1. Chromium sesquioxide (Cr2O3; 1·5 g/day) and [51Cr]chromic chloride [31CrCl3 (0·3 μCi/day)] were compared as continuously administered non-absorbed markers for the correction of faecal recoveries in 14 calcium balance studies each lasting at least 18 days.
2. The mean recoveries of each, 98·4% for Cr2O3 and 101·9% for 51CrCl3, were not significantly different from 100.
3. The two markers reduced the uncertainty in a typical 3 × 6 day calcium balance study to a similar extent (sd = 1·4 mmol/day for Cr2O3 and sd = 1·5 mmol/day for 51CrCl3).
4. 5lCrCl3 is a convenient and satisfactory alternative to Cr2O3 when the laboratory hazards associated with estimating the latter cannot easily be eliminated.