336 J. D. Dana on some points in Lithology. 
crystalline rocks were once actual mud beds or sand beds ; and 
even part of the eruptive rocks may have been so in their 
earlier history. Strongly drawn limits no where exist. Rocks 
are hence of different kinds, not of different species ; and only 
those mixtures are to be regarded as distinct kinds of rocks which 
have a sufficiently wide distribution to make a distinct name 
important to the geologist. Other kinds have to be classed as 
varieties, if worthy of that degree of recognition. 
In the following pages I propose to consider the value of some 
of the distinctive characters which are generally accepted at the 
present time in defining certain kinds of rocks 
some magnetite or titanic iron: and so does doleryte. Diabase. 
to a large extent, is a crystalline-granular rock ; so is doleryte. 
iabase was formerly supposed to be peculiar in containing 
chlorite, but it is now proved, as asserted by Rosenbusch, that 
chlorite is not an essential characteristic, so that diabase may 
be chloritie or not; and the same is true of doleryte. Old dia- 
base was described as differing from the younger rock doleryte 
o glassy portions or grains among the crystal” 
ular diabase, (2) diabase-prophyrite, and (8) glass-bearing 4 
; for dole 
absence of glassy portions, the two rocks are identical. Anal- 
from the Archean to the Tertiary, and of 
