FURTHER NOTES ON EPILOBIUM HYBRIDS 81 



and a bushy little plant is produced 20-30 cm. high. The stems 

 with their leaves are almost glabrous, less hairy even than E. 

 montanum of the same height in the same situation. A slight 

 down is present, however, the hairs being all short, non-glandular, 

 and obliquely placed on the stem and on both surfaces and edges 

 of the leaves." The leaves are broadest about half-way along ; 

 they are usually fringed with small forward-directed teeth, but 

 some are almost toothless. 



(ii) Some of the shoots, usually appearing a little later than 

 the small twiggy ones, grow much more luxuriantly — to a height 

 of 45-60 cm. Their leaves are more distant and much larger, 

 an average good-sized leaf being about 7-9 cm. long x 1*5 cm. 

 broad ; there is a slight tendency to become spathulate, the base 

 of the leaf tapering very gradually and being almost without 

 teeth ; but accurately lanceolate leaves also occur. The stem 

 is often somewhat flexuose, but hard and brittle. The whole 

 shoot is covered with a fine down, somewhat more pronounced 

 than in type (i), consisting of a mixture of erect glandular hairs 

 with a few longer half-erect simple hairs. The general shade of 

 green is also somewhat lighter than in the slender shoots. 



These two types of shoot are possibly due to slight local 

 differences in nutrition among the crowded rosettes ; but they 

 are of sufficient constancy to merit description. 



(iii) In 1912 a third type of shoot was produced by two 

 plants. This has the tall stature of type (ii), but differs in 

 having the leaves more markedly ovate, broadest below the 

 middle, and more deeply toothed ; dimensions, 6-7 cm. x 2-2J cm. 

 The clothing of the stems and leaves is somewhat denser, 

 consisting of mixed long simple and short glandular hairs. 

 The most remarkable feature is, however, the relatively superior 

 development of the flowers. (See below.) 



This type (iii) is specially interesting because, so far as can 

 be seen, it is identical with plants produced by the reciprocal 

 cross, E. montanum 2 X E. hirsutumj , which are flowering this 

 year for the first time ; types (i) and (ii) differ very markedly from 

 the reciprocal cross. It is difficult to account for this at present, 

 for several possibilities are open ; the offspring of the reciprocal 

 crosses may be normally distinct,! and type (iii) may be a bud- 

 sport ; types (i) and (ii) may be atypical and due to special 

 conditions of growth, and so on the other hand may be type (iii). 

 The sharpness of distinction between the three types of shoot is 

 against the theory that they are simply due to ordinary fluctuating 

 variation. In any case, the phenomenon should be a warning 

 against too hasty conclusions that reciprocal crosses give dif- 

 ferent results. 



One of the most curious features of this hybrid relates to the 



* Cf. E. adnatum f. stenophylla $ x E. hirsutwn 2 • Compton, I. c. p. 161. 



t Cf. H. de Vries, " Ueber doppeltreziproke Bastarde von (Enothera biennis 

 und (E. muricata" Biol. Centralbl. xxxi. p. 97, 1911; W. N.Jones, " Species 

 Hybrids of Digitalis" Journ. of Genetics, ii. p. 71, 1912, (fee. 



