Water in Hot Glass. 463 



for compressibility in triplets, and the data are given below 

 with, this reservation. These data accentuate the absence of 

 effective volume elasticity under the conditions stated. 



When the viscous thread has appreciably ceased to contract 

 (1-2 hours), and the tube is allowed to cool very carefully, 

 bubbles make their appearance very much resembling those 

 in a Prince Rupert drop, and probably due to a similar 

 cause *. They begin to form in the axis, and are usually 

 connected by a fine channel. They may grow to a diameter 

 of over | the width of the bore. The formation of these 

 bubbles on cooling is proof that the aqueous silicate is still 

 liquid at the temperature of the vapour-bath (185°-210°), 

 however viscous and incompressible it may have become. 

 The solidifying core of water-glass contracts from the centre 

 outward, and must contract more rapidly than the igneous- 

 glass envelope. It is this last stage (contents cleared again 

 to a pellucid jelly) which I failed to fully observe in my 

 earlier experiments, believing that solidification had set in 

 when the mercury-threads broke off. 



After the tube has passed the final stage with subsidence of 

 contraction, it invariably breaks throughout its length when 

 cold, in such a way as at first sight to suggest expansion on 

 solidification of the aqueous silicate within. It makes no 

 difference how carefully the cooling is performed. If a thread 

 of fusible metal is allowed to solidify in a capillary tube, the 

 latter breaks sooner or later in the same way. 1 do not by 

 any means imply that the aqueous silicate does actually 

 expand on solidifying, for there are other and better ways in 

 which the breakage can be explained. The appearance of 

 bubbles, for instance, is evidence of contraction, and the 

 breakage is rather due to an excessively shrinking core. 



When the cold tube is cut across, the core of water-glass 

 practically fills the tube, and to all appearances is as hard, 

 clear, and firm as the igneous glass surrounding it. There is 

 a difference of refraction between the two glasses sufficient to 

 make the aqueous core apparent under favourable illumination; 

 but for this and the bubbles, the tube would be undistin- 

 guishable from a glass rod. 



If, however, the end of the glass tube is slightly heated 

 above a candle-flame or a small bunsen-fiame, the core soon 

 begins to melt, to swell up enormously in bulk as the result 

 of the frothing which accompanies the escape of steam. The 



* It is to be remembered, however, that whereas the Rupert drop is 

 cooled down from above 1000° C, the present high temperature is only 

 about 200°. 



