In this post, we briefly explore the documented development of the USDA textural triangle and compare historical triangles to the present (post-1951) iteration.
Milton Whitney published the first plot of soil class boundaries in Cartesian coordinates (which become a right triangle because all size fractions must add to 100) in 1911 (Whitney, 1911). The coordinates of this diagram (Whitney, 1911, Fig. 1) are silt fraction (0-100%) on the x-axis and clay fraction (0-100%) on the y-axis. This work was an international landmark in establishing a quantitative basis for soil textural classes.
|Figure 1. USDA Right Soil Textural Triangle, Whitney, 1911.|
Soil scientists involved with survey work at the time were asked to classify soils in the field based on accepted class names in their region. These samples were then sent to the USDA Soils Bureau laboratory, which performed mechanical analysis on 8,664 samples. The results of these analyses were used to provide field limits (and hence class boundaries) for 8 soil textural classes. The approach of matching textural class boundaries based on mechanical analysis to perceived field textures and class names (and not the other way around), has driven the lasting utility of the USDA triangle. The disadvantage of this triangle in terms of field use is that clay and silt are emphasized on the axes (when clay and sand are typically the easiest to estimate in the field), and although the remaining fraction (sand) can be determined from the information in cartesian coordinates, it requires an additional calculation step which is not field expedient.
In 1927, the first equilateral textural triangle (with sand, silt and clay fractions represented on equal axes) was published (Davis and Bennett, 1927, Fig. 2). This triangle contained 10 textural classes (with the addition of Silty Clays and Sandy Clay Loams) and presented a more regular geometric approach to textural class divisions.
|Figure 2. USDA Equilateral Soil Textural Triangle, Davis and Bennett, 1927.|
Some argued that the equilateral approach includes superfluous information (the 3rd axis is not necessary because the last size fraction can be determined from 100% minus the other two), but the field utility of the equilateral triangle for bracketing hard to estimate fractions from the clay estimate alone (see post: Sand Bracketing 101) is unparalleled.
In 1938, the USDA revised the upper size limit of the clay fraction downward from 5um to 2um (Knight, 1938) in order to better coincide with the International system, the perception of textures in the field and the physical and chemical properties of clay size particles (Truog et al., 1937, Shaw and Alexander, 1937, Knight, 1938, Marshall, 1947).
In 1951, the USDA published the current version of the textural triangle (Soil Survey Staff, 1951, Fig. 3), which includes 2 more textural classes (Silt and Loamy Sand), for a total of 12. The triangle also shows "wider" clay ranges for all textural classes due to the conversion from a 5um to 2um upper limit for the clay fraction. This conversion is not a simple one, and requires significant approximations (see Marshall, 1947 and Nemmes et al. 1999).
|Figure 3. USDA Equilateral Soil Textural Triangle - Soil Survey Handbook, 1951|
This triangle retains the field advantages of an equilateral triangle and also shows a more pragmatic approach than the 1927 triangle to matching field perceptions to mechanical analysis. This point is reflected in the "sand drop" - that is, on the sandy side of the triangle, the threshold clay percentages are lower than for the subsequent loam or silt loam classes. This is largely due to work by Marshall (1947), who demonstrated this phenomenon related to the human perception of plasticity (an estimate of clay content) with increasing sand fraction. Essentially, the clay fraction is more effectively plastic when the remainder of the soil is dominated by sands rather than silts, requiring a lower threshold for the perception of a more clayey textural class (i.e. sandy clay loams, sandy clays) Marshall (1947, 2003) has provided 2 hypotheses for why this is the case:
1. Because particle size fractions are weight percentages, non-clay fractions "dilute" the feel of the clay fraction much less when they are present as a few large particles (i.e. sand) than a larger number of smaller particles (i.e. silt), requiring a lower threshold with increasing sand content.
2. A large silt fraction with many silt-size particles close to the 2um boundary also implies that there will be a large number of clay-size particles close to that boundary. These coarse clays do not contribute as much to the cohesive and adhesive forces that are perceived as plasticity in the clay fraction. Therefore, there is a greater proportion of "inactive" clays, making the clay fraction less plastic (and requiring a higher threshold with increasing silt content).
The end result of these 4 decades of hard work (1911-1951, Fig. 5) was a world-renowned textural triangle that is pragmatic, effective in the field and the laboratory, and useful for determining quantitative differences in soil properties by feel alone. A tremendous achievement!
Post Script: There is a 1945 "missing link" triangle that was never published in any official USDA literature but was reproduced in Marshall (1947) (Fig. 5).
|Figure 5. USDA Equilateral Soil Textural 1945, Reproduced from personal communication in Marshall, 1947.|
This triangle reflects the 1938 revision of clay fractions boundaries downward ("wider textural classes"), and the addition of a loamy sand particle size class. Except for the sandy loam "cup" under the loam textural class, however, it does not reflect the now well accepted "sand drop". Marshall states:
"A copy of this diagram was made available to the writer through the courtesy of officers of the United States Department of Agriculture*...*[Footnote] The writer is indebted to...Mr. J.K. Ableiter and Dr. L.T. Alexander of the United States Department of Agriculture for details regarding this diagram."
Davis, R.O.E., H.H. Bennett. 1927. Grouping of soils on the basis of mechanical analysis. USDA Department Circular 419.
Knight, H.G. 1938. New size limits for silt and clay. Soil Science Society of America Proceedings 2: 592.
Marshall, T.J. 1947. Mechanical composition of soil in relation to field descriptions of texture. Council for Scientific and Industrial Research, Australia, Bulletin No. 224.
Marshall, T.J. 2003. Particle-size distribution of soil and the perception of texture. Australian Journal of Soil Research 41: 245-249.
Moeys, J. 2014. The soil texture wizard: R functions for plotting, classifying, transforming and exploring soil texture data. http://cran.r-project.org/web/packages/soiltexture/vignettes/soiltexture_vignette.pdf
Nemes, A., J.M.H. Wosten, A. Lilly, J.H. Oushe Voshaar. 1999. Evaluation of different procedures to interpolate particle-size distributions to achieve compatibility within soil databases. Geoderma 90: 187-202.
Shaw, T.M. and L.T. Alexander. 1937. A note on mechanical analysis and soils texture. Soil Science Society of America Proceedings 1: 303-304.
Soil Survey Staff. 1951. Soil Survey Manual. USDA Handbook No. 18.
Truog, E., J.H. Taylor, R.W. Simonson, M. E. Weeks. 1937. Mechanical and mineralogical subdivision of the clay separate of soils. Soil Science Society of America Proceedings (1): 175-179.
Whitney, M. 1911. The Use of Soils East of the Great Plains Region. USDA Bureau of Soils Bulletin No. 78.