REPORT OF TEE CHIEF ASTRONOMER 771 



SESSIONAL PAPER No. 25a 



degrees of its ' solidification point.' Harker's objection that differentiated rocks 

 are seldom sharply separated by distinct surfaces of contact is not a strong one. 

 In the first place, the separation between such silicate differentiates is in many 

 cases remarkably sharp, especially when we consider the scale of operations in 

 magmatic chambers. Secondly, we could hardly expect the separation to be as 

 perfect between these viscous and highly complex magmatic fractions as is, for 

 example, the separation between phenol and water. In summary, it may be 

 stated that a host of field and laboratory observations favour the application of 

 the liquation (limited miscibility) principle to natural silicate magmas; and 

 tbat not a single fact is known to the writer which conflicts with that assump- 

 tion. The efforts of physical chemists should be spent, not on denying its 

 validity, but in defining the conditions under which the liquation so often 

 demonstrable in nature has taken place. 



3. Gravitative Differentiation. — Gravity is one of the controlling forces in 

 separating crystals from their mother-liquors. Illustrations of this truth are 

 given in the writer's paper on the ' Origin of Augite Andesite and Related 

 Ultra-basic Rocks.'* A striking example has since been reported from the New 

 Jersey diabase, by Lewis.f The sinking of crystals is expected to hare its 

 maximum differentiating effect within volcanic vents, where the agitation of the 

 magma tends to prevent undercooling and to promote crystallization while the 

 magma retains relatively low viscosity. These conditions are chiefly due to the 

 upward passage of gases in volcanic vents, which in this respect are contrasted 

 with dikes, sheets, and laccoliths. The steady or intermittent working of two- 

 phase convection within the lava at surface vents is, as we have seen (page 711), 

 competent to keep the column long within the temperature interval of crystal- 

 lization. The writer believes, therefore, that Darwin's theory of fractional 

 crystallization under the control of gravity, is a permanent acquisition to the 

 petrology of effusive rocks. 



It is generally agreed, however, that differentiation is, as a rule, a splitting 

 into liquid fractions. For example, it is impossible to believe that the drastic 

 differentiation in the Moyie sills or in the Sudbury sheet (see chapter X.) can 

 be due to the settling of solid crystals. 



In volcanic vent or intrusive body gravity must tend to separate the 

 liquated fractions. The lighter always rising to the top of the magma chamber, 

 the geologist will rarely be permitted to see the rock representing the basic, 

 heavier differentiate in subjacent bodies. He may find it in the form of dikes 

 cutting the overlying, more rapidly solidified differentiate. 



The relative importance of fractional crystallization and liquation can be 

 estimated only after the physical chemistry of magmas becomes better imder- 

 stood. Meanwhile, we may use the expression 'gravitative differentiation,' 

 as a name for the chief process in magmatic separation, without therewith 

 implying whether fractional crystallization or liquation is the more active in a 

 given case. 



•Jour. Geology, Vol. 16, 1908, p. 411. 



t J. V. Lewis, Ann. Rep. State Geologist of New Jersey for 1907 (1908), p. 129. 

 25a— vol. iii— 50 



