464 APPENDIX 



what extent this invalidates the comparison of the two sets of 

 determinations. The error which is thus introduced, for the small 

 difference of temperature concerned is, however, negligible, since 

 it was found by actual trial that the difference between the poten- 

 tials measured at 30 degrees and those measured at 34 degrees 

 could barely, with certainty, be detected upon my potentiometer- 

 bridge. Nor is this fact surprising for, exclusive of any possible 

 change in the equilibrium between protein and alkali, the potential, 

 according to the Nernst formula, varies directly as the absolute 

 temperature, and, therefore, only increases by 1 /300th per degree 

 centigrade at 30 degrees. It has been observed by W. A. Osborne* 

 and myself f that there is no evidence of an appreciable shift in the 

 equilibrium between casein and alkali, as the temperature rises, 

 until the temperature of 36 degrees is reached. 



I have mentioned in describing the apparatus employed, that 

 difficulty was encountered in leading off the hydrogen, after it had 

 bubbled through the fluids, through the exit-tubes of the half- 

 element to the outside of the incubator. The difficulty was this, 

 all of the solutions containing casein foam upon passing the 

 hydrogen through them, and to a greater extent the less the excess 

 of alkali in the solutions. This foam, collecting in the exit and 

 connected rubber tubes, gave rise to a pressure which drove the 

 fluid out of the half-element and thus interrupted the continuity 

 of the chain. Accordingly, it was found necessary, not only to 

 abandon the idea of conveying the waste hydrogen out of the 

 incubator, but also to cut off short the exit-tube of the half-element 

 and permit the foam to escape freely into the water-bath (the 

 lower end of the exit-tube was, of course, well above the surface of 

 the water in the bath). After this plan had been adopted no 

 further trouble was encountered from this source. Foaming, may, 

 however, be prevented or greatly reduced by the addition of a few 

 drops of octyl or amyl alcohol. 



All observers who have endeavored to measure the conductivi- 

 ties of solutions containing proteins have encountered the difficulty 

 involved in the precipitation which occurs at the electrodes, 

 particularly in the case of casein, in neutral or very faintly acid 

 solutions. Whetham and -Hardy { in order to minimize the error 



* W. A. Osborne, Journ. Physiol., 27 (1901), p. 398. 



t T. Brailsford Robertson, Journ. Biol. Chem., 5 (1908), p. 147. 



t W. B. Hardy, Journ. of Physiol., 33 (1905), p. 251. 



