Yol. 59.] IN GLASSY IGNEOUS KOCKS. 441 



The devitrification of a volcanic glass implies, however, a stage 

 beyond that of the formation of a chert. In the latter certain 

 groups of adjacent silica-molecules have only to arrange themselves 

 in a definite order l ; but in the former a kind of sifting process is 

 necessary — a change of place as well as of position — the silica- 

 molecules gathering around one centre and those of an alumina- 

 alkali-silicate around another, although the mass itself remains 

 practically solid. 2 



A special type of this latter devitrification might be termed ' patchy 

 or ' splotchy ' (though perhaps poikilitic, as being less English, will 

 be more generally acceptable). In it a number of molecules, perhaps 

 already separated as in a spherulite, are locally re-arranged, so that 

 a granular is superposed on a radial structure. In the absence of 

 a spherulite we may regard this as merely a variety of the granular 

 structure ; although I suspect it to be even then secondary, the rock 

 having already been very minutely devitrified. But when it occurs 

 in a spherulite (sometimes making this far less distinct under crossed 

 nicols than with ordinary transmitted light) we cannot doubt that 

 it is not a primary structure. Here the formation has probably 

 been facilitated by the fact that the component crystallites in small 

 portions of the spherulites (as in a cone with a narrow angle) are 

 already lying nearly parallel, so that only a small change in position 

 is needed to bring them into line as parts of a crystalline grain. 

 Eut if so, we may be asked why a spherulite is not converted into 

 a series of radiating crystals showing in sections as isosceles tri- 

 angles. The most probable answer is, that even in these spherulites 

 the structure is not perfectly symmetrical ; the radii may throw off 

 side-branches (this is often very strongly marked in artificial glasses 

 and slags) which cause confusion, or they may be modified in direc- 

 tion by the shape of the nucleus from which they have started 

 whether that be a crystal previously formed or a cavity. Absolute 

 symmetry is likely to exist but rarely, and this c^stallizing force 

 (if I may so call it) may be restricted in its effects within compara- 

 tively small limits, so that even if the apical part of a triangle (in 

 sections) form one crystal, the basal may have to break up into two 

 or three. Moreover, we have no reason for supposing this re- 

 arrangement to begin at the centre of a spherulite and proceed 

 outward. It is more likely to be set up at a series of independent 

 points, where the slightest difference in conditions may suffice to 

 initiate it. 



One point in secondary (and in some cases also of primary) 

 devitrification has often struck me as remarkable. When a granular 

 structure is moderately coarse, the quartz is easily distinguished 

 from the felspar ; but often in spherulites, and almost invariably in 

 the 'patchy' devitrification, the former mineral is hardly visible, 



1 As we might imagine eggs in a basket arranging themselves with all their 

 longer axes parallel. 



2 The possibility of such a movement was demonstrated by the late Sir 

 William Eoberts-Austen in his experiments on the migration of gold into lead, 

 Phil. Trans. Boy. Soc. a, vol. clxxxvii (1896) p. 383. 



