2 THE JOUKNAL OF BOTANY 



I have further seen quite recently at the South London Botanical 

 Institute a sheet collected in 1903 near Farthing Downs, Surrey, and 

 labelled " _£/. sfricfa^'' by Townseud, that I can only separate from 

 J£. confusa by its white corollas. 



It thus appears that we have in the hilly, silicious moors of 

 Devon and Cornwall, in the Lake District, and possibly on the cal- 

 careous hills of south-east England and elsewhere, a seemingl}^ endemic 

 Euflirasia^ of dwarf, decumbent and freel}^ branching habit, and 

 generally bearing white flowers, which, however, tend on Exmoor to 

 become more or less yellow. The yellow colouring is evidently'" an 

 unstable character, and not a uniform feature as with most forms of 

 -E. minima of the Alps, where, among thousands of plants in any 

 j)articular locality, the flowers are usually exactly alike in colour. 



As E. confusa was founded on the yellow-flowered plant, the 

 commoner white-flowered one must, if separated, take subsidiary 

 rank ; and it seems best distinguished as a form only, which may be 

 diagnosed thus : — 



E. CONFUSA Pugsl. b. ALEIDA forma nova. 



Planta corollis albidis nee luteis, quam typus interdum rohustior. 

 A plant with white (not j^ellow) corollas, sometimes more robust 

 than the yellow- flowered type. 



If it is thus admitted that these yellow and white-flowered plants 

 are but forms of one species, and that the yellow^ colouring is excep- 

 tional and, when present, of varying intensity, it will be seen that 

 E. confusa shows little aflinit}^ with E. minima Jacquin, and requires 

 careful differentiation from E. nemorosa, to Avhich Townsend origin- 

 ally assigned it, and from E. gracilis. The best distinguishing 

 character of E. confusa, when white-flowered, is its essentially dwarf, 

 decumbent, flexuous and much-branched habit — much dwarfer and 

 slenderer than any form of E. nemorosa of normal growth, more 

 branched and leafy than E. gracilis, and distinctly less erect than 

 either of them. Its leaves, both cauline and floral, are narroAver, 

 with fewer and less acute teeth, than what commonly obtains in 

 E. nemorosa, and their arrangement on the stem and branches is 

 more clearly alternate than in that species or in E. gracilis. The 

 corolla, whether yellow or white, is variable in size, on an average ex- 

 ceeding that of E. gracilis and approximating to the tyj)ical form of 

 E. nemorosa. A feature of the flowers is that in dried specimens the 

 style is frequently exserted. The capsule is generally broader than in 

 the two kindred species, but, although commonly emaj'ginate or retuse, 

 its apex is sometimes truncate or occasionally rounded-obtuse. 



In 1920 I collected on the Lynmouth Foreland a slender, much 

 branched Eyebright, with lavender-tinted flowers, that looks inter- 

 mediate between E. confusa and E. nemorosa. Such plants may 

 probably be found in other localities, and may render it difficult or 

 even impossible to maintain the two separate species. 



E. Kerneri Wettst., under which name Mr. Pearsall sent out the 

 DerAventwater plant referred to above, sometimes resembles E. confusa 

 in its profuse branching and numerous small leaves, but I think it 



