EFFECTS OF CROSSHSTG. 39 



require species to cross in much closer or in much more radical 

 relationship than is their long-established habit. 



It is known that very close breeding among poultry, swine, human 

 beings, and in corn is harmful. On the other hand, wheat and 

 other self-fertilized species, though bred up to a high standard of 

 specialization, as in the production of a large proi^ortion of seeds as 

 compared with the remainder of the plant, do not seriously, if at 

 all, retrograde under self-fertilization, the most incestuous kind of 

 inbreeding, carried on for many genei-ations. And the evidence is 

 most conclusive that some of these close-fertilized species may be 

 materially improved by a system of selecting the self-fertilized plants. 

 The best newly originated variety of wheat at the Minnesota experi- 

 ment station came from a single mother plant (or maj^ we call it 

 mother-father plant ?) chosen in 1892. There is some good evidence 

 also that close breeding in some classes of animals is not injurious, 

 and in many families and herds of animals close breeding, or in-and-in 

 breeding, has been ver}^ useful in fixing valuable new types. 



The general statement that cross breeding gives increased vigor, 

 size, and value is also too sweeping, and the limitations of the good 

 effects should be better understood. Crossing does often increase 

 vigor, size, and other good qualities, and it often decreases these 

 qualities, and sometimes in radical crosses the average progeny is 

 exceedingly weak, even being so weak in fecundity as to be sterile. 

 This fact was observed in certain wheat hj^brids at the Minnesota 

 exijeriment station, where in a few generations the hybrid stocks 

 became very weak and finally ceased to produce seeds, while other 

 stocks from the same two individual parent plants were very strong 

 and were the progenitors of some of our most promising new wheats. 

 But, so far as the writer has observed, hj^bridizing increases variation 

 in the first few generations. Swingle and Webber ^ show that many 

 radical hybi-ids vary but little the first generation, but that all hybrids 

 varj^ within a few generations. Usually this variation is both upward 

 and downward, though in some cases of radical crosses none of the 

 progeny are equal to either of the parents, and in other cases nearly 

 all are better. As to whether the average progeny of the cross is 

 stronger or weaker than the mean between the parents — " mid-x)arent," 

 as Galton expi-esses it — depends in part on whether the parents are 

 properly related, or whether the cross is too radical or too close to suit 

 the habits of the parents. Determining what degree of relationship 

 is best in the mating of plants is, indeed, an interesting subject for 

 scientific inquiry. 



In hyl)i-idizing plants to form new varieties, large numbers can 

 usually be employed, and the average qualities of the progeny is a 

 matter of no i)articular consequence. The important feature is that 

 the hybrid stocks vary greatly in tlic desired dii-ection, giving a few 



'Yearbook, U. S. Dept. Agr., 1897, p. 40?! 



