RELATIVE COST TO SOIL FERTILITY OF LONG-TERM CROP PRODUCTION WITHOUT FERTILIZATION

An earlier analysis of yield trends of stubble-wheat in six cropping s ystems, over 35 yr, in a thin Black Chemozemic soil at Indian Head, Sa skatchewan, showed that fertilizer improved soil quality, while absenc e of fertilizer, combined with frequent fallowing, led to soil degrada tion. The inclusion of a legume green manure crop in the rotation fail ed to maintain soil fertility, apparently because legumes do not suppl y P. Because the fertility and stored moisture effects were confounded , we conducted a growth chamber experiment to quantify soil responses to N and P in these six cropping systems. Soil from the top 15-cm of t he rotation phase that had just grown two successive wheat (Triticum a estivum L.) crops was used. Various factorial combinations of ammonium nitrate-N and triple superphosphate-P were applied at N/P2O5 rates up to 200/200 kg ha(-1). Soil moisture was maintained in the available r ange. Regression analysis showed that the fallow-wheat-wheat (F-W-W) a nd continuous wheat (Cent W) systems that had not been fertilized in 3 5 yr, and which had moderate amounts of NaHCO3-P, only responded to N. In contrast, the green manure (GM)- and hay (H)- containing systems, which had also nor been fertilized before, had low levels of NaHCO3-P and responded to both N and P. In the field, the yields of wheat grown on stubble in 1991 rated: Cent W (N + P) > F-W-W (N + P) > F-W-W-H-H- H > Cent W > GM-W-W > F-W-W. However, in the growth chamber the rating was: Cent W (N + P) > F-W-W-H-H-H > GM-W-W > Cent W > F-W-W (N + P) > F-W-W. We suggest that the growth chamber results more accurately ref lect the present fertility status of these soils, because fertility is no longer confounded with soil moisture. Grain yields in the growth c hamber were directly proportional to the previously measured initial p otential rate of N mineralization, indicating the value of the latter parameter as a useful index of soil N fertility.