478 Dr. Carl Bams on the Absorption of 



to the normal compressibilty of water at 200° (so far as 

 known) , I am led on the whole to favour a colloidal hypo- 

 thesis. In an extensive study of the solution of vulcanized 

 indiarubber * in a variety of solvents, the fusion temperature 

 of the saturated coagulated colloid under pressure to a clear 

 solution seemed to be nearly independent of the solvent used. 

 Below this temperature there is saturation and swelling, but 

 no solution. One may therefore regard the opaque stage of 

 the reaction of water on glass as the swelling of a coagulated 

 colloid in presence of its solvent, while the melting-point of 

 the coagulated colloidal glass (as evidenced by the second 

 stage of reaction) would lie below 210°, and probably above 

 150° C. Another question thus thrusts itself forward: 

 whether the property of swelling is retained by colloids at 

 relatively very low temperatures, seeing that glass at ordinary 

 temperatures does not swell in the presence of water. Finally, 

 the viscosity or stiffness of the fused colloid, and its readiness 

 to coagulate, increases as the amount of solvent per unit 

 of volume decreases. 



13. Summarizing the above results, it is best to avoid 

 direct reference to the character of solution, whether colloidal 

 or chemical. An explanation may then be given in terms of 

 the compressibility observed along successive isotherms for 

 the different concentrations of glass solutions at 210°. This 

 at first shows a relatively small value, implying a steep 

 isotherm in a Clapeyron (pv) diagram. Thereafter com- 

 pressibility passes through a relatively enormous value (over 

 five times the initial result), implying a nearly horizontal 

 isotherm. The reaction ends with even smaller isothermal 

 compressibility than the first observed, implying steeper 

 isotherms than the initial curves. Now, although these 

 isotherms are all at the same absolute temperature, the water- 

 glass is becoming continually more concentrated. Hence 

 the corresponding temperatures of the isotherms in van der 

 Waals' sense are continually decreasing. It follows that the 

 reaction considered as a thermodynamic process is a march 

 through the critical region of certain phases of the water-glass 

 examined. 



In the light given by J. W. Gibbs's investigations, one 

 may arrive at clear notions by adopting but two phases of 

 the water-glass for comparison. Many phases may coexist : 

 two are selected in the interest of brevity, and called phase 

 1 and 2 respectively. During the earlier stages of the re- 

 action (dilute water-glass) phase 1 is stable. At the end of 



* Am. Journal, (3) xlii. p. 359 (1891). 



