This study examined the hypothesis that the force generated by myometrial strips from pregnant women is influenced by the smooth muscle content and fibre orientation of the strips and that correcting for these structural variables will provide a more accurate measure of contractility. Myometrial strips (n=72) were contracted by exposure to KCl, oxytocin, U44619 and phenylephrine and maximum responses were recorded. Morphological techniques were used to determine the cross-sectional area of the strips, the area occupied by smooth muscle and the area occupied by smooth muscle longitudinal in the strip. Maximum responses to contractile agents were expressed in terms of these three variables. The mean cross sectional area of strips was 2.01 ± 0.06 mm(2), of which 50% was smooth muscle, and 18% was smooth muscle longitudinal in the strip (n=72). There was much heterogeneity in responses, smooth muscle content and fibre orientation. Correction for morphological variability did not improve the heterogeneity in responses where coefficients of variation among strips from the same donor ranged from 43% to 63% when expressed in relation to longitudinal smooth muscle cross-sectional area. The standard method of preparation of myometrial strips for in vitro recording results in samples that are not structurally uniform. Correcting for the known structural variables does not provide a more accurate measure of maximum contractile responses. Because of the heterogeneity shown here, experiments that are dependent upon accurate estimation of maximum contractile responses require a large number of replicates to reach meaningful conclusions.