The high production rate of in-situ cosmogenic Cl-36 from potassium allows K-rich minerals to be dated with high sensitivity and precision, K-rich minerals, such as K-feldspar and biotite are commonly associated with quartz in which Al-26 and Be-10 are produced, allowing three cosmogenic nuclides, with half-lives ranging from 0.3 Ma to 1.5 Ma, to be measured in a single sample. To calibrate the production rate of Cl-36 from potassium We have measured K-feldspar samples from ice-scoured bedrock from three separate sites, covering a range of attitudes, latitudes and exposure ages, The results from each site are internally consistent, with eight glacial pavements from the Sierra Nevada giving a production rate of 1290 +/- 220 atom/(gK)/yr at 3000 m and 38 degrees N, three Scottish glacial pavement samples giving 301 +/- 27 atom/(gK)/yr at 520 m and 58 degrees N, and two Antarctic bedrock samples giving 1310 +/- 50 atom/(gK)/yr at 2000 m and 70 degrees S. Normalising these production rates to sea level and high latitude, results from the Sierra Nevada and Scotland are found to be in good agreement with each other, but the production rate derived from the Antarctic samples is found to be approximately 35% higher. Repeat measurements and stepwise dissolution indicate that the high Cl-36 concentration of the Antarctic samples is real and cannot be attributed to meteoric Cl-36. We must therefore conclude that either the Cl-36 production rate differ on the 10(4) and 10(6) yr time scales, or that the current altitude scaling factor applied to the Antarctic results underestimates the true scaling factor for the Antarctic atmosphere, The production rates derived above are based principally on K-feIdspars with low chloride contents. Samples with high chloride contents required corrections for Cl-36 produced by neutron capture on Cl-35. The Cl-36/Cl ratio of this component was measured directly in four samples from the Sierra Nevada after crushing the K-feldspar to release Cl-rich fluid from inclusions. The Cl-36/Cl ratios measured appear to be in reasonable agreement with values predicted by recent calibrations of the production by neutron capture on Cl-35, but the final comparison awaits measurements of uranium, thorium and neutron absorber (B, Gd, Sm) concentrations in our samples.
