AN ISLAND IN THE CHILKA LAKE. 263 



relatively small, usually rectangular blocl^s as a rule much broader than deep. 

 This is owing to the fact that it is separated by the planes of stratification into 

 well-defined layers mostly less than a foot deep and traversed in a vertical direction 

 by narrow veins of softer consistency than the bulk of the rock. There is none 

 of the piling up of great blocks into pyramids so conspicuous in the surface 

 formation of many of the hills and headlands of the neighbourhood. 



A small pond has been dug near the middle of the island. Its water-level differs 

 greatly at different seasons and in the early autumn, a considerable depth of rock is 

 submerged which becomes dry in winter. The surface of this rock, unlike that sub- 

 merged in the brackish water of the lake at the same season, weathers black. The 

 water of the pond is very nearly fresh when it is full, and onl)^ slightly saline when 

 it has sunk to its lowest level." 



I have to thank Dr. E. H. Pascoe, of the Geological Survey of India, for the 

 following note on rock- specimens : — 



" The rock specimens from Barkuda Island, Chilka I^ake, consist of two extreme 

 types of quartz- schists with intermediate forms. On the one hand, you have a 

 garnet-quartz-schist or, as T. L. Walker calls it, a garnetiferous quartzite, and, on 

 the other hand, a khondalite. In the former the bulk of the rock consists of quartz 

 while the garnets vary in size and number and, in the specimens under consideration 

 which are evidently from near the surface of the ground, are much weathered ; one 

 specimen shows pronounced foliation. In the khondalites or garnet-sillimanite- 

 schists, the quartz ground-mass is very much crushed, the garnets weathered and the 

 sillimanite mostly in large disjointed parallel needles. These rocks, according to 

 T. L,. Walker who gave them their name, are paragneisses, representing in all prob- 

 ability ancient metamorphosed sediments. Owing to the absence of felspar, the 

 soil derived from the rocks described would consist of a ferruginous sand with scarce- 

 ly any true clay and very little soluble food for plants. Similar schists are found 

 over a great part of the Ganjam Malias in Kalahandi, along the eastern border of the 

 Eastern Ghats in Vizagapatam, in Coorg and elsewhere." 



Changes in Water-leveiv on the Shore of the Iseand. 

 The changes in water-level on the shore of the island are of importance in 

 reference to the fauna of the foreshore. They are of two kinds, seasonal and occasional. 

 The seasonal changes vary in extent from year to year in correlation with the rain- 

 fall in the basin of the Mahanadi, a branch of which enters the northern part of 

 the lake, and also in all probability with the size and position of the sea-mouth, which 

 vary as a result of changes in the currents on the coast and other factors of which 

 at present we know little. In a normal year the water is highest about the middle 

 of September, but occasional floods may raise it temporarily still higher earlier in 

 the year. It begins to rise as soon as the south-west monsoon is well established, 

 sinks as the rains diminish and is at its lowest in June just before they corrimence. 

 The area then left bare on the shore is considerable, but variable in area from year 

 to year. 



