1)6 



Messrs. D. Herman and F. Rutley. 



crystalline group is convex, the convexity being directed towards the 

 interior of the solid. When magnified between 500 and 600 diameters 

 these convex surfaces are seen to be fringed by the projecting termi- 

 nations of the divergent crystalline fibres. 



Specimen K. This is part of a completely devitrified trigonal 

 prism of plate-glass, devitrified by two operations, under precisely the 

 same conditions as Specimen I, and the section has been taken parallel 

 to the basal plane. The general principle of devitrification elucidated 

 by the examination of the preceding specimens may also be clearly 

 recognised in this case, but the crystallisation, after the first envelope 

 was formed, advanced in a somewhat irregular manner, which needs 

 interpretation. The irregularity in the crystallisation of this specimen 

 may be attributed to the fact that there is a flaw in it. The general 

 structure is shown in Plate 3, fig. 8. Here we notice first of all the 

 envelope or devitrified crust, due to the first heating operation, in 

 w r hich there is prismatic structure and a series of divergent crystal- 

 lisations trending inwards. Next comes a similar but coarser series 

 of prisms also normal, or approximately normal, to the sides of the 

 trigonal prism, and in these the divergent crystallisation has also 

 travelled from without inwards. So far there is no deviation from 

 the general principles of devitrification which we met with in the pre- 

 ceding specimens, in fact the crystallisation has proceeded inwards as 

 usual, in directions approximately normal to the limiting planes of 

 the devitrified solid. We now meet, however, with an apparent 

 exception to the general rule, for the three sets of crystalline fasciculi, 

 instead of continuing their course until they arrest one another in 

 three straight lines joining the angles and the centre of the triangular 

 section, are suddenly arrested and enclose an area rudely shaped like 

 a three-rayed star, this being subdivided into three irregular portions. 

 The deltoidal areas are traversed by cracks, and from points along 

 these lines we have groups of crystals diverging on both sides of the 

 lines. They have consequently travelled from within outwards. 



The different areas of devitrification are by no means symmetrically 

 disposed. Diagram, fig. 8, Plate 3, shows, with approximate truth, 

 how the parts of the actual section really occur. It will be seen on 

 reference to this fig are, that at the point a there is a 'crack which 

 extends in a curve towards b. From a point on the curved line ab, 

 about opposite to the middle of the edge in which the crack a occurs, 

 another nearly straight crack passes to c, and from the inner surface 

 of the devitrified crust a third crack extends in a curve from the little 

 fissure a to the point d. The crystallisations diverge on both sides of 

 these three cracks. They are bounded by prismatic joints, which are 

 continuous across the cracks, and each pair of crystalline fasciculi 

 diverges from a common centre situate on the crack and between a 

 pair of prismatic joints. These three distinct areas of crystallisation 



