water where they are likely to be devoured by fishes. There are, however, several species and varieties of Lym- nwa which are especially adapted for life in habitats which are subject to severe wave action and are unable to exist elsewhere (e.g. Lymnva stagnalis sanctcemarice). Marsh Development. The occurrence of groups of aquatic and semi-aquatic plants in and around ponds, small lakes, and streams has a profound effect upon their suitability for molluscan life. The swamps of the Sub-Arctic Region are dominated by Larch (Larix spp.), and with the trees Sphagnum is usually found in some quantity. This moss also occurs in great abundance in the Sub-Arctic bogs, and as has been mentioned elsewhere in this account, has an inimical effect upon the molluscan fauna. Marshes however provide both food and shelter for many species. BIOLOGICAL STUDY OF SUB-ARCTIC MOLLTSCA 177

Force of the Current in Streams. As has been mentioned under (ii) above the majority of the fresh water molluscs which inhabit the Sub-Arctic Re- gion are adapted, for life only in quiet waters. Any considerable increase in the current of a quiet stream results in the elimination of many species.. While it is true that the inclination of the stream bed plays an important part in determining the force of the current in many streams, the climate and vegetation are significant modifying factors. It frequently happens that the backwaters of streams which have a swift current support a rich fauna of molluscs. Desiccation. The occurrence of severe droughts, both periodically or at irregular intervals, is one of the characteristic features of the climate of the Sub-Arctic Region. In many instances these droughts have the effect of reducing the iñol- luscan fauna to a minimum. A number of aquatic species however, are able to withstand a considerable degree of desiccation (see MQzley 1932).

IV. THE VARIATION OF SOME SUB-ARCTIC MOLLUSCA The repeated glaciation to which the greater part of the Sub-Arctic Region was subject during the Pleistocene period may be considered, for practical purposes, to have resulted in the extermination of the whole of the molluscan fauna which inhabited the area previous to that time (see Mozley 1934). It may be of interest to enquire what modification (if any) the species which migrated into this territory in Post-Pleistocene times have undergone since their immigration. Even a cursory review of the fauna reveals that considerable diversity exists in the degree, and possibly also in the nature, of the change which has taken place. Bearing in mind the observations of Crampton (1916), Sumner (1932), and other students in diverse countries, it cannot be said that profound modifications of the original 178 ALAN MOZLEY immigrant stock have taken place in the Sub-Arctic Region. The relatively short period of time which has elapsed since these animals invaded this territory is probably a contribu- try factor in the small amount of change which has taken place, and it is possible that the low temperatures which prevail over the whole of the Sub-Arctic Region throughout most of the year may also have played a part. What- ever the nature of the process may be it does appear that evolutionary change takes place with greater frequency in countries which have a high prevailing temperature than in those which are subject to cold weather during a considerable part of each year. It is conceivable that this may be a simple physical effect of low temperature. In most instances the diversities which appear to have arisen since these animals invaded this territory in post-glacial times involve a change in the relative proportions of the parts of the shell. Except in certain lacustrine varieties these changes are not supposed to have any survival value. It is interesting to observe that within certain limits each family of fresh water molluscs inhabiting this region tends to vary in a characteristic manner. The Valvatide form a good example of this (Mozley 1935).

Aplexa hypnorum (Linné) Aplexa hypnorum is a fresh water gastropod which is found in considerable numbers throughout the Sub-Arctic Region. In northern Asia this species has been found as far north as N.L. 730 30'. In the region lying to the south of the fiftieth parallel it is seldom found in abundance, and what individuals do occur are of small size. In spite of the immense geographical range of this snail it varies between somewhat narrow limits. The principal diversity in the extreme variants, either within a series of shells from a single habitat, or between individuals from far-distant localities, is found in the length of the spire. Even in this character however the extent of variation in A. hypnorum is much smaller than in many other fresh water pulmonates BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 179 which inhabit this region. Supposed sub-species, e.g. A. hypnorum tryoni, have been described on the basis of shells with elongated spires, but none of these appear to possess really distinctive features. Long-spired individuals are almost equally likely to be met with as isolated individuals, or as forming the bulk of a local population. It should be noted however, that while this conclusion is. based upon the examination of numerous specimens, both in their natural habitat and in museum collections, it has not yet been statistically demonstrated. It appears therefore, that Aplexa hypnorum is a species which varies between narrow limits, and what variation does occur is as likely to be met with in one locality as in another. No tendency towards the accumulation of diversities, or segregation into local races has been observed.

Unionidoe The unionid fauna of Sub-Arctic America has been derived from the valley of the Mississippi River. It is a matter of some interest to observe that several different species of these fresh water mussels appear to have been modified to an approximately equal extent by the severe conditions of life in this northerly situation. That is to say, mussels which attain a large size in the drainage of the Mississippi River are represented in the northern part of the continent by dwarf forms which in some instances are sufficiently distinct to have been given specific or varietal names. The degree of dwarfing is approximately equal in each instance (Pis. I and II). The species in which this is true are the following:

Mississippi Form Sub-Arctic Form

Qnadrula quadrula ...... Unnamed dwarf variety

Anodonta grandis footiana ...... A. icennicotti

Las'migona corn planata ...... L. corn planata katherisnr

Ligumia recta latissisna ...... Unnamed dwarf form

Larnp.rilis radiata ...... L. borealis

Lampsilis sitiquoidea ...... L. siliquoidea rosacea and L. superiorensis Larnpsilis ventricosa ...... Dwarf variety (unnamed) 180 ALAN MOZLEY It seems reasonable to suppose that the cause of this dwarfing lies in the general geographical conditions of Sub- Arctic America. It is probable that the great length of the winter is an important factor. This may affect the fresh water mussels both directly through the long period during which they are in a partly dormant condition, and also indirectly through the food supply which must be similarly responsive to the low temperature and darkness of the winter months. To these influences the several species listed above have reacted in a similar way, and to an approximately equal extent. The structural modifications which have resulted may not in a sense, be evolutionary changes. It is conceivable that they are more of the nature of simple physiological results of the unfavorable conditions which these animals have to endure in the northerly territory. If this be true it is not surprising that the morphological changes have taken place to an approximately equal extent in diverse species. Whether or not such reactions as this have any evolutionary significance is difficult to say, but it is evident that at least as long as the environmental conditions remain as at present, the mussels will remain dwarf forms, with characteristics which mark them off from their southern relatives. What would be the result of transferring the northern mussels to the Mississippi River, it is not possible to predict. Such an experiment might not be relevant to the main fact, namely, that in their natural habitat these animals are structurally distinct from the related forms in more southerly latitudes. The above remarks notwithstanding, it is possible that these diversities may differ from some types of evolutionary change in degree only. It is of some importance to note that not all of the fresh water mussels which inhabit Sub-Arctic America are dwarf forms of southern species. In some instances the Ca- nadian shells are of a size equal to that of average specimens of the same species in the Mississippi drainage. This is known to be true of Amblema costata (specimens taken BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 181 about four miles below the Englishman's Rapid on the Berens River, Manitoba), and also of Anodontoides ferussacianus (specimens taken in the Wliitesand River, near Theodore, Saskatchewan). These two species however, are not among the most common mussels of this region, and for the most part have a rather limited or sporadic distribution within the territory. From this it might appear that they are unable to tolerate the unfavorable conditions which are found over the greater part of this region, and so occur only in especially favorable habitats, in which they attain a size equal to that of Individuals in the Mississippi drainage. Lyrnnida A greater degree of variation is to be found in the fresh water gastropods belonging to the family Lymnid (and the closely related Planorbid) than in any other group of molluscs in the Sub-Arctic Region. Although these two families have a wide range of occurrence over the earth's surface, being found even in Australia, they are nowhere represented by such large numbers of individuals, or by so many diverse types as in the northern part of the northern hemisphere.

Lymncea palustris and L. emarginata Two of the most common members of this family in North America are Lymna palustris and L. emarginata. A comparison of the local and geographical variation of these species in the central part of Canada and the United States (Mozley 1935a) showed that while in L. palustris the range of variation in. several shell characters in one locality approximates that which occurs over the greater jart of the geographical range of this species in North America, in L. emarginata there is a well marked tendency for the range of variation in the same characters to constitute only a small part of that found over the whole geographical area inhabited by the species. That is to say, 182 ALAN MOZLEY in the one instance local variation approximates geographical variation, while in the other it covers only a small part of it. In other words L. emarginata tends to split up into local races, whereas L. palustris does not. It is important to note that this difference in the type of variation is so strongly marked that it can be recognized whenever a large series of shells is examined, even in a casual manner. Stu- dents of the systematics of mollusca have bestowed numerous specific and varietal names upon diverse races and extreme variants of each of these two groups. A statistical demonstration of the essential facts has been given in another paper (Mozley, loc. cit.). The explanation of this diversity in type of variation is believed to lie in the local distribution of the species. Within the territory in which the statistical study was carried out, L. palustris occurs in a high proportion of the very numerous ponds and small lakes; for example, it was found in 238 of a series of 315 such habitats examined. These ponds and small lakes are often situated close together, and it is easy to understand that there are frequent opportunities for the mixture of their fauna. Under these conditions the appearance of recessive characters would be sporadic, the variation in all localities would be expected to be very similar, and its range wide. In contrast to this continuous type of variation is that of L. emarginata, which is found as a rule only on the rocky or bouldery shores of large lakes, a type of habitat which is of relatively rare occurrence in the area where the work was done. Such habitats are often situated one or two hundred kilometers apart. There is thus little opportunity for genetic intermixture between the members of different colonies. Under these conditions recessive characters might be expected to appear in the bulk of the population in certain localities, and thus to form the basis of local races. The conclusions noted in the paragraphs above were drawn from a study of these species in central Canada and BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 183 parts of the northern United States. When the scope of the investigation was extended to include the northern Rocky Mountain area it was found that the LymnTm for the most part occupied isolated habitats in that region and there tended to split up into local races some of which have been given specific names, e.g. Lymnaa traskii Tryon (Moz- ley 1930a). Lymncea emarginata is a purely North American species. L. palustris however, is also found in Europe and northern Asia. Several distinct races of L. palustris were discovered in the latter territory and were described as subspecies (Mozley 1934a). In connection with the appearance of local races of this species in northern Asia the following remarks regarding the local distribution of L. palustris in that region as compared with North America may be of significance, " In collecting over any extensive area in Siberia and northern Kazakstan one is impressed by the fact that this species is less common there than in the comparable regions of North America. In parts of Canada Lymncra palustris (or closely related and homologous forms the systematic position of which is not yet clear) is extremely common, and is to be found in abundance in almost any pond, small lake, and marshy area. . . . This is not true in Northern Asia, and though to the casual observer conditions may appear to be very similar, the proportion of ponds and lakes which is populated by Lymnaa palustris is. much smaller—about one in four or five." Obviously if the explanation of the difference in type of variation observed in L. emarginata and L. palustris in central Canada is valid, it may also apply to the diversity of the variation of the latter species in the Rocky Mountain area and in northern Asia described above.

Lym'ncea stag'nalis Linné Lymna stagnalis, is a large fresh water gastropod which is to be found in abundance in most parts of the Sub- 184 ALAN MOZLEY Arctic Region. The geographical range of this species includes the northern part of North America, northern Asia from Kashmir to Siberia, and very nearly the whole of Europe as well as that part of Africa which lies to the north of the Sahara, and Asia Minor. The northern limit of this species in Great Britain is Edinburgh, while in Sweden it is found as far north as Gellivare (N.Lat. 67° 7', E.Long. 21° 40' approx.). In North America the most common form of Lymna stagnalis is known as L. stagnalis jugularis. This variety is found over the greater part of the continent. Several sub-species have been described from North America the most important of which are L. stagnalis sanctcemarice, L. stagnalis lilliana, and L. stagnalis was atchensis. Each of the three latter forms appears to be specially adapted to the special conditions under which it lives and there is little or no tendency for their areas of distribution (either local or geographical) to overlap. In northern Asia there is one generally prevailing and widely variable form which has been referred to under the name Lymncea stagnalis Linné (Mozley 1935). In the course of extensive collecting in this area no other race of this species was found. In northern Europe there is one common form which is very similar to that found in northern Asia, and in addition numerous more or less distinct races which have been described as sub-species. Westerlund (1886) lists twenty- eight of these from the Palarctic Region, while Paiget (1929) recognizes twelve varieties in addition to the typical form in Switzerland alone. The characteristic ratios of the length of the shells to the length of the aperture in these varieties are stated by Paiget to be as follows: BIOLOGICAL STUDY OF SUB—ARCTIC MOLLUSCA 185

Shell Length Aperture Length L. .stagnaljs f. t...... 1.76 to 1.80 var. producta Coib. ...... 1.80 1.84 vulgaris West . ...... 1.84 1.86 palustriformis Kob. ...... 1.86 1.89 arenaria Coib. ...... 1.80 1.90 generally ...... 1.85 1.90 subulata West (= raphidia Bgt.) ...... 1.87 1.90 elophila Bgt. ...... 1.70 1.75 turgida Mke. ...... 1.65 1.69 'borealis Bgt . ...... 1.65 rhondani Kob. ...... 1.50 1.60 intermedia God. (= media Htm.) ...... 1.50 1.60 lacustri Stud ...... 1.37 1.50 bodainica Cless. ...... 1.30 1.36 According to the evidence advanced by Paiget, some at least of these races breed true for several generations. A study of the races of L. stagnalis in other parts of Europe indicates that these or comparable varieties are not always as well defined as they are reported to be in Switz- erland. It appears that the condition described by Paiget, although probably correct as far as the observations of fact are concerned, is only a special instance of the extremely wide variation exhibited by this species. In northern Europe it frequently happens that one body of water is inhabited by what might be regarded as essentially one race of L. stagnalis. In other instances however, the range of variation among the individuals found in a single lake or pond may be very wide, and as far as certain of the shell characters are concerned, may approach the range of variation which is to be found over the whole of the geographical range of the species. The condition which exists in Finland is of special interest. Many different bodies of water in that country were examined in the course of this investigation, and a marked tendency was found for the L. stagnalis in any single lake or pond to be more or less alike, and recognizably different from those in any other Finnish lake. It therefore appears that there are many local races. It is important to note however, that the intergradation of these 186 ALAN MOZLEY races is much finer than is reported to be the case in Switz- erland, and also that the diverse forms do not appear to be specially adapted to the local conditions under which they live. It is not yet possible to advance a full explanation of the diversity in type variation described above, but since this account is based upon a great many observations there can be little doubt that the general nature of the situation is as stated. It seems to be reasonable to suppose that the root of the diversity lies in multiple factor inheritance and that in some manner not at present -understood, a considerable degree of segregation has occurred in some parts of the Sub-Arctic Region, while in others there has been little or none. This matter requires further investigation. Lymncidce in Isolated Mountain Lakes In several different parts of the Sub-Arctic Region there are mountain lakes which by reason of the ruggedness of the surrounding country, and the torrential character of their outlet streams, are isolated from other bodies of water. That is to say there is little or no opportunity of an intermixture of their fauna with that of neighboring lakes. For the most part the fauna of such lakes is meager. In many instances in fact, no mollusca whatever are to be found in them. It is interesting to observe that when these animals do occur in such situations, there is a marked tendency for the inhabitants of each such body of water to resemble closely one another, and to be recognizably different from all the related forms in the surrounding territory. In other words, they form local races. This is true of the Lymnva in Lake Zazurnia situated in the mountain range known as Khamar Daban, to the east of Lake Baikal. The form which inhabits this lake is closely related to Lymnea pereger, but is distinct from it, and has been named L. zazurnensis. Similar instances are to be found in the northern Rocky Mountain region of North America. In Amethyst Lake, Alberta, for example, there is an unique BIOLOGICAL STUDY OF SUB-ARCTIC MOLLUSCA 187

Lym'na, as yet undescribed, but related to L. palustns and L. traskii. It is unlikely that any changes of a fundamental nature are involved in the production of " new " forms such as these. It is probable that the particular combination of characters which mark off these races has arisen through the segregation in these isolated localities of multiple genetic factors which are present in the bulk of the population of related forms in other areas but are not much in evidence. It is possible to place a different interpretation upon the facts outlined here on the basis of a supposed direct influence of the environment upon the form of the shell, etc. On the whole, bearing in mind the wide range of variation of related forms in other localities, and a certain amount of experimental evidence relating to other species (Paiget, loc. cit.), this explanation appears to be an unlikely one.

V. CONCLUSIONS It will be evident from the foregoing account that much remains to be learned about this fauna. Nevertheless sufficient information is available to justify tentative answers being given to the questions mentioned in the Introduction. The molluscan fauna of the Sub-Arctic Region is made of two hundred and twenty-five species and varieties. One hundred and seventy of these inhabit fresh water, and the remaining fifty-five are terrestrial. These species are drawn from a wide variety of different phylogenetic groups. It appears that the fauna is made up to a. considerable extent of especially hardy representatives of groups which are found in greater variety in the territories lying further to the south. There are in addition a very few families, such as the .Lymna3iclie, which appear to find conditions in this region specially favorable, and occur in great abundance. The geographical affinities of the Sub-Arctic fauna are as follows: in both Eurasia and North America the 188 ALAN MOZLEY Sub-Arctic fauna is closely akin to that which inhabits the temperate zone portions of each continent respectively; in addition to this there are a number of species which are characteristic of and common to all parts of the Sub-Arctic Region. It appears therefore, that there have been two series of migrations in this region, viz. (i) from south to north in both Eurasia and North America; and (ii) in a circumboreal direction, by way of an isthmus in the neighborhood of the Behring Strait from North America to Eurasia, and vice versa, and also possibly by way of an isthmus in the neighborhood of Greenland. The distribution of both fresh water and terrestrial species within the Sub-Arctic Region is governed largely by climate and the character of the vegetation. Both of these factors have far-reaching direct and indirect effects. There has been some modification of the original molluscan stock since its migration into this region, but the amount of change is small in comparison with that which has been observed in other parts of the world which have a warmer climate than the Sub-Arctic. Some species have undergone little change, while others have been profoundly modified. In certain instances the change appears to have resulted from a rearrangement or segregation of genetic factors which are present in the bulk of the population, rather than a change of a fundamental character. Certain species which attain a considerable size in the temperate zone are represented in the Sub-Arctic by dwarf varieties. A very few totally distinct species appear to have had their origin in the Sub-Arctic Region.

