Electro- and vectorcardiographic methods describe left ventricle increases only when it has attained a significantly high magnitude, but even in such circumstances the exactitude of such methods is substantially lower than that of echocardiogram (ECHO). On the other hand, in comparison to other age groups, there have been relatively few reports relating electrocardiogram (ECG), vectorcardiogram (VCG) and ECHO with left ventricular mass (LVM) in healthy elderly subjects where increases of the left ventricle mass, if present, would be small or moderate. In this paper LVM as well as LVM index (iLVM) from a group of healthy subjects belonging to a physical training program for elderly, was studied by means of ECHO and computerized ECG and VCG. From ECG, voltage indexes and other LVM associated parameters were extracted; from VCG, planar maximum vectors, areas within VCG loops and maximal spatial magnitude of QRS (SM), were measured. Results of LVM (221 +/- 37.9, g) were higher than figures reported for others groups. Voltage indexes showed normal values, but QRS duration was somewhat prolonged. The best simple linear regression, combining variables from VCG and ECG was maximum horizontal vector (Vmax-Hor) vs Sokolow-Lyon index (SOK) and combining ECHO with ECG or VCG, LVM vs Area inside horizontal loop (AreaHor). A model for estimation of LVM from electrical variables was obtained by multiple linear regression; combining five variables from ECG and VCG. The best model included Sokolow-Lyon index and variables from horizontal and sagittal planes of VCG and spatial magnitude of QRS: LVM = 4.8 SOK-186 VmaxHOR-80 VmaxSAG + 126SM + 340 AreaH + 175.(ABSTRACT TRUNCATED AT 250 WORDS)