349 
in opposite directions. The first problem to be considered is 
whether the compositions and the mutual arrangement of nau- 
jaite and lujavrite can be accounted for by the hypothesis of 
differentiation in place. 
When this problem has been dealt with, the relation of 
naujaite to sodalite-foyaite will be separately considered. 
Finally the origin of the kakortokite mass with its regular 
alternations of white, red, and black sheets will be discussed. 
Origin of naujaite and lujavrite. — It is generally ad- 
mitted that differentiation in place may be produced either by 
fractional crystallization’, with one or more minerals crystalli- 
zing at the coolest border of the magma chamber, while the 
remaining magma by diffusion or convection currents is kept 
homogeneous, or by gravity, by which the first formed crystals 
are mechanically separated from the magma. In both cases 
the normal result of the process is a solid rock which con- 
tains a very large proportion of the first mineral to crystallize, 
and a residual magma showing corresponding impoverishment 
in that substance’. 
In the present case we are taught by observation that the 
naujaite crystallized before the lujavrite. It has also been stated 
that the first formed mineral of the naujaite is sodalite. The 
first question is, therefore, whether the composition of naujaite 
is such as would result if a portion of the agpaitic magma has 
become strongly enriched in sodalite, and if so whether the re- 
maining portion, impoverished in that substance, would assume 
the composition of lujavrite. 
This question must be answered in the affirmative: a 
mixture of 63 per cent. of ‘mean agpaite’ and 37 per cent of 
1G. F. Becker, Fractional Crystallization of Rocks. Amer. Journ. Sc., 4 
ser., vol. IV, 1897, p 257. 
2 As TEALL has pointed out complications may arise when crystals de- 
scending to lower levels are dissolved there (Geolog. Magazine, new series, 
decade 3, vol. II, 1885, p. 119.) 
