[Vor. 7 
44 ANNALS OF THE MISSOURI BOTANICAL GARDEN 
In a study of probable error he grew single plants in each of 10 
bottles of 1200 cc. capacity, employing uniform methods and 
seed of a selected strain of rye. Four concentrations of nutrient 
solutions were used with the different lots. Experiments con- 
ducted during the early months of the year yielded results as 
follows: The greatest individual variation in any one lot 
amounted to never less than 70 per cent. In the weakest con- 
centration the individual variation was 333 per cent. However, 
the probable error of the mean in these cases was only about 
3-10 per cent of the mean dry weights. Comparable differences 
were found in cultures made later in the spring with a pure line 
of barley. 
That considerable variability has been found by others is 
evident from the examination of the tables in any case in which 
the data have been given in detail. Livingston and Tottingham 
(18) give a table from which it appears that while the culture 
RC; yields only a fair growth of roots the dry weight for the 
entire plant is third and for tops is the highest of all in the 
series. This solution, however, differs from RsC, in containing 
one-eighth the concentration of KNO:, 3 times as much 
Ca (H;PO,), and 6 times as much MgSO,. RC, is regarded 
as “the best balanced for young wheat plants of all the 
nutrient solutions so far noted in the literature.” 
The Shive solution, on the one hand, and the Livingston- 
Tottingham solution, on the other, were of course designed with 
the idea of simplification. The 3 salts employed contain all 
the essential ions except iron. Theoretically, the phosphate, 
nitrate, and sulphate ions may be added in the form of the salts 
of either of the 3 bases or cations, K, Ca, and Mg. However, 
the relatively low solubility of Ca as sulphate may seem to 
render it less practicable to use this salt. There remain 7 
possibilities in the selection of salts, and it might appear that in 
the selection of these the only important points might be, first, 
H-ion concentration, and second, the use of a base required in 
relatively low concentration with an anion which may be sim- 
ilarly reduced in strength. 
From the emphasis in the literature, as is well known, an 
important consideration is an appropriate ratio between the 
ionic proportion of Ca and Mg. Other antagonistic relations 
