398 ME. P. TOWNSEND ON AN EBY.THB.EA 



On an Erythreea new to England, from the/Isle of Wight and 

 South Coast. By Pbedebtck Townsend, M.A., P.L.S. 



[Read December 2, 1880.] 



(Plate XV.) 



In the ■ Journal of Botany ' for October 1879 I drew attention to 

 an Erythreea which I had lately found on the downs of Freshwater 

 in the Isle of Wight, and to which at that time I was unable to 

 give a name, though inclined to think it might be Erythreea capi- 

 tata, Willd. I hesitated to refer it to that species, because my 

 plant differed in some respects from the longer description given 

 by Kcemer and Schultes in their ' Systema Vegetabilium,' and I 

 was unacquainted with Willdenow's plant either in the living or 

 dried state. 



On first gathering the Erythreea in the island, I was much 

 struck with its habitual dwarf character, in which it approached 

 a dwarf and capitate variety of Erythreea Centaur ium, which also 

 grew abundantly in the same locality. I soon found it easy to 

 distinguish the former by its narrower leaves, by the shortness of 

 its corolla-tube, and, above all, by its almost free stamens, the 

 filaments of which, without exception, I found to be attached only 

 at the base of the corolla-tube, and to be otherwise perfectly free 

 within it (compare fig. 3i and fig. 6). This latter character 

 seemed to have been hitherto unobserved in any species of the 

 genus Erythreea ; and indeed it would take the plant out of the 

 genus, as described by Grrisebach in his monograph of the 

 order Gentianacese, given in the ' Prodromus ' of De Candolle, 

 where he says "Stamina 5-1, corolla tubo superiore inserta^ (DC. 

 Prod. ix. p. 57). 



It has been suggested to me both by Dr. Svine and Dr. Ascher- 

 son that the comparatively free stamens might represent a ten- 

 dency to, or "the last remains of," dimorphism, like Menyanthes 

 and Limnanthemitm in the same order, or like Erimula &c. ; and 

 if this were the only character by which the plants could be distin- 

 guished, or that a similar tendency could be traced in other spe- 

 cies of the genus, I should myself be disposed to give some weight 

 to the suggestion ; I have, however, carefully searched for other 

 examples in the same genus, and have examined numerous speci- 

 mens, especially of the following species, E. Centaurium, E. pul- 

 chella, and E> Uttoralis (E. linartfolia, Pers.), in all of which I 



