A deterministic method of predicting numbers of each animal class present over time in a beef cow herd with multiple calving seasons per year is described. Breeding cow classes are defined according to the number of generations of descent from the cows in the herd prior to the introduction of a specific heifer replacement breeding policy. Cow classes are further divided into subgroups defined according to age of cow, calving interval and whether or not the cow was last mated to a terminal sire. The method is demonstrated by showing the evolution of herd breed make-up with several alternative replacement female breeding policies including grading-up, two-breed rotation and three-breed rotation breeding policies. For an example herd, the first generation of replacements peaked at approximately 0·4 of the herd at 6 years after the first matings for a new breeding system. Later generations of replacements reached equivalent proportions of the herd to the previous generation approximately 3·3 years later, with peak proportions at approximately 0·25 of the herd. Differences in the relative levels of expression of heterosis for grade up versus rotational replacement policies increased steadily from 5 years onwards with 0·17 of the heterosis expected in a herd of F1 cows remaining 20 years after the first matings of the new breeding system. The two-breed grading-up system showed less heterosis than the three-breed grading-up system only after 8 years, with the difference increasing to 0·16 of the heterosis in an F1 herd after 20 years. Results are also presented showing the difference in contributions of breeds to the herd gene pool over time with the different breeding systems. With the rotational systems, breeds lag by approximately 3 years in the timing of their contribution according to the order of their introduction to the rotation. The effects of faster heifer replacement rates on results were also discussed. It is concluded that short-to medium-term impacts of breed and replacement breeding system choices are likely to be more relevant in practical decision making by suckler herd managers than the characteristics of breeding systems once herds have reached genetic equilibrium.