|
Results
The frequencies of alleles Gld 1B and Gld 1D in variants of
population (IP, NS and NAS) are listed in Table
1. The comparison of variants IP and NS shows the effect
of 4 years of natural selection : the frequencies of Gld 1B4
and Gld 1D3 increase, and of Gld 1B15 and Gld 1D6 decrease.
These differencies between the two populations are
statistically significant (F=10.8**,
7.3**, 9.7**, and 8.3**,
respectively). Consequently, both Gld 1B4 and Gld 1D3 are
adaptive and in contrast, Gld 1B15 and Gld 1D6 both have
some negative effects on the survival or grain production
under the conditions of experiment. The changes in
frequencies of other alleles are not statistically
significant.
It is known that selection can also operate at the level of
twolocus combinations5-7). We studied changes in
frequencies of the most widespread alleles of one Gld locus
in combinations with the different alleles of the second
locus (Table 2). The comparison of
IP and NS variants shows the strong influence of natural
selection on some two-allele combinations. For example, in
NS almost 98% of Gld 1D1 exists in genotypes either with Gld
1B4 or Gld 1B1. The combinations of Gld 1D3 with Gld 1B4 or
Gld 1B1 are also favourable. In contrast, alleles Gld 1D5
and Gld 1B1 interact unfavorably.
The comparison of variants NS and NAS shows the influence of
grain size selection. As one might expect, the effects of
this 4-year-long selection on allele frequencies were
stronger than those of environmental selection (Table
1). There are significant difference between NS and NAS
in the frequency of 4 out of 5 Gld 1D alleles studied, as
well as Gld 1B1 allele (F=4.9* for Gld 1D3 ;
10.7** for Gld 1D1 ; 7.3 for Gld 1D6 ;
31.5*** for Gld 1D10 ; 9.6** for Gld
1B1).
There are many changes in frequencies of two-allele genotype
also (Table 2). Apparently it is
possible to interpret the increase in Gld 1D3 frequency by
favouring artificial selection for combination Gld 1B1-Gld
1D3. In NAS variant, Gld 1D5 exists only with Gld 1B4 or Gld
1B1, but never with three other Gld 1B alleles studied. In
contrast to environmental selection, selection for seed size
causes a several-fold increase in the frequency of the Gld
IBI-Gld 1D5 combination, but reduces the frequency of Gld
1B1 Gld 1D1 genotypes.
Discussion
This is the first report to our knowledge in which
significant changes in Gld allele frequencies in a hybrid
population of T. aestivum after several years of
selection has been described. Different Gld alleles as well
as genotypes are not equal in their adaptive values. Natural
selection was the major force responsible for the patterns
of genetic change which occured in hybrid populations of
barley (2), and maize (3), and for the genetic structure o
wild populations of barley (6, 7) and T. dicoccoides
(8). Very likely, prolamin-coding loci as well as some
enzyme loci used in these studies do not have direct effects
but serve as good genetical markers through gene linkage.
These markers may be successfully exploited in selection
programmes (9). For example, 56% of all modern spring wheat
varieties of the USSR have Gld 1B4, and 72% have Gld 1D3
alleles (Metakovsky, unpublished), although there are more
than 10 different alleles in each of these two loci (4). It
is these alleles which were shown to be the most adaptive in
our study.
Two homoeologous loci studied was shown to be different in
their responses to the selection for grain size. One may
suggest that this result reflects to some extent the special
role of D genome in formation of bread wheat grain.
|