92 PIIOCEEDI^^GS OF THE &EOLOGICAL SOCIETT. 



is no residual glass : a fair qnantity of an earthy dust, and sundry 

 globnlites hare formed, also a nnmber of minute belonites ; bnt the 

 dominant structure consists of crowded •'• brushes '" of an acicular 

 mineral with a general resemblance to that of the first type described 

 above. The most interesting point, however, is that the fragment 

 has three external faces, parts of the front, back, and an end of the 

 plate ; the acicular microlithic aggregates grow out from each of 

 them until they meet, where a sharp divisional line is formed, visible 

 even to the unaided eye, clearly indicating that the crystallites 

 started simultaneously from each exterior face, and grew inwards 

 till they met. The structure is roughly represented in the annexed 

 diagram. Here, as the whole glass is devitrified, the dominant 

 crystallites must be nearly identical in composition with it. 



Section of a fragment of devitrified glass. 



In concluding this subject, I must recall to your memory the most 

 important experiments of M. Daubree on common glass. These, 

 however, were performed in closed tubes in the presence of water 

 (about one thiid of the weight of the glass)*. Here the composition 

 of the glass was considerably altered, quartz crystals were developed, 

 with many belonites of a silicate, with spherulites and a tufted growth, 

 probably of chalcedony, and with a few grains of pyroxene. The 

 irregular crowding of the spherulites is very noteworthy, and the 

 effect of the surfaces of the glass on the grouping of the structures 

 produced by the alteration. A study of M. Daubre'e's remarks and 

 plates appears clearly to indicate that there is a great difference 

 between the results of mere '-heat" devitrification as described 

 above, and those of '' heat-water-pressure '" devitrification. 



The experimeutal evidence above cited indicates that consider- 

 able structural change and possibly some amount of molecular 

 segregation involving actual change of relative position (which is 

 certainly considerable when water is present) can take place when a 

 soHd bo'dv is rendered moderately plastic, but without fusion. This 

 is further illustrated by such facts as the kernel-roasting of copper- 



* 'Etudes Synthetiques de Geologie Erperimentale,' voL i. pp. 159-171. 

 Composition : — 



^ SiO, 68-4 



ALA 4-9 



CaO 120 



MgO -5 



y-aJO 14-7 



100-5 



