Occasional Papers of the Museum of Zoology 23 



macropits. The following day J. H. Williamson started col- 

 lecting high up the quebrada where we had left it the day 

 before, and near its source he collected three males of capi- 

 talis, the only specimens of the species taken at Aroa. We 

 found macropus on all the streams. The surprising thing at 

 Aroa was the absence of caja, common a few miles down- 

 stream at Boqueron and thence to the coast at Tucacas. It 

 is probable that the streams, near the hills as they are at Aroa, 

 were too swift for caja. 



6. Bejiima, Venezuela, about 30 miles west of Valencia. 

 Ivies in a circular plain surrounded by high hills. Through 

 this plain the Rio Bejuma meanders in a sandy or gravelly 

 bed, shallow pools alternating with gentle ripples. Alost of 

 the valley is or has been under cultivation, and the native 

 flora is largely gone. The stream is fifteen to thirty feet wide 

 and bordered along much of its course with wild cane which 

 reaches a heighth of twenty-iive feet or more. In the sur- 

 rounding hills are many small quebradas of the usual swift, 

 rocky type, pools alternating with swift rapids and waterfalls. 

 Plant life immediately adjacent to and in these quebradas is 

 usually varied and luxuriant, but, on the steep hills above, the 

 native forest is usually replaced by coflfee and banana plant- 

 ings. The commonest plant in the forest quebradas of 

 Colombia and Venezuela where we have collected is a divided- 

 leaf palm-like aquatic growing from a foot to three or four 

 feet high among rocks in the stream. It may occur as a single 

 plant, as small, scattered clumps, or in a continuous growth 

 filling the stream bed for a hundred feet or more. On its 

 leaves rest many of the dragonflies of these quebradas. From 

 photographs and my description Mr. Ellsworth P. Killip, of 

 the U. S. N. M., has identified this plant as Cyclanfhus bipar- 

 titus Poit. Where these quebradas debouch from the hills 



