520 
Journal of Agricultural Research 
Vol. VII, No. 12 
as colloidal ferric hydroxid, (3) as soluble undissociated iron compounds, 
and (4) as ionized iron. There was probably a balance between these 
forms of iron, as after filtering the nutrient solutions more precipitate 
formed on standing a short time. 1 From determinations of iron in the 
filtered nutrient solutions it is evident that more or less half the iron was 
precipitated as phosphate and hydroxid. The amount and composition 
of this precipitate probably varied somewhat in the different solutions (9). 
The greater part of the remaining iron was probably present as colloidal 
ferric hydroxid. The available iron, which included the soluble undis¬ 
sociated and ionized iron, was undoubtedly extremely small and was gov¬ 
erned chiefly by the completeness of the hydrolysis of the dissolved iron. 
The amount of iron hydrolized would depend on the reaction of the 
solution, being less in acid solutions, and would also depend on the form 
in which iron was added, being less with the less ionized organic salts. 
The effect of the form of iron and the reaction of the solution on the assimi¬ 
lation of iron by rice is thus easily comprehensible. 
While the amount of soluble or available iron could not be determined 
analytically, some idea of the very small amount present could be obtained 
by comparing the amount of iron absorbed by the plants with the total 
amount of iron added in the volume of solution available during growth. 
This calculation showed that the plants absorbed only one-fifteenth to 
one two-hundredth part of the iron supplied, even in solutions where 
growth was obviously inhibited by lack of iron. If it is assumed that one 
* fifteenth to one two-hundredth part of the iron added to the solutions 
was in a soluble condition, the concentration of the soluble iron in some 
cases would have been from 0.13 to 0.01 part per million. A more prob¬ 
able assumption is that the concentration of soluble iron present at any 
one time was even lower and that as iron was removed by the plant more 
went into solution. It thus appears that in some cases the amount of 
iron in true solution must have been too small for slight differences to be 
accurately determined. It is evident that in certain nutrient solutions 
rice can assimilate sufficient iron’ when the concentration of soluble iron 
is probably less than 1 part in 10,000,000. 
Comparative analyses in Table XII of solutions in which plants had and 
had not grown show that at least the amount of colloidal iron was notably 
diminished by the growth of plants in the solution. The colloidal iron 
was evidently precipitated. 
SUMMARY OF EXPERIMENTAL RESULTS 
The results of the previous culture experiments showed plainly that 
rice in nutrient solutions was not particularly sensitive to an acid or alka¬ 
line reaction per se. Apparently the reaction of the nutrient solution 
affected the growth of rice only through influencing the availability of the 
1 The precipitates in these dilute nutrient solutions were iron compounds, as before the addition of iron 
salts the solutions were perfectly clear. 
