70 APPLICATION OF PEINCIPLES OF HEREDITY TO BREEDING. 



plants m question were grown under favorable conditions, it was 

 easy to demonstrate that leaf lobings are a good Mendelian char- 

 acter, but under certain unfavorable conditions the lobing disap- 

 pears, thus confusing the Mendelian results. Kellogg found a similar 

 case in silkworms. Certain strains of silkworms which produce 

 salmon-colored cocoons when crossed with certain white strains 

 gave results which showed clearly that the salmon color is a Men- 

 delian character. But in some crosses the salmon color became 

 extremely variable, extending all the way from almost pure white 

 to very deep salmon color, thus somewhat obscuring the numerical 

 relations of the colors in the second generation of the hybrid. In 

 some of Correns's work with variegated plants he found a similar 

 character. The variegation, although a good Mendelian character, 

 varied with true green leaves as one extreme, and in some crosses 

 this fact made the number of green leaves in the second generation 

 larger than theory called for. 



CORRELATION. 



A good many cases have been found where two characters wliich, 

 so far as appearance goes, are not physiologically related to each 

 other seem always to be transmitted together, and the breeder 

 frequently gets irregular results because of this coupling or correla- 

 tion of characters. For instance, Tschermak found in Chinese oats 

 that hull-less seed is correlated with long, many-flowered spikelets, 

 and . that these two characters were always transmitted together. 

 In Price and Drinkard's work with tomatoes at the Virginia Agri- 

 cultural Experiment Station they found what seems to be a case 

 of this kind. In one of their crosses one of the parents had green 

 foliage and two-celled fruit. The other parent had 3^ellowish green 

 foliage and many-celled fruit. The hybrid was like the first parent. 

 In the second generation of this hybrid all the plants having green 

 foliage had two-celled fruits, and all those having yellowish green 

 foliage had many-celled fruits, thus indicating that the many-celled 

 condition of the fruit is transmitted with the yellowish green leaf 

 character, at least in this case. Hedrick and Booth, in their work 

 with tomato hybrids, found that dwarf stature was correlated with 

 dark-green rugous leaves, while standard stature was similarly 

 correlated with lighter green smooth leaves. 



Such cases sometimes give the breeder a good deal of difficulty. 

 They are of special importance in relation to theories of heredity, 

 and should always be carefully noted and reported. 



165 



