374 
MERRILL. 
r 
reduces E. roxburghiana Benth. to E, tHphylla DC, including 
most of the synonyms of the two as given by Hooker f. 
Lamarck's description is based on a fruiting specimen col- 
lected in the Philippines by Sonnerat, which is preserved in 
the Herbarium of the Paris Museum of Natural History. Nat- 
urally his original description is incomplete, as the flowers were 
lacking. To the species he also referred Ampdcus angustifolius 
Rumph. Herb. Amboin. 2: 188, L 62, and Rumphius' plate has 
undoubtedly been the basis of interpretation of Lamarck's species 
for most if not all succeeding botanists. The figure very closely 
resembles our Philippine *'Evodia triphyUa/' and may represent 
the same species, but this can be definitely determined only by 
an exhaustive botanical exploration of Amboina, and a critical 
study of the material secured. Sonnerat's specimen must be 
the type of the species, because it is the plant described. The 
reference to Rumphius, which may or may not represent the 
same form as the Philippine plant, cannot logically be inter- 
preted as the type. 
In 1824 DeCandolle * transferred Lamarck's species to Evodm 
as Evodia triphyUa DC, giving only a very short diagnosis. 
This name has been retained until the present time, although 
there is nothing in the original description definitely to indicate 
that the species really belongs in Evodia, Later authors referred 
to it Philippine material, Cuming 1819, which was accepted by 
all botanists considering the species, until very recently K. 
Schumann,^ examining the specimen in the Berlin herbarium, 
noted that it had 8 stamens, and was therefore a Melicope and 
not an Evodia. The Cuming plant, with other Philippine mate- 
rial was designated by Engler as a new species, Melicope luzon- 
ensis Engl., and published under this name by Doctor Perkins 
1. c. 
Shortly after the publication of Melicope luzonensis Engl., I 
examined our Philippine material, which had been referred to 
Evodia triphylla Lam., and found that, so far as flowering 
specimens were concerned, all of the collections had 8 stamens, 
and that all the material was hence Melicope and not Evodia.^ 
Since that time I have examined many Philippine specimens 
in the hope of finding one having flowers with 4 stamens that 
would agree with Lamarck's description of Fagara triphylla, 
but without success. In November, 1911, Doctor C B. Robin- 
i 
I 
' Prodr. 1 : 724. 
• Perk. Frag. Fl. 
• Bur. Govt Lab. 
Phnip. (1905) 162. 
Publ. (Philip.) 35 (1905) 24, 
