The height of dominant trees, the ones that grow without crowding or shading, at any given age is really the best measure of site quality. We used 50 years as the standard age, and estimated from growth curves the height of any stand where age was less or more than 50 years. This actual or estimated height at 50 years is known as site index. A stand, for instance, whose current height and age indicated a probable height of 75 feet at 50 years was assigned a site index of 75.

Obviously, comparison of tree heights of two stands, say 10 and 40 years old, would not be a fair comparison of the richness of two soils; but if the height of the 10-year-old stand were calculated from the curve to what it would be in 40 more years, and the height of the 40-year-old stand were calculated to what it would be in 10 more years, the two heights could be compared on a fair basis. This we did by assigning a site index to each stand.

When we placed the site index of each stand in one of the six positions according to its soil, the group site index averages fell into regular order roughly by steps of 10 from 50 to 100. Black locust stands on sites whose subsoil was slowly drained, sticky, compact, and bluish-colored or drab-mottled below 8 inches averaged only 50 in site index—much too low for profitable black locust growth. Stands on sites with good to fast drainage, with crumbly, loose, reddish-brown subsoil, averaged roughly 100 in site index. This arrangement of soil and tree growth brought an order to soil facts that showed how tree growth responds to soil.

The differences in thrift of black locust stands were pronounced. We found knotty, straggly, tight-barked, runty stands; and straight, tall, fluted stems with bulging roundness that seemed to split the bark. All these differences stood out more and more clearly as we measured trees and looked at the soil. The soil was different, too.

A slow-growing stand in Ripley County, Ind., near Osgood, was on a ridge top where the drying winds swept the moisture right out of the soil. To make the site still worse, only a shallow soil lay over bedrock—not much chance for good timber here; only 32 feet high at 25 years. A fast-growing grove in a little valley churchyard near Paris Crossing, Jennings County, Ind., had a deep, mellow silt loam soil ( Cincinnati silt loam). Its subsoil was well aerated; its color was a golden brown. The site was sheltered from excessive wind; its soil was moist—small wonder that the trees were 13 inches in diameter and 90 feet high at 40 years. We began to use simple relations of soil stickiness, compactness, and color to drainage and aeration to explain tree differences.

These simple relations have great value to forestry, but they were not found in a hurry. As early as the late 1920's, Tom Bushnell, chief of the Indiana Soil Survey, was arranging soils by drainage groups.

Richard Bradfield, working at Cornell University, found that roots of apple trees penetrated the soil only so far as it was well oxidized. If the ratio of oxidation to reduction (measured electrically) was high, the roots penetrated deeply and the trees were thrifty. If the ratio of oxidation to reduction was low, the roots died at a slight depth and the trees were slow growing and unthrifty. Dr. Bradfield, of course, knew about the relation of drainage to aeration; he wanted a quick test for apple soil.

His work adds another link to the chain of site evidence. A well-drained soil is a well-aerated soil, and a well-aerated soil is a well-oxidized soil. But the iron oxides in a well-oxidized soil are red and brown, whereas the iron oxides in a reduced soil are blue or green. Therefore the subsoil of a well-drained, well-aerated, well-oxidized soil is red or brown, and the subsoil of a poorly drained, slowly aerated, reduced soil is bluish or drab. The degree of drainage and aeration accordingly are indicated by subsoil color.

Milne, a South African soil scientist, in 1936 gave the name "catena" (Latin for "chain") to groups of soils varying systematically in drainage. Soil surveyors have worked painstakingly and long, mapping and describing soils, and their work continues. Even now they do not fully agree on what a catena really means; its application is new. Some say it is a hydrologic sequence—a high-sounding expression meaning arrangement by degree of soil moistness; some say it is an order of drainage; but whatever they finally decide, black locust growth defines it as a range of usable soil water.

The arithmetic of soil water is simple. Total rainfall less runoff water less evaporation water equals ground water. Obviously, useful water in soil does not depend alone on how much rain falls, or altogether on how much runs off the surface, or even on how much runs into the soil; but it depends also and importantly on how much is evaporated.

A very little observation soon convinces one that a south slope loses "Wore water by evaporation than a north slope because it gets more direct sunlight, and that a wind-swept upper slope or ridge loses more water by evaporation than a lower slope or cove because more moisture-absorbing air passes over it. But evaporation varies so much that a way of evaluating it by sites had to be found for hilly land. At this stage of the search for an answer to site prediction on abandoned fields, we had attempted to find it by measuring the subsoil. The application of topography, with its three parts—aspect, exposure, and position—in making site predictions remained to be studied.