A plotless vegetation sampling technique (ellipse‐intercept method) is described and its accuracy and precision compared by simulation models with two other line‐based techniques: line‐intercept and a point‐intercept method (which measure species covers). The ellipse‐intercept method provides three measures of plant abundance (percentage cover, density and intercept frequency) and also provides plant size distributions, yet requires no more sampling effort than the line‐intercept method. A field test of the three techniques in succulent karoo, showed that the discriminating power per metre of ellipse‐intercept survey (for all three abundance measures) is the same as that of each metre of line‐intercept survey or each four points of point‐intercept survey. The additional ellipse‐intercept information of plant densities and size distributions is shown to be useful for interpreting vegetation differences and for revealing differences (e.g. in vegetation dynamics) not detectable by measures of species covers alone.