3S0 ME. G. BENTnAM OK OBCHIDE^I. 



coriaceous, very rarely thin, never plicate. In some cases the 

 end of the caudex or its branches becomes erect or pendulous, with- 

 out roots and with more closely distichous small leaves. In a 

 few genera these leafy stems are crowded on a very short creep- 

 ing rhizome, or in a few species the whole stem or rhizome is 

 very short, with a dense tuft of roots, and only one or two 

 leaves, or none at all. The peduncles are always lateral, either 

 evidently axillary or apparently leaf-opposed, or breaking irre- 

 gularly through the leaf-sheaths as already explained. The 

 flowers vary, solitary on the peduncles or more frequently race- 

 mose or paniculate. Of the thirty-two genera here included, 

 five are exclusively American, the remainder limited to the Old 

 AVorld ; all tropical, but a few species extend into South Africa 

 or into eastern extratropical Asia. 



Notwithstanding considerable differences in the floral structure, 

 the genera are difficult to define accurately or to form into 

 natural groups. Characters very constant in some instances 

 are very much the reverse in other genera evidently natural. 

 The best arrangement I have been able to devise is perhaps too 

 artificial, the three following series being founded on the pre- 

 sence or absence of a mentum to the perianth, or of a spur to 

 the labellum ; and even this distinction is not always well marked. 

 The mentum may be exceedingly short in some spurless species, 

 and slightly prominent in some long-spurred ones ; and I have 

 reckoned as spurless a few where, although there is no spur at 

 the base, the labellum has a spur-like gibbosity on the back far 

 above the base. 



Series 1. Columna apoda. Labellum ecalcaratum. — We have 

 here first three American genera, Lockhartia, Centropetalum, and 

 Pachyphyllum, with stems ascending erect or pendulous from a 

 short rooting base, closely covered with the persistent sheaths of 

 small distichous leaves. In Lockhartia, Hook., about ten species, 

 and Pachyphyllum, H. B. & K., six or seven species, the pollen- 

 masses, in those dried specimens which I have been able to 

 examine, have appeared to me to differ from those of Vandea? 

 generally in being more acuminate and free, or connected by a 

 small gland. Reichenbach has, however (in Saund. Ref. Bot. 

 t. 76), figured a LocTchartia with a true Vandeous pollinarium, 

 the pollen-masses affixed to a distinct stipes. Lindley thought 

 he had identified LocTchartia with the previously published Fer- 

 nandezia, Buiz & Pav. ; but Buiz and Pavon's figures of the 



