274 



TWIN HYBRIDS OF OENOTHERA HOOKERI T. AND G. 



on the Mutation theory (German edition, Vol. II, p. 468) such 

 unpaired characters produce constant hybrids, whereas Mendelian 

 splittings require the assumption of an active and an inactive 

 condition of the hereditary factors. This latter view has been changed, 

 later on, by Bateson, into his hypothesis of presence and absence, 

 which seemed a more empirical form of the same idea. But since 

 Morgan has shown that it is just as possible to determine the po- 

 sition of absent factors in the chromosomes of Drosophila as that 

 of present units, it seems to me that the question latency versus 

 absence has been definitely decided in favor of the first alternative 

 and that my hypothesis of the unisexual crosses is still unimpaired. 



In the case of splitting laeta the sexual cells of the second parent 

 must evidently contain a character antagonistic to that of laeta and 

 this must be in the inactive or velutina condition. This is evident in 

 the case of Oe. blandina (Oe. Lamarckiana mut. velutina) and must 

 be assumed for the wild species Oe. Hookeri, in order to explain 

 the complete analogy of the phenomena. The splitting laeta are, 

 as a rule, provided with germs in almost all their seeds and this 

 fact at once excludes the hypothesis of the influence of a lethal 

 factor and of a third type of hybrids, hidden in empty seeds. It 

 confirms the reality of a splitting into two types only. 



I will now give a list of the observed cases, showing the percentage 

 of germs in the seeds and of velutina among the offspring. 



Table 6 

 Splitting of laeta into laeta and velutina. 



The figures for the velutina of the first two crosses are deduced 

 from the tables given above, those for the two following cases from 

 my book on Gruppenweise Artbildung (1913, p. 132) and those for 

 the last two from experiments described elsewhere {Opera VII, p. 

 160). The percentages of germs in the crossed seeds have also already 

 been published (Opera VII, pp. 125, 134, 140 and 201). The 

 laeta letalis are excluded from our table. 



