324 
Journal of Agricultural Research 
Vol. XXIV, No 4 
centrations of all these substances ranged from 500 to 5,000 parts per 
million. The general relations are nearly the same as with the car-^ 
bonate shown in figures 3 and 4. The simple mixture of sodium chlorid 
and sodium sulphate shows additive toxicity, while the other mixtures 
add a much smaller amount or nothing at all to the harmful effects of 
the chlorid. The three and four component mixtures follow the simpler 
chlorid curve closely (fig. 5, D). No antagonism for sodium chlorid is 
evident. ' 
The arrangement of the experiment represented in figure 6 differs 
from that of figure 5 only in having sodium nitrate substituted for 
sodium chlorid. The relative shapes and positions of all the curves are 
so exactly analogous to those in figure 5 that the remarks and conclusions 
given above apply in this case also. 
I^ATER WORK 
With the experience gained in the experiments already described, the 
investigation of this problem was continued in the summer of 1921. 
It was decided to concentrate effort on one soil, containing one salt as 
a base, and to vary the subsequent treatment as much as possible. 
Accordingly, six different trials were carried out with Greenville loam 
