630 NORMAN E. BORLAUG 



MUTATION BREEDING 



S. Kedharnath has given an excellent summary of the present status 

 of mutation breeding. I again caution you, just as I did 4 years ago, 

 not to look for magic in this sophisticated approach to plant breeding. 

 Where there are unexploited known genes for resistance present in nature, 

 as is the case with C. ribicola or Pissodes strobi , why dissipate your 

 limited budgets on such sophisticated approaches until these genes have 

 been utilized. Although there have been hundreds of scientific articles 

 published on the anticipated contributions of mutation breeding to the 

 improvement of crop plants during the past 20 years, the accomplishments 

 have at best been very modest if not insignificant . Most of these research 

 efforts have produced nothing worthwhile. It is my contention that had 

 the amount of money that has been spent on mutation breeding been spent 

 instead on conventional breeding, much more would have been 

 accomplished. 



CONCURRENT IMPROVEMENT OF DISEASE RESISTANCE AND OTHER CHARACTERS 



Ernst Schreiner (1969) has forcefully stated the need for concurrent 

 improvement of growth rate, product quality, and disease and insect 

 resistance. I concur that this is not only desirable but feasible if 

 the breeding program is properly organized. To achieve this objective 

 it is absolutely necessary to develop a broadly based gene pool, and grow 

 and study large populations. The development of a diversified gene pool 

 must begin with the selection of a large number of superior parent trees 

 of the variety or species which is to be improved. All too often the 

 number of parent trees is too small, and consequently the genetic base 

 is too narrow to permit multiple character improvement. The selection 

 of these parent trees should be done not only on the basis of disease 

 and insect resistance but should also take into consideration rapid rate 

 of growth, good tree form, and a wide breadth of adaptation. Similarly 

 when interspecific crosses are to be made to transfer disease resistance, 

 i.e., rust resistance, the "donor" parent trees should be selected not 

 only for their outstanding resistance but also, insofar as possible, for 

 a good combination of other desirable silvi cultural characteristics. 



Although provenance tests of many of the most important forest tree 

 species have clearly established the existence of ecotypic races that 

 differ widely in silvicultural characteristics, the genetic implications 

 of these differences are all too often ignored when breeding programs 

 are being organized. Currently in most forest tree breeding programs the 

 genetic base on which the program is being built is extremely narrow, 

 much more so, for example, than in maize breeding programs. All too 

 often breeding programs have been based upon a few dozen or at the most 

 a hundred parent trees. Is this an adequate sample of the genetic varia- 

 tion that occurs in a species that may cover a range varying from 500 to 

 1,000 miles, and may include tremendous variations in elevation, latitude, 

 and sites? 



The situation is even more unrealistic in the choice of parent trees 

 of exotic donor species. Frequently a very few trees, one to several, 

 nerally growing in an arboretum have been used as being representative 

 F a species. Often such donor species occupy vast forest areas in 

 remote parts of the world. Are these parent trees adequate samples upon 

 which to build viable, long-range breeding programs? Regardless of the 

 accuracy and sophistication of the population genetics that are employed 



