ARMY ANTS — SCHNEIRLA 383 



first this mass is small in diameter, but as more and more ants pour 

 into it the wall spreads outward from the center, and so a symmetrical 

 cylinder results. Since Eclton bivouacs typically grow from the top 

 downward, first in hanging ropes, then joined together as they increase 

 in length, the diameter tends to be largest at the top, and the mass 

 usually tapers inward toward the bottom. E. hurchelU^ in which the 

 workers seem to be stronger in their bivouacking processes, can form 

 taller bivouacs from a surface to the ground, or pouches when the 

 clusters form higher up. I have seen bivouacs of this species at 

 heights of 20 meters or more from the ground, as in the hollows of 

 great trees. The more frequent forming of elevated bivouacs by 

 iurcTielli than by hamatum may be due to the fact that the former 

 species extends its raiding masses into trees in far larger numbers 

 and more frequently than does hamatwm. 



The ecology of the Eciton bivouac is a complex and devious prob- 

 lem. We have tried to understand why Eciton bivouacs, seen in the 

 daytime, generally are found with appreciable overhead shelter, in 

 places less exposed to sunlight and more humid than elsewhere in the 

 general environment. The answer does not seem to be that the first 

 ants to cluster have hit upon a good site through direct sensory dis- 

 crimination of atmospheric differences. At the time when the clusters 

 are actually forming in the evening, various places in the general forest 

 have an equal and minimal illumination and are nearly alike in their 

 temperature and humidity. Investigation of the process leads us 

 to the conclusion that the optimal character of the bivouac is achieved 

 accidentally and indirectly. This is suggested by the fact that the 

 bivouacs are formed near trail junctions, which are established under 

 daylight conditions, when environmental differences in light and dry- 

 ness have some effect upon the behavior of the ants, causing them to 

 avoid places of extreme stimulation. Also, the bivouacs typically 

 hang from elevated surfaces, which are likely to afford some protec- 

 tion to the cluster when it is exposed to the general environment on 

 the following day. 



The interior of the bivouac, where the brood is sheltered and the 

 single colony queen rests, offers an impressively stable environment 

 to these more susceptible members of the community as well as a 

 central resting place for the worker population. The hanging cluster 

 traps a cubic area of atmosphere which does not reach the extremes 

 of temperature and dryness attained by the general forest environ- 

 ment, but in general is somewhat cooler and drier at night and some- 

 what warmer and more humid during the day. This result is achieved 

 rather indirectly, on the whole, although mainly as a result of worker 

 behavior. The indirectness may be suggested first of all by the fact 

 that, when resting, workers cluster more closely together at night in 



