EARLY TESTING AND RECURRENT SELECTION 411 



ing lines would have resulted in a very great saving of time and nionejy- us 

 compared with testing at a more advanced stage of inbreeding. 



RECURRENT SELECTION 



Superficially recurrent selection has a considerable resemblance to the 

 ear-to-row method of breeding. However recurrent selection differs in several 

 important respects. It provides for a much more accurate genetic control, and 

 the plot technic can be modified to give any desired degree of accuracy. Our 

 use of the recurrent selection technic was a direct outgrowth of the work on 

 early testing. It appeared logical to assume that if the individual So plants 

 selected on the basis of test cross performance were a superior group, inter- 

 crosses among this group to provide source material for a new cycle of selec- 

 tion would minimize certain of the limitations arising from continued 

 selfing. Accordingly a group of the best lines from the early testing series 

 were intercrossed to provide material for the evaluation of this method. 



Somewhat earlier, studies were started to compare the relative efficiency 

 of recurrent selection and inbreeding in isolating material having a high oil 

 percentage. At the time the work was started we were of the opinion that 

 this was a new idea. It was some time later that we discovered that essential- 

 ly the same ideas had been published independently by East and Jones (1920) 

 and by Hayes and Garber (1919). In neither of these cases was any extensive 

 use made of the method and no critical data were published. The first de- 

 tailed description of recurrent selection was made by Jenkins (1940). The 

 breeding procedure did not receive a name however until Hull (1945) pub- 

 lished his article dealing with recurrent selection for specific combining abil- 

 ity. 



Because of the shorter time period required per cycle we have much more 

 information on recurrent selection as a method for modifying chemical 

 composition than we have for the modification of combining ability (Sprague 

 and Brimhall, 1949). We shall report in some detail only one study — that 

 contrasting recurrent selection and inbreeding in modifying oil percentage in 

 corn. The source material for this study was obtained from Si ears from re- 

 ciprocal backcrosses involving the single cross wxOs420Xlll. High Oil. Indi- 

 vidual plants were self-pollinated in each backcross population and analyzed 

 individually for oil percentage in the grain. The five ears having the highest 

 oil percentage in each population w^ere planted ear-to-row the following 

 season and all possible intercrosses made among the ten progenies. Equal 

 quantities of seed from each cross were bulked and used as source material 

 for a new cycle of selfing, analyzing, and intercrossing. 



A duplicate planting of the ten ears mentioned above was made in 25 

 plant, ear-row progenies. The phenotypically most desirable plants in each 

 progeny were self-pollinated. At harvest time approximately five ears were 

 saved and analyzed individually for oil content of the grain. The two ears 



