40 Ar/ 39 Ar step heating of single K-feldspar crystals (∼ 500 μ g) in a furnace with a low argon blank and accurate temperature control provide constraints on the thermal history of the Salton Sea geothermal field, southern California. Estimates of 40 Ar loss from detrital K-feldspars, coupled with kinetic information gleaned from laboratory degassing of 39 Ar, allow possible thermal histories to be evaluated. Samples from core and cuttings of sandstones from the State 2–14, Landers #1, Dearborn Farms #1, Magmamax #2, Elmore #3 and Sinclair #3 wells have been analyzed for this purpose. No recent 40 Ar ∗ loss is recognized in any of the 17 crystals measured, even for samples currently at temperatures greater than 275°C. This observation suggests that the present temperature distributions in these wells have not been maintained for greater than 1000 years. However, aggregates (∼ 100 mg) of two K-feldspar samples do reveal recent 40 Ar ∗ loss from the State 2–14 well, suggesting a time scale of 1000 to 5000 years for near-peak thermal conditions. These estimates of heating duration are broadly consistent with earlier results from thermal modelling, but do not support interpretations that paleotemperatures were once hotter than present. Provenance ages span the Tertiary through Middle Proterozoic but cluster at ages of 25, 70 and 1250 Ma. These age peaks are very common K-feldspar K-Ar ages in detachment terranes of southeastern California and southwestern Arizona, both drained by the Colorado River. This coincidence supports earlier suggestions that the sands of the Salton Trough were derived mainly from the Colorado River. Unrealistically low apparent activation energies (< 20 kcal/mol) for argon transport are observed for several single crystals which yield absurdly low calculated closure temperatures (e.g., -100°C). We infer that this anomalous behaviour results from a combination of the laboratory heating schedule and the distribution of diffusion domain sizes.