476 



THE AMERICAN NATURALIST [Vol. XLVII 



flowers, however, were very small (petals 6 mm. long) and 

 the anthers as far as observed were completely sterile. 

 Since this plant apparently could not be selfed, I pol- 

 linated it from a large sister plant of the F 2 with grandi- 

 flora-\\ke flowers. The result was 154 seeds from several 

 capsules which gave 46 plants in the F 3 generation (cul- 

 ture 12.57). Of the rosettes, 40 proved to be large-leaved, 

 exhibiting much variation, but with several plants similar 

 to grandiflora; 6 rosettes bore long narrow leaves. 



From the 40 large-leaved rosettes there developed 

 plants 1.2-1.5 m. high with a foliage of crinkled leaves, 

 and medium-sized flowers (petals 2-2.5 cm. long). Of 

 the 6 narrow-leaved rosettes, 5 developed plants which 

 agreed with the "elliptica" type and 1 became a broad- 

 leaved form similar to the 40 described above. The 41 

 large-leaved plants of the culture evidently took their 

 characteristics largely from the pollen parent of the cross 

 and represented something of a blend. I am at a loss to 

 account for the five individuals of the "elliptica" type 

 unless they came from apogamously formed seed. The 

 "elliptica" type has since appeared in other F 2 genera- 

 tions from the cross grandiflora X biennis, and it appears 

 to be a not uncommon expression of one of the extreme 

 forms which may be thrown in the F 2 of this cross. 



