20 



THE CHLOEANTHAL DIVISION. 



FRANKENIACEiE. 



With those botanists who regard this family as in nearest affinity with Caryophyllacece I entirely concur. The leaves are opposite in pairs ; each flower has 2 bracts at its base, not more, the 

 others being only pairs of opposite leaves ; their calyx, corolla with scales, and stamens agree very closely with Cartjophyllacece ; and their marginal placentation occurs in monstrous specimens of 

 Saponaria officinalis (Linn. Proc. vol. 1. p. 159). The embryo of Caryophyllacece although generally, is not always curved, so that the seeds nearly agree, the tubercle at the chalazal end being 

 perhaps analogous to that of Reaumur iacece. 



Reaumuriace^. 



The affinities of this family are it is generally admitted with Tamaricacece and Ficoidece on the one hand, and Hypericacece on the other, but it being desirable to decide if possible to which of 

 these widely different families they approach the nearest, I have minutely examined the genera for that purpose, and to me it appears that they are so near Tamaricacece that they must be stationed 

 as conterminous with them.* Except in the presence of albumen they scarcely differ from them, having the same placentation and also dehiscence, i.e. through the midribs of the carpels. 



And I also believe that they approach very nearly to Caryophyllacece and their allies, of which they have the remarkable and singular scales adhering in pairs to each petal ; and the ovules of 

 Eichwaldia have long funiculi (nearly as long as the ovule itself), bringing to mind those of Mesembryanthemum and Batis. 



The characters in which they correspond with Hypericacece are : — 1, dotted leaves; 2, unequal-sided petals ; and 3, polyadelphous stamens. 1. The glandular dots (or rather appearances, 

 for they are not well defined,) on the petals and leaves are not resinous, as they entirely disappear from petals and almost from the leaves, by being placed in boiling water, which is not the case 

 with those of Hypericacece ; but supposing they did resemble those of that family, it would still be a character of doubtful value, because in JEgiceras, among Myrsinacece, the leaves and sepals have 

 dark-coloured dots like those of an Hypericum, and they are not destroyed by infusion. I have also observed in Crassulacece dots of a very dark colour and irregular form, (but more numerous and 

 better defined than those of Reaumuria,) imbedded in the tissues of the leaf.f 2. The unequal-sided petals are doubtless the same or nearly so as those of Hypericacece, but this character is 

 indeterminate in the same way as the glandular dots, because the sepals of Mgiceras are quite as unequal-sided as the petals of any of the genera of that family. § I have also observed this character 

 very well marked in the hypogynous scales occurring in Crassulacece, besides which it is common to Apocynacece, a family having no affinity with either of these. As the families in Tables III. & IV. 

 are now arranged, Reaumuriacece show the same kind of analogy to Hypericacece as Myrsinacece to Sapotacece. 3. The polyadelphous stamens of Reaamuria are to some extent like those of 

 Hypericum ; the greater part of them are, however, entirely distinct, and the bundles are 5, being equal in number and opposite to the petals, each consisting of 4 or 5, or occasionally 6 stamens 

 united below but distinct above. 



The wood of the Reaumuriacece has remarkable peculiarities, which I believe assist in explaining their affinities. In the wood of Reaumuria hypericoides there are no medullary rays in the 

 first two rings, so that so far it is like that of Frankeniacece, and it also agrees with that of Frankenia Icevis, in having a reddish-coloured tissue between the rings of wood, which are otherwise of a pale 

 lemon colour ; — but in the external rings of its wood the medullary rays are distinctly marked, and in this difference between the central and external rings of wood it agrees with Polycarpcea 

 Teneriffce, Lam. among Illecebracece. Eichwaldia and Hololachna differ from Reaumuria in all the rings of their wood, having well-marked medullary rays, in which they agree with Tamaricacece. 

 The structure of the wood is therefore intermediate between Frankeniacece and Illecebracece on the one hand and Tamaricacece on the other. The wood of Hypericum is of a very different character, 

 being very compact, with numerous small medullary rays like many ordinary dense woods. 



* This notice of Reaumuriaceae was written before the appearance of the < Genera Plantarum' of Bentham and Hooker, but is allowed to remain because I am not aware that the structure of the wood and the character of the polyadelphous 

 stamens have hitherto been described. 



t The glandular appearances in the petals and in the leaves of ReaumuriacecB are very minute and irregular in form, and are only opacities in the tissues, and appear to me decidedly more like those of Crassulacece than those of Hypericacece. 

 § Whether this character is common to all the species of Mgiceras I do not know, but I never saw petals so unequal-sided as the sepals of the species I examined. 



