OF BATIDE^, CALLITRICHACE^, VOCHYSIACEyE, AND CASSYTHE^. 413 



of petals in Daphnacece ; and it is doubtless among the polypctalous families that the 

 more immediate relations of Vochysiacea are to be sought for, — the vicinity of Rosacea, 

 Le(juminos(B, and their allies (which are regarded as polypetalous forms of Daphnacece 

 and Froteacece) appearing to be the true station of this family. 



The relation of tlie VochysiucecB to Chrysohalanece confirms this view of their affinities, 

 as they may be regarded as differing from Chrysobalanete only in the carpel when single 

 being posterior instead of anterior ; for it is on the posterior side of the flower that the 

 stamens are deficient in both families, and the calyx in Vochysiacew is sometimes so 

 nearly regular tliat the spur is scarcely apparent. 



To the Amygdalece they are doubtless equally allied, agreemg with them in the ovules 

 when suspended having the raphe next the placenta * ; but as the flowers in this family 

 are always regular, the approach between them is less obvious. 



To the Rosacea} they may be compared in the remarkably convolute iBstivation of the 

 coroUa ; to Fomaceai in the occasionally inferior ovary, the o\ailes when erect having the 

 raphe next the placenta, and in the cotyledons being either flat or convolute; and to 

 Cahjcanthacece in the imbrication of the calyx, which occasions the flower-buds to look 

 like leaf-buds ; and in their large anthers. 



3. Note on Cassythese. 



The affinity of this family having been for some time well known to be with the 

 Lauracece, I have only to add that it agrees with them in the ovary consisting of a single 

 carpel t, as is shown by a furrow on one side of the style, which is also of a paler colour 

 (fig. 29), and by the ovule being attached to the same side of the ovary, so as to be in rela- 

 tion with the fui-row in the style, the ovule being pendulous, not from the apex of the cell, 

 but from one side near the apex, as in Lauracecs {vide Ann. Nat. Hist. ser. 2. vol. xi.). 



The carpel is also variable in its position in nearly the same degree as in the Lauracew, 

 being occasionally posterior, and otherwise variable from anterior to lateral ; and it agrees 

 also with them in the ovule having the raphe dorsal (fig. 29), so that there does not appear 

 to be any distinction between them except in parasitism. 



* In those genera where the ovules are numerous, they are suspended, amphitropal, with the raphe next llie 

 placenta, and therefore the foramen is uppermost. 



t The ovary of Sassafras officinale consists of a single carpel {vide Ann. Nat. Hist, iud ser. vol. xi.), but occa- 

 sionally, although very rarely, it becomes dicarpous, with an imperfectly bifid style and parietal placentee, consisting 

 of two ribs on the opposite sides of the ovary, which are alternate with the styles. The ovaries of Laiinis, Ciimti- 

 momum, and Tetranthera have also decidedly the appearance of consisting of but one carpel ; but Sassafras being 

 occasionally dicarpous, makes it probable that Nees v. Esenbeck is right in describing Lauracea: as tricarpous, as 

 when the ovules are two, and attached separately, the ovary would most hkely be compound. 



