C. Barus — Compressibility of Colloids. 289 



Neither for gelatine nor for albumen is an increase of com- 

 pressibility with temperature certainly indicated, whereas this 

 is definitely the case with water. 



It is difficult to state whether these detailed results are real. 

 I will pass it over, mentioning merely that it is always a deli- 

 cate question to decide whether the heat generated on com- 

 pression has all been dissipated. In a cooling water bath, 

 moreover, the residual effects of thermal expansion are apt to 

 be superposed as the compression results. Finally in case of 

 coagulated albumen, threads break off and are lost in the 

 opaque body. 



The endeavor was now made to prepare as concentrated a 

 solution as could be filled into the fine capillary bore. The 

 ingredients were weighed out for a 20 per cent solution, though 

 the concentration may have been less. It coagulated firmly, 

 showing marked elasticity in threads. 



Measurements were first made for the coagulated thread as 

 will be shown below, § 7 et seq. After this the tube was kept 

 at 100° for some hours and examined from time to time. The 

 earlier results contain no additional information and will be 

 omitted. The final data are given in Table 3. Compressibil- 

 ity /3 refers to the total interval from p=0 to the value of p 

 given. 



Table 3. — Compressibility of a 20 per cent solution of gelatine in water. 



d, L 



l/L 



P 



3xl( 



100° 



•0003 



7 



__ 



I7'01 cm 



056 



114 



49 





103 



216 



48 





156 



319 



49 





185 



381 



49 





126 



268 



47 





082 



171 



48 





038 



87 



44 





003 



8 



._ 



These new data practically coincide with the former albu- 

 men and gelatine (10 per cent) values ; if anything the former 

 show greater compressibility. They coincide nearly with the 

 compressibility of water at 23°, and fall definitely below the 

 water compressibility for 100°. Little or no effect is therefore 

 observed by the additional concentration of the new solutions. 



5. Table -1 contains results similar to Table 1 for ether and 

 an etherial solution of caoutchouc (not vulcanized). If the 

 volume increment l/L be laid off in terms of the correspond- 

 ing pressure p, it is found that the curves for the colloidal 

 solution and the solvent ether coincide very nearly both at low 

 temperature and at high temperature. Since in the former 



