130 Notes on the Leaf-cutting Ants of Trinidad. (February, 
these are examined after the insects have been engaged in leaf- 
cutting, they are found full of green leaf juice. Finally we de- 
stroyed the nest by drowning, the common method during the 
wet season. A number of channels were dug in the hillside, all 
constructed to collect the rainwater as it streamed down the hill, 
and to pour it into the nest by one of the entrances. I visited 
the nest during the next rain to see how the plan was working, 
and was surprised to find the water pouring out of an orifice 
twenty yards below the nest. After the rain I examined this 
tunnel and found that it entered the nest at the lowest point, some 
eight feet below the surface. I examined many formicaria subse- 
quently, and invariably found this lower tunnel wherever the in- 
clination permitted its construction. I have no doubt that it is 
constructed as a drain, and that the ants know as much about the 
advantage of thorough drainage as they have been proved to 
know, by many eminent observers, of those of other sanitary 
matters. On opening the mound, some three feet below the sur- 
Fic. 7.—Diagrammatiec v. section of an obey te formicarium, den about 
eight os Sai mound of clayey granules; B, unused. entrance; C, chambers con- 
taining le AN rne corridors; Æ, esens to distant entrances; /, drain 
from koa ae of mi 
face was found a series of hall-like cells, some three feet in their 
larger diameter, connected with each other by short smooth cor- 
ridors. From the outermost of these proceeded the tunnels com- 
municating with the surface by the orifices mentioned above. 
Below there was a second series of somewhat smaller cells, the 
lowest of which was entered by the drain just referred to. 
central chambers were all washed out, but several of the lateral 
chambers had escaped damage. In these were found bushels of | 
leaves, several of the large cyclopean ants, many nurses, larve, 
and an Amphisbena. This lizard is generally a guest of the 
