|
||
PNL Volume 21 1989 RESEARCH REPORTS |
||
|
||
RELATIONSHIP BETWEEN INDUCTION
OF MACRO- AND MICRO-MUTATIONS FOLLOWING MUTAGENIC
TREATMENT
Singh, N.P. and B. Sharma
Division of Genetics
Indian Agric. Res. Institute, New Delhi, India
Traditionally, it has been a common
practice in mutation breeding experiments to advance only normal looking
plants from M2 progenies (nonsegregating for macromutations) to M3
generation. In contrast to the earlier experiments, selection in the
present study also has been applied to the families segregating for
macromutations.
The entire M3 material was divided
into three populations: (T) progenies of three normal looking plants from
each M2 family segregating for a macromutation but no longer carrying the
visible mutation (non-segregating in M3); (II) progenies of the
three best plants in the M2 families identified as promising on the basis
of character means and CV; and (III) progenies of three random plants from
each M2 family not included in the above two populations. Observations
were recorded in these three M3 populations, called macromutational (I),
selected (II), and unselected (III), on five important polygenic
characters, namely, days to flowering, pods/plant, seeds/pod, seed size
(1OO-seed weight) and yield/plant. Efforts were made to maintain an equal
population size in all three groups of populations for uniformity in
comparison.
The comparison of different
populations in M3 (Table 1) for the five polygenic traits studied showed
that genetic variability (CV%) decreased in the selected material as
compared with the unselected and macromutational populations. As expected,
selection in M2 shifted the mean in the desired direction (lower in case
of days to flowering, higher in the remaining four traits), but reduced
variability (CV%). The unselected lots, on the other hand, carried the
entire variability induced in positive as well as negative directions,
which was further magnified in the M3 generation. Most interesting,
however, is the fact that the macromutational group showed an even higher
CV for all the five characters than the unselected material derived from
the nonsegregating M2 families. Therefore, in contrast to the past
recommendations that the macromutational progenies should be excluded from
any analysis for polygenic variability, the present study strongly
suggests that the chances of success for isolation of promising families,
even for polygenic variants, are maximized in the macromutational
populations.
Although selection for polygenic
traits among the normal looking plants has a definite advantage, many such
families may not carry any mutation at all. On the other hand, there are
progenies which have already segregated for macromutations in the M2
generation, in which the mutagen has definitely affected the genetic
material at one or more loci. The chances of the so-called "minor" genes
also being hit in such cases is expected to be higher than in the material
where there is no evidence of genetic damage due to mutagenic treatment.
Therefore, selection of normal: looking plants from the macromutational
progenies would be very effective, as maximum variability was found to
exist in such populations. Concentrated analysis of such progenies can
lead to a great economy in experimentation with simultaneous improvement
in selection efficiency. One more cycle of selection on the basis of
progeny mean and variance in M3 would confirm the potential of these
selections. Such close watch on the |
||
|
||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
56
PNL Volume 21
1989 RESEARCH REPORTS |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
experimental material and rigorous
selection can lead to the isolation of confirmed promising strains in M3
which can be put to initial testing in the form of small plot trials in
the M4 generation. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Table 1. Effect of M2 selection
on induced polygenic variability in M3 generation for different
characters. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||