GERANIUM PURPUREUM 99 



therefore impossible at present to refer this name to any shingle form 

 in particular, and it must therefore be neglected, although var. mariti- 

 mum may have been the plant intended. Lange, in Willk. & Lange, 

 Prod. Fl. Hisp. iii. 531 (1878), includes under it both rock forms and 

 seaside forms in Spain (possibly var. maritimum and the glabrous 

 limestone plant ?) ; while Rouy (Fl. Fr. iv. 94: 1897) merely treats 

 it as the red-stemmed form of the type. This latter view is erroneous: 

 see Smith (Fl. Brit. : 1800), " Herba autumno saturate rubet," 

 and Kay (Hist. 1058: 1688) " Robertianum a rubore eaulium & 

 Horum dictum creditur." 



Bouy's " var. littorale" of the "Forme G. purpureum" (op. cit. 

 97 & 98) refers entirely to a plant occurring on the maritime shingle 

 between Cayeux and le Hourdel (dept. Somme), evidently as well 

 known to French botanists as is the Shoreham plant to us. I have so 

 far been unable to obtain specimens ; it will doubtless be one of our 

 south coast forms, probably actually the var. maritimum Bab. It is, 

 of course, not a form of G. purpureum, as the " petales une fois plus 

 longs que le caliee " indicates, while the " rarement presque orangees " 

 added to the " antheres jaunes " in the description of purpureum was 

 presumably added for the benefit of this G. Robertianum form. 



The G. scopulicoluin Jord. in Billot exs. 3519, nomen nudum; 

 Bouy, op. cit. 97, as subvar. of purpureum v&r.genuinum — a plant from 

 the cliffs of Glanville, — is certainly a form of true G. purpureum, but 

 does not seem to be the var. Forsteri. I cannot separate the few 

 specimens I have seen from the typical form, as Bouy's characters do 

 not seem constant. 



Before giving an analysis of Syme's account, I will give a list of 

 the various forms : — 



1. G. Robertianum var. genuinwm Syme. With glandular and 

 jointed hair fruits dark brown and always downy. ? Growing also, 

 on shingle. 



2. Kingsdown form, procumbent on shingle as above, but flowers, 

 smaller and leaves with narrow segments as in var. maritimum. 



3. Glabrous limestone plant. Straggling, fruit almost but not 

 quite glabrous, shoot quite glabrous. Flowers often small. 



4. Var. intermedium mihi. Peduncles and calyces glandular 

 hairy, fruit glabrous. Plant prostrate, leaves like 2. 



5. Var. maritimum Bab. Usually quite glabrous, fruit light 

 brown, quite glabrous, flowers small, otherwise like 3 and 4. 



6. Var. hispidum Druce. Densely hispid all over. Flowers not 

 seen. Fruits small, dark vandyke-brown, nearly or (?) quite glabrous. 



Var. celticum (Ostenfeld ' in B. E. 0. 1919 Bep. 551: 1920) 

 appears to be another form which breeds true, coming between type 

 and no. 3 above. 



7. G. purpureum Vill. emend. Jordan. 



8. Var. Forsteri mihi. 



Syme (Eng. Bot. ii. 203 : 1864) distinguishes three varieties — 

 var. genuinwm (the normal form of G. Robertianum), var. modestum r 

 and var. purpureum. It is clear from his description that his var. 

 modestum is what is here called G. purpureum, for he says its " calyx 

 has the long woolly .... hairs of G. Robertianum, var. a, though 



