THE BALANOPHORAL DIVISION. 15 



Balanophora fungosa (loc. cit.) and that of B. alveolata (Trans. Linn. Soc. vol. XX. tab. 5, figs. 9 & 10), and I have on that account placed them as an apetalous form of that family with which 

 they agree m their syngenesious anthers and albuminous seed. The position of the carpel is inserted on the authority of Dr. J. D. Hooker (loc. cit.), and I should have expected from the analogy 

 between Cynomorium coccineum and the Onagracece where the stamen is single, that the carpel would in that genus be anterior. An examination of the ovule of Lophophytum* confirmed Dr. J. D. 

 Hooker's inference as to the position of the raphe (loc. cit.), as it has all the appearance of being dorsal in that the funiculus is attached to the upper part of the dorsum of the ovule, which has a 

 three-ribbed triangular appearance, one of the ribs being dorsal and two lateral, it being simply flat and rib-less on that side which is next the placenta, and in these characters it agrees precisely 

 with the ovule of Mysodendron. 



Brunoniacece.. 



The structure of the genus Brunonia has attracted much attention, but as its affinities remain up to the present time entirely a question, the following notes, which are the result of repeated 

 examinations, will I trust, not prove unacceptable : — 



The habit of the species is perhaps most like that of Dipsacece with which they particularly correspond in their dichotomous inflorescence, which may be regarded as a capitate form of that 

 occurring in Morina ; and in the venation of their corolla, in which they differ from Composite, they approach Dipsacece very nearly, as in the latter family the segments of the corolla have a 

 midrib, a part of the branches of which are inflected inwards so as to rejoin it; and in all probability also their affinity with Dipsacece offers the most satisfactory explanation of their superior ovary, 

 for in that family we have the ovary quite free from the calyx, except at its apex giving rise to the " flos superus" with " ovarium liberum." 



In all other parts of their structure, except the involucral leaves of their flower, in which they may further be compared with Dipsacece, and the indusiate stigma in which they approach so 

 closely to Goodeniacece that the structures may be regarded as identical, they agree with Compositce being more nearly allied to the Tubulifloral than to the other subdivisions. 



1. The involucre and calyx. Surrounding the calyx are 4 imbricated involucral leaves, to which occasionally a fifth is added, and it may then be analogous to the involucre with five flowers, 

 occurring in Valerianacece in genera connecting that family with Dipsacece, i.e., supposing the central flower to become rudimentary and each involucral leaf to contain a flower in its axil. The calyx 

 although it is not exactly like the pappus of Compositce, yet differs from it only in the simplicity of its segments, for each of them from base to apex is only a slender hairy fibre, scarcely varying 

 in its diameter. 



2. The corolla agrees entirely with that of Compositce except in the venation of its segments, each of which has a central rib from which proceed 1 or 2 branches which form a secondary rib 

 or vein on each side close to the margin, and unite again with the central rib at the apex. In Compositce the central rib is not perceptible, though traces of it may be supposed to exist, and the vessels 

 of each segment consist of only 2 marginal veins which proceed upwards from the veins alternate with the segments and unite at its apex. I would not however in the present instance attach 

 any importance to the venation of the corolla further than that it may serve to show to which of the allies of Compositce, Brunoniacece may be regarded as most nearly approaching in this character, 

 because the venation of the corolla in Calyceracece,f although its segments have midribs, is very nearly the same as that of Composite, so that Brunoniacece and the latter may be regarded as 

 differing in this character in the same way as Dipsacece (the venation of the corolla of which is before noticed) and Calyceracece. 



3. The anthers are so identical with those of the Compositce, especially in their remarkable jointed filament, which is otherwise peculiar to that family and Melastomaeece, and to which I 

 have given careful attention, that from their structure alone a very near affinity would be suggested. 



4. The ovary. The ovary of the Composite I believe is now generally regarded as compound, and is proved to be so by the peculiarly marked characters of the ovary of Centaurea Cyanus 

 and other species (Ann. Nat. Hist. 2nd Ser. vol. 11, PI. XV. fig. 13). The ovary in several of the genera has 4 well-marked lines on its parietes, 1 anterior and 1 posterior, and 2 lateral, (the anterior 

 being the more strongly marked,) so that a question might arise as to whether it consisted of 2 or 4 carpels, but as the stigmas are only 2 it cannot consist of more than 2. In Centaurea the 

 anterior and posterior lines always correspond with the 2 lobes of the stigma, and I find on tracing up the 2 lines in the parietes of the ovary of Brunonia australis through the style (in unopened 

 flowers where it is short) that they distinctly terminate in the centres of the 2 lobes of the indusium; and the ovary may be further compared with that of Centaurea in showing traces of 2 more lines 



* V. fig. L. Weddellii Ann. Nat. Hist. 3rd Ser. vol. 1, PI. VI. fig. 23. 

 f The segments of the corolla in Calyceracece have both the midrib and the lateral veins, but the lateral veins do not spring from the midrib, as in Brunoniacece and Dipsacea, but pass up the corolla at the commissures of the petals as 

 in Compositce. 



