a new Natural Order. 411 



opinions, but to examine, with proper caution, the grounds upon which they 

 stand. Now the result of such a labour will no doubt, in many cases, strike 

 at the root of ideas to which age and custom give a strong colouring of 

 truth, and which however are not the less really misunderstandings of nature. 

 Thus, to quote the only example of that prejudice which belongs to my pre- 

 sent subject, who does not follow Jussieu in considering Acerince as im- 

 mediately connected with Malpighiacece ? Yet neither habit nor characters, 

 but merely a deceptive resemblance between the winged carpels of some 

 Malpighiacece and those of Jeer, is the ground on which that connexion is 

 founded. Now while such a trifling circumstance, which is almost in all 

 cases only of generic value, is there the object of an exclusive attention, the 

 real signs of the affinity of Acerince with Sapindacece seem to have escaped 

 notice. Of these marks the most important, because the most general, is the 

 position of the disc between the stamens and the petals, — a character which 

 may be observed in all Sapindacece (including Hippocastanece) as well as in 

 Tropceolece (including Limnanthece), and also in Melianthece and Geraniacece ; 

 that is to say, in all the orders which, with the addition of Cochlospermece, I 

 consider as members of the same natural class. 



Having thus traced the outline of the groups with which Melianthece may 

 be compared, and having fixed the place of that order in the last-mentioned 

 class, it remains to justify that opinion by more circumstantial details. First, 

 as to the facies, — a new species of Natalia is so strikingly like some Paul- 

 linice, that I have alluded to that resemblance by giving it the specific name 

 Paulliniotdes. Moreover, Sapindacece include poisonous plants, the leaves 

 of several Paullinice and Serjanice, like those of the Magonice, being used to 

 intoxicate fish ; and among them the Paullinia australis, A. St. Hil., being 

 suspected by M. Aug. de St. Hilaire to be the plant which communicated to 

 the honey of the Lecheguana wasp the noxious effects which that distin- 

 guished traveller has related from his own perilous experience*. Now ana- 

 logous properties n)ay be supposed to exist in Melianthus, judging from the 

 strong narcotic and virous smell of the whole plant; and even a like induc- 

 tion might perhaps make us extend the suspicion to many of the Geraniacece. 

 As to floral characters, the contrast of the quaternary proportion of the 

 * See Aug. de St. Hilaire in M6moires du Mus., vol. xii. p. 893. 



