298 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY. _ 
All of these islands extending along the north coast of Vene: 
not only resemble each other very much but they are also like ¢ 
coast of the mainland. There are to be found much the same sp 
constituting the seashore flora, the flora of the lagoon, of the 
cactus-covered hills, and of the few fertile coconut valleys. This 
well illustrated by almost any part of the north coast. Cartipano 
in a long narrow valley with arid hills on each side. Cumani is on 
sandy plain at the foot of the hills. Guanta is in a small valle 
the appearance of a perpetual drought on every side. La @ 
is on a hillside by the edge of the sea and the hill is a brown and 
baked exposure although it is broken here and there by green va 
and by a green mountain rising above. A short way inland k 
still in the coastal region between Caracas and Valencia and a 
the Lake of Valencia trees are scarce or lacking, the mountains a 
brown and clothed only in small shrubs or in dry grass, and int 
valley are scorching sandy plains with here and there the shade of 
tree. Se 
These islands are similar to the coastal land as naturally 
should be, having been in early times a part of the coast and yet t 
is a vast country behind the coast to which they are not at all 
The mountain region of the Andes, anywhere from one tho 
four thousand meters high, the grassy plains of the Orinoco, and1 
forests to the south present features vastly different in every Tes} 
Unfortunately our knowledge of their flora is very limited. 
plants were described as new from Humboldt’s travels, but since 
Out of the six hundred and thirty-four Margaritan pl 
hundred and ninety-five have not been published as occurring 
where else in Venezuela. Inasmuch as many of these are cosmo} 
plants it shows not the peculiarity of the Margaritan flora but 
amount of work that has been done on the mainland. 
