CORNELL UNIVP^RSITY HiXPERIMENT STATION BULLETINS. 327 



pollinated. The lots into which the original pistillate parent, Giant 

 Purple, was again impressed — Bl, B3, B5 — showed very variable 

 offspring, although there were large numbers of purple and purple 

 striped fruits in the progeny. B3 and B5 were very much alike in habit, 

 color and fruit. There was comparatively little variation in any of the 

 B lots. 



Extent of variation in "C" crosses (1890.) 



Series C originated from a cross of Long White by Black Pekin. This 

 aeries then, is much like A in parentage, except that the pistillate parent 

 has longer fruit. The effect of the staminate parent in giving color to the 

 foliage was more marked than in series A. In no case was there an 

 absence of the purplish tinge of Black Pekin, and frequently' the color 

 was nekrly as deep in the parent. The fruit was of intermediate color, 

 but with the purple predominating. In form, a few of the fruits resem- 

 bled the staminate parent; and many resembled the pistillate parent, 

 while others were wholly distinct. 



Four fruits were again crossed or selected in 1890 from this series : 



CL Pollinated by another Hower from same plant. 

 C3. Polliaated by Black Pekia (original staminate parent). 



C4. Selection, not artificially pollinated. Seen at the extreme left in the illustra- 

 tion. It was very angular at the ends, purple with a metallic-gray apex. 



In these lots, the fruit pollinated from the same plant, CI, gave a vari- 

 able and very unproductive offspring. C3, into which Black Pekin has 

 gone twice, gave only purple fruits. C4 was the one which we particularly 

 desired to fix, for the original fruit had strong points of merit. This fruit 

 gave us 169 plants, none of which, however, were like the parent, and 

 none seemed to possess superior merits. Only 31 of the plants from it 

 produced fruits before the frost, and of these five had green herbage and 

 26 purple herbage. All the C plants wei-e very tall in 1891. mostly 

 dark in foliage, and late. 



The result of all this experiment with secondary crosses and the sec- 

 ond generation of primary crosses, numbering 2,126 plants, shows that 

 they were exceedingly variable, that pollination from the same plant did 

 not fix the types, that very few novel and promising types appeared, 

 that the white and purple colors tended to unite to produce striped 



