CHEMISTRY. 283 



finally the glazing operation — consumes considerable time; but by 

 no means as much as the building up and ''training " of the new mem- 

 branes. In order to secure a membrane which will render satisfac- 

 tory service at high temperatures, it is necessary first to deposit and 

 perfect it at some lower temperature, e. g., 25°, then to reinforce and 

 ''train " it at each successive higher interval of 5° or 10°. The mem- 

 branes in the new cells have already been perfected at 30°, and at 

 this temperature they are being employed for the measurement of the 

 osmotic pressure of glucose. It is intended to determine the pres- 

 sures of glucose solutions at each higher interval until the tempera- 

 tures are reached at which the work on cane-sugar was discontinued, 

 and afterwards to investigate both substances. The cells will not 

 again be allowed to rest at ordinary temperature until the high-tem- 

 perature measurements on both glucose and cane-sugar are completed. 



It was found in the case of glucose at 30° that the ratio of osmotic 

 to gas pressure is unity, showing that at this temperature the osmotic 

 pressures of the solutions conform to the laws of Boyle and Gay- 

 Lussac for gases. Three series of measurements of the osmotic pres- 

 sures of glucose had previously been made — one at 0°, another at 10°, 

 and a third at about 25°. The ratios found were approximately 1.06, 

 1.03, and 1.00 respectively. The high ratio at 0°, which diminishes 

 with rising temperature and becomes unity at 25°, suggests also in 

 the case of glucose an unstable hydration of the solute. 



The osmotic pressure of a series of solutions of mannite have also 

 been measured at 30° during the past year. At this temperature, as 

 in the case of glucose, the ratio of the observed osmotic to the calcu- 

 lated gas-pressure of the solute was found to be unity. 



The investigation of electrolytes has been continued. Much inci- 

 dental evidence had accumulated in past years to the effect that 

 bodies of this class act injuriously upon osmotic membranes, robbing 

 them of their semipermeable character. It was, therefore, appre- 

 hended that the direct measurement of the osmotic pressure of elec- 

 trolytes would be exceedingly difficult, if not quite impossible. An 

 investigation was first made of the action of potassium salts upon 

 membranes consisting of copper and nickel ferro-cyanides, and it was 

 found that the effect of these salts upon the membranes is exceedingly 

 deleterious — probably more so upon the nickel than upon the potas- 

 sium compound. It seems probable at the present time, however, 

 that we shall be able to measure satisfactorily the osmotic pressure of 

 dilute solutions of potassium salts, but not of concentrated ones. 

 The deleterious effect of electrolytes appears to be ascribable to the 

 action of the cations upon the colloidal state of the membrane mate- 

 rial — the colloidal state being apparently essential to semiperme- 

 ability in the ordinary sense. It was, therefore, surmised that the 

 lithium cation, which is supposed to carry with it a large " atmosphere 



