ElUCII V0.\ TSC1II:;HMAK. - EXAMKN \)V. l\ ïlllidlilE llKS KACTI-LUS. 'J.j 



cxplanalioii boUi ot" the production oï new fornis resulting: from liyhridizalion. 

 and also of complicaled cases ol' Mendelian inherilance. 



Many such cases which hâve been mucli worked on by Nilsson-Elile and 

 llie Author, may be explained on the supposition Ihal Ihe parents dilïer in two 

 or three iactors. Il must be supposed Ihat thèse i'actors may exiiibil a certain 

 relalionship, and behave often as a single unil. This connection may be mani- 

 l'estcd eilher by tlie union of I'actors previousiy separaled, or by llie sc])aialion 

 ol" I'actors previousiy unil<'d. 



In the first case a new type is proiiuced by synihelic means, and in Ihr 

 second case new l'ornis arise by analysis. 



But it must not be for^otlen that thcse results, salisl'actory Ihouyh Ihey 

 may be, are not sufficient to place the l'actorial tbeory on a solid basis beyond 

 the reach of ail objections. To obtain this end, it is necessary to make a Iho 

 rough examination of the theoretical expectations, by means of carefuilx plnn- 

 ned experimcnts in re-crossing extracled types. 



The author has examined in great détail tiie descendants of the cross 

 l'i^ifiii arvensc var. î'osea and Pi^um sfifuniin var. iilbi(. In this cross a new 

 i'orin is produced in !■', with red flowers and in \\ the three types red, 

 jiinU and white no\\ered occur in the expected proportions. According to the 

 l'actorial Iheory the white variely carries the factor which gives rise to red 

 llowers, and which is absent in the pink flowered variety. Il therefore follows 

 that among the white descendants from the cross some will not carry this 

 factor, and some will be lielerozygous for it. 'Ihis the author clearly shows to 

 be the case. 



Analogous results hâve l)een obtained by the re-crossing of extracled plaids 

 from other crosses of a bifaclorial nature. 



Among crosses worked ont by the author on thèse Unes are : Pistan sati- 

 l'uiii X P. arvense (without spotted seed coats), and also a case in which three 

 I'actors are concerned viz. Matlhiola incana var. rubraxM. glubia\av. alba 

 in which case a pure violet type is obtained in F, and in F, five types appear 

 in the ratio 27 : 9 : 9 : ,' : 10."^ 



l'iirlher analysis shows thaï Ihe pink llowered jiarent may be represented 

 by the formula S.bC and the white parent by (/ Vx:: the violet ¥^ type is there- 

 fore A R C. To prove the correctness of thèse l'ormulae a very large number of 

 crosses hâve been made both between the parental types and the extracled 

 forms: the results hâve been exactly as foreseen on the aljove hypothesis. 



In conclusion, thèse experimenls, in which some thousands of individuals 

 hâve been recorded. confirm the l'actorial Iheory in the most complète and 

 satisfarlory manner. Il cannol be doublée! that Ihe Iheory thus establi.shed is 

 a great step in the progress of hybridizatiun. 



