504 ON THE rLOWER-STEUCTUEE IN NAPOLEONA, ETC. 



tion of the cup-shaped thalamus, and demands uo further notice ; 

 nor is it necessary to add any remark as to the pistil. 



As to the aflBjiities of Napoleona^ no doubt the nearest relation- 

 ship is with the Brazilian genus AsteranthuSy as has often heen 

 pointed out. Prom an examination of Spruce's specimens of the 

 latter plant, it becomes evident that the main differential charac- 

 ters of AsteranthuSj as contrasted with JSapoleona^ are the follow- 

 ing : — The calyx is many-toothed (20-30), There is but a single 

 corolla, almost exactly like that of N'apoleona ; there is no corona 

 between the corolla and the pistil, the place of th e staminodes 

 being occupied, as it would seem, by numerous filiform sta- 

 mens. I have not been able to ascertain the point with cer- 

 tainty ; but my impression is that the numerous stamens of Aste- 

 TcmtJius are in the first instance, as they certainly are in Napoleona^ 

 collected together into 5 groups or phalanges, each of which 

 represents a compound stamen, as in Malvacece etc. Moreover 

 in Asteranthus the anthers are truly bilocular, and the disk, which 

 is so prominent in Napoleona^ is very much smaller. The style 

 and stigma are very different in size in the two genera. The 

 fruit of Asteranthus is not yet known. 



It is not necessary to discuss the affinities of this singular 

 group at any length. It has been placed between Cucurbitacecd 

 and Fassijloracecs by Jussieu and Palisot de Beauvois, near 

 Symplocacecd by Desfontaines and Endlicher, next to Ehizoplio- 

 racecd by Lindley. Decaisne and Maout consider that the Napo- 

 leonece approach Combretace(B by their " ovaire infere 1-loculaire^" 

 this is clearly an error, as the ovary is only partially inferior and 

 is five-celled • nor does the affinity seem very close, even if allowance 

 be made for this. In a still more remote degree there is an 

 affinity with Calycantliace(B and Granate(B\yf\Ae even NympTiwacece, 

 in most points widely different, are similar in the great develop- 

 ment of the disk. Lastly, Bentham and Hooker place the two 

 genera among Myrtaeece, which seems the most fitting resting- 

 place of any that have hitherto been proposed for them. 



