OENOTHERA GRANDIFLORA AIT. 229 



In respect to the third generation, it was to be expected that it 

 would simply confirm the results of the second, and so I have limited 

 myself to one culture for each of the three main types and to some 

 few for the secondary combinations. Table VII gives a list of my 

 cultures of the second generation. They embraced with some few 

 exceptions 60—70 specimens each, and almost all of these have 

 flowered. As the types of the triple hybrids were exactly the same 

 as in the first generation, no special descriptions will be necessary. 

 In this table + means that a second generation of the type men- 

 tioned in the heading above it has been cultivated, whereas — indi- 

 cates that no culture of the type has been tried. 



A third generation of ovata was cultivated for 0. nanella x grandi- 

 flora. From the reciprocal cross lutea and brunnea were continued 

 during one generation more. I shall treat the crosses of 0. grandi- 

 flora lorea separately, since they split off this mutant type, and 

 deal first with the others. Among the special combinations ap- 

 pearing in the second generation there were three which could 

 clearly and easily be distinguished, but only two of them were fre- 

 quent. I shall designate them by the letters R, T, and L. Among 

 these, R was a return to the rapid production of a stem, without 

 preparatory rosette of radical leaves, which is so characteristic a 

 mark of 0. grandiflora, but which was always dormant in the first 

 generation. The plants were usually slender and small, the leaves 

 broad and dark green, and they flowered one or two weeks before 

 their stouter sisters. In July they reached 10—30 cm. more in height 

 than these, but during the flowering period they were overgrown 

 by them. Their flowers showed the same forms. The progeny of the 

 type R was uniformly so. The type T was easily recognized by its 

 truncate flower buds; these are conical in the parental species and 

 in the triple hybrids. The flowers were correspondingly smaller. 

 The leaves were almost like those of ovata, but strikingly broader 

 in their upper half. The height and stature were also the same. In 

 their progeny they repeated their characters exactly; but split 

 off some specimens of the type R. Type T was remarkably rich 

 in the production of pitchers. The type L combined the characters 

 of the hybrid called lutea with the slender stature, rich branching, 

 and thin flower buds of 0. grandiflora. It produced some specimens 

 of R among its progeny, which was otherwise uniform. A continued 

 study of these and other hybrid types of 0. grandiflora would pro- 

 bably offer the material for an analysis of the characters of this 

 species. In counting my cultures of the offspring of self-fertilized 



