Microscopic Characters of Devitrijied Glass. 



93 



the sections now described, it seems certain that devitrification has in 

 this instance commenced at the surface, and has proceeded inwards 

 in directions at right angles to the different surfaces. Owing to its 

 uniform rate of progression, the different sets of crystalline fasciculi 

 have met along lines which divide the de vitrified mass in a remark- 

 ably symmetrical manner, as shown in fig. 5, Plate 3, which represents 

 one corner of the slab. That unequal rate of progression would cause 

 a deviation from this symmetry is shown diagrammatically in fig. 4, 

 Plate 3, and actually in the deflection of the diagonal line in fig. 1, 

 Plate 1, Section A.* 



Specimen G is a plate of flashed glass, about 2^- mm. in thickness, 

 which has been completely devitrified under conditions similar to 

 those described for Specimen 115, that is to say, it was imbedded in 

 silver sand (previously washed and dried), placed in a kiln, and the 

 temperature gradually increased during a period of eleven days up 

 to a bright red. This heat was maintained pretty steadily for eleven 

 days more, after which the kiln was quickly cooled, and the glass 

 withdrawn. The flashed surface is of a deep blue colour, and is 

 incrusted with grains of sand. The opposite face is mottled with 

 small dull green and greenish-white spots, and has a surface like 

 coarse glazed pottery. Flashed glass was chosen in this case, as it 

 was thought possible that some of the pigment might be carried 

 inwards by the crystallisation. This, however, does not seem to have 

 taken place to any great extent, for on examining a thin section taken at 

 right angles to the broad surface of the plate under a power of 250 linear, 

 the blue layer is seen to have remained on the surface, although 

 its boundary is ill defined, and the bluish tint extends for only a very 

 little distance inwards, gradually fading away. On the outer surface 

 of the coloured layer there has, however, been a considerable dis- 

 turbance of the blue glass, which appears to have been fused, and to 

 have run between the sand grains against w r hich it was bedded, 

 //, Plate 2, fig. 2, forming a cement of blue crystalline sheaves. The 

 crystalline structure of the blue layer is throughout very irregular, 

 consisting of similar sheaf-like aggregates and interlacing crystals. 

 Passing from this layer we find the contiguous glass converted into 

 radiating crystalline groups, separated by sharply defined joint planes, 



* Supplementary Note.— Specimen 115. Thermal conductivity appears to be 

 uniform on the large parallel faces of the plate, both at the margin and at a distance 

 from the margin. The isothermal curves are also circles on the sides of the plate at 

 right angles to the large faces. On a transverse section of the plate which traverses 

 the crystalline fasciculi in directions both longitudinal and transverse, as in Section C, 

 the wax also melts in circles both on the area of the longitudinal and on that of the 

 transverse sections. This accords with the statement of M. Ed. Jannettaz ("Propa- 

 gation de la Chaleur/' " Bull. Soc. Greol. de France," 3 e Serie, t. ix, p. 200) that 

 minerals having a fibrous or lamellar chai'acter do not conduct heat better in the 

 direction of the fibres or of the lamellse than if they had no such structure. 



