14 



THE BALANOPHORAL DIVISION. 



DlCLIDANTHEREiE. 



In the Kew Herbarium I find Diclidanthera placed with Styracece to which it is doubtless nearly allied, and it may almost be regarded as a genus of that family with a completely 

 superior ovary, there being no other difference except in the dehiscence of its anthers and the cells of the ovary containing 1 ovule each, in the latter of which characters it agrees with Humiriacece 

 when the ovule is single. In the dehiscence of its anthers it is perhaps nearest to Aptandra, and the ovary (the cells of which communicate above) and ovules agree very closely with those of Humiria, 

 It is also very near Lonchostoma. 



HUMIRIACECE. 



I have no doubt but that Diclidanthera forms a connecting link between Humiriacece and Styracece, as they may be regarded as scarcely differing, except in the petals and stamens of the 

 former being hypogynous ; and this comparison is further supported by its being generally admitted that they are very nearly allied to Olacacece which are a polypetalous form of Styracece. In one 

 of the species of Saccoglottis the anther opens by valves laterally, bringing to mind the Hamamelidece and approaching also Diclidanthera. The cells of the ovary are opposite the lobes of the 

 calyx, in which I find they agree with Diclidanthera, so that possibly this genus may be regarded as a section of Humiriacece. Their barren stamens are remarkable for being either external or 

 internal to the fertile, their narrow terminations in the former case being not unlike those of the external row of barren stamens of Napoleona, which has been compared to SymplocacecB. The 

 station of this family is therefore decidedly in the Epigynous Division, especially as they do not appear to approach any of the hypogynous families except Olacacece, in which however the ovary 

 is in some instances inferior or partly so, and which from their very near affinity with Santalacem are unavoidably placed in the Epigynous Division. The transverse partition, which separates the 

 ovules in Humiria, is most probably the same which takes place imperfectly in Halesia and allied genera, and I believe the ovule is erect in the upper cell and pendulous in the lower as in Halesia. 



Umbelliferje. 



I suggested in the Trans, Linn. Soc. v. XXII., p. 412, that the single carpel posterior is a character not likely to occur in the Epigynous Division, and on that account regarded Vochysiacece 

 as having no affinity with them. But the single carpel in Didiscus cceruleus being always posterior, leaves no doubt that it is a character common to this Division In this plant the ovaries of the 

 central flowers of the umbel are dicarpous, but the greater part of those of the intermediate flowers, and all those of the circumference, are reduced to a single carpel without exception posterior. 



GlJNNERACEJE. 



In the flower of Gunnera there seems to be no evidence to decide whether the ovary consists of one or two carpels, the stigmas being two and the ovule pendulous from the apex of the 

 cell, but on separating the hardened endocarp, which is quite distinct from the external substance of the pericarp, it is found that it is unequal-sided, and that one side is always more convex, 

 protuberant and ridged than the other. In other similar cases, as in Cannabis, this would be the dorsal suture, and if so the raphe is next the placenta ; and I also believe that I have seen the 

 ovule attached to the opposite, i.e., the contracted side of the ovary. The remains of the withered stigmas show that they are opposite these two ridges on the opposite sides of the ovary, so that 

 of the two carpels, one is abortive being represented only by a stigma. 



Balanophorace^e, 



The affinities of this family have been so satisfactorily explained by Dr. J. D. Hooker (Trans. Linn. Soc. vol. XXII.), that I have only to add that I believe that the compound inflorescence of 

 Calyceracece, in those instances where the involucre contains several flowers, may be compared to the compound inflorescence of Balanophoracem (the female inflorescence only) such as that of 



