4 BULLETIN 99, U. S. DEPARTMENT OF AGRICULTURE. 



Sixty-Day stock. By this system of numbering, the common par- 

 entage of strains is shown at a glance. 



MAKING THE SELECTIONS. 



In both the hybrids and commercial varieties selections of indi- 

 vidual heads were made, the selections being based on size of head, 

 vigor of plant, size and color of grain, and other characters, as de- 

 sired. The seed from each head was planted in a head row, usually 

 a row 5 feet long. These rows were then numbered serially, to 

 indicate their location in the nursery, while the selection number 

 indicated the parentage of the selection. Only those rows which 

 appeared to be of value were harvested. Usually not more than 

 10 per cent of the head selections which are made are retained after 

 the first year's test in head rows. No doubt some valuable strains 

 are lost in this way, but the success of work of this kind depends 

 very largely on the judgment of the breeder. Where space and 

 funds are more or less limited it appears to be better to make a large 

 number of head or plant selections, cull these carefully in the head 

 rows, and put only those that appear to be best in the comparative- 

 yield test, going back to the original stock or drawing on new sources 

 for additional material for head rows as opportunity offers rather 

 than to select a comparatively small number of individuals and 

 retain all these in the test for several years. It is only by handling 

 large numbers of individuals and by culling very careiully that one 

 can hope to gain the desired ends in cereal breeding. 



The strains which were selected from the head rows were tested 

 first at McLean, 111., on the farm of Mr. Deane M. Funk. Later, con- 

 siderable numbers of them were tested at the Iowa and Cornell 

 University stations, and less inclusive tests were made at a number 

 of other experiment stations. The results of these tests are given 

 in the following pages. 



TESTS AT M'LEAN. 



METHOD OF TESTING. 



The tests at McLean, 111., were made in rod rows according to a 

 method devised by Mr. Norton. 1 Briefly, the strains included in the 

 comparative test at McLean were planted in rows 17 feet long and 1 

 foot apart, with every twentieth row as a check row. A pure-line 

 selection of Sixty-Day, 62-11-19, was used as a check in 1907; a 

 selection of this strain, 62-II-19-3, was used in 1908. The rows 

 were 17 feet long, so that each row occupied 17/43560 or approxi- 

 mately 1/2560 of an acre. 2 As there are 512 ounces in a bushel of 



1 Norton, J. B. Notes on breeding oats. American Breeders' Association, Report 3, pp. 280-285, 1907. 

 The method is described somewhat more in detail by II. J. Webber in Plant-breeding for farmers (New 

 York, Cornell Agricultural Experiment Station, Bulletin 251, pp. 318-319, 1908) and by C. W. AVarburton 

 in Improvement of the oat crop (U. S. Department of Agriculture, Bureau of Plant Industry, Circular 30, 

 pp. 6-8, 1909). 



» The firstfraction is equal to 0.00039027 acre, while the second is 0.00039063 acre. 



