There is one very remarkable character common to 
all of these species. I refer to the gynostemium which 
has the stigmas conspicuously confluent beneath the ros- 
tellum and not widely separated on lateral lobes or pro- 
jections that have their origin as terminal outgrowths. 
This character is extraordinarily puzzling because it tends 
to eliminate one of the strongest bits of evidence favor- 
ing the separation of Stelis from typical Pleurothallis. 
Furthermore, in the species of this alliance the stigmas 
may sometimes protrude forming a globular viscid mass 
beneath the rostellum, forcing the rostellar process to 
occupy an obliquely erect position. In the illustration 
of Stelis rubens Schltr. var. ovypetala (Schltr. ) Ames, 
this peculiarity is shown in fig. 4. 
In 1925, I received several specimens from Costa 
Rica bearing a strong resemblance to S. rubens, but 
differing from it in the aspect of the labellum, in the 
stouter vegetative parts and more robust and rigid pe- 
duncle. From Stelis Standley: Ames the flowers differ 
most conspicuously in having the eymbiform apical half 
of the labellum strongly upcurved. This difference is 
clearly indicated in the accompanying illustrations. These 
specimens appear to represent still another species of the 
S. rubens alliance. 
Stelis persimilis Ames, sp. nov. Herba caespi- 
tosa. Caules secundarii erecti vel adscendentes, graciles, 
monophylli. Folia oblanceolata vel cuneato-spathulata, 
apice rotundato tridenticulato.  Inflorescentiae  foliis 
multo longiores. Racemi dense secundiflori, graciles. 
Flores atropurpurel. Sepala lateralia elliptico- vel rotun- 
dato-ovata, valde glandulosa, trinervia, extus per medium 
‘arinata. Sepalum dorsale paulo longius et angustius. 
Petala rhombica, breviter mucronata. Labellum in cir- 
cuitu lanceolatum; parte inferiore incrassata cum. lobis 
[14] 
