Respiratory Stimulants 347 



It is obvious that the conditions are somewhat different in 

 the case of a weak acid such as carbonic acid and its salt, sodium 

 bicarbonate. Carbon dioxide occurs in solution in three forms 

 and the relative amounts are varied by the addition of sodium 

 bicarbonate. Though the acid effects of the solution on the in- 

 teriors of cells placed in such solution are theoretically increased 

 by the increase in the freely diffusible C0 2 and H 2 C0 3 mole- 

 cules, the number of free hydrogen ions is diminished. Yet the 

 effectiveness of these ions is greatly increased by the increase in 

 the common anion HC0 3 . If a differential barrier, impeding 

 the migration of the sodium ion, exists between the blood and 

 site of action of the hydrogen ion in the respiratory center a 

 double mechanism exists for the passage of carbon dioxide from 

 the point of low to a point of high concentration accompanied 

 by increased acidity. Michaelis calls attention to the fact that 

 H and OH ions are adsorbed at phase boundaries more than 

 other ions and also that the addition of salts to solution of acids 

 or bases increases this adsorption. This phenomenon should also 

 be carefully considered in relation to the hydrogen ion concen- 

 tration at the site of stimulation. We hope to show in a later 

 paper the relative importance of the various salt effects of sodium 

 bicarbonate. 



In connection with the specificity of carbon dioxide, we have 

 by means of simple diffusion experiments without the aid of a 

 membrane demonstrated striking effects on the hydrogen ion 

 concentration of carbonate buffer solutions on each other, de- 

 pendent solely on the differences in concentration of carbon diox- 

 ide and sodium bicarbonate and their relative rates of diffusion. 

 A relatively alkaline solution of high carbon dioxide tension 

 exerts an acid effect even on a relatively acid solution of lower 

 carbon dioxide tension. The results are comparable to those 

 obtained by Jacobs in the living cell and in the model cell with a 

 lipoid solvent membrane. Our results, however, since they are 

 obtained without a membrane are independent of any peculiar 

 property of the cell membrane. They suggest the possible signifi- 

 cance of the relative rates of diffusion, of the numerator and 

 the denominator of the buffer mixture not only in buffer solu- 

 tions outside the cell, but possibly inside as well. 



As Fletcher and Hopkins have shown that the rate of escape 

 of carbon dioxide from saturated living muscle, dead muscle, 



