94 



Messrs. D. Herman and F. Rutley. 



ij. Plate 2, fig. 2, which traverse the plate normal to tlie large parallel 

 surfaces. These joints are evidently the boundaries of polygonal 

 prisms, and it is the ends of these prisms which cause the green and 

 white spotted appearance on one surface of the specimen, while the 

 reason that no such marking is visible on the other surface is partly 

 due to the screen of sand grains which covers it, while beneath there 

 would be no such markings until we reached the layer of originally 

 white glass, because the joints do not appear to traverse the irregu- 

 larly crystalline blue layer. Divergent crystallisations, also bounded 

 by prismatic joints, start from the green spotted surface of the plate, 

 and the two sets of divergent crystallisations meet in an undulating 

 line, 11., Plate 2, fig. 2, which approximately divides the plate into 

 two plates of about equal thickness. The joint planes on the opposite 

 sides of this line do not coincide, and the halves of the plate if sepa- 

 rated along the surface, of which this undulating line is the trace, 

 would doubtless present a mammillafced appearance. The general 

 structure reminds one of that of part of a much flattened chalcedonic 

 geode. It will be seen that in this specimen the devitrification has 

 taken place on precisely the same principle as in the thick plate pre- 

 viously described. There has been a prismatic structure developed 

 normal to the bounding surfaces, divergent crystallisation occurs 

 within the prisms, and these crystalline fasciculi advanced in opposite 

 directions until they arrested one another, but the line of arrest in 

 this case is sinuous, while in the preceding specimen the lines of 

 arrest are straight. On examining the section under a power of 

 50 diameters, fine lines, like small scratches, are seen to cut across 

 the divergent crystallisation. Under much higher powers they appear 

 as rod-like microliths, and they lie with their longest axes in all 

 directions, but mostly transverse to the divergent fibres. 



Specimen I is part of a completely devitrified square prism of plate- 

 glass. The devitrification of this specimen was brought about by two 

 separate operations. The whole of the prism, about 4 inches in 

 length, was bedded in silver sand and heated during four days to a 

 temperature gradually increasing from that of the atmosphere up to a 

 red heat, maintained at that for two days more, and then quickly 

 cooled. When cold it was broken in two and found to be regularly 

 devitrified to a depth of about \\ mm., the interior being unaltered. 

 One of the halves was then burnt again, this time for the same period 

 and under exactly the same conditions as Specimen Gr, i.e., bedded in 

 sand, brought gradually to a bright red, maintained steadily at that 

 heat for eleven days, and then quickly cooled. . The faces are of a pale 

 greenish-yellow, have a glazed appearance like that of pottery, and 

 are traversed by a network of very fine cracks. When the specimen 

 is held before a strong light these surfaces present a spotted appear- 

 ance, similar to that seen on the plane surface of other devitrified 



