OBJECTIVE--To investigate possible discrepancies between M mode and Doppler echocardiography in assessing early diastolic filling. DESIGN--Forty seven patients with left ventricular hypertrophy due to aortic stenosis and 26 healthy controls with a similar age range were studied by M mode, Doppler, apexcardiography, and phonocardiography. The patients also underwent cardiac catheterisation. M mode echograms were digitised by a computer. Early diastolic filling in both groups as assessed by the two techniques was compared. SETTING--A tertiary cardiac referral centre with facilities for non-invasive and invasive investigations. SUBJECTS--Patients referred for assessment of aortic stenosis who had left ventricular hypertrophy. MAIN OUTCOME MEASURES--Filling velocities on Doppler and rates of wall thinning and dimension increase on M mode. RESULTS--Digitised M mode indices of diastolic filling (peak wall thinning rate 6.4 (3.0) v 10.0 (3.0) cm/s and peak rate of dimension increase 9.3 (3.3) v 16 (4.5) cm/s) in the patients and controls were consistently different. In contrast, the Doppler A/E ratio and peak E wave velocity were not; they varied widely among patients with left ventricular hypertrophy. In part, this variability was because the Doppler A/E ratio, but not the digitised M mode indices, was very sensitive to the abnormalities of isovolumic relaxation frequently present in left ventricular hypertrophy. The Doppler A/E ratio varied similarly with age in both normal and hypertrophied hearts; in the patients with ventricular hypertrophy the peak rate of dimension increase depended on age only, whereas the thinning rate was independent of age in both the patients and controls. Neither the A/E ratio nor the M mode indices could be related to the left ventricular end diastolic pressure or the peak aortic pressure difference. CONCLUSIONS--When Doppler and M mode techniques are used to assess rapid filling in patients with left ventricular hypertrophy the M mode indices are more consistently abnormal. The two methods measure different aspects of left ventricular diastolic function and should be regarded as complementary rather than interchangeable.