GENESIS OF THE ALKALINE ROCKS 123 
founded on the slowness of diffusion, a process so sluggish that 
homogeneity cannot be established or re-established before those 
phases have been separated by gravity or other forces. 
Question of liquid immiscibility—Bowen decides against the 
hypothesis of liquid immiscibility as a condition for the splitting of 
silicate fractions. His chief reason is the negative evidence of 
experimental melts and of glassy lavas; in neither case have 
immiscible globules been detected, though the artificial or natural 
quenchings have taken place at all stages in the cooling histories. 
In spite of such observations, even on melts “in bombs under high 
pressure of water-vapor,”’ one cannot but question whether melt 
or lava flow truly represents the controlling conditions in a great, 
quiet, magma chamber. ‘There, at pressures of hundreds or thou- 
sands of atmospheres, the magmatic life is to be measured in weeks, 
months, years, or centuries. Bridgman’s work indicates the 
probability of the great undercooling of such a magma by pressure. 
Kuenen has proved that for some liquids pressure raises the critical 
temperature of unmixing, for other liquids the reverse. Is immis- 
cibility among silicate solutions developed by undercooling through 
pressure? Further, are the possibilities of the colloidal state for 
silicates at continued high pressure sufficiently understood ? 
Until these and kindred problems are solved a definite denial of 
liquid unmixing in magmas should be postponed. 
Bowen’s second reason for rejecting that principle is found in 
the high melting temperatures of olivine, magnetite, and other 
components of monomineralic rocks. Temperatures so high can- 
not be easily, if at all, assumed for most magmas. Yet it is not 
ascertained that the phase which has separated was pure olivine, 
magnetite, or any other of the minerals forming the actual mono- 
mineralic mass. The temporary dissolving of a very small propor- 
tion of hydrogen, oxygen, or water in any of these substances gives 
the solution a lower temperature of consolidation than that of the 
corresponding gas-free phase. The writer has long realized the 
difficulty of crediting all the observed serpentinization of dunites 
™P. W. Bridgman, Proc. Amer. Acad. Arts and Sciences, XLVII (1912), 5303 
Physical Review, IIL (1914), 182; zbid., VI (1915), 14; J. P. Kuenen, Phil. Mag., 
VI (1903), 637. 
