Íslenskar landbúnaðarrannsóknir - 01.03.1970, Side 62

60 ÍSLENZKAR LANDBUNAÐARRANNSÓKNIR
TABLE 19
Test of significance of effect of allele parameters on segregation ratios
Line Parameters estimated Due to parameters Remainder
DF X2 Prob. DF X2 l’rob.
1. Reciprocal matings separate
1 None 140 125.348 0.50 > P > 0.30
2 ai, a2, a3, a^ 8 7.859 0.50 > P > 0.30 132 117.490 0.50 > P > 0.30
bi, b^, b3, b4
3 a2, b2 2 5.959 0.10 > P > 0.05 138 119.390 0.50 > P > 0.30
2. Rt ciprocal u •latings added
4 None 98 76.141 0.95 > P > 0.90
5 Cl, C2, C3, C4 4 5.961 0.30 > P > 0.20 94 70.180 0.95 > P > 0.90
6 C‘2 1 5.220 0.05 > P > 0.02 97 70.921 0.98 > P > 0.95
nearly significant amount of the variation.
The combined parameters, c^, c2, , c4,
when all are estimated together, show a
nonsignificant reduction in the variation.
When c2 only is estimated it accounts for
a significant proportion of the variation.
From the parameter x2’s in table 19 one
can calculate some differences of interest
between x2’s-
It is seen from table 20 that none of
the x2’s there are significant. This means
that the only significant differences among
the parameters are those between c2 on one
hand and c1; c3 and c4 on the other hand,
and c2 is the only parameter that is signi-
ficantly different from 1.0.
The question whether the observed signi-
ficance of the deviation of c2 from 1.0 is
real or not must be dealt with critically.
It is conceivable that significant deviations
of this kind could arise merely by chance
when a large number of tests is carried
out on the same material.
On inspection of the individual matings
in table 9 it can be seen that the de-
ficiency is very marked in the matings
A2A3 X A5A5 (x2i = 4.372; P < 0.05) and
A5A5 X A2A5 (x21 = 3.333; 0.10>P>0.05),
the sum of the reciprocal matings giving
X2i= 7.672 (P<0.01). This is the only
significant x2 in combined reciprocal mat-
ings with 20 progeny or more.
The parameters a2, b2 and c2 were there-
fore re-estimated excluding these two mat-
ings. This led to the values given in table
21.
The estimates in table 21 are still below
the expected value of 1.0, but not signi-
ficantly so. The proportion of the varia-
tion accounted for by the new estimates is
shown in table 22.
The proportion of the total x2 account-
ed for by the parameter x2’s in table 22 is
also nonsignificant. This might lead one
to think that the deviations in a2, b2 and
c2 found in table 18 were due to aberrant
ratios which were specific to the A2A5 X
A5A3 and A5A5 X A2A3 matings. Further
inspection of table 9, however, shows that
the mating A2A5 X A2A5 also gives a con-
siderable excess of nongrey progeny (x2i =
2.333; 0.20 > P > 0.10). Although this de-