﻿NEVIDSIA, A NEW GENUS OF ROSACEA. 375 



As I have seen no unexpanded flowers, it is barely possible that there may be petals, 

 which are caducous. But as no traces of their insertion can be detected, it is thought 

 that the flowers are really apetalous. 



The manifest stratum of albumen in the seeds of this ^ilant no more hinders me 

 from referring it to the Order Rosacea;, than it woiild from referring Cercis, &c. to 

 the Leguminosce. Nor can I disregard the clear indications of aflinity in this case, 

 because the presence or absence of albumen has appeared to form the sole absolute 

 technical distinction between the Saxifragacece and the Rosacea. If I am correct in 

 this view, there now remains no single absolute distinction between these two great 

 orders, taken in the largest sense. But, between the Rosacea proper (Rosacea;, Endl. 

 &c.) and the Saxifragca;, the absence of all union of the carpels inter se and with the 

 calyx, the fewness of the seeds, and the distinct development of the stipules in the 

 former, contrasted with the manifestly opposite tendencies of the latter in all these 

 respects, seem to afford well-founded distinctions ; atfd these in the present instance 

 outweigh that derived from the presence of a moderate amount of albumen in the 

 seeds. Upon these grounds I venture to suggest a reversal of the position in which 

 Dr. Hooker sets the two connecting links between these orders ; namely, that " Neillia 

 may be safely referred to Saxifragca;, and Astilbe must, technically, fall into Saxifragea 

 also, though its relationship to Spiraa is all but generic." * In my opinion, the re- 

 semblance of Astilbe to S^nnea Anmciis, close as it is, is one of habit, rather than in 

 the floral characters, which are truly Saxifrageous ; while Neillia — really the more 

 ambiguous of the two — is so Rosaceous in the preponderance of its characters, as 

 well as in aspect, that I think it must be remanded to that order, j* 



Intimately related to Neillia is Zuccarini's Stephanandra, from Japan ; — a shrub 



* In Journal of tlie Proceedings of tlie Linnsean Society, 2. p. .54. 



t I could not, indeed, maintain this upon Dr. Hooker's diagnosis of Neillia (1. c. p. 57) ; but, having exam- 

 ined the two species, I may state that I find the calyx perfectly free from the gynaBcium, and the latter in both* 

 species occasionally consisting of two pistils, perfectly separate and free, although closely sessile in the bottom 

 of the calyx. I may add, that the calyx-lobes are evidently imbricated in [estivation, and the stamens not al- 

 ways numerous, being often as few as 15, or even 10, in N. thyrsiflora. The seeds in ray specimens are all 

 empty, so that I have not verified theu- internal structure. 



However the line be drawn, it is evident that no gi-ouping of the orders can be truly natural which does not 

 bring the Saxifragacem and the Crassulacea; into the same alliance with the Rosacea, or into contiguity thereto. 

 This would have the further advantage of approximating the Cunoniece to the Legnotidete, according to the 

 affinities indicated by Brown and confirmed by Bcntham, and also, in bringing Philadelphus and its relatives 

 (purely Hydrangeous genera in my opinion) nearer to MijrlacecB, better satisfy those with whom " no doubt 

 can exist that these plants have a near relation to Myrtleblooms." 



I 



