A NEW THREE-SALT NUTRIENT SOLUTION 
physiologically the most efficient on the basis of the whole group of 
criteria employed. Of course this is not to be interpreted to mean that 
this solution is the very best possible for this plant, for these climatic 
conditions and for these criteria, since the distribution of our solutions 
on the triangular diagram is rather open, much more so than in the 
case of Shive's series of thirty-six different sets of salt proportions, 
and of course some other set of proportions than those here tested might 
have given still higher plant values. It is not probable, however, that 
either the growth values or the salt proportions of such a hypothetical 
very best solution might have been pronouncedly different from those 
indicated. 
On the triangular diagram, this solution of our series lies very close 
to Shive's best solution for wheat; ours is R6C1, while his is R5C2, and 
solution R5C2 was not tested in our series. 
The relative data show that our single culture with solution R6C1 
gave somewhat higher values than the corresponding average of the 
three controls with Shive's best solution, in all cases excepting that of 
root length. The approximate mean length of the main roots for our 
solution R6C1 is shown to be equal to the control average. The 
variations due to unknown conditions (''individual variations") are 
always so great in this sort of experimentation, however, that no par- 
ticular emphasis should be placed upon the indications that our solu- 
tion R6C1 gave better growth than did Shive's best; for it is to be re- 
membered that the data for our twelve new solutions are derived from 
single cultures, of only six plants each. For the present, it is sufficient 
to say that our three-salt solution R6C1 is apparently 7^5/ as good for 
the growth of young wheat plants (judged by the numerical data of 
table 2) as is Shive's optimal R5C2, despite the fact that the propor- 
tions of the component ions are considerably different in the two solu- 
tions. 
It does not seem desirable to enter into further discussion of these 
results at the present time, on account of the paucity of our data, but 
it may be of value to present briefly our observations on the apparent 
health of the plants in the various solutions. The peculiar morpho- 
logical responses called "magnesium injury" by Tottingham, which 
frequently occur with Shive's optimal R5C2 for wheat, appeared in 
some of the cultures (see the asterisks in the fourth column of table 2), 
and they were pronounced with solution R6C1, the one that appears 
best on the basis of the numerical values. In this set they occurred 
