Íslenskar landbúnaðarrannsóknir - 01.03.1970, Síða 56

54 ÍSLENZKAR LANDBUNAÐARRANNSÓKNIR
ed to brown ewes, none of them brown-
pigmented. He could therefore be regard-
ed as being homozvgous for the gene for
black pigment. When mated to ewes with
white markings, he produced altogether 15
nonwhite lambs, 7 of them without white
markings and 8 with white markings,
which is in agreement with exceptation if
he was heterozygous with respect to white
markings. The zygotic genotype of the ram
has therefore been A4AgB^B^S^S^.
In connection with the colour descript-
ion it should be mentioned that the ram
was described as being of a lightgrey col-
our, similar to that of ordinary honio-
zygous grey rams, but that some of the
features of the mouflon pattern coukl be
observed on closer inspection. He had also
a darker colour on one side of the head.
The author had never the opportunity to
examine the ram, and the colour descript-
ion above was given by Mr. Gudmundur
Jónsson, Kópsvatni, who informed us about
the existence of the ram and who was in
charge of the experiment carried out in
the second year with this ram.
CHAPTER VI
Analysis of observed segregation at the A-locus
As mentioned previously, 441 different
mating types are possible with respect to
locus A, if all genotypes of sires are mated
to all genotypes of dams.
In table 9 are shown the colours of the
progeny resulting from 111 realized matings
with respect to this locus.
As table 9 shows, the amount of informa-
tion available from individual matings
varies tremendously. d’here are thus avail-
able colour descriptions of 390 lambs re-
sulting from the mating A4A5 X AiAg,
while in 65 ol the 111 matings in table 9
less than 20 progeny have been observed,
and in 50 matings 5 progeny or less have
been observed.
Tests of agreement between observed
and expected frequencies of colours in
the matings reported in table 9 are
complicated for two main reasons. The
first of these is the large number of tests
that can be carried out. One can test the
goodness of fit of the observed frequenc-
ies from individual matings and then add
up the results from related matings and
test the overall frequencies and at the same
time the discrepancy between the matings
on which the overall frequencies are based.
The grouping of matings can be done in
so many different ways and so manv tests
carried out that even by chance several
apparently significant deviations are bound
to be observed. As the chance deviations
can not be distinguished frorn genuinely
significant deviations any attempted inter-
pretation of such tests will be highly sub-
jective.
The other difficulty lies in the limited
number of observations behind several of
the matings in table 9, making the tests
insensitive even to large deviations. Cor-
rection for continuity has not been used
in the present study, as this will on the
average give too conservative tests
(Grizzle, 1967). Any grouping of matings
in order to improve the test must necess-
arily be subjective.
From the above considerations it was
clear that a test was needed which utilized
all the available information on segrega-