268 TWENTY-FIRST REPORT. 



and L*.250 for the Kabbit-Ear parent. The coefficient of variability for the 

 hybrid is 17.79, as compared witli 15.00 in the rogue and 8.66 in the type. 

 Altliough the Rabbit-Ear rogue is more variable in regard to width of stipule 

 than its typical parent, nevertheless the difference between the two forms is 

 in the nature of a discontinuous variation. Tt is evident that the greater 

 variability of the rogue dominates in the hybrid offspring. 



A similar condition has been reported l>y ShuU (2, 3). He found in 

 Oenothera that mutational changes were (tften accompanied by increased 

 variability of the affected organs. Thus a statistical study of Oenothera 

 LtnnarclidiKi and its dwarf mutation iiaueUa .showed a great increase in varia- 

 Itility correlated with a decrease in plant height of the latter. An increase 

 in the mean number of branches and total branch length of mut. rubrinervis 

 was accompanied by a marked increase in variability over that found for the 

 same parts in phmts of the parent species (Oe. Lamarckiana ) . 



HEHAVIOK OF GR.\I)rs RAKBIT-EAR ROGUE CROSSED WITH OTHER VARIETIES 



THAN GRADUS. 



Beside the crosses between Rabbit-Ear rogue and Gradus type several 

 crosses were made between type plants of Peter Pan and of Thomas Laxton 

 with (iradus Rabbit-Ear rogues. Peter Pan and Thomas Laxton are varieties 

 dilfering from Gradus in some respects, but like Gradus, In that they both 

 produce Rabbit-Ear rogues. Such rogues differ from those of Gradus only as 

 the type plants of both varieties differ from Gradus, 1. e., chiefly In pod shape 

 and plant height. None of these hybrids have been carried past the first 

 generation, but in every case the Rabbit-Ear characters have been dominant 

 and behaved as in the first generation of crosses with Gradus type. The 

 results arc inchnlcd :il Hie end of Table 1. Iml are not iiieluded in the snni- 

 inai y. 



RAimlT-EAR ROGUES 1\ TETER PAN VARIETY. 



Comparable results were obtained in Ihe llrsl. second and Ihird genera- 

 lions (»r the cro."i!s Peter Pan type x Peter Pan R. E. rogue, and reciprocal. 

 Also the back-cross, Peter Pan type x (P. P. type x P. P. R. E. rog.) — F,, R. E. 

 rog., gave a like Fi and Fa generation. Although the number of individuals 

 was small in each cross and the number of varieties used was not large, yet 

 these results coupled with those of other workers would indicate that the 

 behavior of the Rabbit-Ear rogue, when crossed with type plants of other 

 varieties that produce Rabliit-Ear rogues, is the same, no matter what typical 

 plants ai*e used in the cross. At the same time, although there is no apparent 

 segregation of the rogue factors in such a cross, it has been noted by Bateson 

 and Pellew (4) that such characters as pod shape and color of cotyledons show 

 the expected Mendelian segregation, 



