AN ISLAND IN THE CHILKA LAKE. 273 



probable that the bleached shells are of more recent origin than those of the 

 brackish-water forms found with them. As the country is liable to be flooded the 

 beds have frequently been disturbed. The shells of all species occur in very fine grey 

 sand. 



If we compare these beds at the head of Rambha Bay with those on Barkuda, we 

 notice very little difference so far as the composition of the fauna is concerned, 

 except in the absence of probably adventitious freshwater forms in the latter. The 

 only other species that is not common to the two localities is Sunnetta scripta, 

 of which only a few valves were found on Barkuda. 



The soil of Barkuda is too shallow to admit of any differentiation into distinct 

 layers or strata, but a certain difference may be observed between the shells found in 

 it and those from the sandy shores of the northern and western ends of the island. 

 In the soil the most abundant species is Meretrix casta, while Potamides cingnlatus is 

 relatively and the Cuma and the Area actually scarce. On the shore, on the other 

 hand, the most abundant species is Cuma disjuncta, while the Meretrix is scarce and 

 the Potamides has not been found at all. Furthermore, bivalve shells are seldom to 

 be discovered on the exposed eastern end of the island, on which the Cuma 

 is particularly common. These facts seem to point to a certain diversity of origin 

 among the fossils of the island, but are probably correlated rather with differences in 

 the habits of the molluscs than with differences of date. We know that Cuma 

 disjuncta is essentially a rock-loving species, feeding on the small mussels [Modiola 

 striatula) that adhere to rocks and stones, while all the other species frequent a soft 

 bottom. Meretrix casta cannot live in soft mud but prefers clean sand, while Area 

 granosa is much less particular ; Potamides cingidatus prefers very shallow water with 

 a bottom composed of either sand or sandy mud. It is quite probable, to judge 

 from the condition of the shells, that the deposits of Meretrix on the island 

 do not occupy precisely the situations in which the mollusc originally lived, and 

 all the shells from the shore of the island have the appearance of having been rolled 

 about in the water; but the distribution of the species in a subfossil state corres- 

 ponds on the whole with their habits in a living one. 



The deposits at the head of Rambha Bay are so often disturbed by floods that 

 it is difficult to derive from them any very precise information as to relative dates of 

 origin. The only difference of phase between the shells they contain and those 

 from Barkuda is to be found in the Meretrix, which is larger, thicker and more 

 variable in shape in the beds on the mainland than it is in the island deposits. 

 Moreover, the shells show a greater tendency to be ovoid in the former, owing to 

 their posterior extremities being frequently more produced. In other words, the 

 sublossil shells of the mainland are further removed from the forma typica 

 of the species than those from Barkuda. As Hornell has shown {op. cit., p. 167; 

 1917) conditions are now very favourable for the growth of the forma typica in the 

 outer channel of the Chilka Lake. The beds near Rambha probably represent an old 

 channel which ran southwards and eastwards into the Bay of Bengal at a time 

 previous to the separation therefrom of the lake, but conditions in it must have 



