438 



IV' CONFÉRENCE INTERNAÏlONAf.E DE Cf.NÉÏlQLE. 



al ils normal weight (Fig. 7 a) or pcrha[)s insorted in a 25 mm. boli, whcn il 

 was " compressée! " slill furlher to weighlsas low as 0.0^5 g., wliifh liavo never 

 been rearhed by eilher parenl (Fig. 6). 



Fanlaslic Ihough llie above explanalion niay possil>ly seen, Ihere can ])e 

 lilUe doubt but Ihal il represenis Ihc true slale of affairs. The error from 

 nalural crossing prevenls liie wriler from presenling F- dala derived from 

 families of crilical sizo, bul tiie behavioiir of llie mimerons small F- l'amilies 



30 ifO m m 



Fijj. S. — Meai] Miiiximum I.en^lli of iIh' I.inl. 



raised from Ihis cross shows, afler eliminalion of vicinisis, Ihal Ihe conibined 

 action of ségrégation and anlogenous nvicluation occnis as (lescril)ed aliove. 



TnE LKNGTn OF Tni: LI.NT. 



The inherilance of Ihe mean maximum lenglh of Ihe hnt, delermined 

 by combing ont nol less Ihan five seeds per planl, is siiown in Figs 8 and 9. 

 The inler-Egyplian cross has a definile indication of bi-modalily in F,, thougli 

 we cannol inlerpret this as ségrégation wilhoul évidence from F-;. 



Tiie Egyplo-American cross (Fig. 9) repeals in Ihis matler Ihe behaviour 

 of Ihe seed-weighl. The same wide range behveen the parental extrêmes is 

 shown, wilh indications of five modes. On dissecling for correlated characters 

 we oblain the same séquence of resulls, one only of thèse beiiig shown in the 

 figure; seed-weighl and tint lenglh are positively connected, so that the lieavy 

 seeds (which we hâve seen to be iiew featnres of the F^) raise Ihe lenglh of the 

 segregating short lint to 12() mm. in place of l'I mm. 



I.ONCI.rSIONS 



A. 1. The error from nalural crossing is a serions obstacle to précise 

 work on a-enetics in cotton. 



