416 



NATURAL HISTORY OF PLANTS. 



have mentioned Celastrace/s and Rhamnacece, which are so closely 

 allied to one another. Now-a-days " it is impossible, after all efforts 

 to find on thorough analysis more than two differences between 

 these two groups : the position of the stamens with respect to the 

 petals, and the situation of the raphe. Thus the order Saxi.fr agacem 

 is allied to Rhamnacea and Celastrinea, to the former by Bruniaces, 

 formerly classed in Rhamnece, and now made Hamamelidea? by several 

 authors ; to the latter by D/dongies and several allied genera which 

 were formerly ascribed to Celastrines, and differ from Saxifrages 

 in no essentia] point ; and at the same time by Brexiaces and 

 Pittospores, as we have just shown." Others have approximated 

 Pittosporea to Violariaces, Berberidacea, and Brixacece, which have 

 the same placentation as in most Saxif rag aces. Berberidaces, 

 among others, by its free carpels approaches Dilleniaccs, to which 

 Tetracarpses was formerly referred, and Ranunculaces and Rosaces, 

 whose affinities with Cephalotus we have noticed. "We have also 

 shown that Brexies link most closely Escallonics and several 

 other small series of this order with the little group Pittospores ; 

 and the series Brexies itself has been placed beside Ericaces by Ad. 

 Brongniart, and Berberidaces by Agardh. Ixerba, as well in 

 vegetative organs as in flower, is as near to certain species of Pitto- 

 sporum as to the other genera of its own series, Brexies. It has also 

 been said that between Brexies and Piftosporum " the only essential 

 difference lies in the number of ovary cells, which is the same as that 

 of the petals in Brexies, while most Pittospores have a dicarpellary 

 gynseceum. But this difference disappears in the 4-5 carpellary 

 species of Pittosporum" Now, the approximation we have shown 

 between the Brexies and Pittosperes, on the one hand, and the 

 Saxifrages, with a superior ovary, on the other, leads us to the 

 further affinity of the latter with Celastraces and Rhamnacea. 1 



1 From these numerous affinities it fellows that 

 not one of the characters by which we distinguish 

 the several orders of Dicotyledons is constant in 

 this one. The same may be said of the anatomical 

 structure, which varies with the series or genus 

 under examination. In this respect Cunoniece, 

 Hydrangea, Pittosporece, JSscallonieee, Prexieee, 

 Bauerece, &c, require differential study. The 

 Planes have sometimes been taken for an example 

 of the structure of the wood in Dicotyledonous 



trees. (See Mieb., Mem. Phys. Teg. (1815), t. 

 9, i. — Link., in Ann. Hus., xix. (1812), 340, icon. 

 — Henteey, Microsc. Diet., art. Wood; Elem. 



of Pot., 534, fig. — Schacht, D. Baum, 200 



A. Geat, Introd. to Pot. (1858), 15.— Duchte., 

 Elem. de Pot., 160. In this genus especi.d notice 

 has been taken- of the mode in which the bark 

 peels off (see p. 399, note 5) ; this depends on 

 the presence of sheets of periderm between su- 

 berous plates of unequal thicknesses. The latter 



