A JOURNEY IN BRAZIL. 
should now entertain the idea that the surface rocks are 
everywhere decomposed, and that there is no erratic forma- 
tion or drift here. But upon close examination it is easy 
to perceive that, while the decomposed rocks consist of 
small particles of the primitive rocks which they represent, 
with their dikes and all other characteristic features, there 
is not a trace of larger or smaller boulders in them ; while 
the superincumbent drift, consisting of a similar paste, 
does not show the slightest sign of the indistinct stratifica- 
tion characteristic of the decomposed metamorphic rocks 
below it, nor any of the decomposed dikes, but is full of 
various kinds of boulders of various dimensions. I have 
not yet traced the boulders to their origin ; but the majority 
consist of a kind of greenstone composed of equal amounts 
of a greenish black hornblende and feldspar. In Entre 
Rios on the Parahyba, I was told by an engineer on the 
road that in Minas Geraes iron mines are worked in a 
rock like these boulders. This week I propose to explore 
the Serra da Mantiqueira,* which separates the province 
of Rio from Minas, and may advance the question further. 
But you see that I need not go to the Andes to find 
erratics, though it may yet be necessary for me to go, 
in order to trace the evidence of glacier action in the 
accumulation of this drift ; for you will notice that I 
have only given you the evidence of extensive accumu- 
lations of drift similar in its characteristics to Northern 
drift. But I have not yet seen a trace of glacial action 
properly speaking, if polished surfaces and scratches and 
furrows are especially to be considered as such. 
" The decomposition of the surface rocks to the extent 
to which it takes place here is very remarkable, and points 
* Mr. Agassiz was prevented from making this excursion. 
