There has been a close historical connection between developments in geology and civil engineering. Highway engineers, in particular, not only design and build heavy structures covering relatively small areas of earth material, but they also construct lightly loaded pavement structures covering large areas. In both cases, information is needed about the soil conditions. Detailed soil mechanics investigations can usually be justified for the design of heavy bridge and overpass structures; however, the engineer concerned with roadway design can seldom afford the time and money that are necessary for a detailed survey of a right-of-way, which is often 300 feet wide and several miles long. Proper roadway design, therefore, requires the use of every available aid in order to relate a small amount of test and pavement performance data to the largest possible mass of earth. The principal sources of such assistance are geologic and agricultural (pedologic) soil reports, which provide information on the distribution and character of surficial deposits, and vertical aerial photographs. Significant progress has been made in the correlation of geologic, pedologic, and air photo map units with engineering characteristics of soils. In the beginning, such correlations were based primarily on qualitative field observations; later, representative test data were included; and finally, through the use of statistical methods, some measure of the variability of natural soils was obtained. The statistical correlation of quantitative engineering data with map units is still in its infancy, but it gives promise of providing a rational basis for the design of highway pavements. This review is primarily a discussion of the development of procedures of engineering soil survey and classification which rely upon geology, pedology, and air photo interpretation for exploration, presentation of data, extrapolation of data, and formulation of design criteria. Particular emphasis is given to pedologic classification and its engineering applications.