IMPHOVEMEXT OF CORX 115 



uncommon to find ears of the same variety and selected to 

 the same type which yield 100 per cent more than other 

 ears which, so far as one can tell, appear to be equally good. 

 For ear-to-row tests, a uniform field wide enough to plant 

 as many rows as there are ears to be tested should be pro- 



Figure 4G. — The value of the ear-to-row test. The crates show the yield from 

 two rows which appeared to be equally good. 



vided. The rows may be as long as desired; they are limited, 

 of course, by the number of kernels on the ears to be tested. 

 A fair-sized ear will plant a row from 40 to 50 rods long. 

 Many corn breeders use shorter rows and plant duplicate 

 rows from each ear, to serve as an additional check in 

 determining the best. At harvest time, the product of each 

 ear is carefully weighed and the high-yielding and low-yield- 

 ing ears thus discovered. The corn from the low-yielding 

 rows is discarded, and only seed from the high-yielding rows 

 is saved for future planting. 



The chief objection to this method is that the high-yield- 

 ing rows of corn have been fertilized in part with pollen 

 from other rows. The individual kernels on the selected 



