The solubility products (K-sp) of 1:1 complexes of glyphosate, [N-(phosphonomethyl)glycine], with Mg2+, Ca2+, Mn2+, Zn2+, Cu2+ and Fe3+, were determined in buffered (pH 7.0) distilled water, moist Ottawa sand, sandy loam and clay loam soils, each adjusted to 0.02 M with respect to KNO3. The K-sp values decreased in the order of Mg approximate to Ca > Mn > Zn > Cu > Fe, regardless of the medium in which they were determined. The constants measured in Ottawa sand were similar to those in water, but those in the forestry soils depended upon the type of metal ion involved. The values for the Mg, Ca, Mn and Zn complexes were about 2 to 3 times lower in sandy loam soil than those in water, but those in clay loam were about 3 to 10 times lower. The K-sp of the Cu and Fe complexes were similar to those in water regardless of the soil type used. In a bioassay experiment using tomato plants, immersed in the saturated solutions of the complexes or planted in the sand and soils containing saturated solutions of the complexes, no mortality occurred although slight inhibition in growth was observed in all cases. The most soluble complexes of Mg and Ca caused the most reduction in plant height, while the least soluble complex of Fe caused little growth inhibition. The Zn, Cu and Mn complexes caused some growth inhibition depending on their K-sp values. The larger the solubility product, the greater the concentration of glyphosate ion in solution, and the greater the growth inhibition. In a similar experiment using white spruce seedlings, growth inhibition was insignificant over the 12-d bioassay period.