LIST OF REFERENCES DALL, W. H. 1905. On the Relations of the Land and Fresh-Water Mollusk- Fauna of Alaska and Eastern Siberia. Popular Science Monthly, 1905, pp. 362-366. CRAMPTON, H. E. 1916 et seq. Studies on the Variation, Distribution, and Evolution of the Genus Partula. Publ. Carnegie Inst. Washington. MozLEY, ALAN. -. Molluscs from the Manitoba-Ontario Boundary. Nau- tilus, XXXIX, pp. 121-128. 1930. Further Records of Western Canadian Mollusca. Nautilus, XLIII, pp. 79-85. 'EI T STUDY OF SUB-ARCTIC MOLLUSCA 189

1930a. Reports of the Jasper Park Lakes Investigations, 1925-26. The Mollusca of Jasper Park. Trans. Royal Soc. Edinburgh, LVI, pp. 647-669. 1932. A Biological Study of a Temporary Pond in Western Canada. Amer. Nat., LXVI, pp. 235-249. 1933. The Local and Geographic Distribution of Some Rocky Moun- tain Mollusca. Proc. Malac. Soc. London, XX, pp. 214-221. 1934. Post Glacial Fossil Mollusca in Western Canada. Geol. Mag., LXXI, pp. 370-382. - 1934a. New Fresh Water Mollusks from Northern Asia. Smith- sonian Misc. Coll., 92, No. 2. 1935. The Fresh-Water and Terrestrial Mollusca of Northern Asia. Trans. Royal Soc. Edinburgh, LVIII, pp. 605-695. 1935a. The Variation of Two Species of Lymna. Genetics, 20, pp. 452-465. 1936. The Statistical Analysis of the Distribution of Pond Molluscs in Western Canada. Amer. Nat., LXX, pp. 237-244. PAIGET, J. 1929. L 'adaptation dela Li'mnaea stagnalis aux milieux lacustres de la Suisse romande. Etude biometrique et génetique. Rev. suisse Zool., 36, pp. 2637531. SHELFORu, V. E. 1919. Nature's Mobilization. Natural History, 19, pp. 201-210. SUMNER, F. B. 1932. Genetic, Distributional, and Evolutionary Studies of the Subspecies of Deer Mice (Peromyscus). Bibliog. Geneticcr, 9, 1-106. WEST, GEORGE. 1910. An Epitome of a Comparative Study of the Dominant Phanerogamic and Higher Cryptogamie Flora of Aquatic Habitat, in Seven Lake Areas in Scotland. In Murray, Bathymetrical Survey of the Scottish Fresh-Water Lochs, I, Edinburgh. WESTERLUND, C. A. 1886 et seq. Fauna der in der Palharktisehen Region Lebenden Binnenconchylien. Berlin and Lund. :

Large valve Lamiqona complanata, Cincinnati, Ohio. X 1. Small valve Lasinigona coinpianata kat/icrinn, Birdtail Creek, near Birtle, Manitoba. X 1. iiia!I \;1k \Ir \ iiiitj g, mar Mat Ink, \laiiit,Ia. X I. Photographs iitncic by the Photographer if the Smithsonian Institution under the direction of the author. NATIONAL RESEARCH COUNCIL of CANADA

THE FRESH-WATER MOLLUSCA OF SUB-ARCTIC CANADA

By ALAN MOZLEY

Reprinted from the CANADIAN JOURNAL OF RESEARCH D, 16 : 93-138. 1938 Reprinted from Canadian Journal of Research, Sec. D, Vol. 16, May, 1938 93

THE FRESH-WATER MOLLUSCA OF SUB-ARCTIC CANADA'

By ALAN MOZLEY2

Abstract A systematic account is given of the constitution, distribution and geographical affinities of the molluscan fauna occurring in fresh waters of sub-arctic Canada. The area covered is that part of western Ontario, Manitoba, Saskat- chewan and Alberta lying north of N. Lat. 49°. A total of 111 species and varieties was collected and identified. Types of habitat available in this region for settlement by molluscs have been classified, and the local distribution of the species in them observed. A brief description of seven principal habitat types, with comments on their fauna in other parts of the sub-arctic region, is followed by a series of examples from specific localities. These molluscan associations, while not necessarily typical, are believed to be representative. The study is concluded with a comparison of the molluscan fauna of northern North America with that of northern Eurasia. Three geographical elements in Canadian sub-arctic Mollusca are: a group of circumboreal species, a large number of strictly North American species, and a group characteristic of this region. An explanation of the close relation between the sub-arctic molluscs and those of the Missis'sippi drainage probably lies in the geological history of the region. There appears to have been a greater degree of speciation in North America than in northern Asia; the total number of species and varieties in sub-arctic Canada is 111, in northern Asia it is only 50. The explanation may le partly in the richer source of supply, the greater facility for migration, and the wider range of habitats available in Canada. Thus new species as they arose would find suitable unoccupied habitats more readily. While this is hardly the sole explanation, the existence of some connection between physiography and speciation appears to be reasonable.

The following account deals with the constitution, distribution, and geographical affinities of the molluscan fauna of fresh waters in sub-arctic Canada. In so far as this report is based upon original observations-in the field, it applies to the western part of the Province of Ontario, and to Manitoba, Saskatchewan and Alberta. Field work was carried on in that area from 1924 to 1931. During these and subsequent years the collections in several museums were studied, particularly those of the United States National Museum, and the British Museum (Natural History). This has made it possible to extend the scope of the work to include the whole of sub-arctic Canada: That is, it covers the greater part of the country lying to the north of N. Lat. 49°. Much of this vast territory remains to be explored, so that it is not possible to claim any degree of completeness in this work. Nevertheless sufficient is known about the mollusca of this region to justify an attempt to evaluate the fauna as a whole for purposes of comparison with other parts of the sub- arctic region. Manuscript received September 4, 1937. 2 Fellow of the National Research Council of the United States of America at the Johns Hopkins University. Present address: do Barclay's Bank Ltd., Dar-es-Salaam, East Africa. 94 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D. The first published account of the molluscs of this region, by G. M. Dawson, appeared in 1875. Other faunal lists, by Whiteaves (37, 38), Christy (11), and Hanham (16), followed. A general revision of the knowledge of the constitution of this fauna was undertaken by Dall during the early years of the present century, and was published in 1905. Since that time numerous short papers dealing with this fauna have appeared, and there has been a great advance in the study as a whole. A revision and extension of Dall's work is therefore needed. Systematic Account This part of the paper consists of a list of the species and varieties which have been examined by the author and found to be distinct from 'all other members of the fauna. Definite localities are given, and also in most instances a concise summary of the geographical range of the diverse forms. A great effort has been made to avoid errors in the identification of the species. During the early part of the work very valuable assistance was rendered by several North American malacologists in the determination of specimens. In this connection thanks are due particularly to Mr. F. C. Baker, and to Dr. Bryant Walker. The late Dr. Victor Sterki was kind enough to examine the Sphaeriidae. Whatever changes may be made in the nomenclature of the group in the future, each of the forms included in this list should be recognizable, as each name is known to represent an actual animal which exists in nature in some numbers.

OUTLINE OF THE CLASSIFICATION OF THE FRESH-WATER MOLLUSCA OF SUB-ARCTIC CANADA

Number of Number of

- species and species and varieties varieties

GASTROPODA PELECYPODA Family Lymnaeidae Family Sphaeriidae Lymnaea 26 Sphaerium 12 Family Planorbidae Musculium 5 Planorbis 17 Pisidium 15 Planorbula 3 Family Unionidae Family Physidae Quadrula 1 Physa 4 Amblema 1 Aplexa 1 Fusconaia 1 Family Ancylidae Strophitus 1 Ancylus 1 A nodonta 3 Ferrissia 2 Anodontoides 1 Family Viviparidae Lasmigona 2 Cam peloma 1 Proptera 1 Family Valvatidae Actinonaias 1 Valvata 3 Ligumia 1 Family Amnicolidae Lampsilis 3 Amnicola 5

Total Gastropoda 63 Total Pelecypoda 48

Total fresh-water mollusca 111 MOZLEV: SUB-ARCTIC MOLLUSCA 95

In addition to the forms that have been collected and examined in the course of this investigation and are included in the list, there are a number of records of other species and varieties in the literature. In the absence of specimens it appears best to leave these out of consideration for the time being. There is little doubt however, that the records and descriptions of Canadian mollusca published within recent years by Mr. F. C. Baker are to be relied upon, so in instances in which the author has not had the opportunity of examining adequate series of these forms, the names are listed at the end of the respective families. This critical policy may have resulted in certain valid species and varieties being ignored in the present paper, but it has the advantage of providing a firm basis for the further study of this interesting group of animals.

Class GASTROPODA FAMILY LYMNAEJDAE Genus Lymnaea s.s. Lamarck 1799

Lymnaea stagnalis jugularis Say Lymnaea jugularis Say, Art. Conchology, Nicholson's Encyc. I. 1817. Geographical Range. L. stagnalis is found in Europe, northern Africa, northern Asia, and North America. The variety jugularis is confined to North America, east and west of the Rocky Mountains, from California and New York to Alaska. Northernmost record: Lake Harrison, N. Lat. 700. southernmost record: N. Lat. 37° in Colorado, 41° in Illinois and Ohio. ,ocal Distribution in Canada. This is one of the commonest fresh-water nolluscs in the western part of Canada, and is included in almost every list )f species from this region. Only new records are included in the list given )elow. The previously reported localities are summarized by Dali (13), md others may be found in the more recent papers listed in the bibliography. 4ew records: Ontario—Oba River; Caramat; creek running into northern end )f Long Lac; Nakina; Savant Lake district, Elbow Lake. 'Ianitoba—Lake Winnipegosis; marsh near the Meadow Portage; Grand Rapids. of the Saskatchewan; Atikameg Lake, Mile 17, Hudson Bay Railway; Thrmorant Lake, Mile 32, Hudson Bay Railway; Resting Lake, Mile 130; ake near Mile 137; creek at Mile 180.6; creek near Wintering Lake, Mile 186; ake near Mile 237; Split Lake; lake near Mile 286. lberta—Chestermere Lake, Calgary; Hastings Lake, Tofield; Hoople Lake, ntwistle; lake near Lac la Biche. '.Torth West Territories—Mackenzie River delta. revious records published since Dall (13): Iglukitaktok, Mackenzie River lelta, Dall (14). Southwest side of Mackenzie River, 30 miles above Fort rovidence; at mouth of Hay River, Whittaker (39). Minaki, Ont., Winnipeg iver, Mozley (21). Indian Bay station, Man.; Sturgeon Creek, near Winnipeg; St. Norbert, LaSalle River; Lockport; Clandeboye, Muckle Creek; 96 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Grand Beach, the Grand Marais; Jackfish Creek near Jackfish Lake; Victoria Beach; Treesbank; Souris River; Baldur, slough near Cobbs Lake; Ninette, marsh at northern end of Pelican Lake, slough 5 miles north of Ninette Douglas Lake near Onah; Theodore, Sask., Powells Lake; Ingebright, Mozley (23). Mud Turtle Lake, near Brereton, Man., Mozley (22). White- shell River district, Man., Little Whiteshell Lake, Crow Duck Lake, White- shell River above Jessica Lake; Molson; Beausejour; Portage Ia Prairie; Macdonald; Delta; Ninette, Bone Lake; Basswood, Long Lake; Clear Lake, Riding Mountains; Dauphin; Lake Winnipegosis, marsh near the Meadow Portage; Kuroki, Sask., Fishing Lake; Humboldt, Burton Lake; Yonker, Eyehill Creek, Mozley (24). Birtle, Man., marshes along the course of Birdtail Creek, also in many ponds and small lakes to the north of Birtle; Kamsack, Sask., pond one-half mile west of Kamsack, also in Madge Lake, 20 miles north of the town; Margo, Margo Lake; Kuroki, lake 1 mile west of Whitesand Lake; Wadena, Foam Lake; Paskwegin, pond near Pask. wegin Brook, also in small lake 2 miles south of Paskwegin siding; Elfros, Brich Creek; Quill Lake station, Quill Creek; Watson, Ironspring Creek; small lakes between Kelliher, Touchwood, and Hudson Bay Lake; Dalesford, creek flowing into Lake Lenore 2 miles south of the hamlet of Dalesford, Mozley (25). Marchington River, Sioux Lookout district, Ont., Baker and Cahn (9). 2 miles west of Invermay, Sask.; north shore of Fishing Lake, near Wadena; 9 miles northeast of Goodeve; Sandy Creek waterworks, near Moose Jaw, Russell (36). Kennedy; Sask.; small lakes in the Moose Mountain Forest Reserve; Shonts, Alta., Amisk Creek; Tofield, Beaverhills Lake; Wabamun, Lake Wabamun, also in the marsh between Wabamun and Kapasiwin; Junkins, Chip Lake, Mozley (30). Fossil deposit on the shore of the Pasquaia River, near The Pas, Man., Mozley (31). Freshwater Lake, Charlton Island, James Bay, Richards (35). Habitat. Ponds which contain water permanently, small lakes, marshes on the shores of larger bodies of water, also in small marshes along the borders of sluggish streams. Lymnaea stagnalis jugularis occurs most abundantly in the prairie and parkland regions. It is not particularly common in the forested region, and ha little toleration of saline conditions. It has not been collected in any temporary ponds. Lymnaea s/a gnalis sanctaemariae Walker L'imnaea stagnalis sanc/aemariae Walker, Nautilus, 6 : 31. 1892. Geographical Range. North America, from the Great Lakes region north- west to Manitoba. Northernmost record: Wekusko Lake, N. Lat. 540 50' (see below). Southernmost record: Lake Michigan in N. Lat. 450• Local Distribution. New record: Wekusko (Herb) Lake, 10 miles west of Mile 81, Hudson Bay Railway, Man. Previous record: Lake Brereton, Man., Mozley (22). Habitat. In large lakes, on smooth rocky shores sloping steeply into deep water in situations subject to severe wave action. In calm weather the MOZLEY: SUB-ARCTIC MOLLUSCA 97 snails are to be found near the surface, but during storms they retire to the deeper water. The young individuals live in slightly more sheltered situations than the adults. The eggs of this variety, together with mature individuals, have been found in rock pools formed by the breaking of the waves on the shores, but this is probably not the usual site of egg-laying. Lymnaea stagnalis lillianae Baker Lymnaea stagnalis lillianae Baker. Geographical Range. North America, from New York west to Manitoba. Northernmost record: Shoal Lake, Manitoba, N. Lat. 490 30' (see below).

Southernmost record: Wisconsin, N. Lat. 46 0 . Local Distribution. Previous records: Indian Bay, Shoal Lake, eastern Man., Mozley (24); Hill, Blackstone, St. Joseph, Bamaji, Fitchie, Abram, and Cat Lakes, Bamaji Lake outlet, Sioux Lookout district, Ont., Baker and Cahn (9). Habitat. This variety has been found only on a sandy shore in Shoal Lake, where it was exposed to severe wave action. Lymnaea stagnalis wasatchensis (Hemphill) Baker Lymnaea stagnalis wasatchensis Baker, Chicago Acad. Sci., Spec. Pub. No. 3 152. 1911. Geographical Range. North America, confined to the western mountainous region. Northernmost record: near Fort Anderson, N. Lat. 68 °. Southernmost record: Utah. Local Distribution. Previous records: numerous localities in Jasper Park, N. Lat. 530, Mozley (26, 30). Habitat. The habitat of this variety in Jasper Park is usually among rooted aqutic plants in small lakes. It is occasionally found in larger bodies of water at relatively low altitudes (3000 to 4000 ft.). Lymnaea (Pseudosuccinea) columella casta (Lea) Lymnea casta Lea, Proc. Am. Phil. Soc. 2 :33. 1841. Geographical Range. North America, Ohio and Illinois north to Nova Scotia and Manitoba. Northernmost record: Lake Brereton, Man. N. Lat. 49 ° 50'. Southernmost record: L. casta, Ohio. L. columella, Florida, N. Lat. 27°. Local Distribution. Previous record: Lake Brereton, eastern Man. Habitat. In Lake Brereton this species was found on rocky shores in shallow bays where it was somewhat protected from wave action. Remarks. Dall (13) records L. columella from Lake Winnipeg, and it may be that his record is based upon specimens of the variety casta. Lymnaea (Bulimnea) megasoma (Say) Lymnaeus niegasoma Say, Rept. Longs Exped. 2 :263. 1824. Geographical Range. North America, from Vermont to Manitoba.

Northernmost record: Echimamish Lake district, Map., N. Lat. 57 0 . Southernmost record: Stark Co., Ohio, N. Lat. 410. 98 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Local Distribution. New records: creek running into the northern end of Long Lac; Savant Lake district, Elbow and Heathcote Lakes, small lake south of Elbow Lake; Allenwater, Ont. Previous records: Indian Bay station, Man., Falcon Bay, Mozley (21). Lake Brereton, Mozley (22). Whiteshell River district, shore of Whiteshell River below the unnamed lake 10 miles below Jessica Lake; quiet bay below third rapid below Betula Lake; near the eighth rapid below Betula Lake, Mozley (24). Rennie, Rennie River, Mozley (25). Marchington River; Bamaji Lake; Cat and Hill Lakes, Baker and Cahn (9). Habitat. Lakes, ponds, and streams. Commonly but not invariably found in situations which are free from wave action. On shores and lake bottoms composed of rock, sand, and mud, and also among aquatic plants and in marshes. In favorable situations it may occur in large numbers. Lymnaea (Stagnicola) caperata (Say) Lymnaeus caperatus Say, New Harmony Diss. 2 :230. 1829. Geographical Range. North America, from Maryland, Colorado and Cali- fornia north to the mouth of the Mackenzie River. Northernmost record: 30 miles north of Aklavik, N. Lat. 68 ° (approx.). Southernmost record: Colorado. Local Distribution: New records: Calgary, Alta.; 30 miles north of Aklavik, N.W.T. Previous records: Molson, Man.; Navin; Sifton; Lake Winnipegosis, marsh near the Meadow Portage, Mozley (24). Fort Frances, Ont.; Grande Pointe, Man.; Birtle; Kamsack, Sask.; Elfros, Brich Creek; Kuroki, Van Patten Creek; Clair, Clair Brook; Quill Lake station, Quill Creek; Watson, Ironspring Creek; Lanigan; Kelliher; Touchwood; Dalesford, creeks flowing into Lake Lenore, Mozley (25). Wainwright, Alta., Viking; Kinsella; Irma; Shonts; Tofield; ponds near Cooking Lake; Lindbrook; Edmonton; Junkins, Chip Lake; Peace River district, 1 mile east of Faust; also along the shores of Lesser Slave Lake near Faust; Wanham, pond in the Birch Hills; creek entering Cadotte Lake from the south; 3 miles north of Spirit River, Mozley (30). Habitat. Usually in temporary ponds, but sometimes to be found in marshes. This is a very common species on the prairies and parkland. It is somewhat less abundant in the forested region. Lymnaea (Fossaria) umbilicata (C. B. Adams) Limnaea umbilicata C. B. Adams, Am. J. Sci. 39 :374. 1840. Geographical Range. North America, New Jersey and Maine west to Mani- toba.

