2 PROCEEDINGS OF THE jS'ATiONAI. MUSEUM vol. 66 



unknown genus, all are rather coriaceous dicotyledons. By far the 

 most abundant leaves are those of the white mangrove and the 

 MvrniLsops. 



At least three of the forms — the buttonwood {Conocarpus), white 

 mangrove {Bhisopkora), and black mangi'ove {Aviceunia) — are 

 members of the mangrove association, and indicate more or less tidal 

 muddy coastal swamps. These were in all probability estuary in 

 position which is where they usually find their optimum conditions 

 of growth. The Mimusops^ which is described as new, finds its 

 closest living homologies in forest species. This might be taken 

 to mean that their leaves were river borne, but since they are the 

 most abundant forms in the Oropouche clays it would seem that 

 they must have been growing near at hand in the beach jungle be- 

 hind the mangrove swamps or in the lower valley of the supposed 

 stream that made the estuary. Such an environment would be the 

 natural one for the other members of this flora. 



All of the plants are lowland humid tropical types. All the 

 existing species recorded as fossils occur in the existing flora of 

 Trinidad and those which are described as extinct, have closely re- 

 lated existing species Jn Trinidad and the adjacent coastal region of 

 South America. 



Class MONOCOTYLEDONAE 

 Order ARECALES 

 Family ARECACEAE 



PALM RAY 



Plate 1, figs. 1, 2 



Tlie single basal part of the ray of a fan palm is the only repre- 

 sentative of this class of plants found in the Oropouche clays. 



These rays are linear lanceolate, markedly inequilateral proximad, 

 ■where they are contracted to a petiolar-like attachment on the rachis. 

 The venation appears to be characteristic, consisting of 8 principal 

 regularly spaced longitudinal veins with 3 or 4 thinner parallel veins 

 in each interspace. 



It might be possible to connect the fossil with some recent Triridad 

 palm, but in view of the fragmentary nature of the fossil and the 

 uncertainties involved, it did not seem worth the labor of searching 

 through herbaria in which palms are usually so incompletely repre- 

 sented. 



Type.— Cat. No. 37017, U. S. N. M. 



