PITTIEK — PLANTS FROM COLOMBIA AND CENTRAL AMERICA. 121 



agree better with what is considered to be C. pedata. On the other 

 hand, specimens of C. pedata in the National Herbarium, with short 

 calycinal appendages and large fruits, are not otherwise distinguish- 

 able from the former. In both forms the fruit is either smooth or 

 aculeate. Cydanthera pedata grows in Costa Rica and Central 

 America generally, but several of the National Herbarium specimens 

 collected in that region and labeled as this species belong more 

 likely to the nearly related C. tonduzii, which differs by its female 

 flowers borne on rather long, stiff pedicels and with densely aculeate 

 ovary, and also by the rounded fruits, rather dry, and always covered 

 with rigid spines. The statement made in the "Plantas usuales de 

 Costa Rica" by the present writer, that the caifas proceed from this 

 plant, may be erroneous, as is the name under plate 13, which should 

 read " Elateriopsis oerstedii." Specimen no. 9787 of the Instituto 

 fisico-geografico (IT. S. National Herba- 

 rium no. 471840), the only one origin ally 

 labeled as caifa, was wrongly identified s 



as C. tonduzii; it is certainly C. pedata ^f^ ~^<\ I C^v 





and it came from San Rafael de Cartago, 

 where it grows abundantly, covering at 2^~~;fy 



times the stone walls and bushy hedges W / 



that surround the truck gardens. That t/v 



the identity of our caifa is as stated is 

 confirmed by the fact that in the same 

 neighborhood Oersted collected his nos. 

 58 and 60 (Copenhagen Herbarium) 



which, according to Mr. CogniaUX, be- Fig. '#>.— Male flower of Cyclanthera tenvr 



long to the same species. Nevertheless, isepalM ' Scate 6 ' 



it is not yet quite clear whether the tender young fruits of C. tonduzii 



are not likewise used as a vegetable. As only a short and incomplete 



diagnosis of this plant has been published, its full description is given 



below. 



In addition to the above-mentioned species, four more have been 

 so far reported from Costa Rica, of which it is not known whether 

 the fruits or the tender shoots are used as articles of diet. Wo give 

 below the distribution, general and special, of the. six species, which 

 can readily be distinguished by means of the following key: 



KEY TO THE SPECIES. 



Leaves pedate or digitate. 

 Primary leaflets 5. 



Fruits oblique, smooth or with soft spines only at base C. pedata. 



Fruits almost straight, always eehinate C. tonduzii. 



Primary leaflets 3. 



Leaflets crenate or slightly lobate. Fruits with short 



peduncles C. naudiniana. 



Leaflets dentate, the lateral ones often 3-lobate. Fruits with 



long peduncles C. langaei. 