Northernmost record: Birtle, Man., N. Lat. 50 0 251 . Southernmost record: Burlington, New Jersey. Local Distribution. The only known occurrence of this species in the western part of Canada is at Birtle. Habitat. The specimens were collected in a small marsh along the edge of Birdtail Creek. MOZLEY: SUB-ARCTIC MOLLUSCA 99

Remarks. Mr. F. C. Baker identified the shells upon which this record is based. Lymnaea (Fossaria) parva sterkii Baker Lymnaea sterkii Baker, Nautilus, 19 :51. 1905. Geographical Range. North America, New York to Minnesota, Manitoba, and Alberta. L. parva f.t.—"James Bay ...... Maryland ...... Arizona", Baker (1). Northernmost record: Jasper Park, Alta., N. Lat. 53 °. Southernmost record: Hickman Co., Tennessee. Local Distribution. Previous records: Birtle, Man.; Jacques Lake district, tributaries of the'Rocky River, Jasper Park, Alta. Habitat. In both the above localities this snail was found in marshy areas along the border of streams. It appears to prefer muddy flats on which there is little vegetation. Lymnaea (Fossaria) dalli Baker Lymnaea dalli Baker, Bull. Ill. State Lab. Nat. Hist. 7 :104. 1906. Geographical Range. North America, from Kansas and Arizona through Ohio and Montana to Manitoba and Alberta. Northernmost record: Pyramid Lake, Jasper Park, Alta., N. Lat. 530 Southernmost record: southern Arizona, N. Lat. 32° (approx.). Local Distribution. Manitoba, Saskatchewan, and Alberta. Habitat. Marshes and small lakes, usually among vegetation. This species has not been found in ponds, and appears to have little ability to withstand saline conditions. Lymnaea (Fossaria) modicella (Say) Lymnaeus modicellus Say, J. Phil. Acad. 5 : 122. 1825. Geographical Range. North America, Texas to Nova Scotia and California, north to Manitoba. Northernmost record: Ninette, Man., N. Lat. 49 0 25. Southernmost record: Texas. Local Distribution. Previous record: near Round Lake, Ninette. Habitat. The habitat in the above locality is a small mud flat. Lymnaea (Fossaria) obrussa decampi (Streng) Limnaea desidiosa var. decampi Streng, Nautilus, 9 :123. 1906. Geographical Range. North America, Maine to Illinois, Manitoba, and Alberta. Northernmost record: near Fort Providence, Mackenzie River, N. Lat. 61°. Southernmost record: Illinois. Local Distribution. New record: lake near Mile 137, Hudson Bay Railway. Previous records: Southwest side of Mackenzie River, 30 miles above Fort Providence, Mackenzie River district (as above); Mills Lake at mouth of Horn River, Whittaker (39). Clearwater (Atikameg) Lake, Mile 17, Hudson Bay Railway, Mozley (24). Bamaji, Hamilton, Cat, Pashkokogan Lakes, and 100 CANADIAN JOURNAL OF RESEARCH. VOL. 16. SEC. D.

outlet of Kapikik Lake, Ont., Baker and Cahn (9). Postglacial fossil deposit near Lavenham, Man., Mozley (31). Habitat. Lakes of moderate or large size, usually those which have clear and cold waters. Lymnaea (Fossaria) exigua (Lea) Lymnea exigua Lea, Proc. Am. Phil. Soc. 2 :33. 1841. Geographical Range. North America, Tennessee north to Maine and Mani- toba. Northernmost record: creek running into the Grand Marais, at a point two miles east of Balsam Bay, Man., N. Lat. 500 20'. Southernmost record: Tennessee. Local Distribution. Previous records: Mile 69, G.W.W.D.Ry., Man., Mozley (21). Onion Lake, near Minaki, Ont.; Clandeboye, Man., Mozley (24). Bamaji Lake outlet, Sioux Lookout district, Ont., Baker and Cahn (9). Habitat. Usually in small marshy streams, or on the protected shores of small lakes. Lymnaea (Stagnicola) palustris (Muller) Buccinum palustre MUller, Verm. Terr. 2 :131. 1774. Geographical Range. Europe, northern Africa, northern Asia, North America. In North America from New Jersey, Arizona and California, north to James Bay, and Alaska. Northernmost record: lake at N. Lat. 68 ° 20', W. Long. 141°. Southernmost record: Arizona. Local Distribution. New records: Fishing River station; Barrows Junction; Clear Lake, Riding Mountain; Miles 82 and 130, Hudson Bay Railway; small creek near Wintering Lake, Mile 186; ponds on the tundra north of Mile 510; tundra ponds near Esquimo Point, Man. Moose Jaw Creek, Sask. Deep Creek, 12 miles south of Entwistle; near Lac la Biche, Alta. Previous records: Bernard Harbour, Arctic Coast, in part (?) described by Baker (6) as Stagnicola kennicotti, DalI (14); southwest side of Mackenzie River, 30 miles above Fort Providence, Mackenzie River district; south shore Second Lake, Horn River; Mills Lake at mouth of Horn River; at mouth of Hay River; south shore Great Slave Lake, near Sulphur Point, Whittaker (39). Mile 69, G.W.W.D.Ry., Man., Mozley (21). Indian Bay station, eastern Man.; Winnipeg, Catfish Creek; St. Vital; St. Norbert, La Salle River; Clandeboye, Muckle Creek; Goldeye Lake, near mouth of the Red River; Matlock; Balsam Bay, small pond in creek running into the Grand Marais, 2 miles east of Balsam Bay station; Jackfish Lake and Creek; Steep Rock; Aweme, slough near tamarack swamp; Ninette, marshy shores of Pelican Lake; Theodore, Sask., Whitesand River; Edmonton, Alta., Mozley (23). Cartier, Man., Stony Mountain; Portage la Prairie; Macdonald; Dauphin; Sifton; Molson; Delta, Mozley (24). Grande Pointe, Man., Kamsack, Sask.; Madge Lake, Duck Mountain; Kuroki; Paskwegin, Paskwegin Brook, also in Little Quill Lake; Clair, Clair Brook; Quill Creek; Wynyard; Watson, Ironspring Creek; Kelliher; Touchwood; Dalesford, Mozley (25). MOZLE Y: SUB-ARCTIC MOLLUSCA 101

Jasper Park, Alta., pond between Patricia and Katrine Lakes; Lake Mildred, Mozley (26). Wadena, Sask., north shore of Fishing Lake; Moose Jaw, Sandy Creek waterworks, Russell (36). Langbank, Sask.; Kennedy; small lakes in Moose Mountain Forest Reserve; Regina, Pile of Bones Creek; Bethune, near mouth of the Arm River; Wainwright, Alta.; Viking; Phillips; Kinsella; Irma; Shonts; Tofield, Beaverhills Lake, also in Amisk Creek; 3 miles southwest of Tofield; ponds near Cooking Lake; marsh between Wabamun and Kapasiwin; Junkins, Chip Lake; Peace River district, McLen- nan, Kimiwan Lake, Kinuso, Strawberry Creek near Lesser Slave Lake; Wanham, pond in the Birch Hills 4 miles south of Wanham, also in Cadotte (Ká Kut) Lake; 3 miles north of Spirit River; near White Mountain, Spirit River, Mozley (30). Agassiz clay near Fort Frances, Ont.; postglacial fossil deposit on the shore of the Pasquaia River near The Pas, Man., Mozley (31). Abitibi River, 12 miles north of Iroquois Falls, Ont., Richards (35). This species has a very wide range in the western part of Canada, but is more common on the prairie and parkland than elsewhere. Habitat. Temporary and permanent ponds, small and large lakes, streams. L. palustris thrives in temporary ponds in which there is water for only two months in each year. It occurs in large numbers in permanent ponds, small lakes, marshes, and usually in the more sluggish streams. It has also been found on a sandy shore of Lake Winnipeg in a position only slightly protected from wave action. It has a marked resistance to saline conditions, occurring in considerable numbers in parts of the Little Quill Lake, Sask. The waters of this lake were found to contain about 9700 parts per million of dissolved materials, chiefly sodium chloride, sodium sulphate, and magnesium sulphate (see Mozley, 25). Remarks. Lymnaea palustris is an exceedingly variable species, and in the past different names have been applied to the variant forms. Since the range of variation is so great (see Mozley, 32), it appears best to adopt a single general name for the majority of the animals of this group. Subspecific names may be recognized later. Since large series of the species and varieties recently described by Baker (5-7) have not been available, it has been impossible to study them critically, and they are not included in this account. This does not necessarily mean that they are unrecognizable as distinct races, but merely that the writer has been unable to reach any conclusion regarding them.

Lymnaea (Stagnicola) palustris castorensis Mozley Lymnaea traskii castorensis Mozley, Proc. Malac. Soc. London, 20 : 241. 1933. Geographical Range. North America, a pond formed by beavers (Castor canadensis) near Medicine Lake, Maligne drainage, Jasper Park, Alta. This is the only known locality. Remarks. In the original description this form was given as a variety of L. traskii Tryon. As that "species" appears to be merely a variant of the palustris group it may be preferable to treat castorensis as a variety of palustris. 102 CANADIAN JOURNAL OF RESEARCH. VOL. Jo, SEC D.

Lymnaea (Stagnicola) traskii (Tryon) Limnaea traskii Tryon, Proc. Phil. Acad. 149. 1863. Geographical Range. North America, confined to the western mountainous region. Northernmost record: Twin Lakes, Maligne drainage, Jasper Park, Alta. Southernmost record: California. Local Distribution. Banff, and numerous localities in Jasper Park. (Mozley, 26). Habitat. This species is usually found on open shores of large lakes but has also been collected in ponds.

Lymnaea (Stagnicola) saskatchewanensis Mozley Lymnaea vahlii saskatchewanensis Mozley, Am. Mid. Nat. 13 :236. 1932. Geographical Range. North America, interglacial deposit near Beaubier, Sask. Remarks. Mr. F. C. Baker considers that this should be regarded as a distinct species. Lymnaea (Stagnicola) hedleyi Baker Lymnaea hedleyi Baker, Nautilus, 40 :23. 1927. Geographical Range. North America, the headwaters of the Fraser River, British Columbia. Local Distribution. New record: Lucerne, B.C., Fraser River. Previous record: Red Pass Junction, Fraser River (Baker, loc. cit.). Habitat. This species was found in large numbers on stones on the bottom of the Fraser River in places where the current was strong, but the surface of the water unbroken. Remarks. The new record extends the known distribution of this species upstream to within a few miles of the continental divide at the Yellowhead Pass. Lymnaea (Stagnicola) lanceata (Gould) Limnaea lanceata Gould, Proc. Boston Soc. Nat. Hist. 3 :64. 1848. Geographical Range. North America, Ohio and Wisconsin north to Ontario and Manitoba. Northernmost record: Canyon Lake, Ont., N. Lat. 50 ° (approx.). Southernmost record: Summit Co., Ohio, N. Lat. 41°. Local Distribution. New record: Canyon Lake. Previous records: Mile 95, G.W.W.D.Ry., Mozley (23); several localities on the Whiteshell River, Mozley (24); Lake Brereton, Man., Mozley (22). Habitat. Marshes, particularly those near streams or large bodies of water. Lymnaea (Stagnicola) catascopium Say Lymnaea catascoiium Say, Nich. Encyc., Amer. Ed. 2. 1817. Geographical Range. North America, Maryland, Ohio, and Nova Scotia. Northernmost record: Great Slave Lake, N. Lat. 61°. Southernmost record: Maryland. MOZLEY: SUB-ARCTIC MOLLUSCA 103

Local Distribution. Previous records: Lake Winnipeg, Mozley (23). Habitat.. Exposed shores of large lakes, particularly those which are strewn with large boulders. Lymnaea (Stagnicola) emarginata (Say) Lymnaeus emarginatus Say, J. Phil. Acad. 2 :170. 1821. Geographical Range. North America, from New York, Michigan and Ontario to Alberta. Northernmost record: Grand Rapids of the Athabasca, N. Lat. 56 ° (approx.). Southernmost record: New York State, N. Lat. 45°. Local Distribution. New record: Grand Rapids of the Athabasca. Previous records: Lake Winnipeg, Mozley (23); Lake Winnipegosis; Atikameg Lake, Mile 17, Hudson Bay Railway, Mozley (24). Habitat. The shores of large lakes, particularly those which are subject to severe wave action. Lymnaea (Stagnicola) emarginata angulata (Sowerby) Limnaea angulata. Sowerby, Conch. Icon. 18 : Lim. No. 47. 1872. Geographical Range. North America, Michigan to Manitoba. Northernmost record: Winnipeg River, Man., N. Lat. 50 ° 15'. Southernmost record: Michigan. Local Distribution: Previous record: Winnipeg River, near the mouth of the Whitemouth River, Mozley (21). Lymnaea (Stagnicola) emarginata canadensis (Sowerby) Limnaea canadensis Sowerby, Conch. Icon. 18 : Lim., sp. 45. 1872. Geographical Range. North America, from New York north to Anticosti, and west through Michigan and Wisconsin to Manitoba and Alberta. Northernmost record: Northern Twin Lake, Maligne drainage, Jasper Park,

Alta., N. Lat. 53 0 . Southernmost record: southern New York, Baker. Local Distribution. Winnipeg River, near Minaki, Ont., Mozley (21); numerous localities in Jasper Park, Mozley (26). Lymnaea (Stagnicola) walkeriana (Baker) Stagnicola walkeriana Baker, Nautilus, 39 :119. 1926. Geographical Range. North America, Lakes Superior and Michigan, far western Ontario. Northernmost record: Winnipeg River, N. Lat. 50 °. Southernmost record: Lake Michigan, near Sturgeon Bay. Local Distribution. Previous records: English River, near its junction with the Winnipeg River; Lost Lake, near Minaki, Ont. Lymnaea preblei Dali Lymnaea preblei Da 11, Harriman Alaska Exped. 13 :70. 1905. Geographical Range. North America, far western Ontario, northern Mani- toba and Saskatchewan. 104 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Northernmost record: Clear Lake, Sask., N. Lat. 56 °. Southernmost record: English River, Ont. Local Distribution. Previous records: English River; Knee Lake, northern Man.; Lac lie a la Crosse, and Clear Lake. Remarks. This is a recognizable form which will probably prove to be closely related to L. emarginata. L. binneyi is omitted from this account as it appears difficult to recognize the limits of its variation, some of the named specimens approaching L. emarginata very closely.

Lymnaea randoiphi Baker Lymnaea randolph'i Baker, Nautilus, 18: 63. 1904. Geographical Range. North America, Yukon Territory and Alaska. Northernmost record: lake near Cosmos River, north of the Kowak River, Alaska, about N. Lat. 68°. Southernmost record: Lake Lindeman, Y.T., N. Lat. 60°. Local Distribution. Several localities in the Yukon Territory and Alaska. Remarks. ball (13, p. 71) and Baker (1, p. 453) cite a locality in the Liliooet district of British Columbia. The specimens in the United States National Museum from this locality are altogether distinct from L. randolphi. The locality from Washington given by Dall is also erroneous.

Lymnaea atkaensis (Dall) Lymnaea ovata var atkaensis Dali, Proc. U.S. Nat. Mus. 7 : 343. 1884. Geographical Range. Aleutian Islands, Behring Sea. Local Distribution. Lake on the Island of Atka, Aleutian Chain, near Korovin Bay (Dall).

Lymnaeidae Reported on Good Authority to Occur in Sub-arctic Canada, but not Examined by the Author Lymnaea (Stagnicola) alberta (Baker), Brazeau Lake, Alta., Baker (2). This species has not been recognized among the collections from Brazeau Lake made by the present author. Most of the specimens from this lake, and the valley below it as far north as Isaac Creek, were L. traskii or similar members of the palustris group. Lymnaea (Stagnicola) exilis Lea, reported from several localities in the Sioux Lookout district by Baker and Cahn (9). None have been seen by the author. Specimens of L. lanceata (Gould) from Canadian localities have been compared and found to be identical with the type in the collection of the United States National Museum. Lymnaea (Stagnicola) catascopium kempi (Baker and Cahn (9)), reported from several localities in the Sioux Lookout district, Ont. Lymnaea (Stagnicola) palustris alpenensis Baker, reported from Bamaji Lake, Sioux Lookout district, by Baker and Cahn (9). Lymnaea (Stagnicola) kennicotti (Baker) (6), Bernard Harbour, Arctic Canada. Lymnaea (Stagnicola) palustris ungava (Baker) (6), Fort Chimo, Ungava. MOZLEY: SUB-ARCTIC MOLLUSCA 105

FAMILY PLANORBIDAE Genus Planorbis Geoffroy 1767 Planorbjs (Helisoma) antrosus Conrad Planorbis antrosus Conrad, Am. J. Sci. 25 :343. 1834. Geographical Range. North America, from Alabama and Arkansas north to Massachusetts, Illinois and Manitoba.

Northernmost record: Birtle, Man., N. Lat. 50 0 201 . Southernmost record: Alabama. Local Distribution. Previous record: Birtle, Mozley (23). Habitat. These specimens were collected in Birdtail Creek, near Birtle. Remarks. These shells were P. antrosus f.t., but some individuals approached the variety portagensis Baker.

