46 
PENNSYLVANIA ANORTHOSITES 
a long period of time. As the magma approaches the crystalliza¬ 
tion temperature, the mafic minerals are the first to separate out. 
Almost simultaneously with the metasilicates, calcic plagioclase, 
of composition approximating bytownite (Ab^ Ang), begins to 
crystallize. While the mafic crystals, because of their high speci¬ 
fic gravity, will sink almost as soon as formed, the calcic plagio¬ 
clase will float in the magma because its density more nearly ap¬ 
proaches that of the liquid. Experiment shows that Ab, An has 
a density of 2.733, while that of diabase-glass at room tempera¬ 
ture is 2.763. Heat will cause the liquid to expand more than 
the crystals, the difference of expansion up to 1200° being about 
2 per cent. This nearly neutralizes the above relation. The pres¬ 
ence of volatile constituents in natural magmas, as opposed to dry 
melts, tends to make the crystals very slightly heavier than the 
liquid. While the crystals are suspended in the liquid they grad¬ 
ually change in composition from Ab^ Aiig toward Ab^ An^. The 
loss of the metasilicates has produced an impoverished residual 
liquid which is increasingly more silicic, and in which the content 
of mineralizers present in the original magma becomes more and 
more concentrated. The plagioclases will be enriched in soda 
just so long as the density of the liquid will support them. After 
they have reached the composition Ab^ An^ they are no longer in 
equilibrium with the liquid; hence they will sink and gradually 
form a closely packed mass of plagioclase of approximately labra- 
doritic composition. The lighter residual magma which is 
squeezed out is enriched in albite, orthoclase, and quartz, and upon 
crystallization forms further silicic differentiates, or solidifies as 
one rock type, according to whether the residual magma is cooled 
slowly or rapidly. Bowen interprets the syenite associated with 
the Adirondack anorthosite as a residual differentiate of the gab- 
broid magma. Any later differentiate may have intrusive rela¬ 
tions in respect to an earlier one. Hence the intrusion of the 
syenite into the anorthosite is not inconsistent with this theory. 
It is evident that if anorthosite is formed by such a process as 
this, it should be essentially solid, as Bowen maintains, and devoid 
of interstitial liquid. We should therefore expect to find that 
any movement subsequent to the settling of the feldspar crystals 
43 Bowen: private communication. 
