Variability among individual plants causes low seed yields in field pea (Pisum sativum L.) crops. To quantify this variability, an empirical principal axis model (PAM) was developed which has two components: (a ) a principal axis, representing the relationship between the seed weight (SWT) produced and plant weight (PWT) of individual plants; and (b ) an ellipse, which characterizes the scatter of individual values around the axis. To develop the model, plant-to-plant variability was simulated by systematically changing the mean and standard deviation (s.d.) of frequency distributions for SWT and PWT. Changes in the intercept and slope of the principal axis and the location and shape of the associated ellipse were used to describe the plant-to-plant variability.