Planorbis (Helisoma) antrosus sayi (Baker) Helisoma antrosa sayi Baker, Bull. Wisc. Geol. & Nat. Hist. Surv. 70 :322. 1928. Geographical Range. North America, New York, the Great Lakes region, Ontario to Alberta. Northernmost record: Knee Lake, northern Man., N. Lat. 55 ° . Southernmost record: ? New York. Local Distribution. New records: Knee Lake. Previous records: Indian Bay station, Man., Falcon Bay; Mozley (23). Lake Brereton; Mud Turtle Lake, Mozley (22). Lake Winnipegosis; Mossy River near Winnipegosis, Mozley (24). Madge Lake, north of Kamsack, Sask., Mozley (25). Pyramid Lake, Jasper Park, Alta., Mozley (26). Wabamun, Alta., Lake Wabamun, Mozley (30). Postglacial fossil deposit, near Lavenham, Man., Mozley (31). Habitat. Usually in small streams or those of moderate size.

Planorbis (Helisoma) antrosus royalensis Walker Planorbis bicarinatus royalensis Walker, Nautilus, 23 :9. 1909. Geographical Range. North America, Isle Royale, Lake Superior; far western Ontario. Northernmost record: Lac Seul district, Ont., N. Lat. 500 (approx.).

Southernmost record: Isle Royale, N. Lat. 48 0 . Local Distribution. Previous records: Bamaji, St. Joseph, Hamilton, Cat and Botsford Lakes; outlet of Bamaji Lake, Ont., Baker and Cahn (9).

Planorbis (Helisoma) trivolvis Say Planorbis trivolvis Say, Nich. Encyc., 1st ed., II, pl. ii, fig. 2. 1817. Geographical Range. North America, far northeastern Asia. In North America, from Tennessee and Missouri north to Alaska. Northernmost record: Fort Yukon, Alaska, N. Lat. 66 0 40'. Southernmost record: Tennessee. 106 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Local Distribution. P. trivolvis has been reported from a great many localities in sub-arctic Canada. It is most abundant on the prairies, and least common in the forested region. There are shells in the collection of the United States National Museum from the Seward Peninsula, Alaska, labelled, "Under 12 ft. muck on cobbles." Baker (8) is of the opinion that many of the previous records of P. trivolvis actually relate to other species, but reports that the typical form is found as far north and west as Wainwright, Alta. Habitat. P. trivolvis is usually found in marshy places along the borders of lakes or along the edges of streams. It has little ability to withstand saline conditions, but appears to be able to withstand desiccation to a very slightly greater extent than Lymnaea stagnalis.

Planorbis (Helisoma) trivolvis pilsbryi Baker Planorbis trivolvis pilsbryi Baker, Nautilus, 39 : 117. 1926. Geographical Range. North America, Massachusetts and New York west to Wisconsin, Ontario, and Manitoba. Northernmost record: Landing River, northern Man., N. Lat. 56 °. Southernmost record: northern New York. Local Distribution. New record: Landing River, second portage from the Hudson Bay Railway on the route to Split Lake. Previous records: Indian Bay station, Man., Falcon Bay; Brereton, Lake Brereton, Mozley (25). Marchington River, 'Sioux Lookout district; Sturgeon Lake, Rainy River district, Baker and Cahn (9). Habitat. In marshes, and on submerged vegetation in situations protected from wave action.

Planorbis (Helisoma) corpulentus Say Planorbis corpulentus Say, Longs Exped. 2 :262. 1824. Geographical Range. North America, northern Michigan and Minnesota, Ontario, Manitoba, and Saskatchewan. This species is confined to the forested region. Northernmost record: Lac lie a la Crosse, Sask., N. Lat. 55 0 451 . Southernmost record: northern Minnesota, Baker. Local Distribution. Numerous localities in far western Ontario, Mozley (21), Baker and Cahn (9). Also known to occur at Falcon Bay, Shoal Lake, eastern Man., Mozley (24). Habitat. Lakes and streams. Usually in somewhat protected situations, but not in marshes. Planorbis (Helisoma) corpulentus multicostatus (Baker) Helisoma corpulentum multicostatum Baker, Nautilus, 46 : 7. 1932. Geographical Range. North America, far western Ontario. Local Distribution. Kahnipiminanikok Lake, Rainy Lake district; Cherry Island, Rainy Lake; Root River (McInnes coll.); Hill, Birch and St. Joseph Lakes, Kenora district; swamp of Lac des Mule Lacs, Sturgeon and Abram Lakes, Thunder Bay district, all in Ontario, Baker (8). MOZLEV: SUB-ARCTIC MOLLUSCA 107

Planorbis (Hel'isoma) infracarinatum (Baker) Flelisoma infracarinatum Baker, Nautilus, 46 : 8. 1932. geographical Range. North America, Ontario, Manitoba, northern United states. 540 Northernmost record: Wekusko Lake, N. Lat. 501. southernmost record: "The United States," Baker. Local Distribution. New records: Isle La Crosse, English River (sic), IJ.S.N.M. No. 29231, Kennicott coil., det. Baker; Wekusko (Herb) Lake, 10 miles west of Mile 81, Hudson Bay Railway. Previous record: Basswood River rapids, Rainy River district, Ont., Baker (8).

Piano rbis (Helisoma) cam panulatus wisconsinensis Winslow Pianorbis cam panulatus wisconsinensis Winslow, 0cc. Papers Mus. Zool. Univ. Mich. 180 :5. 1926. Geographical Range. North America, Michigan and Quebec, west to Illinois, Ontario, and Manitoba. This variety is confined to the forested region. 550 Northernmost record: Landing Lake, Hudson Bay Railway, N. Lat. 201. Southernmost record: Michigan. Local Distribution. New record: Landing Lake. Previous records: Indian Bay station, Man., Falcon Bay; Minaki, Ont., Winnipeg River; Star Lake near Redditt; Alice and Onion Lakes near Minaki; English River near its junction with the Winnipeg River, Mozley (21). Snake Lake, near Indian Bay station, Man., Mozley (23). Brereton and Mud Turtle Lakes, Mozley (22). Several small lakes on Duck Mountain near Madge Lake, north of Kamsack, Sask., Mozley (25). Habitat. Small lakes, and also to some extent in those of moderate size. Frequently found on bare stones and rock faces, but not in the most exposed situations.

Piano rbis (Helisoma) cam panulatus rudentis Dali Pianorbis cam panulatus var. rudentis, Dall, Harriman Alaska Exped. 13 :90. 1905. Geographical Range. North America, Michigan to Manitoba. Northernmost record: Knee Lake, Man., N. Lat. 550 Southernmost record: Marl Lake, Roscommon Co., Michigan. Local Distribution. Knee Lake.

Planorbis (Helisoma) cam panulatus davisi Winslow Planorbis cam panulatus davisi Winslow, 0cc. Papers Mus. Zool. Univ. Mich. 180 : 8. 1926. Geographical Range. North America, Michigan to southern Manitoba. Northernmost record: Douglas Lake, near Onah, Man., N. Lat. 490 45g. Southernmost record: Pinnebog River, Huron Co., Michigan. Local Distribution. Previous record: Douglas Lake, Mozley (24). 108 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Planorbis (Menetus) exacuous Say Planorbis exacuous Say, J. Phil. Acad. 2: 168. 1821. Geographical Range. North America, east of the Rocky Mountains, from New Mexico north to Alaska. Northernmost record: left bank of the Yukon River below Fort Yukon

(Pleistocene marl), N. Lat. 66 0 401 . Southernmost record: New Mexico. Local Distribution. Previous records: south shore Second Lake, Horn River; south shore Lake Kakisa, near mouth of Kakisa River, Mackenzie River district, Whittaker (39). Minaki, Ont., Winnipeg River, Mozley (21). Indian Bay station, Man., Falcon Bay, and Snake Lake; Grand Beach; Douglas Lake near Carberry, Mozley (23). Lake Brereton, Mozley (22). Whiteshell River district, Whiteshell Lake and vicinity; small lake on portage between Whiteshell and Crow Duck Lakes; Macdonald; Lake Winnipegosis, near the Meadow Portage, and on the lake bottom near Snake Island; Clear- water (Atikameg) Lake, Hudson Bay Railway, Mozley (24). Small lake one- half mile west of Madge Lake, north of Kamsack, Sask.; Wadena, Foam Lake; Kuroki, Van Patten Creek; Kelliher; Touchwood, Mozley (25). Hibernia Lake near Jasper, Alta., also in Marjorie, the larger Trefoil Lake, an unnamed lake north of Geikie station, Annette, Caledonia, Edna, and Pyramid Lakes, pond in the bed of Pyramid Creek, and other localities, Mozley (26). Small lakes in Moose Mountain Forest Reserve, Sask.; Phillips, Alta.; Tofield, Beaverhilis Lake; Junkins, Chip Lake; Peace River district, creek 3 miles east of Driftpile; Wanham, Cadotte Lake, 1\'Iozley (30). Cat, Bamaji and Kimmewin Lakes, Sioux Lookout district, Ont., Baker and Cahn (9). Habitat. Temporary ponds, small lakes, marshes, on the bottom of large lakes, e.g., Lake Winnipegosis.

Planorbis (Menetus) exacuous megas Dali Planorbis exacuous var. megas Dali, Harriman Alaska Exped. 13 :91. 1905. Geographical Range. North America, described from Birtle, Man. Stated by Baker (3) to be a northern form found principally in Wisconsin, Michigan, and Minnesota. Local Distribution. Cat, Bamaji, and Kimmewin Lakes, Ont., Baker and Cahn (9). Planorbis (Gyraulus) deflectus Say Planorbis defiectus Say, Rept. Longs Exped. 2 :261. 1824. Geographical Range. North America, Massachusetts vest to Nebraska, and north to Manitoba. Northernmost record: Victoria Beach, Man., N. Lat. 50° 40'. Southernmost record: Nebraska. Local Distribution. Previous records: Victoria Beach; Jackfish Lake, east of Balsam Bay, Man., Mozley; Clearwater Lake, Hudson Bay Railway, Mozley (24). MOZLEY: SUB-ARCTIC MOLLUSCA 109

Remarks. Dali (13) records P. deflectum from Great Slave Lake and Alaska. I'he specimens upon which these records are based are in the U.S. National viuseum and are not deflectus. Some shells of this species approach P. acro- icus Férussac.

Planorbis (Gyraulus) arcticus "Beck" Möller Planorbis arcticus (Beck Mss.) Mölier, Index Moll. Groeni. p. 5. 1842. eographical Range. North America and Greenland. orthernmost record: Greenland or Alaska. Southernmost record: Illinois (fossil), Baker. !ocal Distribution. Previous records: Mackenzie River district, south shore )f Fawn and Second Lakes, Horn River; western end Lake Kakisa; south thore Lake Kakisa near mouth of the Kakisa River; south side of Mills bake, Whittaker (39). Sioux Lookout district, Ont., Kimmewin, Bamaji, :at, St. Joseph, Hamilton, Kapikik, Pashkokogan, and Fitchie Lakes, Baker md Cahn (9). rhis is a common species in the territory extending from Ontario to the \'Iackenzie basin. In the past it has been mistaken for P. parvus Say.

Planorbis (Gyraulus) umbilicatellus Cockereil Planorbis umbilicatellus Cockerell, Conch. Exchange, 2 :68. 1887. eographical Range. North America, New Mexico north to Manitoba. orthernmost record: Touchwood, Sask., N. Lat. 51° 30'. southernmost record: Mesilia, New Mexico, N. Lat. 32° 15'. r.ocal Distribution. Previous records: St. Vital, Man.; Mozley (21). Brere- :on, Mozley (22). Grande Pointe; Birtle; Kelliher, Sask.; Touchwood, Mozley :25). Small lakes and ponds in Moose Mountain Forest Reserve; Wain- rright, Alta.; Viking; Shonts; Tofield; ponds near Cooking Lake; Edmonton; Peace River district, 3 miles north of Spirit River, Mozley (30). Flabitat. P. umbilicatellus occurs only in temporary ponds. The explanation )f the rarity of this species in the forested region lies in the fact that the ;mall basins which form its sole habitat are quickly filled with vegetation in :hat region. Planorbis (Gyraulus) crista (Linné) 1autilus crista Linné, Syst. Nat., ed. 10 :709. 1758. eographical Range. Europe, northern Africa, northern Asia, North America. orthernmost record: Fort Yukon, Alaska, N. Lat. 66° 40'. Southernmost record: Illinois, Baker. ,ocal Distribution. New records: Fort Yukon; Athabasca Delta, Alta. Previous records: Virl Lake, near Jasper, Alta.; also in the smaller Trefoil bake, and Lake No. 6, Mozley (26). Tofield, Alta., Mozley (30). Flabitat. Ponds and small lakes. 110 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Genus Planorbula Haldeman 1842 Planorbula armigera Say Planorbis armigerus Say, J. Phil. Acad. 2 :164. 1818. Geographical Range. North America, Louisiana and Georgia, north to New England and Canada. Northernmost record: ? Manitoba, see below. Southernmost record: Louisiana. Local Distribution. Previous records: at mouth of Hay River, Mackenzie River district, Whittaker (39). Grand Beach, Man., the Grand Marais; Ninette, northern end of Pelican Lake, Mozley (23). Sioux Lookout district, Ont., Bamaji, Cat, St. Joseph, Botsford, Hamilton, and Fitchie Lakes, Baker and Cahn (9). Remarks. With the exception of the records mentioned above the reports of this species from Canada must be regarded as doubtful as they may refer to the more recently described P. crassilabris. Planorbula crassilabris Walker Segmentina crassilabris Walker, Nautilus, 20 : 122: 1907. Geographical Range. North America, Michigan, Iowa, north to Canada. Northernmost record: Balsam Bay, Man., N. Lat. 50 ° 30' Local Distribution. Previous records: Waugh (Indian Bay station) near mouth of the Falcon River, Man., Mozley (20). Arnaud; Balsam Bay, small pond in creek flowing into the Grand Marais, 2 miles east of Balsam Bay station, Mozley (23). Lake Brereton, Mozley (22). Grande Pointe, Mozley (25). Junkins, Alta., Chip Lake; marsh between Kapasiwin and Wabamun; Peace River district, Kinuso, Strawberry Creek near Lesser Slave Lake; pond near Mile 196.7, E.D. & B.C. Ry.; small creek near Faust, Mozley (30). Habitat. Ponds, small lakes, and marshes. Planorbula cam pestris Dawson Geographical Range. North America, from South Dakota to the North West Territories. Northernmost record: Mackenzie River, 30 miles above Fort Providence, N.W.T., N. Lat. 610, Whittaker (39). Southernmost record: Coteau Hills, 5 miles northeast of Clear Lake, South Dakota. Local Distribution. Previous records: Mackenzie River district, south shore Fawn Lake, Horn River; southwest side of Mackenzie River, 30 miles above Fort Providence; Mills Lake at mouth of Horn River; south side Mills Lake, Whittaker (39). Beulah, Man., Wade's slough, Mozley (23). Wadena, Sask., Mozley (24). Birtle, Man., Kelliher, Sask.; Touchwood, near Hudson Bay Lake, Mozley (26). Ponds in Moose Mountain Forest Reserve; Phillips, Alta.; Viking; Tofield; Peace River district, 3 miles north of Spirit River. This species has not yet been found in the densely forested region by the author. The precise habitat is not known of the specimens upon which the MOZLEY: SUB-ARCTIC MOLLUSCA 111 records from the Mackenzie River district are based, but in that region this species may be confined to "islands" of grassland in the forest. Habitat. Temporary ponds, rarely in permanent ponds which have a tem- porarily flooded area surrounding them.

Pianorbidae Reported on Good Authority to Occur in Sub-arctic Canada, but not Examined by the Author Planorbis (Helisoma) whiteavesi (Baker (5)), Lac des Mule Lacs, Thunder Bay district, Ont. Planorbis (Helisoma) plexata Ingersoll (Baker (8) ), Harricanaw River; Black Bear Island, Lake Winnipeg; Cormorant Lake, Man.; Athelstane Lake, Ont. Planorbis (Helisoma) hornii Tryon (Baker (8) ), Fort Simpson, Mackenzie River. Piano rbis (Helisoma) subcrenatus Carpenter (Baker (8) ), New Osgoode, Sask.; Wainwright Park, Alta.; Fawn Lake, mouth of Hay River; Mackenzie River, 30 miles above Fort Providence; little lake west end of Great Slave Lake, Mackenzie River district. Pianorbis (Helisoma) trivolvis macrostomus Whiteaves (Baker (8) ), several localities in northwestern Ontario; lake northwest of Cormorant Lake, Man. Pianorbis (Helisoma) cam panulatus canadensis (Baker and Cahn (9)), Bamaji Lake, Sioux Lookout district, Ont. Planorbis (Gyrauius) latistomus (Baker (5)), McAree Lake, Rainy River district, Ont. Planorbis (Gyraulus) circumstriatus Tryon (Baker and Cahn (9)), Winnipeg River, Ont. Pianorbis (Gyraui'us)' deflectus obiiquus DeKay (Baker and Cahn (9)), several localities in Sioux Lookout district, Ont.

