1 1 4 The American Geologist. August, i898 
magmas, distinguished by the relative chemical proportions 
existing between the alkalies, lime, magnesia, alumina, iron 
and silica. Of this the author remarks that, "except the 
fovalitic and the peridotic magmas, the others are arbitrarily 
defined and are in reality only an embarrassment in the study 
of rocks very dissimilar. The theoretical examination into 
the 'kerns' of Rosenbusch, distinguished by stoechiometric 
proportions, leads at once to a mineralogical determination, 
but it reveals then one important lack, viz., the black and white 
micas; the excess or the absence of alumina has not been 
sufficiently taken into consideration, and the history of acid 
and intermediate magmas is singularly obscure." 
As to the "differentiation" of magmas, put forth (1892) 
by Iddings, under the idea that the fundamental magma 
undergoes sundry changes at volcanic centres producing a 
succession of different rocks in a certain order, the author dis- 
credits, first, the assumption that such a fundamental magma 
exists, and second he considers that the facts of observation do 
not warrant such segmentations as have been predicated by 
Iddings. However, the author considers the "capital memoir" 
of Iddings on the origin of eruptive rocks the most instructive 
and the most suggestive. Iddings' criticisms of the views of 
his predecessors are more convincing than his own conclu- 
sions. 
The cause of such differentiations, if they exist, Iddings 
supposes to consist principally in differences of temperature, 
the extremely variable dissolvant, changing with the nature of 
the magma, is influenced by the conductibility of the walls of 
the subterranean chambers which it fills. Pressure is quite 
subordinate to heat. This is essentially the idea of Soret, — it 
IS the theory of a closed chamber (vase clos). In this pro- 
cess of differentiation, as set forth by Iddings in his abscissas 
and ordinates, the relative proportion of the alkalies, potash 
and soda, bears the most important role, being the least 
variable in each family. The first rock to appear, in a family 
of common consanguinity, according to Iddings, is an in- 
termediate one, the later rocks being variations in opposite 
directions toward extremes of acidity and basicity. This in- 
termediate rock may be the result of differentiation from a 
primitive magma, which thus repeats essentially the segmen- 
