Multipoint interval mapping (MIM) and the MAPMAKER/SIBS program (M/S) are two methods of mapping quantitative loci by examining identity by descent (IBD) sharing in a region spanned by multiple microsatellite DNA markers. For the purpose of comparison, we simulated a quantitative trait controlled by a two-locus model, and evaluated the power and genome-wide false positive rate of both approaches. Based on our simulation, we examined the effects of marker density (5 cM, 10 cM and 20 cM) and sibship size (2, 3, 4 and 5) on the power to detect linkage. Our results indicate that a 10 cM map provides the optimal trade-off between power and type I error, and that the power of MIM increases with sibship size and, in general, performs better than MAPMAKER/SIBS. Furthermore, we conclude that using a reasonable sample of randomly ascertained sibships, it is possible to map a quantitative trait locus (QTL) which accounts for 25% of the phenotypic variance.