246 REVISION OF THE NATURAL ORDER HEDERACEJ. 
libus.— Gastonia, Comm. in Juss. Gen. 217 ; Lam. Dict. ii. p. 610 ; De 
Cand. Prodr. iv. p. 256. 
The genus Gastonia has been incorrectly described. De Candolle 
(* Prodromus’), to go no further back, assigns to it from 5—16* petals, 
and double the number of stamens, two stamens being said to be placed 
before each petal. Decaisne and Planchon describe the petals as 3- 
nerved. All three are wrong. The stamens in this, as in all other 
genera of the Order with definite stamens, alternate with the petals, 
and are equal in number with them; and they are always l-nerved. 
What has given rise to the belief of their being 3-nerved is that they 
are very narrow, and are more or less coherent, in some instances not 
separating at all from each other, so that the corolla appears to have 
fewer petals than it really has. The normal number of petals and 
carpels seems to be 10. But there are sometimes fewer by abortion, or 
more by excess. 
Gastonia is closely allied to Grotefendia, a genus also inhabiting the 
islands of Eastern Africa. Both have imparipinnate leaves and very 
narrow petals, but they are easily distinguished from each other, viz. :— 
Gastonia. Pedicelli inarticulati. Flores ecalyculati. 
Grotefendia. Pedicelli articulati. Flores calyculati. 
l. G. eutisponga, Lam. Dict. ii. p. 610; De Cand. Prodr. iv. p. 
256; G. spongiosa, Pers. Ench. ii. p. 20 ; Bois d'éponge, incolarum.— 
Mauritius (Sieber! Fl. Maurit. Exs. ii. p. 197, Carmichael! Bouton! 
in Herb. Kew.), growing in woods, abundantly in Mount ** Pouce.” 
. Species dubia, indescripta :—G. saururoides, Roxb. Cat. Hort. 
Calc. p. 70; Gilibertia saururoides, De -Cand. Prodr. iv. p. 256, in 
the Moluccas. 
Species excluse :— 
G. aculeata, Hortor.— Cuphocarpus aculeatus, Dene. et Planch. 
G. Candollei, Hortor.— Brassaiopsis speciosa, Dene. et Planch. 
G. dentata, Hortor.= Brassaiopsis speciosa, Dene. et Planch. 
G. longifolia, Hortor.— Brassaiopsis speciosa, Dene. et Planch 
G. Nalugu, Lam. (Gilibertia Nalugu, De Cand.)— Leea doa. 
PS teste Wight et Arn. Prodr. ;= L. sambucina, Willd. teste Miq. 
. (?) Oahuensis, A. Gray, Bot. Wilkes, p. 726, is the type of a new 
genus. : 
G. palmata, Roxb.— Trevesia palmata, Visian. 
* “16” may possibly be a misprint for “6,” else the number would be 32. 
