104 



Messrs. D. Herman and F. Rutley. 



manner for a slight distance from the surface, after which an irregular 

 crystallisation has been set up from independent centres; but it 

 should be remarked that difference in the chemical composition is 

 known to influence the mode of procedure, as well as the character 

 of the devitrification. The direction of the prismatic structure 

 always seems to be approximately normal to the surfaces, and the 

 divergent sheaves of crystals advance from the surface inwards by 

 successive growths within the prisms. It seems quite possible that 

 in the absence of such prismatic jointing the whole mass would 

 become spherulitic, or would consist of an irregular felted mass of 

 crystallites. The near resemblance which some of the specimens just 

 described bear to devitrified and partially devitrified obsidians shows 

 how close the structural relationship is, and that, allowing for difference 

 in the conditions under which the process takes place, the principle of 

 devitrification is the same. 



Specimen No. 122c?, a piece of ordinary sheet-glass, which was 

 bedded in white sand and heated during a period of only four days to 

 a temperature gradually increasing from that of the atmosphere up to 

 a blood-red — a temperature somewhat lower than that employed for 

 any of the specimens previously described, shows purely superficial 

 devitrification by the development of globulites and spherules or 

 spherulitoid crystallites, like fig. 11, Plate 3. In this particular 

 crystallite, which is of a pale brown colour, no structure can be made 

 out. It seems merely to consist of an aggregate of globulites, but in 

 other cases bodies of precisely similar form show a decided radiating* 

 crystalline structure, like that of the brown spherules, which occur 

 with them in the same specimen, the only difference between these 

 crystallites and the spherules consisting in the external form or limit- 

 ing surface. It is for this reason that we propose to call them 

 spherulitoid crystallites. Fig. 10, Plate 3, drawn from the same piece 

 of glass, shows part of the network of cracks by which the surface is 

 cut up, and the curious manner in which the globulites have segre- 

 gated along these cracks, so as to leave the fairly well-defined circular 

 and oval spaces in which the globulites are less densely packed. 

 Spherules sometimes occur within these clearer areas, but the latter 

 do not seem to have any necessary connexion with the development 

 of the spherules. 



In Specimen ISo. 1226, superficially devitrified under the same con- 

 ditions as the preceding, a tendency to the formation of perlitic 

 structure is seen around some of the spherules. 



Specimen 126, a piece of rough plate-glass, -J inch thick, bedded in 

 white sand, contained in a small fire-clay pot, and placed in a kiln, 

 the temperature of which was gradually raised during a period of 

 8-| days, by which time a dull red heat, about 650° C, was attained. 

 As it was known by comparative experiments with similar pieces of 



