ANTAGONISM 131 



example, flat-topped curves and also curves with two 

 maxima, as shown in Fig. 52. 



If instead of mixing two equally toxic solutions we 

 keep the concentration of one salt constant while varying 

 that of the other, it becomes very difficult to determine the 

 additive curve, especially when variations in osmotic 

 pressure influence the result. It is therefore difficult to 

 obtain an accurate quantitative expression of antagonisin 

 by this method, and in critical 

 cases it may be impossible to de- 

 cide whether antagonism exists 

 or not. 



Emphasis should be laid on the 

 fact that the growth of parts 

 not in immediate contact with 

 the solution does not furnish a 

 trustworthy criterion of antagon- 

 ism. Thus the leaves of wheat Fm. 52.— Types of antagonism 

 / T ' ^ 1. 1 , -nn curves: the ordinates express 



(which are not m contact with the growth; the abscisssB express the 

 . composition of the mixtures as in 



solution) often grow well at the fig. 49. 



start in solutions of toxic substances because the latter 



are held back by the roots. 



The method of mixing equally toxic solutions has also 

 a great advantage when three solutions are employed. As 

 an illustration of this we may take mixtures of NaCl + 

 KCl -f- CaClg. In the case of wheat it was found that the 

 roots grew equally well in solutions of NaC1.12 M, KC1.13 

 M, and CaCla 0.164 M. Mixtures of these solutions were 

 prepared and the growth of the roots in these mixtures 

 was measured after a period of 30 days. In order to show 

 the results graphically, the composition of the solutions 

 may be conveniently expressed by means of a triangular 

 diagram as drawn in Fig. 53. 



