114 



R. A. GRAHAM 



tomentose subsurface. Still (in correspondence with me) regarded this variety as a very 

 hirsute form, but this would seem to be incorrect although hirsute forms with serratures 

 as described above certainly occur. At its best this variety can be easily told by the 

 saw-like leaf edges, but there is every degree of variation in serrature characters, and no 

 distinction can be adequately described between this and forms with fewer serrations 

 except as a matter of extremes. 



Fraser considered the miniature examples from Braunton Burrows dunes to be this 

 variety, a view with which I have never been able to agree, as the number of serratures 

 on the reduced leaves of these small plants seems to be relatively insufficient for agreement 

 with var. denticulata. Furthermore, these specimens are shaggy with hairs. In my 

 view they are more likely to be reduced examples of var. grandidentata, which evidently 

 has fewer though deeper serratures in its normal state than var. denticulata. 



Briquet's var. lupulina, which would appear to have been chosen to replace the 

 earlier, valid name (Braun's various forms are clearly to be regarded as of varietal grade), 

 has a slightly different description from that of var. denticulata. Those who wish to do so 

 can seize the chance to differentiate between them, but it would seem preferable to regard 

 the two as essentially the same sort of water mint, as Briquet no doubt intended. 



(g) var. pedunculata Persoon (1807, 119). 



Hirsuta, foliis suhinciso-serratis, capitulis lateralihus longe pedunculatis sub- 

 fastigiatis. 



This variety probably owes its state to luxuriance. It is told primarily by the long- 

 pedunculate axillary verticils, which would probably appear as lateral capitula rather than 

 verticils. A further character is the subincise leaves. I possess a specimen from near 

 Westerham, v.c. 17 (D. B. Fanshawe), which has subascending peduncles no less than 

 8 cm. long at maturity - these being the longest I have yet seen. But this example 

 differs from Persoon' s description in being markedly subglabrous throughout, and in 

 having very shallow serratures. 



(h) var. purpurea (Host) Perard (1870, 340). 



Plante plus ou moins velue. Tige et feuilles rougedtres. M. purpurea Host 

 (1831, 141). Folia petiolata, suhcordata . . . Folia, caules, rami, hracteae sunt 

 plerumque colore purpureo tincta. 



This more or less hairy water mint is purely a colour form, dependent no doubt 

 for its colour on the chemical content of the ground water. Such highly coloured forms 

 will be met with quite often, perhaps more commonly amongst subglabrous material. 



(i) var. minor Sole (1798, 23-24, fig. 10). 



For a general description see Sole's text, with which his specimens at the Linnean 

 Society and British Museum agree. In this variety, which is quite easily recognised at 

 its best, the stems are usually reddish ; the leaves rather small (up to 4-2 X 2-8 cm.), 

 ovate, acute, hasally rounded or suhcordate, with neat, rather jagged serratures, sparsely 

 hairy above and more thickly so below; the inflorescence a large, showy, terminal capitulum, 

 with or without axillary verticils; the stamens exserted. The large terminal capitulum, 

 especially during anthesis, affords a somewhat topheavy appearance to the plant owing 

 to its surmounting the rather small leaves. Sole's specimen called Mentha aquatica 

 decumhens appears to be essentially the same sort of water mint as var. minor but with 

 included stamens. (Water mints are essentially erect plants, and a decumbent state is 

 presumably due to an artificial factor). Despite its name, var. minor is not necessarily 

 a small plant. One of the Linnean Society specimens is 32 cm. tall. 



