﻿126 



ME. G. W. TYEBELL OS THE 



[vol. lxxii, 



specific gravity somewhere near 2 - 83. Similarly, Dr. Daly has 

 calculated that olivine, which has a specific gravity of 3 40 at 

 ordinary temperatures, would only have a specific gravity of 

 approximately 3 - 30 when crystallized at 1100° C. 1 The contrast 

 between the specific gravity of crystallized olivine at 1100° C, 

 and that of molten picrite at 1200° C, is sufficiently great to 

 warrant the conclusion that the olivine- crystals, when formed, 

 would sink to lower levels within the magma, and would continue 

 to do so at decreasing speed until the increase of viscosity con- 

 sequent upon cooling inhibited further movement. 



It is conceivable that, under favourable conditions, an ahnost 

 monomineralic layer might be formed in a magma by the sinking 

 of crystals. The favouring conditions would be early crystalliza- 

 tion of a comparatively heavy mineral in a highly fluid magma, 

 unimpeded by the presence of other minerals. The rate of crystal- 

 lization should be rapid, the sizes of the crystals should be large (as 

 the rate of sinking is proportional to their bidk), and the mineral 

 should appear suddenly in large quantity. Olivine and augite 

 frequently satisfy these conditions, and on sinking the} r would 

 form, respectively, layers of dunite and pyroxenite. Iron-ores less 

 frequently satisfy the conditions requisite for the formation of 

 monomineralic layers. In most magmas their crystals are small, 

 and the development of the minerals is feeble, though they may be 

 the first to crystallize. Nevertheless, in a few cases, the formation 

 of an iron-ore rock as a result of gravity -settling has been recorded. 



The sinking of ciystals heavier than the surrounding liquid 

 would probably continue to a diminishing extent, through practic- 

 ally the whole period of crystallization. In this case Dr. Bowen 

 believes that the crystals would tend to sink as a swarm, rather 

 than as individuals, with little tendency to relative movement 

 between the different kinds. 2 The swarm, however, would be 

 dominated by the mineral crystallizing earliest, by the heaviest or 

 largest mineral, or by the mineral crystallizing in the greatest bulk 

 with the greatest speed, according to circumstances. The develop- 

 ment of any one mineral within the swarm would be controlled, in 

 general, by a combination of these conditions. 



Before the density-stratification of the central part of the Lugar 

 sill can be accepted as due to the sinking of early-formed olivine- 

 crystals, it is necessary to explain why augite and iron-ores are 

 not segregated to the same extent. In the first place, the olivine 

 had a start of the other minerals in crystallization. It is a rapidly 

 crystallizing mineral, and it was forming in large quantity from a 

 richly-magnesian magma. Hence it would form the major con- 

 stituent of the sinking swarm. Augite, iron-ores, and labradorite 

 began to crystallize somewhat later, and the two first-named would 

 probably participate in the sinking movement. They would find 

 the magma appreciably more viscous, and the field already largely 



1 Journ. Geol. Chicago, vol. xvi (1908) pp. 404-406. 



3 ' The Later Stages in the Evolution of the Igneous Rocks ' Ibid. (1915) 

 Suppl. p. 15. 



