KDITU H. SAllNDKIiS. — TIIK lilIKKIHNS W 1)01 IlLK KIJtWKRS. iOI 



We hâve Iheii Ihis kiiid of inler relationsliii) betweeii faclors oxliiljilcd in 

 the case of each of llie following' pairs. 



1) X and ^" (tiio Iwo farloi's re(|uircd for sint;icness) in ail llio evorsporting 

 sirains. 



'il P and W (the factors producing palencss of sap colonr and wliile plas- 

 tids respectively) in hétérozygotes of P/AVu' composition. 



." X^' and W the linked couple liehaving as a unil ami lorniing m [lair 

 \\ilh \\ in hétérozygotes of XYj;i/\\' H' composition. 



FurLiier in the pecnliar sulphur-white race, an eversporting race which is 

 heterozygous as regards W as weil as X and V, yielding single whites and 

 double creams with a small percentage of double whites, we appear to hâve 

 anolher instance of this inter-rclationship bctween a linUed couple and a single 

 factor. In this case, however, as in other eversporting forms X and Y behave 

 as independent factors, and the couple is formed by Ihe linking of one of thèse 

 faclors — we will assume it lo be X — wilh W, so Ihat XW forms one membcr 

 of the pair and Y the other. The single zygote of a sulphur-white is the resuit 

 of a union between gamètes of X Y W and j- y w composition respectively, and 

 Ihe gametic output of sucli a zygote is again composed almost entirely of the 

 combinations X Y W and xyw while Xi/W and .i\ iv a|)pi'ar lo be the rarcr 

 ternis in the séries. We may therefore add a loui'lh ia--r to lln' list of inter- 

 related pairs given above, in which the same Ihree factors are concerncd as in 

 case 5, but where Y is the independent faclor and XW behaves as the linked 

 cou|)le. 



lïy Ihis coMceplion of Ihr relaliiiiiship cxisling between Ihe two faclors 

 rerjuircd for the production of a single, together with the fact of Ihe sex-limi- 

 ted distribution of thèse factors to the gameles, we are able lo explain the 

 hitherto puzzling fact thaï eversporting singles conslantly throw an cxcess of 

 doubles. Il now remains to consider the précise amount of this excess, or, in 

 other words, to détermine the value of )! in the schcme for the gamelic ouiput 

 in the eversporting forms given above. 



The proportion of doubles obtainable from eversporting singles of tlu' l'en 

 week class was found on the average to be between b7> and ,'>7 for llie planl 

 an a ic/iole. Ihough a higher proportion was obtained now and again as a chance 

 variation. If the higher number should proveto be correct we may lake n to be 

 equal to 8, and the Irue ratio to be 7 single : 9 double or 50, '25 0/0 of doubles. 

 If, on the other hand. the lower number more nearly represents the Irue pro- 

 portion, the value of /! is probably Ki, which would give a ratio of 7.5 single 

 lo 8.5 double or 5r>.t'25 0/0 of doubles. The fact thaï sm;dl sample sowings, 

 where thèse sowings were made from Ihe seeds of a single pod, frequently gave 

 divergent results renders it difficult on the présent évidence to décide between 

 thèse two values, and until considerably larger records are available we may 

 convenienlly imiicale the proportion of single lo double occurring in Ihese 

 forms by Ihe gênerai expression 1 + x single lo 11 — ./• double where x has sonie 

 value less than I. 



Among the ofTspring of crossbreds derived from malings belween c\ersjior- 

 ling and true-lireeding singles the number of doubles, as previously slated, was 

 in most cases found to be in Ihe proportion usually met with in Ihe case of the 

 récessive, where one pair of allelomorphs only is concerned, /'(': 1 in i. In cer- 

 tain malings however Ihe proportion was very much less, being sometimes as 



IV COM-Én. ISTERN. DE ^.t.Nl'.TUjUE. 2G 



