JACQUES LOEB 591 



collodion bags were filled over night with this crude oxyhemoglobin 

 solution and the bags were used the next day for similar experiments 

 to those described for membranes treated with gelatin. Fig. 5 shows 

 the results of such experiments. The MgCl2 curve begins to rise at a 

 pH of 6.0 while the curve for Na2S04 rises at a pH of about 7.0 or 8.0. 

 Hence when the membrane has been treated with oxyhemoglobin a 

 much lower hydrogen ion concentration is required to induce a posi- 

 tive charge on the membrane than when the membrane has been 

 treated with gelatin. This was to be expected if the reversal of the 

 sign of charge of a membrane treated with a protein depends on the 

 isoelectric point of the protein. According to Michaelis the isoelectric 

 point of oxyhemoglobin is pH = 6.7 to 6.8, and we notice accordingly 

 that the membrane treated with oxyhemoglobin assumes a positive 

 charge when the pH is a little less than 6.7; e.g., 6.0. In Fig. 6 the 

 curve for NaH2P04 shows a distinct rise at a pH of 6.0 and that for 

 CaCla at about 6.0 or a little less. In judging the curves in Figs. 5 

 and 6 the reader must bear in mind that the CO2 of the air lowers the 

 pH during the experiments and although the experiments lasted only 

 20 minutes the errop so caused was noticeable in the neighborhood of 

 the point of neutrality. Thus the beginning of the rise in the curve 

 for Na2S04 was in reality not at pH = 8, but at a lower pH, either 

 7.0 or probably less. 



V. Collodion Membranes not Treated with a Protein. 



When we use membranes not treated with gelatin and repeat the 

 experiments described in Fig. 1, we get altogether different results 

 (Fig. 7). There is no minimum or drop at pH of about 4.0 or 6.0 

 since the sign of charge of collodion membranes not treated with pro- 

 teins is not reversed by acid.^ Collodion membranes not treated 

 with a protein are always negatively (and the watery phase positively) 

 charged within the range of our experiments. Hence a m/256 solu- 

 tion of CaCl2 cannot attract pure water through such a membrane 

 even in acid solution and the curve for the effect of m/256 CaCl2 is 

 flat for the whole range of hydrogen ion concentrations (up to 10"^ n). 

 The curve for m/256 Na2S04 does not reach a minimum at pH 4.0 

 but drops slowly reaching a minimum at pH 2.0 or below, this drop 

 being due to the depressing effect of the high concentration of the cat- 



