548 PREFERENTIAL FERTILIZATION IN 



per cent, for 0. Lamar ckiana mut. distorts 23—32 per cent, and 

 for 0. Lamar ckiana mut. elongata 31 per cent. From this we may 

 conclude that the 0. laeta ovules were fertilized in these cases in too 

 small numbers. The same condition prevails in other instances. 

 If the pollen of these races is brought on to the stigma of other 

 types, having two kinds of ovules, the 0. amphilaeta will also be in 

 the minority. For example, they were reduced to 3 per cent in 0. 

 Lamar ckiana x elongata, to 15 per cent in 0. oblongax elongata, and 

 to 11 per cent in 0. oblongax nitens. 



From these considerations it is clear that, in order to be com- 

 pared with the relations derived from the general rule for mono- 

 hybrid splitting, the figures found for the fertilization of gametes 

 of the type of 0. velutina should be diminished, whereas those for 

 the group of 0. laeta should be increased. Constitutional strength of 

 the races used, individual vigor, and the conditions of the cultures 

 seem to be the chief causes which determine for each case the amount 

 of the deviation. 



Besides these, internal and hereditary factors may play some 

 part in producing this phenomenon. Especially it may be presumed 

 that the lethal factors might be a cause of a weaker constitution, 

 and thereby influence the degree of the deviations from the numeri- 

 cal expectations. In order to try this, I have chosen a hybrid race 

 which has only gametes of the type of 0. laeta, but some modified 

 and the others normal, the latter having the ordinary lethal factor, 

 whereas the first have lost it. 



In a previous article (5) I have described the hereditary behavior 

 of 0. rubrinervis. It is a half mutant from 0. Lamarckiana, in which 

 the 0. laeta gametes have been modified in such a way as to lose 

 their lethal factor and to obtain a factor for brittleness. In the 

 self-fertilization of this race these gametes produce brittle indivi- 

 duals almost without empty seeds, which have been the origin of 

 a new race, called 0. Lamarckiana mut. deserens. If this race is 

 crossed with the parent species we get two kinds of hybrids, the 

 composition of which may be expressed by the formulas 0. (deserens' 

 x laeta) and 0. (deserens' x velutina), both of which are viable. To 

 the former I have given a special name, 0. lucida (5), since it has 

 often been used in crosses or as a control for identifying the charac- 

 ters of other hybrids. This hybrid, 0. lucida, evidently has the 

 formula (deserens' -\-laeta) x (deserens' +laeta), and therefore may be 

 made use of for a comparison of the vital gametes of 0. deserens 

 with the lethal ones of 0. laeta. In the present investigation, however, 



