316 MB. G. BENTHAM ON OBCHIBEJE. 



the under surface at the lower end is always highly viscid, aud the 

 whole pollinarium is readily carried oft* by any insect or other 

 foreign hody with which it may come in contact, giving it a fair 

 chance of being lodged on the stigma of another flower. It is 

 that form of pollinarium in which the pollen-masses and gland 

 are separated by a stipes which has been erroneously described 

 a3 pollen-masses with a caudicle and gland, and has led to con- 

 founding it with the true caudicles and gland of Ophrydese and of 

 Calantlie and some other Epidendrea?. 



The above characters appear to draw a very marked line of 

 distinction between Epidendrea) aud Yandea?; and out of 88 genera 

 and above 2000 species of the former, and 130 genera and about 

 1400 species of the latter tribe, there are not, I believe, half a 

 dozen tolerably well-known species as to which there may be any 

 doubt of ihe tribe to wdiich they should be referred. But the 

 division of the Vandeae into subtribes is much more difficult, and 

 as yet very vague in its results. Habit, and even geographical 

 distribution, has often to be more relied upon than any absolute 

 character ; and the eight following ones which we have adopted 

 for the ' Genera Plantarum ' may very possibly, when the doubtful 

 genera become better known, require considerable modification. 



Subtribe 1. Eulophieje. — The three genera here collected 

 are mostly terrestrial, producing pseudobulbs which bear a few 

 long plicately-veined leaves, and leafless flowering scapes proceed- 

 ing from the rhizome at the base of the pseudobulb. In a few 

 species, however, the flowering-stem itself is leafy, and thickens 

 at the base in a pseudobulb. The habit of the subtribe is there- 

 fore that of most Cymbidiea?, from which it is chiefly distin- 

 guished by the labellum more or less produced at the base into 

 a spur, rarely reduced to a short gibbosity. The principal 

 genus, Eulophia, E. Br., including Orthocliilus, Hochst., has 

 about fifty species, chiefly African, tropical or southern, with 

 a few from the Indo-Australian region, and perhaps one true 

 Eulophia from Brazil. The great mass of these species have 

 simple racemes of rather showy flowers on leafless scapes, but 

 two or three South-African ones have the leafy flowering-stems 

 of the American Galeandras ; and a very few, also African, 

 tropical or southern, have small flowers in branching racemes. 

 Blume proposed uniting with Eidopliia, not only the closely 

 allied Qaleandra, but also Zygopetalum and some other Cyrto- 

 podiea', in which I am unable to concur, the structure of the 