FAMILY PHYSIDAE Genus Physa Draparnaud 1801 Physa gyrina Say Physa gyrina Say, J. Phil. Acad. 2 :171. 1821. Geographical Range. North America, Alabama and Texas north to Great Slave Lake. Northernmost record: Great Slave Lake, N. Lat. 62°, DalI (13). Southernmost record: Texas. Local Distribution. Previous records: Minaki, Ont., Winnipeg River; Sand and Onion Lakes near Minaki; English River 40 miles north of Minaki; Fox Lake, near Wade; Otter Lake near Minaki, Mozley (21). Ninette, Man., Pelican Lake, Mozley (23). Brereton and Mud Turtle Lakes, Mozley (22). Rainy River, Ont., 1 mile below the Canadian National Railway bridge; Grande Pointe; Onah, Douglas Lake, Man.; Margo, Sask., unnamed lake 1 mile northwest of Whitesand Lake; Kuroki, Fishing Lake; Wadena, Foam Lake; Paskwegin, small lake 2 miles south of Paskwegin siding; Kelliher; 112 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Yonker, Eyehill Creek, Mozley (25). Numerous localities in Jasper National Park, Mozley (26). Sioux Lookout district, Ont., Bamaji, Botsford, Hamilton, Pashkokogan, Cat, Fitchie, Kimmewin, and St. Joseph Lakes, Baker and Cahn (9). Scotsguard, Sask.; Blooming; Beaubier, Russell (36). Tofield, Alta., Beaverhills Lake; Deville, Cooking Lake; Peace River district, Kinuso, Strawberry Creek near Lesser Slave Lake; pond at Mile 200.5, E.D. & B.C. Ry.; Faust, Lesser Slave Lake, Mozley (30). Freshwater and Saltwater Lakes on Charlton Island, James Bay, Richards (35). Physa gyrina is particularly common on the prairie and parkland, and also occurs in the Rocky Mountain region. In the forested territory it is replaced to a great extent by P. ancillaria Habitat. Lakes and ponds.

Physa gyrina hildrethiana Lea Physa hildrethiana Lea, Proc. Am. Phil. Soc. 2 :32. 1841. Geographical Range. North America, Alabama, Pennsylvania, Illinois, north to Canada. Northernmost record: Winnipeg, Man. Great Slave Lake (Dall) ? Southernmost record: Alabama. Local Distribution. Winnipeg, Catfish Creek, Mozley (23). Pond in Moose Mountain Forest Reserve, Sask.; Phillips, Alta.; Viking, Mozley (30). Habitat. Temporary ponds and intermittent streams.

Physa ancillaria Say Physa ancillaria Say, J. Phil. Acad. 5 :124. 1825. Geographical Range. North America, New Jersey and Maine, west to Minne- sota and Manitoba. Northernmost record: Madge Lake, north of Kamsack, Sask., N. Lat. 510 30'. Southernmost record: "The Ohio River", Baker (3). Local Distribution. New record: stream entering Lake Nipigon from the east, near Macdiarmid, Ont. Previous record: Madge Lake. Habitat. In the larger lakes, and also sometimes in streams. A common habitat is on smooth rock faces exposed to considerable wave action. The specimens from near Macdiarmid were collected just below a small rapid in the stream. Physa integra Haldeman Physa integra Haldeman, Mon. No. 3, p. 3. 1841. Geographical Range. North America, Ohio and South Dakota, north to Manitoba. Northernmost record: Victoria Beach, Man., N. Lat. 50 ° 40'. Southernmost record: Ohio. Local Distribution. Previous records: Lost and Sword Lakes near Minaki, Ont.; Star Lake near Redditt; Mile 77, G.W.W.D. Ry., Birch River, Man., Mozley (21). Clandeboye, Man., Muckle Creek; Grand Beach, the Grand MOZLE Y: SUB-ARCTIC MOLLUSCA 113

Marais; Victoria Beach; Birtle; Birdtail Creek, Mozley (23). Postglacial fossil deposit near Winnipeg, Mozley (31). Habitat. Lakes and streams. Genus Aplexa Fleming 1820 Aplexa hypnorum Linné Bulla hypnorum Linné, Syst. Nat., ed. X, p. 727. 1758. Geographical Range. Europe, northern Asia, North America. Northernmost record: (in North America) N. Lat. 69° 40', W. Long. 141°. Southernmost record: "South to the vicinity of the Ohio River", Baker (3). Local Distribution. New records: pond on the Arctic coast of the Yukon Territory at N. Lat. 69 ° 40', W. Long. 141°, J. M. Jessup coil.; Bilby, Alta.; Vermilion Lake, Banff, Alta. Previous records: Camden Bay, Arctic coast; also Demarcation Point; pond on Herschel Island; pond at Collinson Point; brackish pond near Teller; Grantley Harbour; Port Clarence, Alaska; pond near Colville mountains, Victoria Island; lake opposite Bernard Harbour; water hole on tundra at Cape Bathurst, Dali (14). Brereton, Man., Mozley (22). Whiteshell River district, Man., above Jessica Lake; Mallard Lake; Little Whiteshell Lake; Zizania marsh below the Whiteshell Lakes; near the Seven Sisters Falls, Winnipeg River; Indian Bay station; Moison; Stony Mountain; Victoria Beach; Treesbank; Steep Rock; Sifton; Yonker, Sask., Eyehill Creek, Mozley (24). Rainy River, Ont.; Grande Pointe, Man.; near Clearwater Lake, Hudson Bay Railway; Birtle; Margo, Sask.; Kuroki, Van Patten Creek, Kuroki Creek; Clair, Clair Creek; Watson, Ironspring Creek; Lanigan; Kelliher; Touchwood, near Hudson Bay Lake, Mozley (25). Lake near Jasper, Alta., Mozley (26). Ponds in Moose Mountain Forest Reserve, Sask.; Lindbrook, Alta.; Deville, Cooking Lake; Peace River district, pond at Mile 200. 5, E.D. & B.C. Ry.; creek near Faust; 3 miles north of Spirit River. Postglacial fossil deposits near Winnipeg, and on the shore of the Pasquaia River near The Pas, Man., Mozley (31). Habitat. Usually in temporary ponds. Occasionally found in lakes and small slow-flowing streams. Physidae Reported on Good Authority to Occur in Sub-arctic Canada, but not Examined by the Author Physajennessi Dali (14). Bernard Harbour, Arctic Canada. P. johnsoni Clench (12). Hot Sulphur Springs, Banff, Alta.

FAMILY ANCYLIDAE Genus Ancylus Muller 1774 Ancylus coloradensis Henderson Ancylus coloradensis Henderson, Nautilus, 44 :31. 1930. Geographical Range. North America, known from one locality in Colorado, and two in Alberta. Apparently confined to the western mountainous region 114 CANADIAN JOURNAL OF RESEARCH. VOL. Jo, SEC. D.

Northernmost record: Two small lakes in the Miette Valley, Jasper National Park, Alta., N. Lat. 53 ° , Mozley (26). Southernmost record: Eldora Lake, Boulder Co., Colorado. Local Distribution. The above-mentioned three localities are the only ones known. Remarks. The occurrence of a member of the genus. A ncylus in North America is of interest since the group in its modern restricted usage is otherwise confined to Europe and northern Asia.

Genus Ferrissia Walker 1903 Ferrissia parallela (Haldeman) Ancylus parallelus Haldeman, Mon. Pt. 2, p. 3. 1841. Geographical Range. North America, northern Ohio, Nova Scotia, Ontario, and Manitoba. Northernmost record: Lake Brereton, Man., N. Lat. 50 ° . Southernmost record: northern Ohio. 'A species of northern distribution", Baker (3). Local Distribution. Previous records: Rennie River, near Brereton, Mozley (22). Whiteshell River, Mozley (24). Pashkokogan Lake, Baker and Cahn (9). Ferrissia rivularis (Say) Ancylus rivularis Say, J. Phil. Acad. 1: 125. 1819. Geographical Range. North America, Ohio and Nebraska north to Canada.

Northernmost record: Birtle, Man., N. Lat. 51 0 251 . Southernmost record: Ohio. Local Distribution. Previous records: Birdtail Creek, Birtle, Mozley. Post- glacial fossil deposit near Winnipeg, Mozley (31).

Ancylidae Reported on Good Authority to Occur in Sub-arctic Canada, but not Examined by the Author Ferrissia fusca (C. B. Adams). Bamaji Lake, Sioux Lookout district, Ont., Baker and Cahn (9).

FAMILY VIVIPARIDAE Genus Campeloma Rafinesque 1819 Cam peloma decisum (Say) Limnaea decisa Say, Nich. Encyc. I, 1817. Geographical Range. North America, New England to Manitoba. Northernmost record: Berens River, above the junction with the Etomami River, N. Lat. 52° 20'. Southernmost record: Tennessee. Local Distribution. Previous records: 'vVhiteshell River district, Man., Whiteshell River near first portage below White Lake; between fifth and sixth rapids below Betula Lake; Berens River several miles above the junction MOZLEY: SUB-ARCTIC MOLLUSCA 115 with the Etomami River, Mozley (24). Fort Frances, Ont., Rainy River below the town, also near the town of Rainy River; Grand Beach, Man., the Grand Marais, Mozley (25). Viviparidae Reported on Good Authority to Occur in Sub-arctic Canada, but not Examined by the Author Campeloma milesii (Lea). Hamilton Lake, Sioux Lookout district, Ont., Baker and Cahn (9).

FAMILY VALVATIDAE Genus Valvata Muller 1774 Valvata tricarinata (Say) Cyclostoma tricarinata Say, J. Phil. Acad. 1: 13. 1817. Geographical Range. North America, from Virginia and the Ohio River to Alaska. Northernmost record: Fort Yukon, N. Lat. 66 ° 40'. Southernmost record: Virginia. Local Distribution. Previous records: Mackenzie River district, south shore Second Lake, Horn River; southwest side of Mackenzie River 30 miles above Fort Providence; western end of Lake Kakisa; south shore of Lake Kakisa near mouth of Kakisa River, Whittaker (39). St. Norbert, Man., La Salle River; Clandeboye, Muckle Creek; Victoria Beach; Treesbank, Souris River; Theodore, Sask., Whitesand River, Mozley (23). Malachi, Ont., Malachi Lake; Whiteshell River district, Man., Betula Lake; between sixth and seventh rapids below Betula Lake; Morris, Morris River; Delta; Ninette Bone Lake; Clearwater (Atikameg) Lake, Hudson Bay Railway; Yonker, Sask., Manitou Lake (empty shells only), Mozley (24). Madge Lake north of Kamsack, Sask., Mozley (25). Sioux Lookout district, Ont., outlet of Bamaji Lake, Fitchie, Kimmewin, Botsford, and Cat Lakes, Baker and Cahn (9). Deville, Alta., Cooking Lake; Wabamun, Lake Wabamun, Mozley (30). Post- glacial fossil deposit near Winnipeg, and also near Portage la Prairie and Lavenham, Man., Mozley (31). Moose Factory, Ont., Butler and Charles Islands (recent and fossil), Shipsands Island, James Bay, Richards (35). Habitat. Usually found in lakes, especially in shallow water, sometimes in large ponds and marshes. Valvata lewisi Currier Valvata lewisi Currier, Kent Sci. Inst. Misc. Pub., No. 1, p. 9. 1868. Geographical Range. North America, the northern part of the continent from the Atlantic to the Pacific. Northernmost record: Bernard Harbour, N. Lat. 68°. Southernmost record: "Southward range not fully known." Baker. Local Distribution. Previous records: Arctic Coast, creek at Bernard Har- bour; lake near Point Williams, Victoria Island, Dali (14). South shore Fawn and Second Lakes, Horn River, Mackenzie River district; southwest side Mackenzie River 30 miles above Fort Providence; south shore Great 116 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Slave Lake near Sulphur Point, Whittaker (39). Numerous localities in Jasper National Park, Alta., Mozley (26). Postglacial fossil deposit on the shore of the Pasquaia River near The Pas, Man., Mozley (31). Habitat. Usually in marshes surrounding relatively large bodies of water. Valvata lewisi helicoidea Dail Valvata lewisi var. helicoidea Dail, Harriman Alaska Exped. 13 :123. 1905. Geographical Range. North America, said to occur with the typical lewisi throughout its range. Northernmost record: Fort Yukon, Alaska, N. Lat. 66° 49'. Southernmost record: Oneida Lake, New York. Local Distribution. Previous records: Mackenzie River district, south shore Second Lake, Horn River; western end Lake Kakisa; Mills Lake at mouth of Horn River; south side Mills Lake, Whittaker (39). Lake Winnipeg, C. H. O'Donoghue coil., Baker (3). Kimmewin Lake, Sioux Lookout district, Baker and Cahn (9).

Valvatidae Reported on Good Authority to Occur in Sub-arctic Canada, but not Examined by the Author Valvata tricarinata perconfusa Walker, Baker and Cahn (9), outlet of Bamaji Lake, Sioux Lookout district, Ont.

FAMILY AMNICOLIDAE Genus Amnicola Gould and Haldeman 1841 A mnicola limosa (Say) Paludina limosa Say, J. Phil. Acad. 1 : 125. 1817. Geographical Range. North America, Texas to New England, Manitoba (and northward ?). Northernmost record: Lake la Loche, N. Lat. 56° 30', Dail (13) = limosa? The northernmost source of shells seen by the writer is near Winnipeg, Man., (postglacial fossils, Mozley (31) ). Southernmost record: Texas. Local Distribution. Previous records: Winnipeg; Lake Brereton (limosa?) Mozley (22). Amnicola limosa porata (Say) Paludina porata Say, J. Phil. Acad. 2 :174. 1821. Geographical Range. North America, Illinois to Manitoba. Limits of range not known. Local Distribution. Previous records: Lake Winnipeg, Victoria Beach, Man., Mozley (24). Sioux Lookout district, Ont., Pashkokogan, Hamilton, Kapikik, Cat, Botsford and Kimmewin Lakes, Baker and Cahn (9). Upper Tintah or Norcross Beach of Glacial Lake Agassiz near Lavenham, Man.. Mozley (31). MOZLEY: SUB-ARCTIC MOLLUSCA 117

Amnicola walkeri Pilsbry Amnicola walkeri Pilsbry, Nautilus, 12 : 43. 1898. Geographical Range. North America, from Ottawa, Ont., to Lake Michigan and Manitoba. Northernmost record: Lake Winnipeg, Man. Southernmost record: Lake Michigan ? Local Distribution. Previous records: Victoria Beach, Lake Winnipeg, Mozley (24). Kimmewin Lake, Sioux Lookout district, Ont., Baker and Cahn (9). Upper Tintah or Norcross Beach of Glacial Lake Agassiz near Lavenham, Man., Mozley (31).

Amnicola winkleyi mozleyi Walker Amnicola winkleyi mozleyi Walker, Nautilus, 39 :6. 1925. Geographical Range. North America, Manitoba and Ontario. Northernmost record: Winnipeg, Man. Local Distribution. Previous records: two postglacial deposits near Win- nipeg, Mozley (31). Remarks. A shell which closely resembles this variety was collected in the La Salle River at St. Norbert, Man., but as it is a single specimen it appears best not to cite this as a definite locality until further collections are made.

Amnicola (Cincinnatia) emarginata (Kuster) Paludina emarginata Küster, Paludina, Conch. Cab. p. 50. 1852. Geographical Range. North America, Kentucky and Arkansas north to Canada. Northernmost record: "N. Lat. 51 ° on Hudson Bay", DaIl (13). The northernmost source of shells examined by the writer is Winnipeg, Man. Southernmost record: Arkansas. Local Distribution. Previous records: Mackenzie River district, southwest side of Mackenzie River 30 miles above Fort Providence; south shore Lake Kakisa, near mouth of the Kakisa River; south shore Great Slave Lake near Sulphur Point; south side Mills Lake, Whittaker (39). Postglacial fossil deposit near Winnipeg, Mozley (31).

Amnicolidae Reported on Good Authority to Occur in Sub-arctic Canada but not Examined by the Author Amnicola lustrica decepta Baker, Botsford Lake, Sioux Lookout district, Ont., Baker and Cahn (9). A mnicola (Cincinnatia) cincinnatiensis judayi Baker, Hamilton Lake, Sioux Lookout district, Ont., Baker and Cahn (9). Amnicola (Cincinnatia) emarginata canadensis (Baker), Beaver Lake and Little Lake, Mackenzie River district, Baker (3). 118 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D. Class PELECYPODA FAMILY SPHAERIDAE All the species of this family which are listed below were identified by the late Dr. Victor Sterki. The records of Baker and Cahn (9) are included, but not those of other earlier authors. It should be noted that Baker and Cahn's material was also examined by Sterki. As there is little reliable information in the literature regarding the range of the members of this family, no geographical notes are given.

Genus Spliaerium Scopoli 1777 Sphaerium sulcatum (Lamarck) Cyclas sulcata Lamarck, An. sans Vert. 5 :560. 1818. Local Distribution. Previous records: Indian Bay station, Man., Falcon Bay; Rainy Lake, mile east of Wade, Ont., Mozley (21). Theodore, Sask., Whitesand River, Mozley (23). Sioux Lookout district, Ont., Botsford Lake, Baker and Cahn (9). Postglacial deposits near Lavenham, Man., Mozley (31). Sphaerium crassum Sterki .Sphaerium crassum Sterki, Nautilus, 14 :140. 1910. Local Distribution. Backwater of the Winnipeg River, near Minaki, Ont.; Falcon Bay, Shoal Lake, Man., Mozley (21). Sphaerium solidulum (Prime) Cyclas solidula Prime, Proc. Boston Soc. Nat. Hist. 4 :158. 1851. Local Distribution. Previous records: Mackenzie River district, western end of Lake Kakisa, Whittaker (39). Whitemouth River, Man., near its junction with the Winnipeg River, Mozley (21). Postglacial fossil deposit near Winnipeg, Mozley (31). Sphaerium stamineum (Conrad) Cyclas staminea Conrad, Am. J. Sci. 25 :342. 1834. Local Distribution. Postglacial deposit near Winnipeg, Man., Mozley (31). Sphaerium emarginatum (Prime) Cyclas emarginata Prime, Proc. Boston Soc. Nat. Hist. 4 :156. 1851. Local Distribution. Hamilton Lake, Sioux Lookout district, Ont., Baker and Cahn (9); postglacial deposit near Winnipeg, Man., Mozley (31). Sphaerium torsum Sterki Sphaerium torsum Sterki, Annals Carnegie Mus. 10 : 434. 1916. Local Distribution. Postglacial deposit near Winnipeg, Man., Mozley (31). Sphaerium acuminatum (Prime) Cyclas acuminata Prime, Proc. Boston Soc. Nat. Hist. 4 :155. 1851. Local Distribution. Victoria Beach, Lake Winnipeg, Man., Mozley. Post- glacial deposit near Winnipeg, Man., Mozley (31). MOZLEY: SUB-ARCTIC MOLLUSCA 119

Sphaerium vermonlanum Prime Sphaerium vermontanum Prime, Proc. Acad. Nat. Sci. Phila. p. 128. 1861. Local Distribution. Previous records: Mackenzie River district, western end of Lake Kakisa, Whittaker (39). Sioux Lookout district, Ont., Botsford Lake, Baker and Cahn (9). Sphaerium striatinum (Lamarck) Cyclas striatina Lamarck, An. sans Vert. 5 :560. 1818. Local Distribution. Previous records: Mackenzie River district, western end of Lake Kakisa, Whittaker (39). Whitemouth River, Man., near its junction with the Winnipeg River, Mozley (21). Postglacial fossil deposit near Winnipeg, Mozley (31). Sphaerium occidentale Prime Sphaerium occidentale Prime, Proc. Acad. Nat. Sci. Phila. p. 295. 1860. Local Distribution. Ditch near Mile 69, G.W.W.D. Ry., Man., Mozley (21). Sphaerium tenue (Prime) Cyclas tenuis Prime, Proc. Boston Soc. Nat. Hist. 4: 161. 1851. Local Distribution. Previous records: Mackenzie River district, south shore Second Lake, Fawn River; southwest side of Mackenzie River 30 miles above Fort Providence; south shore of Lake Kakisa near mouth of the Kakisa River; south side of Mills Lake, Whittaker (39). Traverse Bay, near Victoria Beach, Lake Winnipeg, Man., Mozley (23). Sioux Lookout district, Ont., Bamaji Lake, Baker and Cahn (9). Sphaeriurn notatum Sterki Sphaerium notatum Sterki, Nautilus, 41: 55. 1927. Local Distribution. Postglacial deposit near Winnipeg, Man., Mozley (31).

