388 



THE AMEBIC AX NATURALIST [Vol. XLIII 



trated by a concrete example. Let the matter of improv- 

 ing maize by selective breeding be taken. Suppose that 

 a breeder starts with a promising variety of yellow dent 

 com. This variety, while promising, has never been im- 

 proved by breeding at all. The ears are only fair in 

 respect to size and shape. The principal aim of the 

 breeder of this corn, let us suppose, is to increase the 

 "earliness" (t. e., shorten the time of maturing), but at 

 the same time he wishes to improve the other character- 

 istics of the corn— size and shape of ear, relative pro- 

 portion of corn on cob, and yield per acre. In making 

 field selections of plants to furnish seed for testing out 

 by the "ear-row" or other method, the breeder will en- 

 deavor to select with reference to as many as possible of 

 the points enumerated above in addition to "earliness." 

 Further it will be desired after the ears have been har- 

 vested and dried to take careful account in selecting the 

 seed of at least the following points : Shape of ear, length 

 of ear, circumference of ear, condition of tip and butt of 

 the ear, kernel shape, germination and proportion of 

 shelled corn to cob. Xow if a plant happens to be especi- 

 ally early, even though it be relatively poor in respect to 

 these ear characters, it is likely to be selected to furnish 

 seed, in spite of these defects. But it is possible to devise 

 a formula for a selection index number which shall give 

 whatever weight may be previously agreed upon to each 

 of these several variable characters of the corn which 

 have been enumerated. Having settled upon a particular 

 formula, the selection of seed then becomes essentially a 

 purely mechanical matter so far as the characters in- 

 cluded in the formula are concerned. The value of the 

 index will be determined by the relative contributions 

 from each of the included variables. If the breeder cal- 

 culates such a selection index number, and takes no ears 

 with an index below that standard, it will then be pos- 

 sible for him to select with reference to a series of char- 

 acters in an unbiased and impartial manner. 



