276 Records of the Indian Museum. [Vol. XXII, 



suffering from a very serious drought at the time of my visit 

 (September, 1918). It had an abundant and healthy vegetation 

 of submerged weeds and plants with floating leaves (lyimnan- 

 themum and water-lilies) were also abundant and healthy. The 

 molluscs were taken among the leaves in great profusion. Their 

 habitat was of a type more common in Lower Bengal than in 

 Madras. 



Tliis exhausts the number of true races with which I am 

 acquainted and we may now turn to the four well-marked phases 

 of the species. 



Phase annandalei, Kobelt, 

 (Plate II, figs. 5-8.) 



lijoS. Vivipava annandalei. Kobelt. Xachr. ilalak. Ges. l.X, p. l6i. 

 u)OQ, Vivipai-a annandalei, Kobelt, in .Martini and Chemnitz. Condi. 

 Cat. I. 21 (21, p. 296. pi. 57. figs. II. 12. 



Kobelt refers to this form as "cine krilische jorm.'" I was 

 prepared to accept it as distinct until I had become acquainted 

 with its habits and had ascertained the fact that in some ponds 

 (e.g. the tank in the ^Museum compound, Calcutta) it graded insen- 

 sibly into the typical form of V . beiigalensis, or rather into a small 

 but not otherwise peculiar phase thereof. Kobelt' s description 

 and figures were based on somewhat exceptional specimens of 

 relativelj' large size and proportionately broad shell. .Such indivi- 

 duals occur occasionally but are by no means typical of the phase. 

 On pi. II four shells are shown. Fig. 5 represents one of Kobelt's 

 type-series, which is from Vizagapatam in the north-east of Madras. 

 The other three (figs. 6-8) are more typical. The shells examined 

 by Kobelt were, moreover, old specimens and had lost the trans- 

 lucency characteristic of the phase. 



The shell is always very thin and light and usually small. 

 The more elongate type is the commoner of the two that occur, 

 but the type-series chances to belong to the other. Apart from the 

 thinness and translucency of the shell, the most characteristic 

 features are the gradual increase in size of the whorls, the shallow- 

 ness of the suture, and particularly the shape of the aperture, 

 which is distinctly subrhomboidal and subangulate at its anterior 

 extremity. The d-ark bauds are sometimes a little incrassated. 

 Some shells of the phase come very near to some of the Assamese 

 race balteata. The animal is usually very pale in colour, but occa- 

 sionally almost as dark as that of the forma typica. I have noticed 

 that in living specimens kept in an aquarium it gradually becomes 

 darker. 



This phase is found in tht territory both of the forma typica 

 and of the race eburnea, but I can find no difference between 

 specimens from Calcutta and those from Hyderabad, Deccan. 

 It is commoner in the vicinity of both cities and almost always 

 occurs in pools of rather foul water used for domestic purposes. 



