1. Lithium has been used by Sanchez-Castillo et al. (Clin. Sci. 1987; 72 , 81–6) for tracking sodium in order to monitor the domestic use of salt. The present study was designed to simplify the original protocol, which is too cumbersome for epidemiological studies. 2. A preliminary study conducted on nine volunteers showed that sharp modifications of lithium intake are detected within 24 h in urinary excretion. The average intake over a 5-day period could be predicted by the mean excretion of the last 3 days. Lithium recovery established from 1 baseline day and 3 plateau days was 95 ± 6% (mean ± sd), not significantly different from the recovery obtained by the integrated excretion curve over 12 days. 3. A simplified protocol was therefore developed, reducing to 4 the 12 days of urine collection originally required. 4. This simplified protocol was tested on five households (14 adults and five children). All domestic salt was substituted with lithium-tagged salt. Urine samples were collected on 1 baseline day and on 3 plateau days. The lithium/sodium ratio of all urine collections (both complete and incomplete) could be used to estimate the proportion of discretionary sodium. In our sample it was about 29% of total sodium intake. 5. The simplified protocol can thus be proposed for population studies.