346 MUTATION AND PLANT BREEDING 



phenotypic penetration, or the aim of the mutation experiment is 

 indeterminate, the advantage of screening groups instead of single 

 plants will often be decisive. For that reason, more and more muta- 

 tion experts recommend selection first in M 3 (24, 25, 26, 29, 49, 50, 

 103), a trend definitely stimulated by the promising experiments on 

 induced mutations in quantitative traits. 



The rapid increase in size of population from Mi to M ;J , can 

 partly be eliminated by proper sampling. Since mutation in a uniform 

 parental material is a chance event but seeds traced back to one and 

 the same germ line more or less interrelated, the sampling should 

 be based on as many independent inflorescence units and as few seeds 

 from each such unit as possible (24, 25, 56). The more seeds taken 

 per independent unit, the lower relative frequency of mutant indi- 

 viduals will appear in the next generation, but the higher will be 

 the chance to detect all mutations induced in the given material. The 

 last statement, however, has a limit set by the segregation ratio 

 between normal and mutant types. In barley, for example, the aver- 

 age chlorophyll mutant frequency per segregating X a head progeny 

 has been found to vary between 13 to 18 per cent, depending on 

 tillering (28, 68). The corresponding figure for X 2 is about 20 per 

 cent (78). Within the 99 per cent probability limit, therefore, not 

 more than about 30 vital seeds per unit should be tested. The simple 

 system of sampling by means of dense sowing under unfavourable 

 conditions, resulting in few seeds per plant, can be used also in M 2 , 

 but this has to be weighed against the risk of interplant competition 

 (107). Negative or positive mass selection in M 2 is another method to 

 keep down size of population. Natural selection under more or less 

 extreme conditions as to overwintering, disease infection, day length, 

 vegetation period, etc., can also be of great help. Highly sterile plants 

 may be discarded if only point mutations are desired. Selection on the 

 basis of morphological properties may generally be effective for physi- 

 ological features as well as due to the high chance for character 

 association. 



Unless the goal of the breeding project is not very determinate, 

 one of the greatest difficulties with practical mutation experiments is 

 to decide which mutants should be selected and incorporated in 

 future work. Induction of mutation as an independent method of 

 plant breeding is only restricted to cases where the plant is vegetative- 



