394 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1955 



shown by simple tests such as displacing a leaf or other surface over 

 which the ants run, which always disrupts traffic until the pathway 

 unit is replaced or the break is rescented. The chemical trail is first 

 laid by ants in the advance group, which indicate their entrance into 

 new terrain by crawling slowly forward in a hesitant, meandering 

 way. Each pioneer moves only a short distance beyond the end of 

 the scent trail, but in this excited advance it rubs its abdomen against 

 the ground, reflexly releasing a secretion (anal gland?) which starts 

 a new section of trail. The trail is thus extended in relay fashion by a 

 succession of individuals, each moving out only a few centimeters 

 before returning to the rear. There are no particular "leaders" in 

 the activity, but each newcomer contributes its bit to the roadmak- 

 ing (Sclnieirla, 1941). 



In a similar manner the path of a hurcheTli swarm is scented by 

 pioneers operating all across the front of the swarm. The difference 

 is that a liamatum pushing party usually contains only a few dozen 

 individuals and a narrow trail is formed directly. The hurcheUi 

 swarm is a large growing body of manj?^ tens of thousands, in its first 

 advance scenting a broad pathway which is finally reduced to a rela- 

 tively narrow trail when the swarm has passed far beyond. This 

 striking difference in raiding patterns no doubt is the product of fac- 

 tors such as different olfactory sensitivity, glandular properties, and 

 general level of excitability of ants in the two species. 



Another important species difference concerns the nature of the 

 prey. The relatively small raiding groups of liamatum do not make 

 the clean sweep of small forest life that the swarmers achieve. Most 

 of the hamatum booty consists of soft-bodied insect young, the larvae 

 and pupae of other ants and wasps in particular. Only when their 

 raids are largest and the foragers most excitable do they capture 

 adult insects in any numbers. Normally, the occupants of invaded 

 insect nests may even escape the column raiders by explod- 

 ing in all directions from the nest exit at the first invasion, carrying 

 their brood off before many of the raiders arrive on the scene. Well 

 after the soldadas have gone, they return. But although the column 

 raiders do not make as clean a sweep as the swarmers and are more 

 selective in their captures, they are far from ineffective. Their mul- 

 tiple treelike systems usually cover an area up to 200° around the 

 temporary nests to a distance of 300 meters or more, and great amounts 

 of booty are taken. In all probability, it is the fact that their types 

 of prey are largely different and competition only partial in this 

 respect that has made possible the coexistence of these two species of 

 wide-ranging carnivorous ants as terrestrial forms in the American 

 Tropics. 



There are two scenes that recur most vividly in my reminiscing 

 about army ants. One is a great swarm raid in full swing with ant- 



