Abstract :
Aquadratic polynomial model withtwo randomregression coefficientswas usedto describe
the growth pattern of 130 single male lambs from the Churra Tensina breed which were
raised under four different management systems. Treatments were: lambs in drylot with
dams indoors (DR-IND), lambs in drylot with rationed grazing ewes (DR), lambs supplemented
in creep feeders continuously with dams on pasture (GR + S) and lambs with dams
continuously on pasture (GR). Lambs from DR-IND and DR were weaned at 49.4±5.0 days
old. All lambs were weighed at birth and thereafter at weekly intervals until slaughter at
the target live-weight of 22–24 kg (23.3±1.5 kg).
The model fitted accuratelyweights (R2 = 0.96). The effect of treatment on the linear coefficient
was not significant (average 267 g/day; P > 0.05). However, the quadratic coefficient
varied across management systems (0.027±0.12, 0.52±0.17 and 0.22±0.12 g/day2 in DRIND,
DR and GR+ S, respectively, vs. −0.26±0.12 g/day2 in GR; P < 0.001). Although standard
errors increased at higher ages, the residual variancewas 43% reduced when using two random
regression coefficients. Around 60% of the variation in the data set was accounted for
by allowing the intercept and the slope to vary across individuals.
It was concluded that the model was suitable to describe a nearly linear growth pattern
in light lambs and accounted for mean live-weight adjustment within lots resulting from
removal of lambs that achieved the slaughter live-weight earlier.