Genus Musculium Link 1907 Musculium transversum (Say) Cyclas transversa Say, New Harmony Diss. 2 :356. 1829. Local Distribution. Previous records: Matlock, Lake Winnipeg, Man., Mozley (23). Postglacial deposit near Winnipeg, Mozley (31). Musculium truncatum (Linsley) Cyclas truncata Linsley, Am. J. Sci. 6 :234. 1848. Local Distribution: Previous records: Hamilton Lake, Sioux Lookout district, Ont.; Skunk Lake, near Minaki, Ont., Mozley (21); marsh bordering Lake Winnipeg, near Victoria Beach, Man., Mozley (23). Musculium rosaceum (Prime) Cyclas rosacea Prime, Proc. Boston Soc. Nat. Hist. 4 : 155. 1851. Local Distribution. Previous records: Mackenzie River district, south side Mills Lake, Whittaker (39). Sioux Lookout district, Ont., St. Joseph, 120 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Botsford, Hamilton, Kimmewin, Kapikik and Pashkokogan Lakes, Baker and Cahn (9). Musculium ryckholti (Normand) Cyclas ryckholti Normand, Notes sur Quelques Nouvelles Cyclades, p. 7. 1844. Local Distribution. Previous records: marsh on shore of Lake Winnipeg at Victoria Beach, Man.; small lake 1 mile northeast of Ninette Sanitorium, Ninette, Man., Mozley.

Musculium securis (Prime) Cyclas securis Prime, Proc. Boston Soc. Nat. Hist. 4 :160. 1851. Local Distribution. Previous records: Alice Lake, between Minaki and Wade, Ont., Mozley (21). Lake Brereton, Man., Mozley (22). Sioux Lookout district, Ont., Fitchie Lake, Baker and Cahn (9).

Genus Pisidiuni C. Pfeiffer 1821 - Pisidiurn corn ressurn Prime - Pisidiurn cornpressum Prime, Proc. Boston Soc. Nat. Hist. 4 :161. 1851. Local Distribution. South side Mills Lake, Mackenzie River district, Whit- taker (39). Birdtail Creek near Birtle, Man.; Whitesand River. near Theodore, Sask., Mozley (23). Lake Annette near Jasper, Alta.; Lake No. 2 near Geikie, Alta., Mozley (26). Postglacial deposit near Winnipeg, Man., Mozley (31). P. compressurn var. Pashkokogan and Botsford Lakes, Sioux Lookout district, Ont., Baker and Cahn (9). Pisidiurn fallax Sterki Pisidiumfallax Sterki, Nautilus, 10 : 20. 1896. Local Distribution. Whitesand River near Theodore, Sask., Mozley (23). Postglacial deposit near Winnipeg, Man., Mozley (31). Pisidium punctatum Sterki Pisidium punctatum Sterki, Nautilus, 7 :99. 1895. Local Distribution. Hamilton Lake, Sioux Lookout district, Ont., Baker and Cahn (9). Pisidium variabile Prime Pisidium variabile Prime, Proc. Boston Soc. Nat. Hist. 4 : 163. 1851. Local Distribution. Lake No. 5 near Geikie, Alta., Mozley (26). Botsford Lake, Sioux Lookout district, Ont., Baker and Cahn (9). Pisidiurn variabile brevius Sterki Pisidiurn variabile brevius Sterki, Nautilus, 19 : 118. 1906. Local Distribution. Hibernia Lake and Caledonia Creek, near Jasper, Alta., Mozley (26). Pisidiurn minusculum Sterki Pisidium minusculum Sterki, Nautilus, 20 :17. 1906. Local Distribution. Kimmewin Lake, Sioux Lookout district, Ont., Baker and Cahn (9). MOZLEY: SUB-ARCTIC MOLLUSCA 121

Pisidium adamsi Prime Pisidium adamsi Prime, Smithsonian Misc. Coll. 145 :16. 1865. Local Distribution. Lake Brereton, Man., Mozley (22). Cat and Fitchie Lakes, Sioux Lookout district, Ont., Baker and Cahn (9). Pisidium neglectum Sterki Pisidium neglectum Sterki, Nautilus, 20 87. 1906. Local Distribution. Jacques Lake, and the stream which drains it (not Jacques Creek), Rocky River drainage, northeast of Jasper, Alta., Mozley (26).

Pisidium scutellatum Sterki Pisidiurn scutellatum Sterki, Nautilus, 10 66. 1896. Local Distribution. Mackenzie River district, south shore of Fawn and Second Lakes, Horn River; southwest side of Mackenzie River 30 miles above Fort Providence; south shore Lake Kakisa, near mouth of the Kakisa River; south side Mills Lake, Whittaker (39). Lakes Annette and Patricia, near Jasper, Alta., Mozley (26). Bamaji, Hamilton, Pashkokogan and Kimmewin Lakes, Sioux Lookout district, Ont., Baker and Cahn (9). Pisidium roperi Sterki Pisidium roperi Sterki, Nautilus, 12 : 77. 1898. Local Distribution. Pond near Mud Turtle Lake, Man., Mozley (22). Marsh on shore of Lake Winnipeg, near Victoria Beach, Mozley (23). Lake No. 8, near Jasper, Alta., Mozley (26). Pisidium subrotunduin Sterki Pisidium subrotundum Sterki, Nautilus, 20 :19. 1906. Local Distribution. Small lake near Ninette, Man., Mozley (23). Pisidium splendidulum Sterki Pisidium splendidulum Sterki, Nautilus, 11: 113. 1898. Local Distribution. Lake No. 10, near Jasper, Alta., Mozley (26). Pisidium tenuissimum Sterki Pisidium tenuissimum Sterki, Nautilus, 14 :99. 1901. Local Distribution. Skunk Lake, near Minaki, Ont., Mozley (21). Lake Edith and the larger Trefoil Lake, near Jasper, Alta., Mozley (26). Pisidium rotundatum Prime Pisidium rotundatum Prime, Proc. Boston Soc. Nat. Hist. 4 :164. 1851. Local Distribution. Hamilton and Kimmewin Lakes, Sioux Lookout district, Ont., Baker and Cahn (9). Pisidium ferrugineum Prime Pisidiumferrugineum Prime, Proc. Boston Soc. Nat. Hist. 4 :162. 1851. Local Distribution. Hamilton Lake, Sioux Lookout district, Ont., Baker and Cahn (9). 122 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Sphaeriidae Reported on Good Authority to Occur in Sub-arctic Canada, but not Examined by the Author (Identifications by Sterki) Sphaerium tumidum Baird, south shore Great Slave Lake, near Sulphur Point (?); near Fort Wrigley, Mackenzie River district, Whittaker (39). Musculium jayense Prime, south shore Fawn Lake, Horn River, Mackenzie River district, Whittaker (39). Pisidium indianense Sterki, south shore Second Lake, Horn River; at mouth of the Hay River, Mackenzie River district, Whittaker (39).

FAMILY UNIONIDAE Genus Quadrula Rafinesque 1820 Quadrula quadrula Rafinesque Obliquaria (Quadrula) quadrulaRafinesque, Ann. Sci. Phys. Bruxelles, 5 :307. 1820. Geographical Range. North America, the Mississippi and St. Lawrence drainages; Canada east of the Rocky Mountains. Northernmost record: Red River at Winnipeg. Southernmost record: eastern Texas. Local Distribution. Red River at Winnipeg.

Genus Amblema Rafinesque 1819 Amblema costata (Rafinesque) Amblema costata Rafinesque, Ann. Sci. Phys. Bruxelles, 5 :315. 1820. Geographical Range. North America, the Mississippi and St. Lawrence drainages; Canada east of the Rocky Mountains. Northernmost record: Saskatchewan River, Lake Winnipeg. Southernmost record: Alabama River ? Southern distribution imperfectly known. Local Distribution. New record: near the Englishman's Rapid, Berens River, Man., at a depth of 12 metres. Previous records: Red, and Black Rivers, Man.

Genus Fusconaia Simpson 1900 Fusconaia flava (Rafinesque) Obliquaria flava Rafinesque, Ann. Sci. Phys. Bruxelles, 5 : 305. 1820. Geographical Range. North America, the Mississippi drainage; Canada east of the Rocky Mountains. Northernmost record: Nelson River, Man. Southernmost record: West Virginia ? Local Distribution. Red, Roseau, Souris, and Nelson Rivers; Lake Winnipeg.

Genus Strophitus Rafinesque 1820 Strophitus rugosus (Swainson) Anodon rugosus Swainson, Zool. Ill., Ser. I, II, pl. 96. 1822. MOZLEV: SUB-ARCTIC MOLLUSCA 123

Geographical Range. North America, the Mississippi drainage; Canada east of the Rocky Mountains. Northernmost record: mouth of the Hay River, N. Lat. 600 51'. Southernmost record: Arkansas. Local Distribution. New record: Birdtail Creek, near Birtle, Man. Previous records: Red, Whitemouth and Saskatchewan Rivers; Lake Win- nipeg; Playgreen Lake. Fossil records: Lake Agassiz clay near Mile 93, G.W.W.D.Ry., Man.; postglacial gravel bed near Birtle, Man. Remarks. Baker (3) states that specimens from Arkansas and elsewhere belong to the variety pavonius of Lea. Genus Anodonta Lamarck 1799 Anodonta grandis Say Anodonla grandis Say, New Harmony Diss. 2 :341. 1829. Geographical Range. North America, Mississippi drainage and upper St. Lawrence drainage; Canada east of the Rocky Mountains. Northernmost record: Lake Winnipeg ? Southernmost record: Texas. Local Distribution. New record: Morris River, near Morris, Man. Previously reported from the Fairford River and other localities, most of the records probably being based upon the sub-species mentioned below. A nodonta grandis footiana Lea Anodonta footiana Lea, Proc. Am. Phil. Soc. 1: 289. 1840. Geographical Range. North America, Mississippi drainage; Canada east of the Rocky Mountains.

Northernmost record: Lake Kakisa, N.W.T., N. Lat. 61 0 . Southernmost record: Michigan ? Local Distribution. New record: Birdtail Creek near Birtle, Man. Previous records: numerous lakes in far western Ontario; Red and Souris Rivers, Man.; Battle River Lake, Alta. Fossil record: postglacial gravel bed near Birtle. Remarks. The records of A. grandis from Lake Winnipeg may refer to this variety. Anodonta kennicotti Lea Anodonta kennicotti Lea, Proc. Phil. Acad. 5 :56. 1861. Geographical Range. North America, St. Lawrence, Hudson Bay, and Mackenzie drainages. Northernmost record: outlet of Lake Winnipeg. Southernmost record: Lake of the Woods ? Local Distribution. New records: Lake Brereton, Man. (previously reported as A. grandis footiana); Lake Winnipegosis. Previous records: numerous lakes and streams in far western Ontario; Lakes Winnipeg and Manitoba; Grand Rapids of the Saskatchewan; Ekwan River, Keewatin. 124 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Genus Anodontoides Simpson 1898 A nodontoides ferussacianus (Lea) Anodonta ferussaciana Lea, Trans. Am. Phil. Soc. 5 : 45. 1834. Geographical Range. North America, Mississippi and St. Lawrence drainages; Canada east of the Rocky Mountains. Northernmost record: Whitesand River, near Theodore, Sask. Southernmost record: Tennessee. Local Distribution. New record: Whitesand River, as above. Previous records: Lake of the Woods; Lake Winnipeg.

Genus Lasmigona Rafinesque 1831 Lasmigona corn planata katherinae (Lea) Unio katherinae Lea, Syn., p. 35. 1838. Geographical Range. North America, the northern part of the Mississippi drainage?, Lake Superior, and Canada east of the Rocky Mountains; Hudson Bay drainage. Northernmost record: Lake Winnipeg; "Nelson River drainage", Dali. Southernmost record: Lake Superior; northern part of Mississippi drainage? Local Distribution. New records: Birdtail Creek near Birtle, Man.; Moose Jaw Creek near Moose Jaw, Sask. Previous records: Winnipeg River at Minaki; Lake of the Woods, Ont.; Roseau, Souris, Red, Assiniboine, and Nelson Rivers, Man.; Lake Winnipeg; lower Saskatchewan River, Shoal, Shell, and Battle Rivers, Battle River Lake, Alta.; Carrot River near Arborfield, Sask. Previous fossil record: postglacial gravel bed near Birtle.

Lasmigona corn pressa (Lea) Symphynota cornpressa Lea, Trans. Am. Phil. Soc. 3 : 450. 1829. Geographical Range. North America, New York to Nebraska, and from Kentucky to Manitoba and Saskatchewan. Northernmost record: Carrot River near Arborfield, Sask. Southernmost record: Kentucky. Local Distribution. New fossil record: postglacial gravel bed near Birtle, Man. Previous records: Boulder and Missinaibi Rivers, Ont.; Carrot River. Remarks. This species has been reported from the Sheyenne River south of Devils Lake, North Dakota, by Winslow (40). This stream is in the drainage basin of the Red River.

Genus Proptera Raflnesque 1819 Pro ptera alata megaptera (Rafinesque) Metaptera megaptera Rafinesque, Ann. Sci. Phys. Bruxelles, 5 :307. 1820. Geographical Range. North America, Mississippi and (in part) Atlantic drainages; Canada east of the Rocky Mountains. MOZLEY: SUB-ARCTIC MOLLUSCA 125

Northernmost record: Red River, Man. Southernmost record: northern Alabama. Local Distribution. Previous records: near Lake of the Woods, Ont.; Red River. Remarks. Following Baker (3), the above name is used for the species formerly known as Lam psilis alatus (Say).

Genus Actinonaias Fischer and Crosse 1893 Actinonaias carinata (Barnes) Unio carinatus Barnes, Am. J. Sci. 6 :259. 1823. Geographical Range. North America, Mississippi drainage; Canada east of the Rocky Mountains; for details of range in the vicinity of the Great Lakes see Baker (3, p. 221). Northernmost record: Assiniboine River, Man. Southernmost record: southern Michigan; Arkansas ? Local Distribution. New record: Assiniboine River at Treesbank, Man. Previous records: Roseau River; Assiniboine River at Millwood, Man.

Genus Ligumia Swainson 1840 Ligumia recta latissima (Rafinesque) Unio latissima Rafinesque, Ann. Sci. Phys. Bruxelles, 5 :297. 1820. Geographical Range. North America, Mississippi and Atlantic drainages; western Canada east of the Rocky Mountains. Northernmost record: Red River, Man. Southernmost record: Georgia and Arkansas. Local Distribution. New record: Assiniboine River at Aweme, Man. Previous records: Lake of the Woods, Ont.; Roseau, Red, and Assiniboine Rivers, Man. Fossil record: postglacial gravel bed near Birtle, Man.

Genus Lampsilis Rafinesque 1820 Lam psilis ventricosa (Barnes) Unio ventricosus Barnes, Am. J. Sci. 6 :267. 1823. Geographical Range. North America, Mississippi and Atlantic drainages; Canada east of the Rocky Mountains. Northernmost record: Nelson River, Man. Southernmost record: Oklahoma. Local Distribution. New record: Nelson River, at the point where the Hudson Bay Railway crosses that stream for the first time north of The Pas, Man. Previous records: Roseau and Red Rivers; Lake Winnipeg. Previous fossil record: postglacial gravel bed near Birtle, Man. Remarks. L. ventricosa is represented in this part of North America by a dwarf variety, but no useful purpose would be served, at least for the present, by describing it as a new sub-species. • 126 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

Baker (3) regards the great majority of the previous records of this species in Wisconsin as belonging to the variety occidens (Lea). Having examined the type specimen of this variety in the collection of the United States National Museum in conjunction with his own and other specimens, the author is not convinced that it is a distinct race.

Lam psilis siliquoidea rosacea (De Kay) Unio rosacea De Kay, Zool. N.Y. 5 :192. 1843. Geographical Range. North America, Mississippi and Atlantic drainages; Canada east of the Rocky Mountains. Northernmost record: Great Slave Lake. Southernmost record: northern Indiana. Local Distribution. New records: Whiteshell River, several localities below Betula Lake; Birdtail Creek near Birtle; Valley River, 1 mile above its mouth; Mossy River near Lake Winnipegosis; Lake Winnipegosis, all in Manitoba. Previous records: this species has been reported from many localities throughout western Canada to the east of the Rocky Mouutains. Remarks. Gravid specimens were collected in Birdtail Creek on November 11, 1924, and in the Red River near Winnipeg on November 1, 1925.

Lam psilis superiorensis (Marsh) Unio superiorensis Marsh, Nautilus, 10 :103. 1897. Geographical Range. North America, from the region of Lake Superior to the Mackenzie River district. Northernmost record: Lake Kakisa, N. Lat. 610. Southernmost record: the upper Great Lakes region. Local Distribution. Previous records: numerous localities in far western Ontario. Lake Kakisa, and the mouth of the Hay River, Mackenzie River district, Whittaker (39).

