270 TWENTY-FIRST KErORT. 



Ill additiuu to the intermediate described as having type-like foliage and 

 rogue pods, Bateson and Pellew describe another form that corresponds to 

 what the present author calls an "intergrading Rabbit-Ear rogue." Their later 

 communication (5) does not make it clear \Yhich form of intermediate pro- 

 duced a majority of type plants, or that there was any difference between 

 the two forms of intermediates in this respect. 



Three exceptions to the nonnal behavior found in crosses between rogue 

 and type plants were noted by Bateson and Pellew. The first exception was an 

 Fi plant from the cross E. G. rogue x E. G. type. Instead of being a rogue 

 at maturity the plant had type stipules and rogue pods. In succeeding gen- 

 erations it behaved genetically like a type plant, except that some of the off- 

 spring w'ould occasionally have curved pods. The second exception was in the 

 cross D. A. I'ogue x D. A. type, which gave an Fi generation of three rogues 

 and two intermediates. The latter were comparable to the intermediates foimd 

 In Early Giant. The offspring of the intermediates have not yet been described. 

 The third exception was in the nature of a bud-sport. An Fi rogue from the 

 cross D. A. rogue x E. G. intermediate gave off from the main stem a branch 

 which w^as type-like, while the rest of the plant was a thorough rogue. Seed 

 saved from the rogue parts gave only rogues in the next generation, while seed 

 saved from the type parts gave some type-like plants in addition to many 

 rogues. The authors give no data in regard to the behavior of the hybrids 

 when Rabbit-Ear rogues are out-crossed with varieties that do not produce 

 Rabbit-Ear rogues. 



The production of rogues by type plants has been thought by Bateson iiiid 

 I'ellew (4) to be connected with some instability of the germ plasm related 

 to the great lateral extension of the foliar parts of type plants. It is siigg(>slo<1 

 that type plants may be con.sidered a mosaic of rogue and type tissue, typo 

 tissue giving rise to the type gametes and rogne tissue to the I'oguc gametes. 

 The behavior of the Fi generation resulting from crosses between type and 

 rogue plants is explained by assuming that in the seedling stage somatic segre- 

 gation of the type and rogue factors takes place so that at maturity the type 

 elements are excluded from the germ lineage. The weakness in a hypothesis 

 involving somatic segregation lies in the fact that no mechanism is known 

 l)y which somatic segregation of factors can take place. 



AVhite (1) has suggested that the rogues are similar to what 10mors<iM 

 ((.!) calls a "recurring somatic mutation." A somatic mutation, in the sen.se 

 in which Emerson uses the term, implies that the difference between the type 

 and the mutation is a difference due to a factor change, or point mutation in 

 :i cliromosonu'. taking place during ontogeny. Such .i miiliilion would be coni- 

 p:inil)h' to tlio.se found by Morgan and bis coworkers (7) in Drosophila and 

 would be expected to show INIendelian inlu-iitMnce. The data at hand are not 

 snliicicnt to .-illow sm-li :in iiitcfiu'ctalion al presejit. 



