Dr. CORREA on two Genera of Plants, t£c. 219 



ways acknowledged as ftiictly natural. To the fagacity and pro- 

 found fcience of JufTieu we are of late indebted for the conftant and 

 almofl: exclufive characters which diftinguifli this family, and cir- 

 cumfcribe its affinities. He remarked, that the feeds in this natu- 

 ral order contain a crooked embryo without perifperm ; that their 

 placentation is always parietal, in a fruit which in confequence 

 mUffl: be moftly unilocular*. 



The Crateva marmelos of Linne, and the Crateva balangas of Koenig, 

 I have obferved in the herbarium of the Right Hon. Sir Jofeph 

 Banks, and I have received the fruits of both from him. Upon ex- 

 amining the fruits 1 have found that their feeds contain a ftraight 

 embryo with a fmall radicula, and flefhy, large, plano-convex coty- 

 ledons ; and that their placentation is central, in a multilocular 

 fruit: they cannot therefoie be fpecies of Crateva. The further 

 examination of the other parts of their frudlification confirms this 

 firft opinion, and, fhewing how far they differ from the Crateva in 

 other important points, gives us a clue to find their proper place in 

 the natural fyftem. 



But before I proceed to the defcription of the frudlification of 

 thefe two plants, as I intend to deviate in fome manner from the 

 common method of defcribing, I mufl give the reafons which per- 

 fuade me of the utility and perhaps necefTity of the alterations I 

 adopt, and ihow that fingularity, or fpirit of innovation, are not my 

 motives, but that the prefent flate of fcience requires, in fome man- 

 ner, this change of method. 



Of the fix divifions in the Linnacan method of defcribing genera, 

 four relate to the flower, and exift at the fame period, viz. the calyx, 

 coroll, ftamina, and pifiill ; the other two exift after the decay of 

 the preceding, viz. the pericarp and the feed. They are the off- 



• JuJJleu Gen. PI. p. 246. 



F f 2 fpring 