Unionidae Reported on Good Authority to Occur in Sub-arctic Canada, but not Examined by the Author Anodonta marginata Say, Schist Lake, Sioux Lookout district, Ont., Baker and Cahn (9). This locality is in the Hudson Bay drainage. It is of interest as possibly indicating a westward migration of this characteristic Atlantic coast species. Baker and Cahn state that the single specimen collected was typical in form and color.

The Local Distribution of the Molluscan Fauna The general procedure in this work has been to ascertain the types of habitat which are available in this region for settlement by molluscs, and then to observe the distribution of the species in them. It has been considered desirable to define briefly the position which this part of Canada occupies .with reference to the adjoining regions. A brief summary of the various zoogeographical sub-divisions of North America is therefore given. 127 I MOZLEV: SUB-ARCTIC MOLLUSCA

In most of the previous work on the flora and fauna of North America, attention has been directed chiefly to the details of distribution rather than to general geographical considerations. It may not be out of place to devote a paragraph to showing that the various schemes of classification which have been proposed are not mutually exclusive, as has sometimes been supposed, but are complementary.

THE ZOOGEOGRAPHICAL SUB-DIVISIONS OF NORTH AMERICA It is now generally agreed among biologists and geographers that North America forms a distinct zoogeographical region. Regarding the sub-division of this continent however, there is wide divergence of opinion. These differences appear to have arisen, at least in part, from the diverse points of view from which the subject has been approached. The "Life Zones" of Merriam (19) afford an example, being based upon a conception very similar to that of the "heat budget," i.e., the total quantity of heat required to produce mature individuals of a given organism and to ensure reproduction. Merriam's Life Zones have the disadvantage of being somewhat arbitrary and inflexible. A useful alternative division of the continent is into Arctic, Sub-arctic, and Temperate belts or regions. This takes into account the facts which formed the basis of the division of North America into 'Holarctic and Sonoran portions by Lydekker (18), and there can be little doubt that there is a change in the character of the fauna of the continent in the neighborhood of N. Lat. 500• Those who approach the subject from the standpoint of ecology place much importance upon the natural developmental cycles of the vegetation, and the associated changes in the fauna. It is also desirable that the existence of widespread natural regions or prevailing landscape types should be kept in mind. According to this conception, the northern part of the continent at least may be divided into a number of natural areas, in which the climate, soils, flora, and fauna are similar within sufficiently narrow limits to give the country an essential homogeneity and characteristic appearance. Examples of these natural regions are the barren ground, northern coniferous forest, parkland, and prairie. It is likely that there is a degree of truth in each of these viewpoints. The region dealt with in this account lies in the sub-arctic region, covers four natural regions or major landscape types, and includes four of Merriam's Life Zones. A consideration of some of the habitats of this region from a developmental standpoint is to be found in the work of Bird (10).

TYPES OF HABITAT FOUND IN SUB-ARCTIC CANADA AND THEIR MOLLUSCAN FAUNA Up to the present, most of the published accounts relating to the fresh- water mollusca of this region have been in the nature of faunal lists. Little attention has been paid to the local distribution of the species.. Several studies of the habitat relationships of the mollusca have been made in the United States, but only one short account has been published of such work, 128 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

carried out within a radius of several hundred miles, in western Canada. Most investigations in other parts of the world have been concerned with the distribution of the mollusca in a single body of water, and the geographical aspects of the subject have been neglected. Although much important information has resulted from ecological studies, many of them tend to convey an impression, whether intentionally or not, of the existence of a high degree of uniformity in the fauna of many types of aquatic habitats. This is erroneous. The fauna of such situations is not stereotyped, nor is the association of the various species a haphazard affair. In another paper (Mozley (34) ) it has been demonstrated that this association may be the result of definite, though complex reactions, and can be measured and mathematically expressed. The diversity in the constitution of the fauna of similar habitats is very noticeable throughout the sub-arctic region. Under different geographical conditions it is apparently possible for one species to occupy the position of another in the environmental complex, with considerable precision. This probably involves the substitution of the one species for the other in the food chains. Investigations along these lines promise interesting results. Within the territory included in the scope of this paper there are four well-marked natural regions, known as the Prairie, Parkland, Forest, and Barren Ground. These are comparable respectively to the Steppe, Forest- Steppe, Taiga, and Tundra of northern Asia. The Prairie, which is situated along the southern boundary of western Canada, is characterized by the predominance of grassy vegetation and the absence of trees. Many of the large lakes of this region are saline, and the most favorable habitats for fresh-water mollusca are small marshy lakes. Ponds are numerous, but, as a result of the semi-arid climate, many of them contain water only during the spring of each year. The Parkland, which is found to the north of the Prairie, is a narrow belt of groves of aspen (Populus Iremuloides) with intervening grassy glades. As far as fresh-water organisms are cbncerned, conditions on the Parkland are very similar to those on the Prairie, but many of the Parkland waters are less saline. Ponds are somewhat more numerous, and they contain water for a longer period each year than do those on the Prairie. The general conditions of life in these two regions closely resemble those on the Steppe of northern Kazakstan, and the Forest-Steppe of Siberia. For the most part, the species of animals which occupy similar situations in Asia and North America are different, although frequently of the same or related genera. Ecologically they are closely comparable. The Forest, which forms a broad belt running across the whole of the country from Ungava to the Rocky Mountains, presents conditions which are unlike those of the Prairie and Parkland. The waters of this area are neutral or acid in reaction. In many instances they have a low content of dissolved mineral salts, and are relatively rich in organic matter. Sphagnum invades many of the smaller bodies of water in this region and appears to have an unfavorable effect upon the molluscan fauna, or at least to be associated with MOZLEY: SUB-ARCTIC MOLLUSCA 129 such an effect. In this respect many of the small lakes and ponds of this region resemble those of northern Sweden and Finland. It is a notable fact that when such waters are aerated, as in streams with rapids and falls, they appear to be capable of supporting a much richer molluscan fauna. Large rock-bound lakes with clear and cold waters are not uncommon in the forested region, and they have a characteristic molluscan fauna. The Barren Ground, which lies to the north of the forested area along the Arctic coast of Canada, is a treeless area which closely resembles the Tundra of Fennoscandia and Siberia. Several fresh-water mollusca are found in this region in abundance. The habitats of fresh-water mollusca in sub-arctic Canada maybe grouped under seven descriptive headings, namely: Temporary ponds. Ponds which contain water permanently, and small shallow lakes. Large lakes having outlet streams. Lakes without direct outlet streams. Intermittent streams. Permanent streams of the non-mountainous region. Streams of the Rocky Mountain region. This grouping of habitats is somewhat arbitrary, and takes no account of the fact that each pond, lake and stream has its own individual characteristics. There are also numerous habitats which are not easily placed in one of the above classes. Nevertheless it is desirable to have clear-cut conditions as the basis of the classification, even though this tends in some instances to exaggerate the importance of extreme conditions. The plan of presentation adopted in the following account is to give a brief description of some of the salient features of the principal habitat types, with comments on their fauna in other parts of the sub-arctic region. This is followed by a series of concrete examples in which the molluscan fauna of certain specific localities is listed. It should be borne in mind that these molluscan associations are not necessarily typical, but taken as a whole they are believed to be representative. All the information is original and has been drawn from a large number of observations.

TEMPORARY PONDS The presence of a large number of small, shallow ponds is characteristic of many parts of the Prairie and Parkland. These usually occupy little depressions in which water from the melting snow collects in the spring, and remains through the months of April, May, and in some instances a part of June. After this they are dry until the following spring. Such ponds are characterized by a short aquatic phase in the spring followed by progressively drier conditions, and are subject to low temperatures during the winter. A description of the aquatic fauna and general conditions in one pond of this type has already been published (Mozley, 27). Similar ponds occur on the Steppe and Forest-Steppe of northern Asia. It may be worth noting that such 130 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

ponds in Canada and Siberia draw most of their water supply from melting snow, and this may be of significance in comparing similar habitats in South Africa and Australia which are fed by rains. Apart from certain molluscs, namely Planorbis umbilicatellus, and Planorbula cam pestris, the animals which are most chal-acteristic of these situations in Canada are phyllopod crustacea such as Lepidurus, Branchipus, Limnetis, and Estheria. Examples A small pond between Lake Brereton and Mud Turtle Lake, eastern Man. Aplexa hypnorum. A small pond near the shore of Mud Turtle Lake. Planorbis umbilicatellus. A small pond near Stony Mountain, Man. Lymnaea palustris, Aplexa hypnorum. A pond 4 miles southwest of Winnipeg, Man. Lymnaea caperata, Planor- bula cam pest ris, Aplexa hypnorum. S. A pond near Spirit River, Alta. Lymnaea palustris, L. caperata, Planorbis umbil'icatellus, Aplexa hypnorum. 6. A pond near St. Vital, Man. Lymnaea palustris, L. caperata, Planorbis exacuous, P. umbilicatellus, Planorbula cam pest ris, P. crassilabris, Aplexa hypnorum.

SMALL FRESH-WATER LAKES AND PONDS Under this heading are included ponds which contain water permanently, and small pond-like lakes. In regions where there are few outcrops of the bedrock, such bodies of water generally have low muddy shores, and are invaded to a greater or less extent by marsh plants such as Scirpus and Typha. Under these conditions the predominant molluscs are Lymnaea stagnalis jugularis, L. palustris, Planorbis tr'ivolvis, and Physa gyrina. In parts of the forested regions the lakes have rocky shores, and under these conditions the tendency towards marsh development is less marked. The species found most commonly in these rock-bound lakes are Planorbis campanulatus wisconsinensis, P. corpulentus, and Physa ancillar'ia. In shallow places where there is decaying vegetable matter Lymnaea megasoma is sometimes to be found in abundance. A particular type of pond, formed by beavers (Castor canadensis) is of considerable local importance as a habitat for molluscs in the Rocky Mountain area (Mozley (28)). The small lakes of Canada closely resemble those of Siberia. In the forested part of the latter country, however, a greater proportion of the small basins are filled to such an extent with vegetation that they are no longer suitable habitats for fresh-water mollusca. Ponds on the flood plains of streams are less important habitats for mollusca in Canada than in northern Asia. Among the fresh-water gastropods which are common to these two countries are Lymnaea stagnalis and L. palustris. The species which in Canada most frequently accompanies L. stagnalis in ponds and small lakes is Planorbis trivolvis. In Siberia this position is taken by P. corneus. Lymnaea palustris MOZLEY: SUB-ARCTIC MOLLUSCA 131 is not as common in Siberia as in Canada, and possibly it is displaced to some extent in the pond fauna of the former country by Lymnaea pereger, a species which does not occur in North America. Examples A pond at Mile 95, G.W.W.D. Ry., Man. Lymnaea lanceata. The smaller Trefoil Lake, near Jasper, Alta. Planorb'is crista. Lake Dorothy, Miette Valley, Jasper. Physa gyr-ina. Rainy Lake, near Wade, Ont. Sphaerium sulcatum. A pond on Duck Mountain, 20 miles east of Kamsack, Sask. Lymnaea palustris, Planorbis trivolvis. A pond formed by beavers on Tekarra Creek, Jasper. Planorbis arcticus, Physa gyr'ina. Skunk Lake, near Minaki, Ont. Sphaerium truncatum, Pisidium tenuissimum. Sword Lake, near Minaki. Planorbis cam panulatus w'isconsinens'is, Physa integra. Virl Lake, Miette Valley, Jasper. Planorbis crista, Physa gyrina. Low-lying ground near Round Lake, Ninette, Man. Lymnaea humilis modicella, L. obrussa exigua. Alice Lake, near Minaki. Planorbis campanulatus wisconsinensis, P. arcticus, Musculium securis. A pond formed by beavers on a small stream crossing Buffalo Prairie, Athabasca Valley, near Jasper. Lymnaea stagnalis wasatchensis, L. traskii, Planorbis arcticus, Physa gyrina. A pond near the shore of Hudson Bay, east of Fort Churchill, Man. Lymnaea palustris, Planorbis arcticus, Physa gyrina, Aplexa hypnorum. A pond on Moose Mountain, southern Sask. Lymnaea stagnalisjugularis, L. palustris, Planorbis trivolvis, P. exacuous, Physa gyrina. A small lake near Cottonwood Creek, Jasper. Lymnaea stagnalis wasatchensis, L. emarginata canadensis, L. dalli, Planorbis tr'ivolvis, P. hirsutus, P. arcticus, Pisidium roperi, P. splendidulum. "Lake No. 1", Miette Valley, Jasper. Planorbis exacuous, P. hirsutus, P. arcticus, Physa gyrina, Ancylus coloradensis, Valvata lewisi, Pisidium sp. Lake Mildred, near Jasper. Lymnaea stagnal'is wasatchensis, L. palustris, L. obrussa decampi, Planorbis arcticus, Physa gyrina, Valvata lewisi. Douglas Lake, near Onah, Man. Lymnaea stagnalis jugularis, L. palustris, Planorbis cam panulatus davisi, P. trivolvis, P. exacuous, Physa gyrina. The larger Trefoil Lake, Jasper. Lymnaea stagnalis wasatchensis, Planor- bis trivolvis, P. exacuous, P. arcticus, Physa gyrina, Pisidium tenuissimum. Pelican Lake, Ninette, Man. (i) On Potamogeton near the centre of the lake, Physa gyrina (abundant), Lymnaea palustris (less common); (ii) in Typha and Scir pus marsh around the shore of the lake, Lymnaea 132 CANADIAN JOURNAL. OF RESEARCH. VOL. 16, SEC. D.

stagnalis jugular'is, L palustris, Planorbis exacuous, P. arcticus, Planor- bula armigera, Physa gyrina; (iii) in a moist meadow on slightly higher ground near the marsh, Lymnaea caperata, L. parva var., Aplexa hypnorum. LARGE LAKES HAVING OUTLET STREAMS Large lakes are numerous in sub-arctic Canada, and their outstanding characteristic is that they are of moderate depth (commonly up to 30 m.), and that their shores are subject to severe wave action. In places where the shores are protected either as a result of their configuration or the presence of islands, the molluscan fauna is similar to that found in smaller bodies of water. The most typical species on exposed shores are Lymnaea stagnalis sanctaemariae, L. emarginata, and L. catascopium. Fresh-water mussels are to be found on the bottom of many lakes, the commonest species being Lam psilis siliquoidea rosacea and A nodonta kennicotti. Few comparable habitats are to be found in Siberia, where there is little or no adaptation in the pulmonate fauna to meet lacustrine conditions such as is found in Canada. The fauna of Lake Baikal, while lacustrine, is of such a special type that no parallel can be drawn between it and any Canadian lake. Lymnaea stagnalis sanctaemariae of Canada is a variety which occupies a habitat very similar to that of L. stagnalis laCustris in Europe, and the shells of the two varieties have a certain resemblance to each other. Lymnaea stagnalis lillianae may also have a parallel form in Europe, e.g., the large variety of L. stagnalis which inhabits Lohja Lake in southern Finland. Examples Maligne Lake, 30 miles east of Jasper, Alta. Altitude 5555 ft. Lymnaea traskii. Pyramid Lake; near Jasper. Altitude 3867 ft. Lymnaea stagnal'is wasatchensis, Planorbis antrosus sayl, P. exacuous, P. hirsutus, P. arcticus, Physa gyrina, Pisidium spp. Atikameg Lake, Mile 17, Hudson Bay Railway, Man. Lymnaea emarginata var., L. obrussa decampi, Planorbis exacuous, P. deflectus, P. hirsutus, Valvata tricarinata. Shoal Lake, eastern Man. (i) Exposed rocky shores of Indian Bay, Physa ancillaria; (ii) sandy shore of Indian Bay somewhat exposed to wave action, Lymnaea stagnal'is lillianae; (iii) protected shore of Indian Bay, Lymnaea obrussa exigua; (iv) Falcon Bay in quiet water, on sand bottom and in small marshes, Lymnaea stagnalis jugularis, Planorbis campanulatus wisconsinensis, P. trivolvis pilsbryi, P. antrosus sayi, P. exacuous, P. hirsutus, P. arcticus, Planorbula crassilabris, Sphaerium crassum, Anodonta kennicotti. Lake Brereton, Man. (i) On exposed rocky shores subject to severe wave action, Lymnaea stagnalis sanctaemariae, Planorbis antrosus sayi, P. cam panulatus wisconsiflensis, Physa gyrina; (ii) on partly protected rocky shores, Lymnaea columella casta, Planorbis exacuous; (iii) on MOZLEY: SUB-ARCTIC MOLLUSCA 133

partly protected sandy shores, chiefly on Potamogelon, Amn'icola l'imosa; on Potamogeton and other plants in bays protected from wave action, Lymnaea megasoma, Planorbula crassilabris, Pisidium adamsi; in Typha marsh at the mouth of the Rennie River, Lymnaea megasoma, L. lanceata, Planorbis antrosus sayi, P. exacuous, P. arcticus, P. hirsutus, Physa gyr'ina, Ferriss'ia parallela, Musculium secur'is. Lake Winnipeg, near Victoria Beach, Man. (i) On shore stones and boulders subject to severe wave action, Lymnaea emarginata, L. catascopium; (ii) sandy shore, species cast up from deeper water, Valvata. tr'icarinata, Amnicola limosa porata, A. walker'i, Muscul'ium transversum, Anodonta kennicotti, Lam psilis siliquoidea rosacea, L. ventricosa; (iii) small shallow sandy bay containing very little vegetation, Lymnaea palustris, Sphaerium tenue; (iv) large marsh partly cut off from the lake proper, Lymnaea stagnalis, L. palustris, L. dalli, Planorbis trivolvis, P. hirsutus, P. arcticus, Planorbula crassilabris, Physa gyr'ina, Aplexa hypnorum, Valvata lewis'i, Musculium ryckholti, M. truncatum, Pisidium roperi. Note: Cam peloma decisum was not collected in this marsh, but occurs in the Grand Marais, a similar habitat on the shore of Lake Winnipeg about 15 miles south of Victoria Beach.

