Íslenskar landbúnaðarrannsóknir - 01.03.1977, Page 24

22 ÍSLENZKAR LANDBÚNAÐARRANNSÓKNIR
Table 8 . The difference in weaning weight between twins in sets of
like and unlike sexes.
Difference in Difference in
Harmonic uncorrected corrected
number weaning weight weaning weight
R(R)-R(E) 4637 -0,348 ± 0,075 -0,323 ± 0,059
E(R)-E(E) 4829 -0,100 ± 0,073 -0,170 ± 0,057
Table 9 . ReRression coefficients on age of lamb (days).
Trait Regression coefficient ± S.E.
Weaning weight 0,197 ± 0,003
Carcass weight 0,083 ± 0,001
Dressing percentage 0,021 ± 0,002
AGE OF LAMB
The linear regression coefficients estimated
for the effect of the age of lamb on weaning
weight, carcass weight and dressing percen-
tage are presented in table 9. The regression
coefficient for weaning weight is higher than
that found by Jónmundsson (1971), but
in his smdy the age of lamb at weighing was
used instead of age on 1 October as in this
study.
By assuming a birth weight of 4 kg the
calculated weight gain from birth to weaning
is 255 g/day, so that the weight gain is
falling in the autumn.
The lambs are normally weighed just be-
fore they are sent to the slaughterhouses and
in each flock this often takes place twice
in the autumn with a time lag of 2—4 weeks.
The first group is often taken directly from
the mountain pastures and is most'ly ram
lambs, but the rest of the lambs are grazed
on cultivated pasture. For 32780 lambs that
have recorded days of weighing one used
acmal age at weighing instead of age in
1. October in the model. The actual age at
weighing has greater variation than age on
1. October (standard deviation 8.97 days
versus 5.74 days). Then a significant
(P<0.01) curvilinear affect of the age on
weight is obtained. The model that used
actual age accounted for 2.1% less of the
variation in weaning weight than the model
that used age on 1. October.
The age effect was found to be quite