232 BOTANICAL GAZETTE [SEPTEMBER 
By this method we find that the effect of 60 A is equal to the 
effect of 42 C, while the effect of 40 B is equal to the effect of 56 C. 
The additive effect of (60 4+40 B) is therefore equal to the addi- 
tive effect of (42 C-+56 C=) 98 C, which is seen from the figure 
to be 11.5. 
The values of the additive effect thus obtained are plotted in 
fig. 2. It will be seen that these values do not differ greatly from 
6.0. the value of an additive effect which is con- 
stant (and equal to the effect of 100 A or 
too B). Unless, therefore, the two dilution 
curves (as plotted in fig. 1) diverge widely, 
there will be no great error in regarding the 
additive effect as constant (and equal to the 
effect of 100 A or 100 B); this error will in 
fact ordinarily be less than the experi- 
mental error. 
In case there are several salt solu- 
tions to be mixed, we may draw the 
corresponding dilution curves and 
average the ordinates 
of these curves at 
various elevations to 
obtain points through 
rea | 
— 8 lu 
0 10 30 50 70M.M. 
Fic. 3.—Curves showing growth in various dilutions of two unequally toxic 
solutions of salts, M and N: the abscissas represent growth; the ordinates represent 
the number of cc. of the salt solution which are taken and diluted to 200 cc. to make 
the culture solution in which the plants are grown. 
which a curve may be drawn which shall serve the same purpose 
as the curve C in fig. 1; or an arbitrary curve (for example, a parab- 
ola or hyperbola) may be drawn for this purpose. 
When the two salt solutions are not equally toxic, we often 
find cases in which a constant relation exists between the amounts 
of the two solutions which (diluted to 200 cc.) produce the same 
