GERANIUM PURPUREUM 95 



and Fawcett, 1882). Somersetshire: Leigh woods. Channel Islands: 

 Jersey and Guernsey. Waterford : Dungarvan (Britten and Nichol- 

 son, 1882, correctly named Lebelii). Cork: on walls and on a 

 cottage roof (Carroll). A weed in Chelsea Garden (Smith herb. ! ; 

 this is G. Robertianum /3, Fl. Brit. ii. 732 : 1800). It is clearly a 

 plant of southern distribution in this country. 



We have next to consider G. purpureum Forster in E. B. Suppl. 

 264-8, of which Evans says " He figured a plant which appears to be a 

 small-flowered form of G. Robertianum; but the whole matter is so 

 doubtful that For&ter ? s .purpureum would be best entirely neglected." 

 The figure shows peculiar small flowers like those of G. lucidum, a 

 glandular hairy calyx, and a strongly wrinkled fruit. The colour of 

 the anthers is neither shown nor stated, but the other characters 

 certainly indicate G. purpureum rather than G. Robertianum. The 

 original" drawing was first named " G. Rail,'" to which was added 

 "purpureum" " W. B(orrer), Augt. 1, 1829." In the printed text 

 we read : " the specimen here figured was communicated by Mr. Borrer, 

 from Stokes Bav, Hampshire " ; specimens from that locality are in 

 Herb. Borrer at Kew and in Herb. Forster in the National Herbarium. 

 These are certainly a form of G. purpureum, and the specimens from 

 Borrer in Sowerby's herbarium confirm this determination. Forster 

 also says " I have gathered it at Aberystwith, Cardiganshire." _ The 

 specimen of this in E. Forster's herbarium has, however, the fruit and 

 larger petals of G. Robertianum, and is near the var. maritimum ; it 

 should be collected again, as the material is very poor. 



Attached to the Borrer specimens in Herb. Forster is a note from 

 Borrer : " I have found this only at Stokes Bay and at Middleton 

 near Bognor," and another, " Geranium — drawn by Sowerby as 

 G. Raii. How Lindley's character applies I know not, nor whether 

 this or the Shoreham one be Ray's plant." Hearing that Mr. J. E. 

 Little had gathered a peculiar plant at Middleton in Sussex, and 

 remembering Borrer's note, I immediately wondered if the long-lost 

 G. purpureum of Forster had turned up again. I was able to visit 

 the spot in 1919, when, on the bare shingle, I found two patches of 

 plants identical with Forster's figure, the one about twenty yards long 

 consisting of few specimens, the second, fifty yards long, with hundreds 

 of plants. With the exception of a solitary specimen close up to the 

 sand-hills at Littlehampton, I saw no others between Bognor and 

 Littlehampton. The anthers were bright yellow, the fruits exactly as 

 in typical G. purpureum (both glabrous and hairy forms were present), 

 and the plant was clearly a prostrate form of G. purpureum analogous 

 to the var. maritimum of G. Robertianum (the " Shoreham one " of 

 Borrer). I was unable to find any tall and typical G. purpureum 

 on the arable land close by the shingle. I have grown a prostrate 

 maritime form of G. Robertianum in my garden, where it spread to 

 the gravel paths and maintained itself unaltered for seven years. I 

 therefore presume that this peculiar form of G. purpureum would do 

 the same and name it as a distinct variety : — 



Var. Forstert, nom. nov. ; G. pmpureumT. F. Forster in E. B. S. 

 2648 (1830), perhaps including Dillenius's synonym, but excluding 

 Sherard and Villars ; Babington, Man. ed. 3, 62, no. 12, y (1851) 



