Microscopic Cliaracters of Devitrified Glass. 



95 



solids. Under the microscope it is seen that the crystallisation has 

 advanced as nsual from the surfaces inwards. After passing through 

 a distance of about mm. from the surface there has been a pause, 

 marked by a fairly well-defined line, indicating the extent of the 

 devitrification produced by the first heating to a red heat. This line 

 is not straight, but has a series of slight convexities directed inwards, 

 each convexity being bounded by joint planes normal to the surface. 

 The prism therefore had first of all a devitrified envelope, the inner 

 surface of which was mammillated, and each mammillation was the 

 termination of a small prism. As the crystallisation advanced from 

 the inner surface of this envelope, a fresh series of less numerous 

 joints was developed, giving rise to a coarser prismatic structure, and 

 between these joints we see in section a beautiful divergent crystal- 

 lisation, each divergent group originating on the inner surface of 

 the first crystalline envelope, a single prism sometimes containing only 

 one such group, at others several. The general direction of these 

 prisms is normal to the surfaces of the devitrified specimen, and the 

 lines of arrest would join the opposite angles of the square section, 

 were it not that in this particular slice an- irregular pentagonal area 

 occurs, against four of whose angles the lines of arrest abut. This 

 irregular pentagon is a transverse section of another set of divergent 

 crystallisations, whose longest axes would diverge from the axis of 

 vision, and they evidently emanated from one of the basal planes of 

 the large devitrified square prism, or from a transverse fracture as the 

 prism was broken across after the first heating. Had the specimen 

 been a cube, a section taken parallel to two of its faces and passing 

 accurately through the centre of the cube,, would merely have shown 

 two continuous lines of arrest joining opposite angles and intersecting 

 in the centre of the square section, assuming, of course, that the 

 crystallisation advanced equally from all six faces. Such a structure 

 would divide the cube into six equal four-sided pyramids, as indicated 

 in the diagram, Plate 3, fig. 7. In the specimen before us the crystal- 

 lisation has advanced rather irregularly, and the lines of arrest are 

 consequently not continuous straight lines, but continuous series of 

 straight lines, a repetition,, in fact, of the conditions indicated in the 

 diagram, fig. 4, Plate 3. 



Specimen H is portion of a similar square prism of plate-glass, 

 heated gradually for six days to a red heat under exactly the same 

 conditions as the first operation on Specimen I. It differs from the 

 preceding specimen in having been devitrified for only a slight 

 distance from the surface. A section of the crust through one of the 

 angles presents an appearance precisely similar in character to that of 

 the crust of Section E 3 (a six-sided prism), figured on Plate 2. These 

 are groups of divergent crystals which pass from the surface inwards, 

 and are separated by prismatic jointing. The inner surfaces of each 



