MR. BENTHAM ON L0GANIACE2E. 87 



one appears to extend over the Moluccas, and two in Singapore. 

 It is possible that some others may exist in herbaria, confounded 

 with Psychotrias and Chasalias. In some cases indeed, there is 

 really considerable ambiguity ; for although the fruit be always 

 entirely superior, the ovary is sometimes semi-adherent. In the 

 plant distributed by Wallich as Pcederia ternata, n. 6248, the 

 ovary is really adherent, but separates so readily from the calyx, 

 especially between the ribs, that, if not dissected with great care, 

 it appears free. I had thus, in a note to the ' Niger Flora,' re- 

 ferred it to Gcertnera, although it be in fact a true Bubiacea, very 

 near to Psychotria, if not a genuine species of that genus, 



A singular anomaly is mentioned by DeCandolle in regard to 

 the G. cdlycina, that two of the stamens are opposite to the lobes 

 of the corolla, whilst the three others are normally alternate. I 

 have only been able to examine unopened buds, in which I cannot 

 perceive any such irregularity ; but it may become developed as 

 the corolla expands, in which case it cannot have any organic im- 

 portance. The large coloured calyx, however, gives to the plant 

 a peculiar aspect, which may justify the maintenance of the section 

 JEtheonema, as proposed by DeCandolle. 



Although I perfectly agree with Endlicher, Blume, and "Wight, 

 in the propriety of uniting Sykesia with Gcertnera, yet the com- 

 paratively short tube of the corolla, the hairs at its mouth, and 

 the somewhat higher insertion of the stamens, may justify the 

 maintenance of two of its species as a section analogous almost as 

 much to Psychotria as to Chasalia, and showing a strong approach, 

 as to flowers, to Pagamea. 



In one of the new species described below, G. rosea, Thwaites, 

 two of the stamens appear to be frequently abortive with short 

 filaments ; but I have had too few flowers to examine to ascertain 

 whether this anomaly is constant or only accidental. 



There are two doubtful genera usually placed at the end of 

 Loganiacece : Codonanthus, G-. Don, which has now been ascer- 

 tained to be a Convolvulaceous plant not generically distinct from 

 Prevostia ; and Anabata, "Willd., so imperfectly described, that it 

 may be equally well attributed to half-a-dozen of the most distinct 

 monopetalous orders. 



The genus Chwtosus, which I had at first referred to Loganiacece, 

 is correctly placed by Alph. DeCandolle in Apocynece. 



Dr. Lindley, in the ' Vegetable Kingdom,' attaches Legnotidece 

 to the end of Loganiacece, as their nearest probable allies ; but this 



