72 ME. bentham's synopsis or legnotidejb, 



from the American Cassipoureas except in habit and inflorescence, 

 in its almost dioecious flowers, and in the rather more dilated 

 stigma. Still the habit is so distinct, that it cannot be joined 

 with Cassipourea, unless, as before mentioned, the whole six 

 genera are united into one. The leaves appear to be deciduous, 

 and have none of the coriaceous texture which all other Bhizo- 

 pJiorecd acquire, at least when full-grown ; the loose cymes are like 

 those of several LythrariecB. The perigynous disk is cupular, 

 as in Cassipourea, but much more developed. 



Daotylopetalum. 



This genus is founded on a single specimen gathered on the 

 island of Nossi-be, ofi" the north-west coast of Madagascar, by 

 Boivin. With the habit, inflorescence, and most of the characters 

 of the American Cassipoureas, I could not include it in that genus 

 on account of the structure of the ovary, which has no dissepiments 

 and only two pairs of pendulous ovules round the central axis, and 

 the stamens are only twice the number of the petals. The fruit 

 is unknown. 



Cassipotjeea. 



This genus must remain restricted to the three tropical American 

 species enumerated in DeCandolle's ' Prodromus.' Additional 

 specimens have proved to me that the C. serrata I described 

 among Schomburgk's plants was founded on a specimen of 

 C. macrophylla with remarkably toothed leaves, and the C. qiiadri- 

 locularis of Spruce's plants belongs also to the same species, which, 

 as well as the two others, occurs occasionally with four instead of 

 three cells to the ovary. The number of stamens varies in all the 

 species, either four times or five times that of the petals ; and in 

 one specimen only of the C. elliptica, the one figured in Hooker's 

 'Icones,' vol. iii. t. 280, does it appear to be generally (but not 

 always ?) three times that of the petals. 



The genus Anisophyllum, Don, or Tetracarpcea of Grardner, which 

 I had formerly considered rather as an anomalous Bhizophoracea 

 than as a Hamamelidea, certainly differs essentially from the former 

 in its alternate leaves, its inflorescence, and more divided styles. 

 Its immediate affinities must, however, remain uncertain until the 

 whole group of small orders clustered round the Saocifragacecs 

 shall have been more carefully reviewed. 



The following is the more technical synopsis of the genera and 

 species : — 



