462 M. E. COLLETT 



with the relative toxicity of their salts. I have not experimented 

 with salts, but there are interesting data to be had from other 

 experiments. The order of toxicity to Lupinus seedlings (Kah- 

 lenberg and True and True, '96) is as follows: 



Acids equinormal, salts equimolecular. 



Acids: HCl > benzoic > salicylic > formic = propionic > 

 butyric. 



Salts : Na — benzoate > salicylate > formate > propionate > 

 butyrate > chloride. 



The order for equinormal acids is almost the same as that 

 obtained in my experiments: that is, HCl is more toxic than 

 benzoic, and benzoic in turn is more toxic than the fatty acids. 

 Since there is little difference in dissociation among the salts, 

 their order, unlike that of the acids, is a measure of the relative 

 toxicity of the anions. Salts of the cyclic and of the fatty acids 

 are more toxic than NaCl and must therefore have more toxic 

 anions. These are salts of the very acids which in solutions of 

 equal ionic concentration proved to be most toxic to cilia. 

 Practically the same conditions hold in the haemolysis of blood 

 corpuscles (Fiihner and Neubauer, '07, and Hoeber, '10). 



Acids: HCl > formic > acetic > propionic. 



Salts: Na — Salicylate > benzoate > formate > acetate > 

 butyrate. 



The similarity between these experiments with equimolar salts 

 and my experiments with acids in solutions of equal Ph makes 

 it seem probable that the acids w^hich are most toxic in 

 these solutions owe their effectiveness to the anions as well as 

 to the H ion. The fact that a different order of toxicity obtains 

 in normal solutions may perhaps be explained as follows : If an 

 acid is not highly dissociated, even though it has a slightly toxic 

 anion, it is Ifess toxic than a more highly dissociated acid with a 

 non-toxic anion (for instance, butyric as compared with HCl). 

 If, however, the anion is very toxic, as is the case with salicylic, 

 the acid in spite of its slighter ionization may equal or even 

 exceed a more completely ionized acid such as HCl in toxicity. 



Another method of investigating this point has been suggested 

 by Klocman ('14). To various concentrations of a weak acid a 



