Water in Hot Glass. 465 



There is good reason to suppose that the aqueous silicate 

 remains homogeneous from the time of incipient solution to 

 the eventual occurrence of a viscous glassy coagulate, liquid 

 enough at 200° to admit of the formation of internal bubbles 

 on cooling and contraction, but at the same time viscous 

 enough to keep similar bubbles in shape and position without 

 cooling. A thin thread of water (-01 to *01 centim. in diam.) 

 is undergoing lateral diffusion into the glass, and concentra- 

 tion difference is virtually confined to the cylindrical surface 

 of contact between igneous glass and water-glass, widening 

 as the action goes on so slowly, that the much more liquid 

 water-glass is free to remain homogeneous. The latter 

 should therefore be identical in composition with the original 

 glass, but for the incorporated water. 



In a final experiment (tube No. 7) 1 put a solution of 

 cobalt nitrate into the tube, rather with the expectation of 

 finding blue water-glass as the result. No such action 

 occurred ; instead of it, the water diffused into the glass as 

 usual and the cobaltic nitrate was left as a granular scum in 

 the axis. Chemical decomposition and incorporation of the 

 cobalt did not therefore occur at 200°. 



3. Instead of tabulating the large number of observations 

 made (in all seven tubes were employed), it will conduce to 

 clearness to present the work graphically. In so doing, the 

 data for the former tubes may be included, for reference. 

 Time is laid off in intervals of 20 minutes (between vertical 

 lines) along the abscissas, while the ordinates indicate the 

 changes of length of the column of water in intervals of 

 1 centim. It is not convenient to specify the full length of 

 the thread in the chart ; but a datum for the length of the 

 cold thread at about 20 c Cwill be given both in the latter and 

 in the tables containing original and final bore, and similar 

 specifications for each capillary tube. These diameters were 

 measured with the cathetometer, the tubes being cut across 

 and looked at endwise. Slight heating increased the sharp- 

 ness of definition between the original glass and the solid 

 core of water-glass within. 



The chart also contains the mean compressibilities j3, for 

 pressure intervals of about 100 atm., no attempt to obtain ft 

 as a function of pressure being made for the reasons stated in 

 § 2. Since /3=(v/Y)/p, or the decrement of volume per 

 unit of volume per atmosphere, or practically decrement of 

 length per centim. of column per atmosphere, the abscissas as 

 above are successive time-intervals of 10 minutes each, while 

 the scale of ordinates is a change of /3 of -000100, Several 



