166 Geologij of EicliDioiid County. A. Y. 



serpentine at Tompkinsville overlies the granitic rocks as it does 

 at Jersey City. Tliis can only be definitely ascei'tained by bor- 

 ings, as the contact of the two rocks cannot be observed. We 

 have the negative evidence, however, that were the serpentine 

 older than the granite, the latter would ])robab1y be found in 

 greater quantity, and in more localities than it really is. Hence 

 the probability is that the two rocks have the same relative ver- 

 tical position on 8taten Island that they have at Jersey City : 

 and they are so indicated on the accompanying maps and sections 

 (Plates "XV and XVI). 



Fifth. — Ellis and Bedloe's Islands, in the Upper Bay. are 

 directly between the two outcrops on the line of strike, and are 

 said to be underlaid by gneiss ; but no very definite information 

 is obtainable on this point. 



As to the origin of the serpentine rocks, I have no new theory 

 to advance, and consider the one which regards them as meta- 

 morphosed highly magnesian limestones to be more in accord- 

 ance than any other with the facts as observed. The reasons 

 for this opinion are as follows : — 



Fird. — It is highly improbable that they were igneous in 

 origin, because they contain about fourteen per cent, of water, 

 are associated with gneissic rocks which we know^ are metamor- 

 phic, and are stratified, — although the stratification can only be 

 distinguished at a few places, and there not very plainly, on ac- 

 count of the cleavage i)lanes which cut the rock in all directions. 



Second. — They are certainly not unchanged sediments, be- 

 cause there are no magnesian silicates known to be formed as 

 sediments on such a large scale as these strata present ; therefore 

 they must be either metamorphosed sediments or metamor- 

 phosed metamorphic rocks. 



Third. — These rocks could not have been sandstones or shales, 

 becausQ they would have become quartzites or feldspathic rocks 

 by metamorphism ; and while serpentine certainly is the result 

 of the decomposition of hornblende in some cases, the extent of 

 the formation on Staten Island Avould render this method of 



