ARMT ANTS — SCHNEIRLA 403 



at least 70 miles during the interval between marking and recovery. 

 How much longer this queen may have functioned before the inter- 

 vention of supersedure or normal death cannot be said. 



How new colony queens are produced in the dorylines was a long- 

 standing mystery (Wheeler, 1913, 1925). The specific solution for 

 ecitons was obtained mainly through work on the Island. Almost all 

 the broods produced at regular intervals in the colonies throughout 

 the entire year are all-worker broods; however, at long intervals 

 smaller broods of many males and a few queens are produced. Such 

 broods arise most frequently in the first third of the dry season, 

 apparently resulting from the initial effective impact of dry weather 

 upon the queen (Sclmeirla, 1948, 1949). Apparently on Barro Colo- 

 rado only a part of the colonies produce such broods in any one season. 

 As an example, in a total of 52 colonies of Eciton hamatum and 

 hurchelli studied on the Island during 6 months in 1947-48, only four 

 were found with sexual broods. These broods contained no workers, 

 but only males and young queens (Schneirla and Brown, 1950, 1952). 

 It is probable that the queen lays the usual number of eggs in such 

 cases, mainly infertile eggs, but that in the early stages their number 

 is greatly reduced. The total of individuals matured is roughly 

 1,500 males in hamatum and about twice that number in hurchelli^ 

 with indications that only about six of the young queens emerge as 

 callows. The normal nomad-statary process continues in a regular 

 mamier throughout the period of sexual-brood production, essentially 

 as with a worker brood. The one difference is that the nomadic phase 

 is shorter by nearly one week, depending upon a shorter period of 

 larval development in sexual than in worker broods. 



Through long-term studies on the Island, we have established the 

 fact that a two-way colony division is the normal outcome of the pro- 

 duction of a sexual brood by an Eciton colony. This process was inves- 

 tigated in detail on colony '46 B-I (Sclmeirla, 1949) , colonies '48 H-12 

 and H-27 (Sclmeirla and Brown, 1950, 1952) and colony '52 B-I, 

 which were studied through periods of 3 months or more in each in- 

 stance, and in still other cases as well in the different years. Our re- 

 sults indicate that the sexual brood, from its early larval stage, has a 

 strong attraction for a considerable section of the colony. Tests show 

 that this unique attraction is based upon the distinctive odor proper- 

 ties of this brood. The effect leads to a counterattraction competing 

 more and more widely with the drawing power of the regular queen, so 

 that her position in the colony becomes somewhat ambivalent for large 

 sections of the workers. 



Colony division thus evidently develops through conflicting odor at- 

 tractions which arise even while the colony still operates as a miit. The 

 division process is more apparent when the fully developed young 



