SUMMARY

Weights of beef cows between 19 and 84 months of age, recorded on a monthly basis during a selection experiment, were analysed using random regression models. Only phenotypic animal effect were considered.

Curves fitted included orthogonal (Legendre) polynomials of age, fitted up to an order of 22, segmented quadratic polynomials and Fourier series approximations. There were strong seasonal effects on weights of cows, and due to short calving seasons, a high degree of confounding between age and month of recording.

Data exhibited cyclic seasonal fluctuations in variances which could only partially be attributed to scale effects. Random regression models proved capable of modelling such pattern of variation adequately, but required high orders of fit and thus a large number of parameters.

Keywords: Random regression, growth curve, mature weights, beef cattle.

Download paper (Word document, 4 pages)