482 Philippine Journal of Science 1920 
The solution represented by the right apex itself (R1S5) is by 
far the poorest of the series, giving tops only 62.5 per cent as 
long as those given by solution R4S1, which is the best of the 
series. This means that solutions having relatively low partial 
concentrations of the calcium salt are the best. None of the 
five best sets of salt proportions has over one-seventh of its 
molecular total made up of calcium nitrate. The three solutions 
having four-sevenths or five-sevenths made up of calcium nitrate 
(which are also characterized by low partial concentration of 
magnesium sulphate) all lie in the poor group. 
The root lengths were also obtained for series 4, and they are 
shown by their relative values on the diagram of fig. 2, but they 
bring out no clear generalization. In this case unity (1.00) 
represents an actual value of 14.5 centimeters. 
Although the 3-salt solutions of the present study are planned 
by volume-molecular rather than by osmotic salt proportions 
(the latter having been used by Shive), yet it will be of interest 
to compare the generalization stated above for these young rice 
plants, with the outcome of Shive’s series for wheat (1915), 
buckwheat (1917), and soy bean.?’ 
The triangular diagrams in question are shown in fig. 3, in 
which the diagram of fig. 1 is repeated for the sake of com- 
parison. Considering the main regions of good and poor gr owth, 
as marked on these diagrams, for wheat it appears that the 
best sets of salt proportions are near the center of the triangle, 
with the three salts in somewhat nearly equal proportions, while 
the poorest sets have very low partial concentrations of mono- 
potassium phosphate and calcium nitrate and very high partial 
concentrations of magnesium sulphate. For buckwheat low or 
medium partial concentrations of calcium nitrate together with 
medium partial concentrations of monopotassium phosphate Te- 
present the best sets of proportions, while the poorest sets have 
low partial concentrations of magnesium sulphate together with 
medium and high concentrations of monopotassium phosphate. 
For soy bean [these data have been derived from Shive’s table 
of results given (1918) on page 109] it appears that low partial 
concentrations of monopotassium phosphate clearly limit the 
region of good balance, while high ones represent poor balance. 
It is remarkable that the diagram for soy bean is very similar 
to that for rice, except that good physiological balance for S°¥ 
bean seems to require a very low supply of monopotassium phos- 
*Shive, John W., Toxicity of monobasic phosphates towards soy-beans 
grown in soil- and solution-cultures, Soils Sci. 5 (1918) 87-122. 
