GONADAL HORMONES AND SOCIAL BEHAVIOR 



1243 



which evoke aggression seem to vary among 

 species and may inchide such as close prox- 

 imity, training, sex hormone, and pain re- 

 sulting from attack. The pattern of response 

 may depend on the stimulus situation, the 

 strength of the stimulation received, and the 

 physiologic state or level of response thresh- 

 old. 



In general, in studies of aggressiveness, 

 conditions are best controlled when the ani- 

 mals are observed in pairs or in small groups 

 and this has long been done by many in- 

 vestigators. Paired encounters were used by 

 Maslow (1936) to determine dominance re- 

 lationships among subhuman primates. If 

 the pairings were made between unac- 

 quainted individuals, they were called ini- 

 tial encounters or initial contacts. These 

 were used by Collias (1943) to determine 

 tlie factors which make for success in estab- 

 lishing dominance in chickens, and by Brad- 

 dock and Braddock (1955) in their work on 

 the fish, Betta splendens. Pairings of mice 

 were used to analyze the effects of thiamine 

 deficiency on fighting success (Beeman and 

 Allee, 1945) and to condition individuals to 

 win or to lose encounters (Ginsburg and 

 Allee, 1942) . Pairings of chimpanzees were 

 made in order to ascertain the effects of the 

 female sexual cycle in female pairs (Craw- 

 ford, 1940) and between mates (Yerkes, 

 1940; Young and Orbison, 1944). Corres- 

 ponding techniques served to test the effects 

 of gonadal hormones in chickens (Allee, 

 Collias and Lutherman, 1939; Allee and 

 Collias, 1940) and of androgen in mice 

 (Beeman, 1947). 



Support for the opinion that staged initial 

 pair contests in neural areas, or cages, give 

 better estimates of levels of aggressiveness 

 than rank in a social order or the frequency 

 and intensity of aggression comes from a 

 recounting of what has been found during 

 work with chickens. In a flock of chickens 

 the frequency of pecks delivered by an in- 

 dividual on others is not a measure of the 

 individual's native aggressiveness. In the 

 determination of a peck-order the tabula- 

 tion of pecks delivered by each bird usually 

 shows no apparent correlation with rank in 

 the social order. Those in the top rank have 

 more individuals to peck and therefore the 

 highest rate of pecking may be expected. 



However, the highest rates of pecking by 

 one bird on another may occur between 

 birds at any level above the lowest ranks. 

 These interindividual interactions have been 

 called "antipathies." Unexpected toleration, 

 i.e., low rates of pecking, also may be found 

 at any level. Antipathies may develop when 

 the flock is assembled as a result of a hard 

 fight, or later when a revolt is unsuccessful. 

 Toleration may follow a passive submission 

 in the initial meeting of unacquainted birds. 

 Furthermore, the rate at which one bird 

 pecks another may vary from week to week 

 according to incidents which arise. The 

 peck-order is learned and the laws of re- 

 inforcement and extinction apply. 



Rank in the hierarchy, or the number of 

 individuals dominated, may be used as a 

 measure of aggressiveness, and may agree 

 with the results of paired encounters (Guhl, 

 1953). However, the reliability of esti- 

 mates based on rank in the hierarchy or the 

 number of individuals dominated may de- 

 pend on the conditions under which the flock 

 was first assembled. At the first meeting 

 the birds engage in initial encounters by 

 pairs which meet at random. Fatigue may 

 set in early for those engaged in lengthy 

 fights, and such individuals may refrain 

 from fighting, or lose subsequent encoun- 

 ters. If new individuals are added after the 

 peck-order is formed, they usually assume 

 low rank. A significant correlation was 

 found by Guhl and Allee (1944) between 

 seniority and social position. These results 

 lead to the conclusion that stimulus situa- 

 tions and other factors (Allee, Collias and 

 Lutherman, 1939) which make for winning 

 encounters need to be controlled and that 

 this can best be done in initial paired en- 

 counters. 



When groups of animals were employed 

 in the study of aggressive behavior the de- 

 sirability of certain practices became ap- 

 parent. Sanctuary (1932) found, when 

 strange hens were added to organized flocks, 

 that the fewer the newcomers in relation 

 to the residents, the greater was the dis- 

 advantage to the introduced hens. Flocks 

 of equal size could be combined with the 

 least disparity. In another study social 

 organization was kept in a state of flux by 

 regularly shifting hens from isolation to a 



