Íslenskar landbúnaðarrannsóknir - 01.09.1976, Qupperneq 52

50 ÍSLENZKAR LANDBÚNAÐARRANNSÓKNIR
Table 2. Analysis of variance for autumn weight of lambs.
Sexes combined Male lambs Female lambs
Source of ----------—-------- ------------------------ ----------------------
variation DF MS DF MS DF MS
Total............................ 13176 27,94 6398 29,94 6739 21,50
Ageoflamb............................... 1 21423,68“ 1 12466,88** 1 9190,76**
Ageofdam ............................... 8 1978,26** 8 1212,78** 8 803,49**
Typeofbirth............................. 5 24514,54** 5 14603,49** 5 10136,13**
Sex .................................... 1 28398,18**
Remainder........................ 13161 15,12 6384 17,09 6725 12,94
R2 0,46 0,43 0,40
** Pg0,01
Vi = <xiijkI)!
3ijkl birth date of lamb of the 1
autumn weight, the k* sex, the
type of birth and the i
age group of dam
X j = average age of lambs
X^ = average of squared age of lambs
X0 = average birth date
o
e..., = randorm term
ykl
The restrictions
2ai=2bj=2ck
i j k
= 0, were imposed.
By leaving out regression of autumn
weight of lamb on squared age of lamb,
Model I was reduced tO' Model II and by
further leaving out regression of autumn
weight of lamb on birth date of lamb Model
II was reduced to Model III.
In addition of the above reductions, all
the analyses were also done on a within sex
basis so that the ck element of the model
was eliminated. A total of 9 different ana-
lyses were therefore carried out.
Results and discussion.
The effects of regression of autumn weight
of lamb on squared age of lamb and of
autumn weight of lamb on birth date of
lamb were both found to be significant
(P<0.05). The multiple correlation coeffi-
cient, R2, on the other hand, was found to
increase by only 0.005 by including these
regressions in the model. They were therefore
excluded in the final analysis.
The analysis of variance which resulted
from the use of Model III is shown in tablc 2.
The table shows that a relatively large
proportion of the total variation in autumn
weight of lambs can be explained by the
systematic environmental effects which are
included in Model III. For the material as a
whole, R2 is thus 0.46, which is higher than
what has been found in comparable Nor-