702 BRITTLE RACES OS OENOTHERA LAMARCKIANA. 



The crosses were made in 1920 and 1921, and 60 specimens of 

 each of the cultures have flowered. The first two cases are explained 

 by the formula (laeta + tarda) x blandina, giving laeta x blandina or 

 laeta rediviva, which is almost like Lamarckiana, and tarda x blan- 

 dina, which must be expected to be like velutina x blandina, having 

 the stature of the latter form, and showing an excess of individuals, 

 according to the general rule for both blandina and velutina. 



The cross scindens x candicans was made in order to prove directly 

 the absence of one of the lethal factors in the tarda gametes. 0. 

 candicans, the formula for which is 0. (candicans + velutina) x 

 velutina (7), has only pollen of the type of velutina, containing the 

 normal lethal of these gametes. If tarda possessed the same lethal 

 their combination would yield barren grains, but no viable plants of 

 either type. Barren grains were almost absent, and 87 per cent of 

 the type of tarda were produced. This percentage shows the same 

 deviation from expectation as in the two previous crosses, but the 

 main result is that tarda x velutina gives viable plants, proving that 

 the lethal of the latter is absent in the former. The other type of 

 the hybrids was laeta x velutina, giving ordinary Lamarckiana. 



Secunda and elongata are derivatives from 0. simplex (6). The 

 viable part of their pollen consists of laeta gametes only. Combined 

 with 0. scindens, it must produce laeta x laeta or barren grains and 

 tarda x laeta or scindens. The barren grains were to be expected in 

 the same percentage as in 0. simplex itself after self-fertilization, and 

 such was exactly the case. The cultures on the beds were uniformly 

 scindens, which, as has been stated, cannot be distinguished in the 

 garden from 0. Lamarckiana. 



A most interesting cross is that with 0. deserens (4), which is a 

 uniform race, with only one kind of gametes. These contain the 

 factor for brittleness, but have no lethal. Combined with velutina 



Table V. 

 Crosses of 0. scindens with 0. deserens. 



they produce 0. rubrinervis, from which the strain has originally been 

 split off. The combination with scindens must therefore be expected 



