302 MR. G. BENTHAM ON OB.CHIDE3E. 



to Panisea. The labellum is shortly saccate at the base, as in 

 C. prolifera, Lindl., not flexuose, as Reichenbach may have been 

 led to conclude from the sketch in Herb. Lindl., taken from an 

 accidentally injured flower. 



JLcrochane, Lindl., a single species from Sikkim, has the habit 

 of some of the larger species of Bulbophyllum ; but the anther is 

 one-celled at the time of dehiscence, with two globular pollen- 

 masses, as inVandeae. These pollen-masses are, however, con- 

 nected by a bipartite lamina, apparently a pollinary appendage, 

 almost as in Bletieae, and not a production of the rostellum. The 

 genus thus shows some aflmity to each of the three groups, but 

 rather more to Dendrobieae than to either of the others. 



Chrysoglossum, Blume (including Diglyphosa, Blume), four 

 species, and the closely allied Collabium, Blume, one species, all 

 from the Malayan archipelago or the eastern provinces of India, 

 have a peculiar habit connected with that of the Dendrobieas. 

 The anther-cells are very distinct and parallel as in that tribe, 

 and there is no stipes or gland to the two pollen-masses, although 

 in their globular shape they much resemble those of Vandese. 



Subtribe 5. Erie^s. — The inflorescence is lateral, as in Den- 

 drobieae (axillary, pseudoterminal, or on independent leafless 

 scapes), but the pollen-masses, always eight in number, four in 

 each cell, have their points or short caudicles more or less con- 

 nected by a pollinary appendage or viscum, varying in amount, 

 often so scanty that the genera have usually been placed in 

 Lindley's tribe Malaxideae ; whilst in many cases it is so conspi- 

 cuous that Blume, Reichenbach, and others have transferred them 

 to Epidendreae. With the exception of the small American 

 genus Ccelia and a single African Pachy stoma, they are all limited 

 to the Indo-Australian and South-Pacific regions. 



The principal genus Eria, Lindl. (Dendrobium, Blume ; Octo- 

 vneria, Don, not of R. Br.), includes about 80 species, more varied 

 in aspect than most large genera of Orchids, and yet very gene- 

 rally admitted with little diversity of opinion as to the limits 

 to be assigned to it, except as to individual species which have 

 occasionally been proposed as distinct genera, but subsequently 

 restored to Eria by Lindley, Blume, or Reichenbach. Eor our 

 1 Genera Plantarum ' I have adopted the ten following sections, 

 almost entirely the same as those characterized by Lindley and 

 others either as sections or as genera : — 1. Porpax, Lindl., inclu- 

 ding Aggeianihus and Lichenora, "Wight, dwarf plants with the 



