RXPERIMENT StATIOX HEpORTS. S03 



t)lanl> call l»c sown s'ldv h\ .sitlc. niaiiv of them arc easilv eliminated on 

 observation. They have weak straw, or various characteristics indicating 

 that they would produce unpractical commercial varieties if increased. 

 The improvement of an old variety rests upon the isolation of its various 

 strains and the increase of the })est. 



The vakie of the plant row series, as a means of determining yield 

 is varioush' estimated by the different crops breeders in the various 

 states. "Whether the plant row series is accurate as a yield series or not, one 

 thing is certain: any one who has had one of these series planted with 

 frec^uent checks for comparison would have no doubt in his mind regarding 

 the inferiority of certain strains. He would discard them on sight. The 

 attitude that has been taken in this work is that the strains that appear 

 to ])e equal or superior to the check are the only ones worthy of con- 

 tinuing. Others are easih' eliminated in the 1-100 acre series that usu- 

 ally follows the plant-row test. By this means only promising strains 

 enter the variety series. 



PURE LIXE WORK. 



Beginning wdth 1906, a new method of breeding small grains has been 

 in use. This plan consists of making large numbers of plant selections, 

 from as many commercial varieties as were available, and to compare the 

 selections with each other in a series of individual plats. The plant-row 

 series with frequent checks is the most satisfactory form of the series. 



Selections were made from five new sources in 1906, seventeen new 

 sources in 1907, another seventeen sources in 1908, two new sources in 

 1909, and forty-eight new sources in 1910. Thus new sources of seed, 

 coming from farmers' seedsmen, experiment stations and the U. S. De- 

 partment of Agriculture (the latter collected from all oat growing coun- 

 tries from around the world) have been added to the list until now (1919), 

 there are 150 numbers. Xewh' selectioned plants are added to the list 

 annually, and tested out in plant-row series. 



This has resulted in the elimination of manj- original sources of seed.^ 

 Others were represented by just one pure line apiece when thej' enter the 

 variety series. 



YIELD TESTING. 



Twenty years ago the method of testing the yielding al)ility of varieties 

 was very crude as we view it today. Usuallj- the experimenter depended 

 upon one plat to the variety, and he may, or may not, have had checks in 

 the series. Perhaps he had a check on either side of the field and one in 

 the center. These plats were so nearly square that it was not possible 

 to duplicate results. The variation in soil between plats was usually 

 greater than the variation among the yielding abilities of the varieties 

 under test. Such work is of comparatively little value under ^Michigan 

 conditions. 



3 The Big 4 Oats was discarded in 190S after two years testing. Big 4 Oats had been 

 received from two sources at that time. Long's White Tartar. Picketts' American Banner and 

 Swedish Select, were also dropped in 1!)0S. The Sixty Day Oats was dropped in 1911. In 

 that year the Station had over two hundred pure lines of oats under test. These represented 

 the choice of those that had been gathered during the years 1900 to 1911. The last strain ot 

 the following varieties were discarded in 1912 : Tartar King, Minnesota No. 26, Progress, 

 Joanette, Chinese, Garton No. 364, Garton No. 691, Clvdesdale, S. P. I. No. 21,672, S. P. 1. 

 No. 22,005, S. P. I. No. 23,29.5, S. P. I. No. 24,846, S.' P. I. No. 24,847, Wvoming, Spanish, 

 White Hungarian, White Kirche, Anderbeck, Beseler II, Probestier, Beseler I, Beseler III, 6 

 varieties from Spain, Duppau, 4 varieties from Germany, Early Sensation, Fichtellbirge, Side 

 Oats, and Black Oats. 



