Correlation of Soil Tests for Available Phosphorus and Potassium with Crop Yield Responses to Fertilization1
- O. H. Long and
- L. F. Seatz2
Separate yield responses to applied P and K were correlated with soil tests for these elements on 72 experiments with corn, 18 with cotton, 35 with small grains, 28 with legume hays and 11 with permanent pastures. Sodium perchlorate in N/10 perchloric acid was the extracting agent.
Results are presented in graphic form showing soil-test values for P and K plotted against crop yields obtained without P or K expressed as percent of yields where phosphate or potash was applied.
Two methods of approach are used in showing the degree of correlation. One method is concerned with “prediction accuracy” at only the lower values of soil P or K. The other method is concerned with the regression curves and correlation coefficients for the experiments as a whole.
Responses of corn to phosphorus fertilization on loess-derived soils correlated least satisfactorily with the soil test where a predicted yield response was obtained in only 17% of the tests. Permanent pastures showed the best correlation; all experiments were on soils low in phosphate and responded to applications of P. The prediction accuracies obtained with other crops were as follows: corn, on limestone-derived soils, 64%; cotton, 54%; small grains, 70%; and legume hays, 63%.
Responses of cotton to potassium fertilization indicated the best correlation with the soil test, the prediction accuracy being 60%. The prediction accuracies obtained with the other crops were: corn, 44%; small grains, 47%; legume hays, 33%; and permanent pastures, 57%.
The correlation coefficient (r) was not significant with respect to P on any crop; it was significant with respect to K only on corn.Please view the pdf by using the Full Text (PDF) link under 'View' to the left.
Copyright © .