24 BULLETIN 1311, TJ. S. DEPARTMENT OF AGRICULTURE 
ratios is +0.81 ± 04 16 . This indicates a fairly close relationship, and 
it would appear that most of the factors affecting the silica ratio 
affect the lime plus soda ratio in the same direction, or that the two 
ratios are mutually dependent. It may be that the silica, like the 
lime and soda, is more easily removed 17 from the colloid than the 
iron and aluminum (relative to the quantities present) and that 
continued leaching tends to deplete the colloidal residue of silica 
and lime and soda; or it may be that the conditions which would 
determine a low silica ratio in the colloid at the time of formation 
would also determine a low lime plus soda ratio. 
It can be seen from Table 6 that there is no relation between the 
silica ratio and the mean annual temperature, nor between the latter 
and the lime plus soda ratio. The correlation coefficients are re- 
spectively — 0.22±0.09 and — 0.06+0.10. There is, however, some re- 
lation between the two ratios and the rainfall. The correlation be- 
tween the silica ratio and the rainfall is —0.58 ±0.07, and the correla- 
tion between the lime plus soda ratio and the rainfall is — 0.52±0.07. 
There is, therefore, an undoubted, but not a close relation between the 
general chemical composition and the rainfall to which the colloid 
is at present exposed. It appears that rainfall is a far more im- 
portant climatic factor than temperature in determining the chemical 
composition of the colloids studied. 
The relation between the general chemical composition and the 
rainfall is probably causal. It is not possible to state whether the 
rainfall affects the composition of the colloidal matter chiefly by 
causing the formation of a colloid of a certain composition, or 
whether the composition is determined chiefly by the amount of 
leaching after the colloid has been formed 18 ; for the continued action 
of soil solutions on the chemical composition of pure colloidal matter 
of soils is not known. In the consideration of leaching effects, the 
colloid can not be considered apart from the whole soil, and it is 
very probable that in the case of younger soils which still contain a 
higher proportion of the original minerals, the colloidal material 
which is now being formed is not ver}^ different in composition from 
that previously formed. In very mature soils, which contain few 
mineral particles other than quartz, there would be little effect of 
the soluble products of mineral decomposition on the colloid, and in 
such old soils the effect of continued leaching on the composition of 
the colloidal matter, after formation, would reach its maximum. 
The fact that there is not a closer correlation between rainfall and 
composition of the colloid indicates that other conditions apart from 
rainfall cause variations in the composition of the colloidal matter. 
The character of the original and residual parent materials are prob- 
ably important factors. The degree of correlation between rainfall 
and composition of any particular series of colloids would doubtless 
depend upon the selection of these colloids ; the correlation might be 
much greater or less than that found in the 45 here examined. Ac- 
10 The writers are indebted to C. L. Stewart and G. C. Haas of the Bureau of Agricul- 
tural Economics for advice in calculating the correlation coefficient. 
17 The silica may be removed in a coJloidally dispersed condition as well as in true so- 
lution. 
18 Of course it is hardly to be expected that the constituents of the colloidal matter 
would be dissolved in the same proportions as they exist in the colloid as it is formed, 
and it is known that the colloid is not insoluble. It is evident that some change in com- 
position of the colloidal matter must take place in the long ages after it is formed, but 
the magnitude of this change is unknown.. 
