A new method for the semiobjective evaluation of left ventricular wall contraction is proposed that uses a percent reduction of areas. The calculation model uses orthogonal coordinates, an internal reference system, and the concept of nonuniform reduction of the cavity, with its obliteration at the apex. From a normal patient population (ten males, four females; aged 45 +/- 13 years), a "normal reference curve" of contraction was obtained. A second patient group (56 males, two females, aged 56 +/- 7 years) with definite contraction abnormalities was then analyzed. A third group (eight males, one female, aged 56 +/- 10 years) was studied before and after nifedipine administration (1 mg over 2 min IV). Specific algorithms compare individual patient contraction curves with the normal reference curve and compute the score point system of contraction analysis. Data are automatically codified and filed during introduction. This saves time and allows a rapid retrieval of data grouped under different coded features. These features include the location and type of wall contraction abnormalities. This method of reading ventriculograms has been shown to be in accordance with the diagnosis of experienced observers and can detect small contraction changes that are induced pharmacologically. The method is proposed as a valuable tool for semiobjective evaluation of left ventricle wall contraction and to facilitate data collection in multicenter collaborative studies.