ULMACEAE 



187 



TREMA Loureiro 



TREMA AM BOI N ENSIS (Willd.) Blume Mus. Bot. 2 (1856) 61 quoad, 

 syn., excl. descr. ! 

 Celtis amboinensis Willd. Sp. PI. 4 2 (1805) 997. 



Sponia amboinensis Decne. in Nuov. Ann. Mus. Paris 3 (1834) 498. 

 Trema virgata Blume Mus. Bot. 2 (1856) 59. 

 Sponia virgata Planch, in Ann. Sci. Nat. Ill 10 (1848) 316. 

 Cortex piscatorium Rumph. Herb. Amb. 4: 125, t. 61. 



Amboina, Hitoe messen, Robinson PI. Rumph. Amb. 328, November 5, 

 1913, in light forests, altitude about 100 meters, locally known as rufut. 



Hasskarl, Neue Schliissel (1866) 84, suggests that Cortex 

 piscatorium may be Sponia timorensis Decne., which is the only 

 previously suggested reduction of the Rumphian description and 

 figure. Sponia timorensis Decne. = Trema timorensis Blume is 

 manifestly closely allied to Trema virgata Blume= Trema am- 

 boinensis as here interpreted (not of other authors), which 

 Lauterbach * confines to Timor Island, with the var. pallida 

 (Blume) Lauterb. in Amboina. The Rumphian figure and de- 

 scription are manifestly applicable to Trema virgata Blume, the 

 figure presenting equilateral leaves which are not cordate at the 

 base and lax inflorescences, while the description definitely states 

 that the leaves are: "ad factum rugosa, sed non Lanuginosa." 

 The actual Amboina specimen cited above, presents a form with 

 rather small leaves, but otherwise agreeing very closely with 

 the figure. 



As to the synonymy given above, it is to be noted that Sponia 

 amboinensis Decne.= Trema amboinensis Blume was based on 

 Celtis amboinensis Willd., and Willdenow's description is very 

 definitely applicable to Trema virgata Blume, not to Trema am- 

 boinensis as currently interpreted. The type was a specimen 

 from Amboina, and the leaves are very definitely described as 

 "scabriuscula" with no mention of the indumentum so character- 

 istic of Trema amboinensis auct., while they are also very de- 

 finitely described as equilateral at the base ; in fact this character 

 is the one on which the species was primarily distinguished from 

 its congeners. It is very evident that Trema amboinensis of all 

 modern authors is not the same as Celtis amboinensis Willd. on 

 which it was manifestly based, but the type has been consistently 

 misinterpreted. Trema amboinensis of modern authors should 

 be reduced to Trema orientalis (Linn.) Blume, at least as a 

 variety; while Trema virgata Blume, generally recognized as 

 a valid species, becomes a synonym of the true Trema amboi- 



* Engl. Bot. Jahrb. 50 (1913) 317. 



