Water in Hot Glass. 469 



sibility of water at 210°. After this, the action of water on 

 the glass seems to have ceased, no further contraction 

 occurred even up to 4 h 40 m , when the tube burst at 100 atm. 

 The chief result obtained from this tube is therefore the 

 certain evidence of an eventual subsidence of the reaction of 

 the water on the glass under given circumstances. The 

 rapidity of reaction, the occurrence of floating cavities partly 

 full of mercury, partly of steam (?), the gradual reabsorption 

 of this steam (?) by the glass on increasing pressure, are to be 

 noted. During the continued volume-contraction compres- 

 sibility j3 increased rapidly at first, reaching a maximum, and 

 then fell off rapidly to the probable normal value for water at 

 210°, with the subsidence of further volume-contraction. 



From my recent experiments on the compressibility of 

 coagulated colloids*, I infer that the water-glass has now 

 reached a similar stage. A convex meniscus forced upward 

 by pressure changes to a conoid with apex sharpened upward. 

 Removing pressure restores the convex form. Hence the 

 column now possesses definite rigidity, and its viscosity- 

 requires long intervals of time to be put in evidence. Sharp 

 axial lines, "05 millim. in diameter, true canals, together with 

 bubbles strung on them like beads, were left after cooling. 

 The axial lines are present at 210°, and then appear as the last 

 vestige of the opaque or granular stage which precedes the 

 clear stage. If there is really something granular about the 

 former, and if the contraction of volume is a cementation of 

 these granules into a homogeneous clear column of aqueous 

 colloidal silicate, then the large final compressibilities are in 

 a measure accounted for ; but the compressibilities for tubes 

 1 and 2 are real in so far as all observations were made in 

 triplets, and the fiducial zero was regained after high pressures 

 (400 atm.). The granular hypothesis is thus quite inad- 

 missible even if the interstices between the granules were 

 supposed to contain non-saturated steam ; for pressures in- 

 creasing to above 400 atm. would soon wipe out the maxi- 

 mum vapour-tensions of 11-12 atm. at 185°, and less than 

 20 atm. at 210°. 



At the close of the experiment the diameter of the mercury 

 thread seemed to be somewhat over one half that of the core 

 of aqueous silicate. Since the bore increased from *24 to 

 •71 millim., the mercury thread did not thicken much on 

 moving up. 



A comparison of compressibility f3 and volume-reduction 



* Am. Journ. (4) vol. vi. p. 285 (1898). 

 Phil, Mag. S. 5. Vol. 47. No. 2SS. May 1899. 2 K 



