W. LAWRE.NCt: HALLS. 



LMIEIIITANCE OF MFASl REABLE CIIAHACÏERS OK EOÏTO.N. 455 



The inler-Egyplian cross shown in Fig. i was made between parenls of 

 almost idenlical seed-wcighl. The F, seed was ralher heavy, and the Iwo 

 Fj familles bolh show thrce modes in Ihe curve. Thèse modes are cerlainly 

 significant, bul the nature of thaï significance is not clcar. 



The Egyptian-Upland cross of Fig. 5 seemed to behave quile simply, 



100 ISO 



l-'i^. ."i. — Mcaii Seed-Wciglil. 



the heavy seed being inlensifie<l in F,, wiiile Ihc F, graph indicates a simple 

 ségrégation of one lighl seed with three heavy seeds. The lighl seeds hâve 

 been brcil lo F^, and pure heavy seeds to Y^. The chief obscurily at Ihis 

 stage related lo the increased weighl in F,, since il was plainly not sufficienl 

 lo ascribe tliis to " vigour "' of the bybrid. 



The real interesl of the character begins wilh such a cross as the one 

 shown in Fg, also made between Egyptian and American, bul wilh parenls 

 of almost equal seed-weight. The F, seed weighl leapt up to nearly 

 double thatofeilhcr parent. The 1", did not give the slighlesl indication of 

 a 5 : 1 grouping, but ranged over a modal curve between the Iwo extrêmes of 

 parent and F,. A prolonged examinalion of the slalisiical data finally led to 

 the conclusion Ihat this complex curve was essentially a séries of 5 : 1 curves 

 overlapping each olher; this conclusion appeared salisfaclory until the Avriler 

 realised thaï no heavy seed had ever been involved in the cross; a heavy seed 

 had therefore made ils appearance in Fj, where none had been before. and 

 was moreover behaving in a simple Mendelian fashion. 



Aulogenous Fluctuation. — The complète story of this inheritance of seed- 

 weight is a slriking example of the opération of a form of lluctuafion which ha? 

 not... lo the besl of the writer's belief... been clearly differenliated previously, 

 and lo which he has given the désignation aulogenous. The external manifes- 

 tation of any given characterislic in an individual (simple fluctuation being 

 eliminaled) is controUed by corrcialion with other characlers of the individual 

 lo a grcater or less extenf . 



Thus, in a highly coin|ilcx 1",. those plants with large bolls hâve a higher 

 niean seed-weight than those \\illi sniall bolls, Regarded from one point of 

 view, the diameler of the boU acts as a crypiomere upon the seed weighl. 

 The same may be shown by plotting the corrélation diagram, bul for présent 

 purposes il will be sulticienl to dissect the F, graph inlo smaller groups 

 composed of plants which |)ossess some character in common. 



Examining llic F,. gi;iph iii liiis way we oblain llic following resulls : 



