460 N. L. BOWEN 
THEORETICAL DISCUSSION 
A theoretical discussion of the crystallization of rocks can, in 
the present state of knowledge, lead to but few definite conclusions. 
It is not at all clear why there should be in any rock an order of 
cessation of crystallization. If the final crystallization is that of a 
eutectic mixture there should be no such order. It may be pointed 
out, however, that the simple eutectic results only when each 
constituent lowers the melting-point of every other constituent. 
Solid solution often brings it about that there may be a raising, 
and with rocks, in which, as a rule, solid solutions (mix-crystals) 
are very common, it cannot be said what the combination of pos- 
sible “‘raisings”’ and ‘“‘lowerings’’ may result in. 
Theory, then, helps very little in making any prediction as to 
what ought to happen. It can, however, be safely stated that 
in any system the tendency of special richness in a certain con- 
stituent will be in the direction of causing that constituent to 
begin to crystallize in the early stages. A “constant order of 
crystallization”’ capable of almost universal application to all 
rocks, whatever the variation in mineral proportions (Rosen- 
busch’s rules), is from this point of view out of the question. 
The order stated is, however, based on considerations already 
discussed and is really only the order of cessation of crystallization. 
Why a certain order should be sufficiently common to be noticeable 
is not apparent, but it does seem more reasonable that this should 
be true of the order of cessation of crystallization for various rocks 
than of the order of beginning of crystallization. 
ORDER OF BEGINNING OF CRYSTALLIZATION 
If it is admitted that the “order of crystallization”’ as commonly 
stated for any rock is merely the order of cessation of crystalliza- 
tion, the question arises as to whether there is no clue to be had 
to the order of beginning of crystallization. 
In laboratory practicet a method of following the course of 
crystallization in a mixture at high temperatures is to hold the 
mixture in a furnace at a definite temperature for a period of time 
tJ. H. L. Vogt, Die Silikatschmelzlésungen, 1, 106; E. S. Shepherd and G. A. 
Rankin, Am. Jour. Sci., XXVIII (1909), 293; C. H. Desch, Metallography, 1910, 210. 
