98 THE JOURNAL OF UOTANY 



at least in part, the portion of Herb. Roper which contains most of 

 the specimens referred to by Arnold being kindly lent me by 

 Mr. Roberts, the Director of the Brighton Museum. This form 

 remained absolutely constant in prostrate habit, coloration, preference 

 for tbe gravel paths, etc., until G. Robertianwm and Q-. purpureum 

 were introduced into the garden, when intermediate forms (? hybrids) 

 occurred. As it remains constant from seed and is intermediate 

 between the type and var. maritimum it may be called var. inter- 

 medium. I have seen it from Dorset (Chesil beach, Kiddelsdell, 19 L2, 

 see B. E. C. 1912, Rep. 240) : Somerset (Porlock, E. S. Marshall, 

 3174) ; Devonshire (Lynton Foreland) and Glamorgan (Barry). 



A form with the same habit, small narrow segmented leaves, etc., 

 but with fruits as hairy as in the type, occurs at Kingsdown and is 

 the plant mentioned by Syme under purpureum (specimens in 

 H. M.B.): tbe same appears to grow on tbe seashore in Arran 

 (A. Ley, leg. 1883, as var. modestum). I have seen what appeared 

 to be typical G. Robert ianum straggling on bare shingle in E. Suther- 

 land (near Dunrobin), becoming reduced and procumbent but with 

 normal flowers ; and similar plants occur in Shetland (stony shore 

 near Boddam, 1890, Beeby). The Kingsdown plant (which has small 

 flowers) may therefore be merely an extreme state of the type occurring 

 on shingle, but the relationship of all these can only be settled by 

 cultivation. 



The var. liispidum Druce in B. E. C. 1916, Rep. 406 (seen in 

 Herb. Barton), from Berry Head, Devonshire, is remarkably hairy. 

 Could this be G. Rail Lindley ? 



The var. maritimum Bab. Man. ed. 3, 62 (1851), refers in particular, 

 to judge from his herbarium, to the Shoreham form, to which it is 

 here restricted. I have gathered this in quantity on Shoreham beach, 

 where it is astonishingly uniform. The shoot is usually (not always) 

 very red, sometimes practically red all over, the leaves smaller than in 

 type, with narrower and more distant segments. The whole plant is 

 almost glabrous, sub-eglandular, the petals 8 (-9-10) mm. long, 

 rather narrowly obovate, the carpels very pale yellow-brown, ultimately 

 pale tawny, always quite glabrous, with a single wrinkle (excluding 

 the apical crest) or rarely a second reticulation deep enough to be 

 called a wrinkle, the reticulations few, about 4 (3-5) arising from each 

 side of the keel. This appears to be the commoner shingle form in 

 Sussex, and is the var. purpureum of most English botanists; I have 

 also seen it from Dorset, Somerset (E. S. M, 3344), Glamorgan. It 

 is remarkably like the glabrous limestone form, except that the latter 

 has a darker-coloured (r 1 smaller) fruit with a few hairs to be seen on 

 it under high magnification. 



Hornemann's account of his var. rubricaule (in Dansk (Ec. PI. ii. 

 226: 1837) is "var. Red-stemmed stinking Cranesbill (Geran. 

 Robertianum rubricaule) with widespread, at the joints swollen, 

 red-shining stems. This is often found on broad sea-beaches (Strand- 

 bredderne) between pebbles." (I give a literal translation, as I am 

 not well versed in the Danish language.) I learn from Copenhagen 

 that no authentic specimens are traceable in Herb. Hornemann ; it is 



