Abstract
The usefulness of soil tests (five each for nitrogen phosphorus and potassium) in predicting crop response to fertilizer addition was assessed by correlating test values with yield data from 70 NPK fertilizer field trials using maize (Zea mays L.) in the West Indian islands of Antigua, Dominica, Montserrat and St. Kitts. The analytical methods were also calibrated against field response data and a critical test value (CTV) for each method determined. Above the critical value responses to fertilizer addition are unlikely; below this value responses may be expected. It is suggested that for routine measurements of farmers' soil samples from the region under discussion the following tests would be the most suitable. Nitrogen: total oil nitrogen with a CTV of 0·14 per cent. Percentage base saturation (CTV = 53 per cent) also gives as good an estimate (64 per cent correct predictions) of response to nitrogen fertilizer as the other tests for soil nitrogen. Phosphorus: either Olsen's 0·5 M sodium bicarbonate extract with a CTV of 7·5 p.p.m. P or a modified Bondorff method (1 g soil shaken for 16·5h with 50ml 0·2N sulphuric acid) using a CTV of 70 p.p.m. P. Both these methods gave 71 per cent correct predictions of crop response to added phosphate fertilizer. Potassium: exchangeable potassium removed in a neutral 1 N ammonium acetate leachate. This test with a CTV of 0·25 mequiv. K per 100g soil gave 83 per cent correct estimates of crop response to potash fertilizer. When taken separately the data from the sit in the much wetter island of Dominica led to selection of different tests and modified critical values. The critical values obtained in this study were compared with those determined for other soils in the Caribbean and other parts of the world.
