44 GEORGE HARRISON SHULL 



that could make the pure-line method of corn production practical. No credit 

 is sought for the fact that I made these four-way crosses some years prior to 

 the similar combinations made by Dr. Jones. They are presented here only 

 because they belong in a historical account. 



In the last section of Table 2.7 I have entered five families which have the 

 form of four-way crosses, but in which the single crossings used were Fo in- 

 stead of Fi. Only the first of these five families actually involved four differ- 

 ent inbreds, the others being partially iterative, in that only three inbreds 

 contributed to each. A comparison of the double crosses both of Fi and F2, 

 with the corresponding single crosses, is instructive. Comparison of the sum- 

 mary of section C with that of section K shows the double cross families 

 slightly inferior to the single cross families, as indicated by a 1 per cent higher 

 grain- row number and 6 per cent higher yield of the single cross families 

 over the double cross. Comparing sections L and E, it is to be noted that the 

 double cross retains the vigor of the F2, instead of declining to the vigor of 

 the F3 families produced by the usual methods, as seen in sections G and H, 

 Table 2.7. 



In 1911 1 realized that the effective exposition of the important discoveries 

 we were making required photographs of prepared exhibits. A number of such 

 exhibits were set up and photographed, and have been presented in lantern 

 slides on many occasions. I have included the most instructive of these here. 



Here the detailed account of these studies must end, for although they 

 were continued in 1912, 1 have been unable to locate the field and harvesting 

 notes including grain-row counts and weighings for the 1912 cultures. These 

 1912 cultures were especially designed to explore the evidences of Mendelian 

 segregations in the F2 and the F3 families, with respect to grain-row num- 

 bers and yields. They included 11 families of the breeding Fi X self, 8 families 

 of Fi X sib, 21 F2 X self, 10 Fo X sibs, and five families of F3 X self. There 

 was also an interesting pair of approximations to eight-way combinations or 

 quadruple crosses produced by reciprocal combinations of the four-way 

 crosses included in the 1911 cultures. While these had the form of quadruple 

 crosses, they were imperfect in that one of the inbreds was repeated, so that 

 only seven different inbreds were represented, instead of eight. This was in- 

 evitable since I initiated only seven inbred lines in the beginning of these 

 experiments. 



The 1912 crop completed the experimental work with corn at the Station 

 for Experimental Evolution, and I spent the next year in Berlin, Germany. 

 In a lecture I gave at Gottingen about three weeks before the beginning of 

 the first World War the word heterosis was first proposed. I used the occasion 

 to discuss the bearing of the results of these studies on the practical work of 

 breeders of various classes of organisms, both plant and animal. I stressed 

 the point that the breeder should not be content, as had long been the case, 

 to seek merely to avoid the deterioration incident to inbreeding, but should 



