DECCAN TRAP. 193 



But this does not seem of very great importance, for there is no reason 

 why one of the fresh- water lakes above spoken of may not have existed 

 prior to the first pouring- out of the Deccan trap, and have been filled 

 up ; its deposits being thereby preserved from denuding action to the 

 present time. 



Although very many sections of these marls, clays, and sandstones 

 were examined carefully, it was only in three that any organic remains 

 were found ; the one just referred to occurs at the village of Todihal on 

 the south bank of the Krishna (four miles north-west-by-north of Bilgi, 

 and fifteen miles north-north-east of Kaladgi) ; the second a mile west- 

 north-west of Mamdapur, and six and a half miles south-east-by-east 

 of Gokak; the third at Uparhatti, a mile north of the second 

 locality. 



The Todihal section is about two-thirds of a mile south-east-by-east 



of the village. The section is not very distinct 

 ¥hysa bed at Todihal. . .„.,,. 



owing to the quantity of soil (a mixture of decom- 

 posed trap and regur) covering the edges of the several formations, but 

 by making some excavations, the lie of the beds was ascertained suflS.- 

 ciently to understand their real relations. 



The outcrop of the marl is so much masked that attention would 

 not have been attracted to it, but that a specimen of Fhysa Prinsepii 

 was seen lying loose on the surface. Even then it was not till after I 

 re-visited the section and made several excavations that the source whence 

 the Thysa came was ascertained with perfect certainty. 



The thickness of the marl bed is from 6 to 8 feet at the spot where 

 I made the excavations. The fossils contained in the freshly-exposed 

 marl were extremely brittle and difficult to extract entire, but those 

 which had been weathered out were in very good condition. Only one 

 or two showed any trace of silicification, the test being unaltered except 

 in color in most cases. A large percentage of the shells were found to be 

 2 a { 193 ) 



