24 NATURAL H [STORY OF PLANTS. 



The family thus understood "by concatenation" has many 

 affinities. By the isostemonous Uuoni/mece with ascending ovules, 

 it approaches the Penceacece, from which we shall find it differs 

 chiefly in the organisation of its gynsecium ; and the Bhamnacece, 

 from which we shall distinguish it by one absolute character— the 

 isostemony of the latter, with oppositipetalous stamens. By the 

 Bmew, and also by the Hippocratem and EuonymecB^ it presents 

 great resemblances to certain Euphorhiacece. But in the latter, 

 which never have more than one or two descending ovules in each 

 cell, the micropyle is directed upwards and outwards, whilst, in the 

 corresponding cases, it is interior and superior in the Celastracew. 

 The latter also approach the Staphylece (which are Sapindacece) and 

 the Ilicinect} ; but these last have been rightly referred to the families 

 with gatnopetalous corolla ; and the former, closely allied as they are 

 to the Celastraceoe, especially the Hippocrate(B, are distinguished 

 from them either by the. independence of their carpels, by their 

 composite leaves, by the organisation of their fruit, by the form 

 of their floral receptacle, and consequent mode of staminal insertion, 

 by the isostemony of their androecium, or especially by several 

 of these characters combined. The impossibility of establishing, by 

 one or more technical characters, an absolute difference between the 

 Celastracece and the various groups withwhich we have just compared 

 them, arises from the fact that they themselves have not a single 

 character which is not sometimes wanting. When their ovules are 

 definite in number, they are ascending with the micropyle primarily 

 exterior, or descending with the micropyle interior, but they may be 

 neither descending, nor ascending, nor definite in number. Their 

 floral receptacle is often convex or plane, and the insertion is then hy- 

 pogynous ; but the receptacle may, here and there, become extremely 

 concave ; ' which entails the" perigyny of the perianth and androe- 

 cium. Their aerial branches are ordinarily woody ; ^ but this character 

 may sometimes be wanting in the exceptional type of Stackhousia. 



^ As in Mortonia, and, to a less degree, Fer- und Schling. Ffl. Tuting. (1827), § 76. On that 



ro«ert«, incliiding C«j-yospen»e«m, of wMoh it has of Emnymus: Lindl. Introd. i. 213. Olitee 



been rightly said that they are Mhamnacea, {Stem Dicot. 25) says that the organisation of 



except that their stamens are alternipetalous. the woods of Salvadora deserves the attention 



2 The structure of the Celastracece is espe- of hotanists. We have pointed out in our 



cially interesting in the climhing species, as Monogr. des Buxacece, the structure of the 



Gelastrus, where we have seen the woody axis tranches of Sarcomcea (7), of the stems of the 



divided into three lohee, the separation being Boxes (8), of the rhizomes Paehijsandra (10), of 



indicated externally by furrows spirally crossed the roots, leaves, etc. (t. 2, fig. 1-12). On the 



(A. Jtjss. Malpigh. 117). On the stem of C. Sox, see also Schacht, Ber Baum, 196. 

 scandens, see H, Mohl, Veb. d. Bau der Banken- 



