494 R. H. RASTALL 
found frequently as sulphide segregations, either with or without 
copper, nickel, and cobalt. Thus it is seen that although iron as 
an element is common to both ends of the series, nevertheless it 
occurs typically in a different state of combination in each. 
Having thus arrived at the fundamental generalization that 
there are two principal groups of ore segregations of very early 
and very late consolidation respectively, we must now proceed 
to consider in more detail what particular type of differentiation 
may be applied in explanation of each. 
At the present time five different sets of causes are commonly 
put forward in explanation of the varied set of phenomena com- 
prised under the head of differentiation; including in this term not 
only the production of heterogeneity in single intrusions, but also 
the separation of minor magmas of varying composition from one 
original magmatic solution. These may be summarized as follows: 
a) Marginal concentration of minerals of high freezing-point 
by diffusion of liquid molecules immediately preceding and during 
crystallization. 
b) Differentiation by gravitational sinking or rising of crystals 
in a solidifying magma. 
c) Separation of a magma into two or more immiscible or 
partially miscible layers. 
d) Assimilation by stoping with its attendant fluxing effects. 
e) Squeezing out of liquid residue from a crystalline sponge or 
network. 
This is not the place to discuss the evidence as to the competence 
of each of these processes as a general factor in differentiation. It 
is probable that all are applicable in different cases and under 
different sets of conditions: we have only to consider which of 
them can be invoked to explain the various types of ore occurrence 
here already briefly alluded to. It is evident that in the two great 
groups of ores of early and late consolidation very different physical 
conditions must prevail, especially as regards temperature and 
presence of the so-called mineralizing agents or fluxes. In the 
early group the temperatures must be very high, and we have reason 
to believe that in the basic rocks fluxes are unimportant; hence it 
appears that any one of the first three categories should be specially 
applicable. 
