ONAGRACE.E. 9 



Var. a ill borders of Avoods, damp places and on rocks. Sparingl}^ 

 but generally distributed from Somerset and Hants to Orkney. 



Var. 3 in damp woods and by river-sides. Much rarer than 

 var. a, and probably either planted or escaped from cultivation. 

 It occurs in several places in Shropshire ; in North Wales ; Hamp- 

 shire ; by the banks of the Swale, Ptichmond, Yorkshire ; CoUinton 

 Woods, near Edinburgh, the latter being the only place in which 

 I have seen it growing, except in gardens, where it is very commonly 

 cultivated under the name of Erench- Willow or Rose-Bay. 



England, Scotland, Ireland. Perennial. Summer 

 and Autumn. 



Var. a with the rootstock emitting numerous thick rather short 

 stolons. Stem erect, 1 to 4 feet high, roundish, smooth, simple 

 or branched. Leaves 2 to 6 inches long, alternate, crowded, entire 

 with callous denticulations, glaucous beneath with the veins con- 

 spicuously reticulated. Eiowers about 1 inch across, very numerous, 

 in long compact racemes ; the lower ones from the axils of bracts 

 resembling the leaves, the upper from small linear leafy bracts. 

 Pistils with a few hairs at the base, at first shorter than the stamens, 

 at length a little exceeding them ; stigmas 4-partite, with the seg- 

 ments at first erect, afterwards revolute. Pedicels ascending, shorter 

 than the calyx-tube, which is slightly curved outw^ards, and about 

 ■J inch long. Sepals strap-shaped, acuminate, free to the base, dull 

 dark purple. Petals deep purplish-rose, the lower ones narrower 

 than the others, and further apart. Capsule 1^ to 3 inches long, 

 curved and spreading when young, afterwards straight and nearly 

 erect. Seeds fusiform-clavate, faintly tubercled. 



I have had great hesitation in reducing var. (d to the rank of a 

 variety, but the characters by which it is distinguished from var. a 

 are such as misjht be caused bv cultivation. Thus the looser soil 

 of a garden would permit it to run more extensively at the root ; 

 this again would tend to produce imperfect seeding, which would 

 account for the shorter pods. Indeed, it has been said that var. 3 

 never produces perfect seeds ; but I have frequently found them 

 full-grown, though I have not tested their capability of germina- 

 tion. The breadth of the lamina of the petals varies much even in 

 the wild plant, and the form of the buds appears to depend on 

 the petals, so that these two characters can only be considered as 

 one. Var. 3, however, appears to be always a taller and stouter 

 plant, the stems being 3 to 6 feet high, and the leaves are broadest 

 near the base, and taper from thence tow^ards the apex. 



Whether it be the effect of cultivation or not, the characters 

 remain constant when the conditions of growth of the two forms are 

 reversed. Var. 3 has grown in Collinton Woods for a very long 

 time, but it still retains all the characters of the garden plant; 



VOL. IV. c 



