438 PROF. T. G. BONNET ON DEVITRIFICATION [Nov. I903, 



indicating that the material of the one which it appears to penetrate, 

 was in a rather more gelatinous condition than when the ordinary 

 4 hebraic ' type is produced. Mechanical resistance, as I have 

 already pointed out, 1 facilitates an actinolitic growth in the direc- 

 tion of the strongest force of crystallization. If a crystal in 

 development encounters an insurmountable obstacle, it is either 

 diverted or compelled to fork ; and, if the obstacles be both small 

 and numerous, the process is repeated again and again. This, as 

 has already been pointed out, is the explanation of frost-fronds, 

 dendritic markings, and the like ; and the more minute and nume- 

 rous the obstacles (as on a roughened surface) the more the branches 

 appear to curve (for a curve, to use mathematical language, is the 

 limit of a polygon). Thus, the fact that spherulites often assume 

 a lobed or root-like growth in their outermost parts, may be ex- 

 plained by the increasing viscosity of the glass from which they are 

 being formed. 



When crystals (for example, microliths) are forming in a magma, 

 these will continue to be enlarged, provided the temperature remains 

 high enough, until the necessary constituents are exhausted. The 

 residual magma is then either a single mineral, such as augite in one 

 case or quartz in another, or more often a mixture which, by the 

 process of crystallization, has gradually become richer in water, and 

 the presence of this last may itself determine whether a compound 2 

 is eutectic. When such an one has been formed, the temperature 

 will fall for awhile without producing further crystallization ; and 

 this, when it occurs, will not come about by the gradual separa- 

 tion of a single mineral, but by simultaneous formation of all the 

 components. 



A mineral cannot be idiomorphic without having ' had its own 

 way ' during crystallization, so that a rectilinear and a curvilinear 

 (or irregular) outline to the mineral constituents of a cooled rock 

 mean differences in the history of solidification. Divergent groups 

 of crystals, as we have seen, are indicative of opposition, the nature 

 of which is implied by the character of the branching. Whether the 

 latter be microscopic or megascopic probably depends on whether 

 crystallization commences independently from many centres (that 

 is, whether the conditions throughout the mass are very uniform), 

 or whether growth begins around certain nuclei — such as previously- 

 formed crystals of felspar, quartz, etc., and the fact that a pegmatitic 

 or graphic structure is so often a ' groundmass characteristic ' must 

 not be overlooked. 



Granular Structure. 



We come next to the different forms of granular structure. In 

 this also, so far as the boundaries are concerned, there is a recti- 

 linear and a curvilinear type. When the former occurs, the 



1 See my remarks in Quart. Journ. G-eol. Soc. vol. xlvii (1891) pp. 103-105. 



2 Thus a subsequent loss of water would mean supersaturation by the other 

 minerals, and might account for their separation. 



