278 BOTANICAL GAZETTE [October 



August. The remaining plants were pulled out as large rosettes 

 in July. 



0. atrovirensXO. cana. — Represented by 55 plants, among which 

 8 densa and 17 laxa have flowered. They were in all respects like 

 the twins of the corresponding cross of 0. Lamarckiana. Just 

 as in this cross, some specimens had cordate petals and others had 

 linear ones, repeating the cruciata type. But on the first plants 

 stray flowers with narrower petals were found, from tirne to time, 

 indicating a high degree of fluctuability rather than a splitting into 

 two constant and uniform types. 



The reciprocal cross yielded only 23 seedlings, 20 of which 

 were yeUow and died very early, and the 3 remaining ones were 

 very weak, reaching only a height of 40-60 cm. when they flowered. 

 They had the type of the gracilis of the corresponding cross of 

 0. Lamarckiana. They were not mentioned in table IX. 



0. canaXO. biennis L. — Among 70 plants of this culture, one- 

 half had the type of 0. Lamar ckianaX biennis, and of these 15 have 

 flowered. The other half were evidently cana. Of these, 28 had 

 the stature of typical 0. cana, but with some marks which indi- 

 cated an influence of the father. The foliage was less gray, a 

 darker green, with broader bracts, and more dense spikes with 

 smaller flowers than in the other crosses. The stigmas were sur- 

 rounded by the anthers, insuring natural self-fertilization, as in 

 O. biennis. Besides these intermediate types there were 7 dwarfs, 

 which had the gray, narrow, and pointed foliage of 0. cana, and 

 which in the table have been calculated together with the high 

 specimens of the cana type. 



The reciprocal cross yielded only laeta and velutina, together 

 57 plants, of which about one-half of each type have flowered. 



0. canaXO. Lamarckiana. — The two types of this culture were 

 exactly the same as in the self-fertilized offspring of the mutant. 

 There were only 19 specimens, of which 6 were cana, 11 Lamarck- 

 iana, and 2 nanella. The dwarfs combined the marks of cana with 

 those of nanella and have been calculated in the table with the 

 cana specimens of taU stature. 



The same cross had been made in 1907, the seeds being sown 

 in 1913. In this case there were 50 offspring, among which 26 



