362 Memoirs of the Indian Museum. [Vol. V, 



M. jenkinsi, Preston, the latter of M. striatula, Hanley ; M. taprobanensis , Preston; 

 M. emarginata , Reeve ; M. annandalei and M. celator, Preston. In practically every 

 series of specimens we have examined, either from the Gangetic delta or from the 

 Chilka Lake, there is a complete transition between these two groups, and many 

 of the specimens identified by Blanford and by Mr. Preston as M. striatula have the 

 lower margin straight, while in Reeve's figure it is much more nearly so than in Han- 

 ley's. In fact, so far as it is possible to say without seeing those of M. striatula and 

 M. emarginata, we believe that the types could be arranged in the following order 

 so as to form an almost complete series in this respect : — 



1. M. jenkinsi. 4. M. taprobanensis. 



2. M. cochinensis. 5. M. emarginata. 



3. M. striatula. 6. M . celator. 



7. M. annandalei. 



With the concavity of the lower margin in these forms a relative narrowing and 

 elongation of the whole shell is often correlated and in those types in which this mar- 

 gin is straightest, the relative depth of the shell is greatest. M. taprobanensis , how- 

 ever, is a rather broad form. 



The type of M. celator is remarkable for its abnormal outline as seen in dorsal 

 view and for the thickened and eroded condition of the antero-superior region of the 

 shell. We find precisely similar shells in a number of our series and also others in 

 which abnormalities of a similar nature occur in other parts. Photographs of abnor- 

 mal shells and of the types of four species described by Mr. Preston are included in 

 the series figured on pi. xv, figs. 7-18. 



The development of radial costae is an extremely variable character, but the 

 surface is less frequently quite smooth than in M. undulata. In many specimens 

 the costae are quite as fully developed as in the var. crassicostata of that species, but 

 they are never branched as in M. subramosa, Hanley. Another variable character 

 is the development of concentric growth-lines ; abnormal specimens occur, especially 

 in forms resembling M. celator, in which they are greatly accentuated. Sculpturing 

 of the shell is not, however, correlated in any way with its shape. It seems to us 

 impossible to recognize Branchyodontes , Swainson, even as a subgenus. 



The colouration is also very variable, but the variation is not quite of the same 

 nature as in M. undulata, the different colours being as a rule more diffused as well 

 as duller. In some specimens, however, zigzag transverse purple lines and longitu- 

 dinal striae can be detected, but the purple is usually less red and the ground-colour 

 of a bluer green. Young specimens are as a rule brighter than adults and fully 

 developed shells are sometimes of an almost uniform dull brown. 



Perhaps the best illustration we possess of correlation between different forms of 

 shell and their environment is a worm-eaten log covered with mussels of this species. 

 It has long been in the Indian Museum and almost certainly came from the Gangetic 

 delta. Among the shells from this log are some that are relatively short and broad 

 and have the lower margins perfectly straight, while others exhibit every degree of 



