104 Mr. B. Clarke on the Structure and Affinities of 



tula frequently reduced to 4, 3, or 2. Ovule pendulous, 

 orthotropal, the secundine visible through the foramen of the 

 priraine, from which it is so readily separable as to appear 

 not at all adherent. (PI. VI. figs. 9, 10, 11, 12 & 13.) 



By the ordinary rules, therefore, by which affinities are deter- 

 mined, there exists only a distant analogy between Platanese 

 and the Urtical and Amental families, of which they have so 

 much the habit, their inferior radicle also being without parallel 

 among them, where the seed is pendulous ; and the only fami- 

 lies to which it appears possible to consider them as nearly 

 allied are Tiliacese and Aceracese, with the latter of which, as 

 they exhibit a tendency to a polygamous inflorescence, Platanese 

 may be associated, the habit also being very similar, so much so 

 that an American species has been called ' Sycamore*.' 



And while Aceraceae approach Platanese in their tendency to 

 diclinism, there is also a tendency on the part of Platanese to 

 become polygamous, consisting in the male capitula producing 

 carpels; but at the same time that the carpels are formed, the 

 surrounding stamens become monstrous and distorted, often 

 bearing an ovule, so that none of the flowers become perfectly 

 hermaphrodite; and Platanese agree further with Aceracese in 

 their unsymmetrical flowers, the number of the carpels rarely 

 showing any correspondence with that of the floral envelopes. 



But as the genus Christiania among Tiliacese is completely 

 apocarpous, having only small petals (or barren stamens) con- 

 cealed by the ovary, Platanese may, notwithstanding their near 

 agreement with Aceracese, be a connecting link between Phyto- 

 laccacese and the former, and this would not be inconsistent 

 with their affinity to Aceracese, as Phytolaccacese nearly approach 

 Aceracese through Petiveria and Seguiera. 



They may also be regarded as having nearly the same kind of 

 relation to Proteacese as Rosacese to Leguminosse, agreeing with 

 them in their apocarpous ovary (the carpels when two in Pro- 

 teacese being quite separated) and inferior radicle. Possibly, 

 also, they may be compared with the Abietinse in their crested 

 anthers and pendulous orthotropal ovule, bearing the same rela- 

 tion to them as Casuarina to Ephedra, 



Altingiace^. 



The solitary genus Liquidambar on which this family has been 

 founded has been placed with the Amentaceous tribes, it would 

 appear partly on account of a remarkable similarity in the habit of 



* Most probably P. racemosus. There are certainly more than two 

 species of PlatanuSy if not more than three, as shown by the very different 

 sepals of the male flowers. 



