276 



H. G. BAKER 



48. MER. : Betw. Barmouth & Llanaber, 1926, Britton (K); Barmouth, n.d., Bowman (Cam.). 



50. DENB. : Plants probably referable to this subspecies grow between Pensam and Llandulas 

 according to a record by Dallman (1911). 



58. CHESHIRE : Stations will certainly be found in this county where Lord de Tabley (1899) men- 

 tions the occurrence of the species on maritime shingle. 



60. MID LANCS. : Wheldon and Wilson (1907) record G.r. var. purpureum (Forster) " as having 

 been found on the north shore of Bare by F. A. Lees. 



69. WESTMORL. : Slack Head, 1919, Adamson (B.M.). 



74. WIGTOWN. : Stranraer, 1883, Druce (B.M.) (Ox.); nr. Drummore, 1909, Druce (B.M.) (Ox.); 

 Marrock Bay, 1909, Smith (Ex. Hiem). 

 100. CLYDE ISLES : Whiting Bay, Arran. 1883, Ley (B.M.) (WNH) (Cam.); Glen Catacol, Arran, 

 1947, Cuany (Cam.). 



107. E. SUTHERL. : Probably on shingle near Dunrobin (see Wilmott, 1921). 



111. ORKNEY : no details, Gillies (B.M.); St. Ola, Mainland, 1922, Johmton (K) (B.M.). 



112. ZETLAND : Boddam, 1924, Druce (Ox.). 



H. 2. N. KERRY : Records from Rossbehy, Cromane & Castlemaine Harbour for G.r. var. purpureum 

 in Scully (1916) probably refer to this subspecies. 

 4. MID CORK : Carrabinny, 1890, Dixon (B.M.). 



9. CLARE : Ballyvaughan, 1948, Baker (Bak.). Probably also in the Aran Isles (see Praeger, 1909). 

 16. W. GALWAY ; var. purpureum Forst. is recorded from Inishbofin by Praeger (1909) and is 



probably this subspecies. 

 39. ANTRIM : Giant's Causeway, 1890, Dixon (B.M.). 



It is unfortunate that the population at the type locality on Shoreham Beach is now 

 very small, but although plants have been collected there by a horde of botanists during 

 a complete century, it is most likely to be the disturbance of the habitat during the second 

 world war which is responsible for the decline in the numbers of plants. A similar 

 interference has operated along the whole length of these south-eastern beaches and 

 increasing urbanisation is taking a further toll. 



G. ROBERTiANUM subsp. CELTicuM Ostenfeld (1920), Rep. Bot. Soc. & E. C, 5, 551-2. 

 G. rohertianum var. celticum (Ostenf.), Druce (1928), British Plant List, ed. 2, 21. 



Type locality : Calcareous cliffs near Ballyvaughan, Co. Clare, Ireland. 



This subspecies was described by Ostenfeld nine years after he collected fruits on 

 the limestone of the Burren, near Ballyvaughan in western Ireland, during the Inter- 

 national Phytogeographical Excursion in 1911. He grew four generations in Copenhagen 

 and found it to be constant in its characters. Seeds were collected at the same time by 

 Druce who then cultivated the plant in Oxford. 



The most striking morphological feature of the subspecies is its colour. Vegetative 

 anthocyanin is restricted to the region of the nodes on the stems and to the bases of the 

 petioles. The laminae of the leaves have a clear green colour and the rather short petioles 

 have a coppery green tinge. The rather large flowers are paler than in the subsp. 

 rohertianum due to a widening of the white streaks on the dorsal surface of the petals 

 while the anthers, which are usually rather darkly red or purple, show up conspicuously. 

 In material from the Burren the stigmata are pink but in plants from the Glamorgan 

 population the stigmata are white. The fruits are hairy and relatively large (a feature 

 which is correlated with stout fruiting calyces and shorter, stouter beaks). Most of the 

 plants (which are decidedly intolerant of frost) flower as annuals and appear to favour 

 a more humid atmosphere than that tolerable to the other subspecies. 



An interesting difl"erence between Welsh cliff-top material and plants from the 

 Burren is that the former are prostrate in cultivation while some at least of the latter 

 have an ascending habit. As one might expect there are differences in hairiness between 

 the populations, the Welsh material being strongly hairy, the Irish much less so. 



