Tritium in the UK environment causes low radiation doses to the public, but uncertainty exists in the dose coefficient for the organically bound component of tritium (OBT). This can affect the assessment of effective doses to representative persons. Contributing to that uncertainty is poor knowledge of the body retention time of OBT and how this varies for different OBT compounds in food. This study was undertaken to measure the retention time of tritium by volunteers after eating sole from Cardiff Bay, which may contain OBT from discharges from the GE Healthcare Ltd plant. Five volunteers provided samples of excreta over periods up to 150 days after intake. The results, which are presented in raw form to allow independent analysis, suggest retention of total tritium with body half-times ranging from 4 to 11 days, with no evidence (subject to experimental noise) of a significant contribution due to retention with a longer half-time. This range covers the half-time of 10 days used by the ICRP for tritiated water. The short timescale could be due to rapid hydrolysis in body tissues of the particular form of OBT used in this study. Implications for the dose coefficient for OBT are that the use of the ICRP value of 4.2 x 10(-11) Sv Bq(-1) may be cautious in this specific situation. These observations on dose coefficients are separate from any implications of recent discussion on whether the tritium radiation weighting factor should be increased from 1 to 2.