SULLIVANTIA OHIONIS. 41 



celled ovary ; from Saxifraga, by its pentandrous flowers ; and from 

 Boykinia, by its less adherent calyx, persistent petals, and very 

 short stamens, to which I may now add the imbricative aestivation 

 of its corolla. For in B. aconitifolia the petals are convolute in 

 aestivation ; which character, if it shall be found to hold good in the 

 two Oregon species (a point that the advanced state of my speci- 

 mens does not allow me to verify), will abundantly confirm the 

 genus Boykinia.* 



Tab. VI. Sullivantia Ohionis, of the natural size. Fig. 1. A cymule, 

 in fruit. Fig. 2. A flower. Fig. 3. Same, with the calyx laid open. 

 Fig. 4. Cross-section of an unripe capsule. Fig. 5. Vertical section of the 

 same. Fig. 6. A seed. Fig. 7. Longitudinal section of the same, display- 

 ing the embryo. All the analyses more or less magnified. 



• This discovery strengtiiens the view I had formerly ventured to take, in ap- 

 pending Pldladelphem to the order Saxifragaceae [Fl. N. Amer. 1, p. 594), 

 although other botanists think that there is only " some collateral relationship " be- 

 tween them. Excepting in the more numerous stamens, Philadelphus differs from 

 SaxifragacesB-HydrangeEe only in the valvate calyx and convolute estivation of the 

 petals ; the very characters which are unexpectedly exhibited in a true Saxifragea 

 by Boykinia. Of course I follow De Candolle and Zuccarini in referring Deutzia 

 — which has valvate petals and definite stamens— to the suborder Hydrangea; 

 as also Decumaria, the petals of which, I believe, are not imbricated (as de- 

 scribed by Endlicher), but valvate, with induplicate margins, like Deutzia, and 

 which is very closely related to Schizophragma, Zucc, an undoubted Hydrangea- 

 ceous genus. (Vide Fl. N. Amer. 1, p. 593.) Lindley, however, in his Vegetahh 

 Kingdom, still comprises both Deutzia and Decumaria in his order Philadelpha- 

 cese, although the ordinal character he assigns suffices to exclude them. 



