84 THE JOURNAL OF BOTANY 



of work this summer, as it may be possible to isolate unit- 

 characters — a result which would immensely simplify the task 

 of classifying such a " variable hybrid " as the present. 



E. hirsutum Linn. ? x E. parviflorum Schreb. $ . 



The cross was made in 1911, and F 1 , consisting of three 

 plants, was growing in 1912. 



As is to be expected, this hybrid has been reported 

 frequently, the parents growing together very commonly and 

 intermediates occurring with them. But as E. hirsutum and 

 E. parviflorum are exceedingly variable species according to 

 conditions of life, and as some of their forms closely resemble 

 one another in certain particulars {e.g., hairiness, shape of leaf, 

 terete stem), the determination of a supposed hybrid is a matter 

 of special difficulty. Moreover, as some of the ovules are appar- 

 ently fertile, we have to consider not only 'F 1 , but subsequent 

 generations — a fact which may account for the reputed variability 

 of this hybrid (apart from growth-forms). 



Haussknecht's description (/. c. p. 64) of this hybrid was made 

 apparently from wild specimens so identified ; it accords so well 

 in essential points with experimentally produced plants that it 

 may be taken as correct. The colour of the flower is stated to be 

 somewhat lighter than that of the common form of E. hirsutum ; 

 I find, however, that the two flowers are indistinguishable in 

 colour. 



The most important features of the F 1 may be subjoined : — 



(a) The anthers are completely contabescent ; the capsules 

 are of a good size (5-6 cm. long), and contain a few good 

 seeds through cross-fertilisation, though the great majority of 

 the ovules abort. 



(b) There are no long runners such as E. hirsutum produces. 



(c) The clothing is of the type of E. parviflorum; i.e., the 

 plant is thickly beset with long simple hairs below, and with 

 a mixture of long simple and short glandular hairs in the region 

 of the inflorescence. E. hirsutum has the admixture of glandular 

 hairs throughout. 



(d) The flower is smaller than that of well-grown E. hirsutum, 

 much larger than that of E. parviflorum, purple. (But many 

 forms of E. hirsutum with smaller flowers occur.) 



(e) The leaves are not so prominently toothed as in E. 

 hirsutum, but much more so than in E. parviflorum. They 

 are connate at the base but not decurrent. 



Thus most of its characters are shared with E. parviflorum, 

 and the only critical features are the size and colour of the flower, 

 the intermediate toothing of the leaves, the emptiness of the 

 anthers, and the semi-sterile ovules. I hope to study the off- 

 spring of the crosses 'F 1 xE. hirsuUnn and F ± x E . parviflorum 

 as well as some others this summer. 



Some interesting points of comparison may be made between 

 this hybrid and E. adnatum i. stenophylla $ x E. hirsutum 2 . 

 In the former purple flower colour is dominant over rose ; in 



