350 MUTATION AND PLANT BREEDING 



The above-mentioned examples are taken from the work of my 

 Svalof colleague Hagberg (46). Since duplication followed by differ- 

 entiation must be considered as one of the most constructive phases 

 in evolution, works of this kind may soon drastically change our ideas 

 of induced mutation in plant breeding. The principle of producing 

 well-defined duplications will have to go via the induction of trans- 

 locations involving the same two chromosomes but with different 

 positions of the breakage points. The combination of two such trans- 

 locations will imply a duplication of the segment between the breaks. 

 For the proper use of this technique in plant breeding, a large set of 

 well-defined translocation lines and a detailed gene map are essential. 



Induced chromosomal rearrangements may be valuable in many 

 situations. They may be useful in splitting up gene blocks and char- 

 acter associations which had an evolutionary advantage in the wild 

 plant but are undesirable from the demand of man. In the opposite 

 way, valuable gene blocks may be inverted to prevent crossing-over 

 or moved inside localized chiasmata. A fascinating way of using 

 radiation experiments for incorporation of small segments from non- 

 pairing chromosomes is well demonstrated by Sears' (9.°)) already 

 classical transfer of leaf-rust resistance from Aegilojjs umbellulata to 

 common wheat. A similar approach is successfully accomplished by 

 Elliott (19, 20) in transferring stem rust resistance from Agropyron 

 elongatum to wheat, and the attempts to transfer bunt resistance 

 from the same donor species are progressing (60, 63). Also, in Europe, 

 this technique is now applied (unpublished by McKelvie, and 

 Wienhues-Ohlendorf). 



The great merits of induced chromosome mutation are definitely 

 in a well-defined chromosome engineering work. Repeatedly, how- 

 ever, suooestions have been put forward to use recurrent radiation 

 experiments for the diploidization of artificial polyploids, where 

 differentiation of identical or too closely related chromosomes was 

 thought to improve disomic pairing (26, 48, 71). No definite report of 

 success along these lines is yet available, if we except the observation 

 by Vettel (115) in Triticale that morphological X-mutants with long, 

 dense heads had an improved seed setting. The Svalof work along this 

 line, including 4x barley, flax, and rye, does not indicate any succes- 

 sive improvement with recurrent radiation, but there is some hope 

 that suddenly types with improved fertility may appear. In the light 



