ARMY ANTS IN BRITISH GUIANA. 309 



much larger insect than those described above, as large as or even 

 larger than the E. (L.) yraedator queen, which, according to Lueder- 

 waldt, measures 33 mm. The great development of the thoracic and 

 abdominal spiracles in the females and males of the Dorylinse seems 

 to me to be an adaptation to the peculiar conditions in which the 

 sexual phases of these insects live, for when the colony is resting these 

 phases are envelopecUn such a dense mass of ill-smelling workers that 

 some provision would seem to be necessary to increase the supply 

 of oxygen. Emery calls attention to the fact that all the known 

 female specimens of Dorylvs have their appendages more or less muti- 

 lated and surmises that this may result from their being dragged along 

 during the migrations by the workers over the rough ground and 

 pebbles. All the Eciton females I have seen, however, had perfect 

 legs and antennae. The longer and more powerful legs and propor- 

 tionally smaller gaster in these insects, as compared with the Dorylus 

 queens, suggest that the former are not so much dragged as piloted 

 or urged along by the workers when the colony is migrating. On the 

 other hand, the male Ecitons found in columns of workers are often 

 more or less mutilated, i. e. dealated. I have observed this in E. (A.) 

 jjiloswn {vide infra) and Wilh. Miiller noticed it in E. burchelli. This 

 dealation is due not to a tearing of the wing membranes but to their 

 tendency to weaken and break off at the base, precisely as in the 

 recently fecundated female ants of other subfamilies. Even cabinet 

 specimens of Eciion males are sometimes liable to lose their wings at 

 the slightest touch. This, too, is probably an adaptation, enabling 

 the workers readily to deprive their own or adopted males of their 

 wings, thus preventing them from escaping from the colony and 

 facilitating the fecundation of the young queens. 



Eciton hamatum Fabricius. 



At Kartabo this species, though common, is less abundant than 

 bvrchcUi. Its columns are also clearly less populous and aggressive, 

 though the soldiers and larger workers can sting and bite severely. 

 Mann noticed in Brazil that hamatvm " is a timid species in compari- 

 son with some of the others, such as E. vagans." Both bvrchelli and 

 hamatum have white-headed soldiers, wjth very similar long, hook- 

 shaped mandibles (Fig. \(i), but the head in the latter is smooth and 

 shining, instead of opaque, as in the former species. 



