OBJECTIVE Analysis of uterine contractility in vitro is usually confined to measuring a few traditional parameters of uterine contractility, such as contraction amplitude, frequency and area under the curve. In this paper, we describe parameters that provide additional information obtained from the traces of force and its first derivative. We propose an improved contractility index which is less dependent on variability between samples. METHODS Standard organ bath recording of myometrial contractions in the presence or absence of oxytocin on samples of human myometrium obtained from 26 patients at Caesarean section. The parameters were obtained from the plots of first derivative vs. contractile force (phase portrait plot). CONCLUSIONS Oxytocin (1nM) significantly increased the contraction amplitude (Fmax), area under the curve, maximum rate of contraction (CVmax), decreased the maximum rate of relaxation (RVmax) and had no statistically significant effect on the duration of contraction (measured as full width at half amplitude, W50). In addition to the above effects, 10nM oxytocin increased the contraction duration (P=0.0036, n=24). The fraction of force developed at the time of CVmax showed no change at any concentration of oxytocin, while the fraction of force remaining at RVmax was decreased in a dose dependent manner. The least variable (i.e. showing lowest P values in paired Student's t-Test) parameters were the Fmax and CVmax/RVmax. When non-paired t-Test was applied, P value of the CVmax/RVmax remained low, while the variability of Fmax increased reflecting the sample-to-sample variations. The product of the Fmax and CVmax/RVmax, which we propose as uterine contractility index (CI) showed low P values in both paired and non-paired t-Tests. We conclude that the phase plot analysis provides useful additional information on contraction/relaxation properties of human myometrium and the CI is suitable for characterising the contractility of uterine samples with different connective tissue content.