Evidence of a large sex-linked effect accounting for 25% of the divergence between mouse lines selected for body weight has been described previously. A marker-based study was undertaken to determine the number and map positions of the putative X-linked quantitative trait loci (QTLs). An F2 population was generated from a reciprocal F1 between an inbred low line derived from the low selection line and the high selection line. To enable inference of marker-associated QTL effects on the X chromosome, an analytical technique was developed based on the multiple regression method of Haley and Knott. The analysis of data on 10 week weight indicated a single QTL of large effect situated at about 23 cM from the proximal end of the chromosome, with a peak LOD score of 24·4. The likelihood curve showed a single well-defined peak, and gave a 95% confidence interval for the QTL location of 8 cM. The estimates for the additive genotypic effects in males and females (half the differences between hemizygous males and between homozygous females) were 2·6 g in both cases, or 17% and 20% of the 10 week body weight in males and females respectively. Dominance effects in the females were found to be non-significant. No significant X-linked effect on carcass fat percentage was detected, but a single X-linked QTL appears to explain almost the entire X-linked body weight effect.