COMPARISON OF NEAR-INFRARED SPECTROSCOPY AND OTHER SOIL-NITROGEN AVAILABILITY QUICK TESTS FOR CORN

Our ability to predict N fertilizer needs for corn (Zea mays L.) is im proving, but more accurate and convenient tests are still needed. This work compared a new quick test for soil N availability using a near-i nfrared spectrophotometer (NIRS) with three published quick tests for predicting soil N-supplying capability (NSC) and relative corn grain y ield. The other tests were the pre-sidedress nitrate test (PSNT), nitr ate concentration (at-plant NO3), and absorbance at 200 nm of a 0.01 M NaHCO3 extract (UV-200 test) of 0- to 20-cm soil samples taken at pla nting. Soil samples taken at planting from 95 field experiments in Pen nsylvania were analyzed at reflectance wavelengths from 400 nm to 2500 nm with NIRS. The coefficients of determination were the same (R2 = 0 .49) for both linear and quadratic regressions of NSC and NIRS test va lues. The abilities of the four tests to predict NSC and relative corn grain yield were compared using data from 90 of the 95 experiments. T he R2 values for linear and quadratic regressions between soil test va lues and NSC ranged from 0.49 to 0.58 for the NIRS, PSNT, and UV-200 t ests; for the at-plant NO3, test, R2 was lower (almost-equal-to 0.40). Eliminating sites where corn directly followed a legume, R2 values fo r quadratic regressions between test values and NSC increased to almos t-equal-to 0.60 for the NIRS, PSNT, and UV-200 tests. The PSNT test wa s slightly better than the other tests in predicting a grain yield res ponse to N fertilizer, but this advantage lessened when first-year-aft er-legume sites were eliminated. No test could accurately predict rela tive grain yield or NSC in fields responding to N fertilizer (R2 = 0.0 8-0.36). The NIRS test offers a convenient, rapid, and inexpensive alt ernative to the PSNT for predicting whether humid-region corn fields w ill respond to N fertilizer.