(i.e., widest in the middle) occur in the Central American part of 

 the range (formerly called Epidendrum lamellatum Lindl.), 

 while those with obovate-oblanceolate petals (i.e., widest in the 

 upper third) are known from the South American mainland. 

 This same character separation in the shape of the petals is 

 manifested also in D. elegans, albeit the Central American range 

 does not seem to extend beyond the Chiriqui mountains. For 

 these restricted Panamaniam plants with obovate-oblanceolate 

 petals, the name D. isthmii Schltr. was once proposed. It is quite 

 probable that we are dealing here with a group of plants which 

 currently undergo incipient divergence or speciation on both 

 sides of the isthmus of Panama. Unfortunately the information 

 so far available does not afford a more decisive conclusion. 

 Another species closely related to this complex is D. buenaven- 

 turae which appears to be limited in distribution almost entirely 

 to the Cauca and Magdalena valleys. Dimerandra Rimbachii 

 and D. tenuicaulis are both endemics in western Ecuador, espe- 

 cially along the coastal lowlands. Dimerandra carnosiflora here 

 described is the southernmost in distribution, known only from 

 Peru and the adjacent Brazilian border areas. Dimerandra 

 latipetala is unique in floral characters and is essentially limited 

 to Nicaragua and Panama west of the isthmus. 



ACKNOWLEDGMENTS 



I am ever so grateful for the cooperation shown by the cura- 

 tors of the many herbaria who have so very kindly loaned the 

 materials necessary for this study. My thanks also to the many 

 institutions whose directors have given freely of their considera- 

 ble knowledge to assist me in this endeavor. 



An especial note of appreciation is due to Dr. Leslie A. Garay, 

 Curator of the Oakes Ames Orchid Herbarium, for his many 

 hours of patient assistance and his generous sharing of the her- 

 barium's specimens as well as giving free access to his own pri- 

 vate files. Without his expertise and unfailingly excellent advice, 

 this paper would not have been possible. 



203 



