APPENDIX. xli 



mises to be ornamental ; tliat one, however,'"^ is a noWe 

 plant, with spherical heads of pink flowers six inches in cir- 

 cumference, and a compact handsome foliage. 



Finally, there is an exogenous plant to which I must par- 

 ticularly allude, in consequence of its very singular structure. 

 This, which seems to be herbaceous, has a round purple 

 stem, clothed near the base with linear striated alternate 

 leaves, and dividing at the top, which is nearly leafless, into 

 a corymbose compact panicle of bright yellow flowers. The 

 calyx is superior and four-toothed ; there are four petals with 

 an imbricate aestivation ; within these are eight stamens with 

 linear bi-locular anthers ; there is no trace of disk ; the 

 ovary is one-celled, and has four membranous wings opposite 

 the petals ; there are four short erect styles, each with a dis- 

 coloured rounded stigma, and the ovules are one or two, 

 anatropous, hanging by short funiculi from the apex of the 

 cavity, one on each side of a slender cord which passes from 

 the apex to the base of the cell. Some of these characters 

 are so much those of Combretacege, that the genus might 

 appear referable to that order, if its habit were not opposed 

 to such an approximation, which its four styles render still 

 more objectionable, notwithstanding the correspondence of 

 its winged fruit with that of Pentaptera. It may also be 

 compared with the genus Quinchamalium, usually referred 

 to Santalacese, with which its unilocular ovary and general 

 habit very much agree, especially if we suppose that in the 

 plant under consideration the cord that separates the two 

 pendulous ovules answers to that which bears the ovules at 

 its own apex in Quinchamalium ; the four distinct styles 

 however, and the absence of a disk, are materially at variance 

 with that genus, although not with some other plants of the 

 same order. It is however among those degenerate forms of 

 Onagracese, known by the name of Haloragece, that the most 

 immediate afiinity of the plant is probably to be sought, and 

 especially with Cercodea, by some authors referred to Halo- 

 ragis ; with that genus it corresponds in having a winged 



(193) PiMELEA spectahilis; foliis ojjpositis linearl-oblongis acutis sessilibus glaucis 

 ramisque glaberrimis, capitulis sphsericis sessilibus multifloris, calycis limbo 

 sericeo : tubo villosissimo, involucri foliolis ovatis acuminatis coloratis. 



F 



