We investigated the repeatability of two recently reported colour M-mode Doppler measures of left ventricular diastolic function. 'Delay of peak velocity' is the time interval from peak velocity of early filling at the mitral tip level to peak velocity in the apical region. The 'velocity of flow propagation' is the speed of the mitral to apical spread of the early diastolic filling. Two examiners independently and blindly made measurements by both methods on 32 digital recordings twice, using custom software. Inter- and intra-observer repeatability were given in terms of limits of agreement, where 1 denotes perfect agreement. Limits of intra-observer agreement were 0.72-1.43 for measurements of 'delay of peak velocity', and 0.39-2.28 for 'velocity of flow propagation'. Limits of inter-observer agreement were 0.68-1.42 and 0.19-2.04 respectively. The correlation between the methods was weak (r = 0.45), suggesting that they may detect different properties of early diastole. We suggest that the reliability of the methods may be improved by replicate measurements performed blindly. There are reasons to express the velocity of flow propagation on a logarithmic scale.