The persistence and retention of active ingredients in granules of Thimet 15G (phorate 15% by weight), Dyfonate 10G (fonofos 10% by weight), Counter 15G (terbufos 15% by weight), and Furadan 10G (carbofuran 10% by weight) were determined in silt loam and organic muck agricultural soils typical of the lower Fraser River valley (BC, Canada). In June 1995, treatment bags made of polyester cloth (7.5 × 7.5 cm) containing granules of a single insecticide, either alone or with soil, were placed during spring planting in the bottom of the furrow and retrieved periodically until April 1996. The parent component of each insecticide declined monotonically except for carbofuran (logistic decline). In the silt loam (organic muck) soil, the average June-to-October first-order rate constants and half-lives were 0.009 (0.010)/d and 80 (71) d for fonofos, 0.012 (0.009)/d and 58 (82) d for phorate, and 0.032 (0.015)/d and 21 (47) d for terbufos; the half-life of carbofuran was 129 (97) d. By December, the average amounts of fonofos and phorate in silt loam (organic muck) were 26% (range: 17–40%; 14% [range: 3.4–21%]) and 21% (range: 15–30%; 10% [range: 5.0–24%]) of the initial amounts of active ingredients measured at time zero, respectively. By April, the percentages dropped to 16% (range: 7.8–24%; 2.3% [range: 0–7.7%]) and 7.3% (range: 1.9–25%; 0.6% [range: 0–1.9%]). During this period, about 95% of the active ingredients were granule bound, the rest remaining in the bag. Only low levels of terbufos and carbofuran persisted in both soils from December to April of the following year. Results indicate an enhanced probability for poisoning of waterfowl and raptors because of the high levels of active ingredients retained on granules of all four insecticides in both soils in the fall. The risk of acute poisoning by phorate and fonofos continued though the winter.