7'22 STUDIES ^N GENERAL PHYSIOLOGY 



bivalent kation, but the life of the embryos was very con- 

 siderably prolonged. 



The shortness of the spawning season limited the number 

 of my experiments, so that I decided to bring my experi- 

 ments upon the annihilation of the poisonous effects of a 

 pure sodium-chloride solution to a close, and to carry the 

 remaining experiments only far enough to decide if we are 

 dealing here, in the main, with the same condition of affairs. 

 That, I believe, is undoubtedly the case, so that I feel my- 

 self justified in making the following statement : The salts 

 of monovalent kations (Na, Li, K, NH^) with monovalent 

 anions (Cl, NO S , CH 3 COO) exert a toxic effect at certain 

 concentrations. This toxic effect can be annihilated through 

 the addition of a small amount of a salt having a bivalent 

 kation. For NaCl, proof has been brought forward that 

 trivalent kations exhibit even a much more energetic anti- 

 toxic effect than bivalent kations. Further experiments 

 are yet to be made, to decide if the poisonous effects of the 

 other salts (LiCl, KC1, NH 4 C1) can also be done away with 

 through the addition of such small amounts of trivalent ka- 

 tions as suffice for NaCl. 



8. While the preceding experiments show an undoubted 

 influence of the valency of the ions upon their antitoxic 

 effects, it was now necessary to prove that the sign of the 

 electrical charge was the second determining variable. I 

 instituted a large number of experiments in which I 

 attempted to annihilate the poisonous effects of a -|m NaCl 

 and a ^m KC1 solution by the addition of salts having a 

 univalent or bi- or trivalent anion. The antitoxic effects of 

 the following salts are investigated; KOH, NaBr, Nal, 

 NaHCO 3 , Na 2 CO 3 , NaSO 4 , Na 2 HPO 4 , sodium citrate, 

 K 2 SO 4 . Extensive quantitative experiments were made 

 with Na 2 SO 4 , K 2 SO 4 , NaHCO 3 and Na 2 HPO 4 . The 

 results were negative throughout. In the best cases 1 per 



