GERANIUM PURPUREUM VILL. AND G. ROBERTIANUM L. : I 163 



Fig. 2 shows the distribution in map form. This account of the distribution of this 

 subspecies is far more complete than that given by Wilmott (1921) and differs consi- 

 derably from the overall picture presented by Druce (1932) and Warburg (1952). These 

 differences and the many incorrect references to be found in local floras are due to the 

 extensive confusion which existed in the older literature between G. purpureum Vill. and 

 the prostrate maritime race of G. rohertianum (which was frequently miscalled G. purpureum 

 by British authors prior to 1921 and occasionally since). The originator of this mis- 

 conception in Britain appears to have been Syme (1864). The confusion has been increased 

 by a beautiful photograph of prostrate G. rohertianum growing on shingle near Rye, 

 Sussex, which appears in Sir Arthur Tansley's great work (1939) with the statement 



that " Geranium purpureum Vill is a characteristic species of the south-eastern 



shingle beaches." 



The earliest Sussex record of G. purpureum subsp. purpureum is from Cockbush Com- 

 mon on the west side of Chichester harbour. As pointed out by WoUey-Dod (1937) 

 " Cooper (1834) gives a short description of it a't Cockbush and Middleton (i.e. Clymping 

 Sands) indicating that he found the true species there." However, we do not need only 

 to trust Cooper's judgment, for through the years there have existed unnoticed four small 

 specimens collected on June 3, 1833 by the Rev. G. E. Smith. At one time these plants 

 rested in the Henslow Herbarium and they now form part of the Cambridge University 

 collection. They were collected at the same time as prostrate maritime G. rohertianum 

 and were mounted on the same sheet. Their true nature was appreciated for the sheet 

 is labelled " Geranium rohertianum with G. purpureum. Cockbush Common. Sussex." 

 In view of the fact that no subsequent collection had been made in this storm-damaged 

 area, it is to be feared that the population is now extinct. 



A rather similar story attaches to the other West Sussex record. In this case, erect 

 specimens were collected in 1930 on shingle at Clymping (Middleton) between Little- 

 hampton and Bognor at the station for the prostrate subsp. forsteri (see p. 000) by Mr. 

 J. E. Lousley. Some of the specimens which were distributed give the impression of 

 having grown upright in shade and, when I visited the population between 1945 and 

 1948 only plants shaded by concrete blocks or the wooden groyne were found growing 

 erect. Nevertheless, Mr. O. Buckle (in litt.) tells me that he knows of a small population 

 of G. purpureum with an ascending habit growing near Clymping so, until tests are made 

 on these plants in culture, the record must be taken as correct. 



Subsp. purpureum appears to be absent from Hampshire and very rare in Dorset, 

 despite the first record for England at Swanage (see Wilmott, 1921), but begins to be 

 locally common in South Devon. Many of the numerous localities given by Martin and 

 Fraser (1939) are, however, incorrect, and the species occurs in only four areas, all but 

 one of them coastal. The exception is Hiern's specimen from Drewsteignton. All 

 records from North Devon are erroneous. The Cornish populations are all relatively 

 well-known and this subspecies is probably more common around Newquay and by 

 the estuary of the River Camel than elsewhere in England. 



The erroneous records by E. S. Marshall and others having been discounted, it 

 seemed for a long time that the only Somerset station was in Leigh Woods. Here the 

 species was first seen growing under a wall by Miss A. Carpenter in 1837 whose specimen 

 of a shade form is at Cambridge. It was first recognised, cultivated and commented 

 upon by Moss (1912) and, since then has been repeatedly distributed, by J. W. White 

 and others. Whether this has contributed to its apparently decreased numbers is problem- 

 atical. 



A record for Cheddar Gorge by H. M. Drummond-Hay was backed by a specimen 

 collected in 1871 and lodged in the Boswell-Syme collection at the British Museum 



