Íslenskar landbúnaðarrannsóknir - 01.03.1970, Blaðsíða 73

COLOUR INHERITANCE IN ICELANDIC SHEEP 71
TABLE 33
Effect of colour genotype of ewes on variation in number of lambs
per ewe lambing. Two year old ewes
Source of variation DF SS MS F
1. All genotypes together
Between farm-years 114 93.0606 0.8163 6.18**
Between genotypes 9 5.8191 0.6466 4.89**
Interaction farm-years x genotypes 219 43.3309 0.1979 1.50**
Error 2140 282.8583 0.1322
2. White vs. nomuhite genotypes
Between farm-years 114 93.5326 0.8205 6.05**
Between genotypes 1 5.4537 5.4537 40.19**
Interaction farm-years X genotypes 89 17.4616 0.1962 1.45**
Error 2278 309.0931 0.1357
3. Among white genotypes
Between farm-years 96 78.6923 0.8197 6.71**
Between genotypes 3 0.0926 0.0309 0.25
Interaction farm-years X genotypes 51 9.7603 0.1914 1.57**
Error 1910 233.2010 0.1221
4. Among nonwhite genotypes
Between farm-years 107 31.9812 0.2989 1.38*
Between genotypes 5 0.4241 0.0848 0.39
Interaction farm-years X genotypes 79 15.9577 0.2020 0.94
Error 230 49.6573 0.2159
■) P < 0.05; **) P<0.01
pected to show a decrease in number oi
larnbs per ewe lanrbing it the deficiency
of grey larnbs sliown in tables 18 and 19
was due to prenatal mortality of embryos
carrying the A2-allele.
The interaction between farm-years and
genotypes is found fronr the between sub-
class sums of squares from which are sub-
tracted surns of squares for farm-years (un-
acljusted) and sums of squares for geno-
types (adjusted). This interaction is signi-
ficant in analyses 1, 2 and 3 of tables 33,
but the magnitude of the effect is low and
does not invalidate any of the conclusions
above regarding the overall effect of geno-
types on fertility. It seerns likely that the
interaction is, at least partly, a reflection
of a scale effect, occurring at low average
fertility, whicli reduces within subclass
variation when only ewes with 1 or 2 and
only very rarely 3 lamb are included in
the analysis. If that is the case the signi-
ficant interaction is a reflection of reduc-
ecl error variation rather than variation in
reaction of genotypes from one farm-year
to another. It is seen from table 33 that