FIELD CROPS 491 



rows 5 feet long and 1 foot apart. If a considerable number of rows 

 are planted, it will be found convenient to run them in three series, 

 with a narrow space between the series, as shown in the diagram. 

 If weeds are numerous, one or two hoeings may be necessary. At har- 

 vest time the plat should be gone over carefully, and those rows 

 which appear to be low in yield, or are particularly subject to lodg- 

 ing, disease, or undesirable qualities of any kind, should be discarded. 

 Those of outstanding value should be harvested and thrashed sepa- 

 rately, and retained for further testing. 



The next year the seed from these rows should be planted in 

 rows 17 feet long and 1 foot apart, planting every tenth row of a 

 standard variety or of the bulk seed from which the original selec- 

 tions were made, for comparison and for the detection of soil varia- 

 tions. Each of the short rows of the preceding year should have pro- 

 duced enough seed for two or more of these 17-foot rows. The several 

 rows of any particular strain should be planted in different parts of 

 the plat so as to equalize any variation in the soil. Their location 

 should be carefully noted, so that they may be compared with each 

 other and the seed combined after harvesting and weighing. These 

 17-foot rows contain approximately one-sixteenth of a square rod, or 

 1-2560 of an acre. At a common rate of seeding in sections where 

 pats are an important crop, 2% bushels to the acre, one-half ounce 

 is sufficient for one of these 17-foot rows. At harvest time the plat 

 should again be carefully studied and only the most promising strains 

 retained. Each row should then be harvested, thrashed, and weighed, 

 and the weight recorded. 



The test the third year is along similar lines. Two or more 17- 

 foot rows should be planted of each of the strains which yet remain, 

 and the check rows should be used as before. In addition, however, 

 plats should be planted of several of the most promising strains, so 

 that they may be increased as rapidly as possible. 



The fourth year the few remaining strains are again tested as 

 before, and plats of considerable size should be planted of those with 

 the best records. After the harvest of this year, all should be dis- 

 carded except those of outstanding excellence. These best strains 

 should now be in sufficient quantity for field tests, and if of real value 

 should be distributed to neighbors and tested under varying con- 

 ditions to demonstrate their general adaptability. If the strain 

 proves its excellence over a considerable area, a name should be given 

 it, to prevent confusion with other varieties. 



New strains selected either from the general crop or from the 

 row tests may of course be introduced at any time by starting them 

 in the 5-foot rows and adding them to the general series of 17-foot 

 rows the following year. 



A permanent record should be kept of the different strains. This 

 record should show the essential facts regarding the performance of 

 a given strain from the time the original selection is made. For 

 convenience, each selection should be given a number, and the num- 

 ber should be retained until the strain is discarded or given a perma- 

 nent name as a variety worthy of distribution. If selections are 



