791 



the course is the simplesl ; for K,SO, it is the most capricious. In 

 one respect tliey agree, in that tiiey are approximately symmetrical 

 on either side of a horizontal line, which for KNO, lies at about 





o — 



3:^ 



p 



o 

 o 

 o _» 



2 ° 



Oq 1 a J V y É. f fi' 9 10 11 II 13 /V IJ- !♦ 

 concKCl xo,oin. 



Fig. 4. 



0.00035 n. KOH (tig. 3), for KCl perhaps al 0.00045 n. KOH (fig. 

 4) and for K,SO, at 0.0040 n. KOH (fig. 1), although the axis of 

 symmetry for KNO, and for KCl is not horizontal. This symmetry 

 is too striking to be neglected, especially in the case of KNO, 

 (fig. 3) and of KjSO^ (fig. 1).. It seems to me an indication that the 

 irregularity in this portion of the curves is not caused by experi- 

 mental errors; moreover the sixfold repetition of the experiments 

 also warrants this. 



The symmetry was probably caused by the salts being active both 

 with their cations and their anions. The part played by anions and 

 cations, when the plasma colloids were positive, was probably 

 reversed when the colloids were charged negatively. 



The complication in the course of the curves consists in the 

 presence of maxima; with KjSO^ (fig. 1) clearly five such may be 

 distinguished, one at 0.00040 n. KOH and then two on either side; 

 at 0.00075 n. KOH and al 0.00005 n. KOH there is in each case a 

 low summit, at 0.00090 n. KOH and at 0.00005 n. H,SO, a high 

 one. In the ease of KCl (fig. 4) there are perhaps also five. A con- 



