Soil-sampling errors and advisory analyses

Twenty-four sampling units from 50 fields of differing soil type and past manurial history have been examined separately for pH and 1% citric soluble phosphorus and potassium to determine the errors which can occur in obtaining a representative sample for analysis. 2. Soils which had received lime and fertilizers within 3 years of the sampling date showed appreciably greater sampling errors than those which had not. 3. The sampling errors were greater than the analytical errors and were such that the common practice of classifying soils into six fertility groups was not justified. Three major groups are more reasonable and practical for routine advisory purposes.