Comparison of Soil Tests for Assessing Boron Toxicity to Alfalfa

Abstract

AbstractThree soils from Colorado that varied in texture and organic matter (OM) content were treated with sodium borate (Na2B4O7·10H2O) at rates of 0, 10, 20, and 40 mg B kg−1 soil. Alfalfa (Medicago sativa, ‘Ladak’, ‘Rambler’, ‘Riley’, and ‘Agate’) was planted in the soils in the greenhouse and cut twice at the 30% bloom stage. Plant tissue digests and soil extracts (ammonium bicarbonate‐diethylenetriaminepentaacetic acid [AB‐DTPA], saturation, hot water, and mannitol‐CaCl2) were analyzed for B using inductively coupled plasma atomic emission spectroscopy (ICP‐AES). Alfalfa yield was significantly (α = 0.01) reduced with increased B application rate in the Ascalon sandy loam and Ulysses loam soils, whereas no yield reduction was observed in the Keith silt loam soil (Aridic Argiustoll, Haplustoll, and Argiustoll, respectively). Yield response for the pooled soils were poorly correlated (r2 from 0.23–0.58) to the four soil extractants; however, yield reduction due to B toxicity corresponded to critical plant B levels of 850 mg B kg−1 and 975 mg B kg−1 for the first and second harvests, respectively. The inclusion of NH4OAc‐extractable Ca into the regression equation of soil B vs. yield significantly increased the correlation (r2 = 0.77). Multiple linear regression analysis utilizing Mallow's Cp statistic increased the adjusted coefficient of determination (R2) to 0.87 for all of the soil B extractants upon inclusion of soil variables (pH, % clay, OM content) into the equation. Although NH4OAc‐extractable Ca did not appear in any of the best subsets, this was probably the result of it's high correlation with percent clay (r2 = 0.98) and OM content (r2 = 0.98).