LAKES WITHOUT DIRECT OUTLET STREAMS Lakes of this type are the reservoirs of inland drainage basins .which have lost their connection with the sea as a result of changes in climate since the last period of glaciation. As there is no outlet stream to carry off the mineral salts leached from the soils of the basin and carried into the lake by its tributary streams, and the channels of underground drainage are insufficient to do this, these bodies of water are becoming increasingly saline. Such lakes are found in considerable numbers on the Prairie and Parkland, but it is only under special circumstances, such as are found subsequent to forest fires, that they occur in the forested region. Similar lakes are found in northern Asia, and the mollusca which in that region have the greatest tolerance of saline conditions are Lymnaea palustris saridalensis, L. palustris kazakensis, and Planorbis planorbis. Examples Little Quill Lake, Sask. Total salt content in 1928, 9688 parts per million. Water analysis given by Mozley (25). Lymnaea palustris. Lake Lenore, Sask. Total salt content in 1928, 3936 p.p.m. Water analysis given by Mozley (30). Lymnaea palustris, Planorbis arcticus.

INTERMITTENT STREAMS Many of the smaller channels of drainage in this region contain no water during the period of midsummer drought, and this, coupled with the great force of flood waters when the snow melts in spring, has an unfavorable effect upon the molluscan fauna. In some instances however, where the inclination of the bed of the intermittent stream is slight, so that the freshet has not too 134 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

great a scouring effect, a few species find suitable conditions. Where this is so, large numbers of individuals may occur. Examples A small intermittent stream draining a group of poplar groves near Birtle, Man. Planorbis umbilicatellus. Paskwegin Brook, a tributary of Little Quill Lake, Sask. Lymnaea palustris. Catfish Creek, near Winnipeg, Man. Lymnaea palustris, Physa gyrina hildrethiana. A shallow ditch near Mile 69, G.W.W.D. Ry., Man. Lymnaea palustris, L. obrussa exigua, Sphaerium occidentale. A small intermittent stream near Lanigan, Sask. Lymnaea palustris, L. caperata, Aplexa hypnorum.

PERMANENT STREAMS OF THE NON-MOUNTAINOUS REGION There is great diversity in the conditions found in the permanent streams of this area, and hence there are corresponding diversities in the molluscan fauna. It is therefore difficult to give a brief description which will apply to them all. An important characteristic of all, however, is that as they flow through fairly level country the force of the current is only moderate. In many instances this permits the growth of aquatic plants on the bottom, and the development of small marshes along the banks. Under these conditions numerous species of molluscs are to be found. The commonest forms on the Prairie and Parkland are Lymnaea stagnalis jugularis, Planorbis trivolvis, and Lam psilis siliquoidea rosacea. Lasmigona corn planata katherinae is abundant in certain streams. In the forested area Lam psilis superiorensis is often the predominant species. In pool just below rapids on a stream entering Lake Nipigon, Ont., at about N. Lat. 490 25', W. Long. 800 8', near Macdiarmid, Ont. Physa ancillaria. Eyehill Creek, near Yonker, Sask. Lymnaea stagnalis jugularis, Planorbis trivolvis, Aplexa hypnorum. Clair Brook, a tributary of Little Quill Lake, Sask. Lymnaea palustris, Planorbis trivolvis, P. arcticus, Aplexa hypnorum. A small brook running into the Grand Marais, Man. Examined at a point 2 miles east of Balsam Bay. Lymnaea palustris, L. obrussa exigua, Planorbis arclicus, Planorbula crassilabris, Physa gyrina. Muckle Creek, near Clandeboye, Man. Lymnaea stagnalis jugularis, L. palustris, Planorbis arcticus, Physa integra, Valvata tricarinata. La Salle River, near St. Norbert, Man. Lymnaea stagnalis jugularis, L. obrussa exigua, Planorbis arcticus, Valvata tricarinata, Amnicola sp. Jackfish Creek, east of Balsam Bay, Man. Lymnaea stagnalis jugularis, L. palustris, Planorbis arcticus, P. hirsutus. MOZLEY: SUB-ARCTIC MOLLUSCA 135

Whiteshell River, below Cross Lake, Man. Lymnaea stagnalis jugularis, L. megasoma, L. lanceata, Planorbis exacuous, P. hirsutus, Aplexa hyp- forum, Cam peloma decisum. Birdtail Creek, near Birtle, Man. (i) On bottom of stream, Planorbis antrosus, Ferriss'ia rivularis, Physa integra,Lasmigona corn planata kather- mae, Lam psilis siliquoidea rosacea, Anodonta grandisfootiana, Strophitus rugosus; (ii) in small marshy areas along the banks of the stream, Lymnaea palustris, L. umbilicata, L. parva sterki'i. A backwater of the Winnipeg River, near Minaki, Ont. Lymnaea stagnalis jugular'is, L. ernarginata canadensis, Planorbis corpulentus, P. cam panulatus wisconsinensis, P. exacuous, P. arcticus, Physa ancillaria, Sphaerium crassum, Lasmigona corn planata kather'inae, Lam psilis superiorensis, A nodonta kennicotti. Red River, near Winnipeg, Man. Lasmigona complanata katherinae, Lam psil'is siliquoidea rosacea, L. ventricosa, Ligurnia recta latissirna, Proptera alata megaptera, Anodonta grandis footiana, Stroph'itus rugosus, A mblema costata. -

STREAMS OF THE ROCKY MOUNTAIN REGION It has been pointed out in a previous paper (Mozley (28)) that the conditions of existence in the rivers and brooks of the Rocky Mountain region are not favorable for mollusca. The reasons for this are intimately connected with the climate and topography of the region, and especially with the irregularity of flow in most of the streams, their low temperature, poverty as a nutrient medium, and the high inclination of their beds. Fresh-water mollusca have been found in only one stream in this region, viz., Caledonia Creek, near Jasper, Alta. Conditions in this stream are somewhat unusual for this mountainous part of Canada. Caledonia Creek drains a small lake, which acts as a reservoir, settling basin, and site for food production. Valvata lewis-i and Pisidiurn variabile brevius were found in the upper part of this stream in considerable numbers. The Geographical Affinities of the Fauna The molluscan fauna of the northern part of North America has much in common with that of northern Eurasia. European elements, as distinct from Eurasian, are absent. The fauna of sub-arctic Canada is less rich than that of the more southerly parts of North America, but nevertheless has certain positive characteristics of its own. There are three geographical elements in this fauna, namely: A group of circumboreal species, viz., Lyrnnaea stagnalis, L. palustris, Planorbis crista, P. hirsutus (= albus), Aplexa hypnorurn and probably one or more of the Sphaeriidae. In addition, there is one North American species, Planorbis trivolvis, which has been found as a fossil in far northeastern Siberia (see Mozley (33) ). A large number of strictly North American species, the great majority of which are characteristic of the Mississippi drainage, or at least are found 136 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D.

abundantly in that region. An exception to this general rule is Anodonta niarginata which has long been regarded as characteristic of the Atlantic coast drainage, but is now known to occur in certain parts of the Mississippi drainage. (iii) A group of species and varieties which is characteristic of sub-arctic Canada, being either wholly confined to it, or else found most commonly within its boundaries. This element includes the following species, Lymnaea palustris castorensis, L. hedleyi, L. preblei, L. randoiphi, L. atkaensis, Planorbis corpulentus, P. corpulentus mult'icostatum, Planorbula cam pestris, Amnicola winkleyi mozleyi, Anodonta kennicotti, Lam psilis sueriorensis, and Lasmigona corn planata katherinae. The explanation of the close relationship of the molluscan fauna of sub- arctic Canada with that of the Mississippi drainage lies in the history of the region. By far the greater part of northern North America has been subject to severe glaciation within recent times. The preglacial fauna of this region was either exterminated or driven southward before the ice sheet. After the retreat of the ice the newly exposed territory was repopulated by plants and animals which migrated northwards from the interior of the continent. General geographical factors tended to exclude the possibility of an extensive faunal migration from the Atlantic and Pacific coast regions. This was particularly true of the aquatic organisms, there being direct drainage connections between central Canada and the Mississippi River system, and little or no connection with the Atlantic and Pacific drainages. The aquatic animals which migrated into this region from the south were subsequently trapped in the northern drainage, and thus lost all possibility of interbreeding with the southern fauna. To this separation is probably due, at least in part, the continuance of the faunistic peculiarities of the northern drainage. The total number of species and varieties of fresh-water mollusca known to inhabit sub-arctic Canada is 111. In northern Asia, apart from Lake Baikal, there are only 50 species and varieties of these animals (Mozley (33) ). As both these faunas have been examined by the same investigator there is some reason for believing that the totals are comparable. In several groups of these fresh-water mollusca there appears to have been a much greater degree of speciation in North America than in northern Asia. The Lymnaeidae of the two areas afford an example. In sub-arctic Canada, there are 26 species and varieties belonging to this family, whereas in Siberia there are only 12. The distinctly greater degree of specialization for life in certain types of habitat (e.g., temporary ponds, see Mozley (33) ) in North America as compared with northern Asia may possibly be a related fact. In any event it is reasonably certain that even though the climate and general types of natural regions in the two territories are similar, there is much greater diversity of fresh-water molluscs in the Canadian fauna than in that of Siberia. The reason may lie partly in the somewhat richer source of supply which was available in Canada, and the greater facility of migration from south to north in North America, than from west to east in Eurasia. Another important factor may have been the greater variety of suitable aquatic habitats PLATE I

3 4

4 64 ••44

I? A A

12 13

4m1

1. Lyinuaea lanceata Gould ); ('(in yon Lake. On!. 2. Lvm noel jo/i ,,son I (Baker); Yoko Park. 3. Lymnaea Izedlevi Baker; Fraser River, Lucerne, B.C. -I. Lrmnaea vukonensis Baker; Cizena, Alaska. 5. Lynznaea palusiris (Mid/er); pond on Duck Mountain, near Madge Lake. Sask. 6. Lymnaea traskii Tryon): 1 mile west of head o; Brawau Lake, Alta. 7. Lvmnaea palustris: Paskwegin Brook, Sash. 8. Same as 5: Birtle, Man. 9. Lvmnaea stagnalis sa,z.cta.emarzae Walker; Lake Brereton, Man. 10. Saizze; St. Mary's River, Saul! Ste. Marie. (Bryant 11 a/her coll.). 11. Lvmnaea stagnalis jugularis Say; Long Lac, on!. 12. Same; Mile 237, Hudson Bay Railway. 13. Same; Wintering Lake, H.B. Ry. MOZLEY: SUB-ARCTIC MOLLUSCA 137 which is available in Canada for settlement by molluscs. Only the northern part of Siberia was glaciated during the Pleistocene period, whereas very nearly the whole of sub-arctic Canada was covered by ice. This has resulted in a somewhat more diversified landscape in Canada as far as the aquatic habitats are concerned. In other words, the ponds, lakes and streams of Canada offer a wider range of habitats than do those of Siberia. The general impression which is gained in travelling through northern Asia is that the landscape, particularly in the south, is more mature than that in Canada, and that the climate for some time past may have been less moist. It should be noted, however, that this conclusion may have been influenced to some extent by the fact that the studies in Siberia were carried out during a period of drought. These differences between the Canadian and Siberian landscapes, if true, might be interpreted as having some effect upon the fauna, in that new species and varieties as they arose would have less chance of finding unoccupied habitats which were suited to their special requirements. This can hardly be considered to be the sole explanation of the condition described, but it is probably an important contributing factor. A somewhat similar conclusion has been reached by Baker (4) regarding the postglacial molluscan fauna of North America, who states that, "Previous to the Glacial Period the country had been reduced to base-level and probably few lakes existed, the physiography being one of rivers with dendritic form of drainage, like the driftiess area in Wisconsin today. After the last invasion, the Wisconsin, the country was greatly changed; in place of rivers there were lakes, swamps, and sluggish rivers. The fauna reacted to this change to such an extent that where previously there had been but one or two varieties in a species, as many as ten developed which were peculiar to the newly glaciated country. Many entirely new species were evolved which have not occurred in any glacial deposits yet examined. The change affected some species more than others, but all have been affected to a noteworthy degree." Some exception might reasonably be taken to the wording of this statement, but there is little doubt regarding the facts upon which the interpretation is based. On the whole, the conclusion, in so far as it relates to the existence of some relationship between physiography and speciation, appears to be a reasonable one. Acknowledgment Owing to the fact that the author is abroad, the galley proof of this paper has been corrected by Mr. A. LaRocque of the Geological Survey, Ottawa, Ontario. References This list includes only those papers which are referred to in the text. A full bibliography may be found by consulting the papers of Baker and Cahn (9), Dali (13), and Mozley (30). BAKER, F. C. The Lymnaeidae of North and Middle America, recent and fossil. Chicago Acad. Sci., Spec. Pub. No. 3. 1911. BAKER, F. C. Fresh-water Mollusca from Colorado and Alberta. Bull. Am. Mus. Nat. Hist. 41: 527-539. 1919. BAKER, F. C. The fresh-water Mollusca of Wisconsin. Wisconsin Geol. Nat. Hist. Sur., Bull. 70. 1928. 138 CANADIAN JOURNAL OF RESEARCH. VOL. 16, SEC. D

BAKER, F. C. Influence of the glacial period in changing the character of the molluscan fauna of North America. Ecology, 11:469-480. 1930. BAKER, F. C. New species and varieties of Helisoma and Gyraulus from Canada. Nautilus, 46 : 6.9. 1932. BAKER, F. C. New species of Lymnaeidae from British America. J. Wash. Acad. Sci. 23 : 520-523. 1933. BAKER, F. C. Two new Canadian Lymnaeas. Can. Field-Naturalist, 48 :69-70. 1934. BAKER, F. C. The fresh-water mollusc Helisorna C'orpu!entum and its relatives in Canada. Nat. Mus. Canada, Bull. 79. 1936. BAKER, F. C. and CAHN, A. R. Fresh-water Mollusca from central Ontario. Ann. Rept. 1929, Nat. Mus., Canada, 41-64. 1931. BIRD, R. D. Biotic communities of the Aspen Parkland of central Canada. Ecology, 11:356-442. 1930. CHRISTY, R. M. Notes on the land and fresh-water Mollusca of Manitoba. J. Con- chology, 4 : 339-349. 1885. CLENCH, W. J. Three new species of Physa. 0cc. Papers Mus. Zool., Univ. Mich. No. 168. 1926. DALL, W. H. Land and fresh-water Mollusks. Harriman Alaska Exped. 13. 1905. DALL, W. H. Mollusks, recent and Pleistocene. Rept. Can. Arctic Exped., 1913-1918. 8, Part A. 1919. DAWSON, G. M. Report on the geology and natural resources in the vicinity of the forty- ninth parallel, etc. Brit. North Am. Boundary Comm. 349-350. 1875. HANHAM, A. W. A list of the land and fresh-water shells of Manitoba. Nautilus, 13 1-6. 1899. LAROCQUE, A. Lasmigona corn pressa (Lea) in the Hudson Bay drainage. Can. Field- Naturalist, 50 : 5 1-52. 1936. LYDEKKER, R. A geographical history of mammals. Cambridge. 1896. MERRIAM, C. H. Life zones and crop zones of the United States. U.S. Biol. Survey Bull. 10. 1898. MOZLEY, ALAN. Segmenlina crassilabris Walker in Manitoba. Can. Field-Naturalist 39 : 83. 1925. MOZLEY, ALAN. Molluscs from the Manitoba-Ontario boundary. Nautilus, 39 :59-61. 1926. MOZLEY, ALAN. Molluscs from the Lake Brereton district, Manitoba. Can. Field- Naturalist, 41: 59-61. 1927. MOZLEY, ALAN. Some molluscs from western Canada. Nautilus, 40 :56-63. 1927. MOZLEY, ALAN. New records of western Canadian Mollusca. Nautilus, 41 : 13-18. 1928. MOZLEY, ALAN. Further records of western Canadian Mollusca. Nautilus, 43 :79-85. 1930. - 26.JMozLEY. ALAN. Reports of the Jasper Pirk Lakes Investigations, 1925-1926. The Mollusca of Jasper Park. Trans. RoSoc. Edinburgh, 56 : 647-669. 1930. MOZLEY, ALAN. A biological study of a temporary pond in western Canada. Am. Naturalist, 66 : 235-249. 1932. MOZLEY, ALAN. A new variety of Lyrnnaea from the Rocky Mountain region of Canada. Proc. Malac. Soc. London, 20 :241-243. 1933. MOZLEY, ALAN. The local and geographic distribution of some Rocky Mountain Mol- lusca. Proc. Malac. Soc. London, 20 :214-221. 1933. MOZLEY, ALAN. Moilusca from the provinces of Saskatchewan and Alberta, Canada. Proc. Malac. Soc. London, 21: 138-145. 1934. MOZLEY, ALAN. Postglacial fossil Mollusca in western Canada. Geol. Mag. 71: 310-382. 1934. 32 MozLEv, ALAN. The variation of two species of Lymnaea. Genetics, 20 : 452-465. 1935. 33 MOZLEY, ALAN. The fresh-water and terrestrial Mollusca of northern Asia. Trans. Royi Soc.. Edinburgh, 58 : 605-695. 1935. Moziy, ALAN. The statistical analysis of the distribution of pond Mollusca in western Canada. Am. Naturalist, 70 :237-244. 1936. RICHARDS, H. G. Recent and Pleistocene marine shells of James Bay. Am. Mid. Nat. 17: 528-545. 1936. 36, RUSSELL, L. S. Pleistocene and post-Pleistocene molluscan faunas of southern Sask- atchewan. Can. Field-Nat. 48 :34-37. 1934. \VEITEAVES, J. F. List of fresh-water Mollusca from Manitoba and the valley of the Nelson River. Geol. Surv. Can. Rept. Prog. 1878-1879, pp. 61-62c. 1880. \VHITEAVES, J. F. Mollusca. Geol. Surv. Can. Rept. Prog. 1879-1880, pp. 75-76c. 1881. WHITTAKER, E. J. Fresh-water Mollusca from Mackenzie River basin, Canada. Nautilus, 38 :8-12. 1924. WINSLOW, M. L. Mollusca of North Dakota. 0cc. Papers Mus. Zoo1. Univ. Mich. No. 98. 1921.