118 NATURAL HISTORY OF PLANTS. 



number of the carpels of which the gyntecium is formed ; one in the 

 latter and two in the former. And even this character is not absolute. 

 It is indeed exceptional that one cell and one ovule is observed in 

 the Thymelcœ ; i but in certain species of Phaleria, a genus of 

 Aquilarieœ, there is nearly as often one ovarian cell and one ovule 

 as two. 2 



This last character shows us plainly enough that, if the affinities 

 of this family with the Lauraceœ, Hernandieœ, Protaceœ were alone 

 perceived by early botanists, it is because they had to compare with 

 them scarcely any but the Thymelece — that is, types with uni- 

 carpellar gynsecium and parietal placentation ; 3 but we must 

 now inquire to what families the TJiymelaceœ ally themselves 

 by their highest types, those whose gynajcium is formed of 

 more than one leaf and contains two cells, complete or incom- 

 plete. These are the Penœaceœ, the Rhamnacece (especially the 

 Colleticœ) and the Celaslraceœ. The 'tube' of the flower, we have 

 ropeatedly said, appears to have the same morphological significa- 

 tion in the Thymelaceœ and in the plants of these families which 

 have a perigynous andrœcium. But in the Rhamnaccœ and in 

 those of the Penœaceœ which have in each cell only two ovules, the 

 latter are always ascending. Moreover, the Rhamnaccœ are dis- 

 tinguished by their oppositipetalous stamens, and the Penœaceœ, by 

 the quite special organization 4, of their ovarian partition. The 

 Celastraceœ are generally hypogynous; and, in this case, they 

 nearly approach, by their entire floral organization, one of the 

 Thymelaceœ scarcely perigynous, such as Octolepis. But in those of 



1 Peddiea has been cited as sometimes having present day to understand the opinion of 

 two ovules in one and the same cell, and De authors who placed them among or after the 

 Martius has seen two or three ovules and as Thymelece, perhaps on account of their induviate 

 many stones in Funifera utilis. fruit. The Troteacece frequently have one 



2 H. Bn. Adansonia, xi. fasc. 10. The gy- ovule like the Thymelece; it is then either 

 nsecium of Aqmlaria is sometimes tricarpellar. orthotropous and descending, or anatropous and 



3 The Lauracecc are everywhere distinguished ascending ; which is never seen in the latter, 

 from the Thymelece : 1. hy the perianth formed The stamens, always the same in number as 

 of two or more verticils ; 2. by the character, the divisions of the perianth in the Proteaceœ, 

 quite peculiar, of their valvicide anthers; 3. by are superposed to those divisions, whilst in the 

 the position of the micropyle, which, in the isostemonous Thymelaceœ (except in Schœno- 

 descending ovule, is interposed between the biblus, a genus still imperfectly known) the 

 hilum and the placenta. The Jlernandicea are stamens alternate with the sepals, and, when 

 Lauracece, and have besides, as we have seen they are opposite, as in Pimelca, they are 

 (vol. ». 449), a double perianth, free stamens fewer. 



and an inferior (adherent) ovary, surmounted 4 See p. 96, fig. 60-63. 



by an epigynous disk. It is difficult in the 



