May 3, 1912] 
method must, therefore, be set aside. The 
important crop-producing properties must 
be correlated with some group of charac- 
teristics of the soil which are more easily 
recognized and grouped, and by this means 
arrive at individual types of soil which 
have some measure of agricultural unity 
and value. The more perfect this correla- 
tion of properties the more satisfactory 
will be the result. The characteristics of 
the soil which have been most often 
selected for correlation with its crop- 
producing properties are the geological. 
Second. Herein lies the second group 
of difficulties, namely, in establishing the 
correct relation between crop-producing 
power and geological origin, and also in 
the accurate separation of soil material 
along geological lines, due to the com- 
plexity of the geological processes and 
materials involved. Attention should be 
directed here to a point often overlooked, 
namely, that the separation of soils along 
geological lines is secondary to their sepa- 
ration according to crop-producing power. 
The method is an incident to the result, 
however interesting it may be as a part of 
the science of geology. Much confusion in 
work has resulted from the failure to sense 
this distinction. The soil type is a geolog- 
ical element in the structure of the earth 
and as such is legitimately investigated by 
the geologist. Unfortunately, for agricul- 
tural purposes, the geologist has seldom 
carried his separations far enough, due 
probably to the fact that he has not had 
the agricultural point of view and, there- 
fore, has not appreciated the kind of dis- 
tinctions which are demanded for farming 
purposes. This may be termed the agro- 
nomic point of view. 
Principles which have been used in the 
Classification of Soils—The soil has been 
classified from many points of view. Hm- 
SCIENCE 
679 
mons! studied the soils of New York be- 
tween 1835 and 1840 and made somewhat 
extensive chemical analyses of the soils in 
the several regions into which the state was 
divided. This is probably the most exten- 
sive of the early surveys and combined the 
chemical composition with geological origin 
in arriving at the agricultural separation, 
a method generally employed by investiga- 
tors working in this field. The physical 
properties of the material were also recog- 
nized as important, but these several fac- 
tors were not correlated in any definite 
way. Owen and Peters? made a very com- 
prehensive study of the chemical composi- 
tion of the soils of Kentucky. Hilgard* 
has made the largest collection of the chem- 
ical analyses of soils available in America, 
in connection with the investigation of the 
cotton soils, and the soils are classified in 
the field and indicated on maps with refer- 
ence to the native vegetation. The mode 
of formation and physical properties re- 
ceived consideration. Shaler* in an article 
published in 1891 called attention rather 
definitely to the relation between the proc- 
esses of soil formation and the character 
of the material and pointed out correlated 
types of plant growth. 
In foreign countries a number of men 
have studied the problem of the practical 
classification of soils and of these the Rus- 
sians have been most successful. Dokou- 
1Hmmons, Ebenezer, ‘‘The Soils of New York. 
Natural History of New York,’’ Agriculture, I., 
207-360 (1843), 1846. 
? Owen, D. D., Geological Survey, Kentucky, An- 
nual Repts., 1855-1875. Peters, Robert, ‘‘Chem- 
ieal Composition of Soils, Marls, Ores, etc., Chem- 
ical Analysis A,’’ Pts. I., IL., III., 1875-1888. 
SHilgard, E. W., ‘‘Cotton Production,’’ 10th 
Census, V. and VI., 1880. 
4Shaler, N. S., ‘‘Origin and Nature of Soils,’’ 
12th Ann) Rept. U. S. G. S., Pt. I., 219-3845, 
1890-91. 
