AFFINITIES OF BALANOPHORE^. 23 



proximity to AristolocMcB, an Order with which Balanophoreai have not the slightest affi- 

 nity. The arguments employed by Blume, EncUicher, Lindley, &c., for combining these 

 Orders into one group, are also employed for removing both from Phsenogams; the 

 strongest reason for allying them being, not that they present characters iu common, 

 but that neither of them is considered to be allied to any other known Order of Phte- 

 nogams. 



Of the authors who consider Balanophorem to be Phsenogamic, the majority refer them 

 to Monocotyledones ; Richard, and others following him, placing them in the neigh- 

 boui'hood of Aroidew. 



Griffith places them ia Dlcotyledones, and suggests theii* being the homogeneous em- 

 bryonate group of JJrticece ; to which the structui"e of the Dlstyli is quite opposed : nor 

 indeed does he endeavour to support this hypothesis by any arguments, but merely 

 throws it out as a suggestion. 



In my opinion, the arrangement of the vascular bundles in the rhizomes of Kelosis and 

 Lmigsdorffia is sufficient evidence of these plants being exogenous ; for these, as I have 

 elsewhere shown, are altogether exogenous, differ little from the stems of Menispermece 

 and other anomalous, but still undoubted. Dicotyledons, and resemble no known Endogen 

 in structure or arrangement. * 



In endeavouring to determine the affinities of Balanophorece, I shall disregard the nega- 

 tive characters, as those may be termed which are foim^ded on the imperfection of organs ; 

 and I shall take the most perfectly developed species, as the best expositors of the typical 

 structure of the Order. In so doing, I believe I am obeying a maxim supported by an 

 attentive study of the natural system ; for there are few Natural Orders, however perfect, 

 that do not present structurally incomplete genera or species, many of which in point of 

 development of their organs might rank below many Balanopiliorece and some Cryptogams ; 

 but which, nevertheless, are not departures from the type of their Order, but simply less 

 developed forms of it. 



Balanophorece have an adherent perianth in all the genera where this organ is deve- 

 loped, and an epigynous stamen in Cynomorimn, the only species in which hermaphrodite 

 flowers occur. These characters indicate a position amongst the epigynous Calycijiorce ; 

 a group which, though far from being well limited as a natural class, is in our present 

 state of knowledge one of considerable value, as comprehending many nearly allied natural 

 families. Amongst them, the most direct relation of Balanophorew is certainly with 

 Saloragea, and especially Gunnera ; with which it presents many important characters 

 in common, especially the valvate perianth and stamens opposite its lobes, and near which 

 I would place it in the linear series. 



A detailed comparison of the individual organs of a family so eminently polymorphous, 

 with those of its allies, can alone establish its affinities : the more conspicuous of them 

 arc : — 



1. Between Cynomorimn and Hippuris, in the one epigynous stamen, on an ovarium 

 consisting of one carpel with a solitary pendulous ovule and simple style. 



2. Between the Bistyli and Haloragece a relation is established through the Australian 

 genus Loudonia, which is a very peculiar form of the latter Order, having four styles, and 



