96 KING: KADAPAH AND KARNÜL FORMATIONS. [PART tr. 
soon found that the thin beds of the Banaganpilly group are cropping 
out from between the limestones and the Kapapaus below. Again, 
further north, about Sultanpoor on the right bank of the Toongabudra, 
the beds which Mr. Foote considers as of this group are again overlapped 
by these limestones, though only to a small extent. The limestones 
show at the village of Shaitancotta on the right bank of the river where 
they rest directly on granitoid gneiss, while a short distance south the 
grits of the Banaganpillys begin to show and thicken out somewhat in 
the direction of Sultanpoor. Opposite this point, on the left bank of 
the river, the Nerjee limestones rest on the grits at Hoopulpad, but as 
they bend round west of that village and run down under Alumpoor, 
they are again seen lying directly on the gneiss. 
From what can be seen of the present workings and those which 
AK c are old and deserted, the diamonds, or the seams 1n 
which the gems occur, are confined to the lower 
beds, possibly to a sub-division of the group which, however, we have 
not been able to distinguish with any certainty. At Banaganpilly one 
can see whence the diamonds are obtained, and partly make out whence 
they must have been obtained in other now deserted workings; but 
there are many other localities with apparently identical rocks where 
diamonds might be expected to be found, but where the people have 
never tried for them and, what is more, where they do not believe they are 
to be found. 
The true diamond-bearing beds are low in the group, and a closer 
examination than we could make may even eventually show that they 
are separable from the strata which overlie them where the seam of 
quartzites is thickest. There is no separation by shales and limestones 
as an intermediate band: it can only be a separation by overlap. For in- 
stance, the lower beds of the group may not have been deposited over the 
whole of the bottom of the basin, but here and there in patches, the upper 
beds having overlapped these—which would be one way of accounting 
© 96) ۱ 
