Íslenskar landbúnaðarrannsóknir - 01.03.1970, Qupperneq 85

COLOUR INHERITANCE IN ICELANDIC SHEEP 83
entiation (Finney, 1968, personal com-
munication).
In the estimation procedure, u ancl v
were first estimated as separate parameters
for singles and twins, while s was estimat-
ed as a common parameter for both singles
and twins. In the first estimation 5 para-
meters were thus involved, ug, uT, vg, vT
and s, wliere the subscripts refer to singles
and twins, respectively. The first estima-
tion gave very low and non-significant
values for us and vg, so in a second analysis
they were both assumed to be 0, and only
uT, vT and s were estimated in tlie second
analysis.
rFhe results from the estimation are
shown in table 44, and the reduction in
X2-vaues due to parameters in table 45.
The results in tables 44 and 45 indicate
strongly that selective fertilization and
selective mortality operate actively in mat-
ings involving genotypes with combina-
tions of alleles Ax and Ag. The significant
parameters obtained from these matings
are also in agreement with deviations from
expectations described earlier. The value
obtained for s in matings from table 39 B
is in agreement with the lowered fertility
of white ewes as shown in tables 32—37.
The value for vT from the same matings is
in agreement with the discrepancies found
between males and females with respect to
the proportion of white lambs (table 28),
and tlie value for uT, vT and s explain
most of the excess variation found in table
39 B. Tlie presence of selective forces of
the type described here, could give rise to
decreased variation in segregation ratios as
found in the AXA5 X AXA5 niatings.
The diíference between singles and twins
in the matings in table 39 B with respect
to u and v is of particular interest. This
differences suggests that selective fertiliza-
tion due to differential activity of sperm
carrying alleles Ax and Ag respectivelv is
mainly active when two ova are shed. One
might assume that variation in ovulation
time is greater in ewes shedding two eggs
than in ewes shedding only one egg. If
this was so the ewes shedding two eggs
sometimes shed an egg very early and sorne-
times very late relative to time of mating,
compared to ewes shedding only one egg.
If the different types of sperrn are assumed
to have different peaks of activity in rela-
tion to tirne of mating, those with an early
peak would fertilize early shed eggs more
frequently and those with a late peak
would fertilize late shed eggs more frequ-
ently. Increased variation in ovulation time
would therefore be expected to accentuate
any differences in peak activity of the
different types of sperm. It should also be
borne in mind that the matings in the
present study usually took place only once
during the heat. This would be expected
to increase the chances of detection of
clifferent peak activity of different sperm
types, whereas the peaks would overlap if
several matings occurred during the same
heat.
It is not clearly established what the
relation is between variation in ovulation
time and the number of eggs shed. Donald
and Read (1967) found, however, that the
length of the heat period in the highly
prolific Finnish Landrace ewes was con-
siderably longer than that reported for
other less prolific breeds (Robinson, 1959),
and Land (1968, personal communication)
found for the same breed sorne indication
of a positive relationship between length
of heat and number of lambs born. The
finding in the present study that u and v
are significant in twins and not in singles
may therefore well have a meaningful bio-
logical explanation.
It should finally be pointed out that the
effect of parameter v is assumed only to be
found in white ewes. This makes it necess-
ary to postulate a difference between white
and nonwhite ewes with respect to either