A double-blind controlled Ca supplementation trial was conducted for 6 months in thirty-four 7-year-old Chinese children from Hongkong and Jiangmen, China. The children were randomly allocated to the study group (n 17) or control group (n 17), and a CaCO3 tablet (300 mg Ca) or a placebo tablet was taken daily. True fractional Ca absorption (TFCA) was evaluated before and after the trial using stable isotopes: 8 mg44Ca mixed in 100 g chocolate milk was given after an intravenous injection of 0·75 mg 42Ca. There was no significant difference in baseline TFCA between the study group (60·6 (SD 11·4)%) and the controls (58·2 (SD 9.0)% P = 0·55). Serum 25-hydroxycholecalciferol levels were comparable between the two groups (P = 0·71). After 6 months, TFCA of the study group (55·6 (SD 12·7)%) was significantly lower than that of the controls (64·3 (SD 10·7)% P = 0·015). By comparing the individual changes in TFCA after the trial between the two groups there was a non-significant reduction in TFCA (5·03 (SD 12·4)% P = 0·11, Wilcoxon signed-rank test) in the study group (60·6–55·6%), whereas a significant increase in TFCA (6·17(SD 7·7)% P = 0·004, Wilcoxon signed-rank test) was observed in the controls (58·2–64·3%). The differential in TFCA between the two groups after 6 months was significantly different (P = 0·001), and remained significant after adjustment for baseline dietary intakes, weight and height by multiple-regression analysis (P = 0·003). If the mechanism of TFCA from chocolate milk in response to the treatment effects is similar to that from the total diet, then our results suggest that children with adequate vitamin D status can adapt to a change in Ca intake by adjusting the efficiency of TFCA. In corollary, children on habitually-low Ca diets have a higher TFCA than the counterparts with higher Ca diets.