392 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1955 



not in the sense of human leadership ; but the swarm at any stage is 

 "directed" collectively in a complex manner through the activities of 

 all ants participating in the raid ( Schneirla, 1941 ) . 



Very different are the raids of E. hamatum, which I have studied as 

 an example of the Eciton "colunm-raiding" pattern (Schneirla, 1933, 

 1938). This species is a frequent sight in the Barro Colorado forest. 

 Operations in its relatively narrow files are marked by large, shiny, 

 white-headed majors and smaller kliaki-colored "soldadas," the former 

 colliding frequently, the latter overrun frequently, according to the ebb 

 and flow of traffic in two directions. 



This species, also a true terrestrial Eciton^ builds up and carries 

 through its raids in the daytime, and typically emigrates after dark. 

 In the morning, when one follows along in the direction taken by most 

 of the booty carriers, and takes the correct turn at junction points, 

 other things being equal he may be able to locate the bivouac. There 

 are other cues, useful although far from infallible, such as the fact that 

 the base of a "Y" junction usually is on the bivouac side. If no booty 

 carriers are seen, and the junctions are irregular, the problem is more 

 difficult. In the afternoon, and particularly in late afternoon, booty 

 carriers are likely to be numerous, but usually they seem to be traveling 

 in equal numbers in the opposite directions, and the bivouac seems to 

 be hidden behind a screen of f arflung and confused lines of traffic. 



The answer is found when the raid is studied as a developing phe- 

 nomenon. For this the columns must be traced out at regular times dur- 

 ing the day, and transitions investigated appropriately. A hamatum 

 column raid, like a burchelli raid, also generally begins at daylight with 

 radiations in all directions around the bivouac, then soon becomes di- 

 rectionalized. But in this pattern, in the nomadic phase, two or more, 

 frequently three, trail complexes develop, each with its principal 

 trail leading from the bivouac to a peripheral zone of branching col- 

 umns. The sketch in figure 3 represents one hamatum raid mapped at 

 midmorning. Starting from the bivouac, we set out to follow one of 

 the three main routes on which ants hurry to and from the bivouac. 

 Booty-laden ants are almost all returning to the bivouac and for some 

 distance the trail is followed by a single column, unbroken by branches 

 other than short ones in places where the ants continue to raid up 

 trees or in masses of vegetation. Earlier in the morning there were 

 branch columns here in profusion, but most of these dropped from use 

 once the area had been largely worked out. The single meandering 

 base column is likely to continue outward for 50 meters or more, then 

 to have a major division; and in all likelihood whichever branch we 

 follow will have further divisions at shorter intervals as it is traced 

 outward. In effect, the ants have established a complex spreading sys- 

 tem of columns resembling a tree, in which a single trunkline from the 



