sizes may have different food or shelter requirements (Rasa, 1969; Emery, 1973). 

 Several studies have reported that in pomacentrids, small individuals often 

 share the territory of an adult conspecific (Clarke, 1970; Emery, 1973; Sale, 

 et al . , 1980), either because the adult tolerates them or because of "topological 

 deception" (Sale, et al . , 1980) whereby they can avoid the adult by using space 

 in which they cannot be pursued. The low number of chases given and received 

 by these small individuals implies that they are not attempting to compete 

 aggressively with the larger fish for the space they share. Large size in male 

 bicolors appears to be one of the factors associated with high reproductive 

 success (Schmale, 1980). In order to establish the significance of size and 

 how it relates to aggression in this species, detailed studies on the feeding, 

 reproduction, and use of space and shelter are necessary. Account should be 

 taken of fish size, stage of maturation (juvenile, adult; reproductive, 

 nonreproducti ve) , and sex. 



The present study demonstrates both similarities and differences between 

 field and laboratory results on the dominance relations of the bicolor damselfish. 

 Previous work on this species (Myrberg, 1972aJ focused on two distinct time 

 periods: reproductive and nonreproducti ve. The distribution of aggression 

 between colony members was found to differ markedly between these two periods 

 in the laboratory phase of the study. The field phase of the study covered 

 only the reproductive period and was much less detailed than the laboratory 

 phase because of the inability of the underwater camera to record all aggressive 

 interactions between colony members, especially those involving very small 

 fish. The present study is considered equivalent to the nonreproducti ve period 

 of the Myrberg study, since little courtship activity and no egg-guarding 

 behavior by males were observed. 



The present field analysis and Myrberg's laboratory results for the 

 nonreproducti ve period show the following similarities: There is a strong size- 

 dependent linear dominance hierarchy with a very low percentage of reversals; 

 the highest ranking individuals did the most chasing and the lowest ranks the 

 least; the middle ranks received the most chases and rank 1 and the lowest 

 ranks the least. 



However, there were two differences between the two studies. First, the 

 rate of interactions was markedly higher in the laboratory study, in which a 

 rate of 10 chases per individual per hour was recorded as opposed to the 1.5 

 chases per individual per hour in the present study. This is a common difference 

 between field and laboratory studies and may be due to: (a) the inability of 

 small individuals to escape those chasing them in an aquarium, (b) the lack of 

 interspecific interactions in the laboratory, and/or (c) the lack of space 

 available per individual in an aquarium, although both the present study and 

 Myrberg's laboratory study had similar fish densities of 0.2 m 2 (Myrberg) and 

 0.1-0.2 m 2 (this study). 



Second, the relative distribution of chases given by each rank to each 

 other rank below it in the hierarchy differs between the two studies. In the 

 laboratory (Myrberg, 1972aJ, the distribution of chases among individuals showed 

 no clear pattern, except for one fish, YB, which directed successively fewer 

 chases at individuals ranked progressively further away from itself. This 'YB' 

 pattern of chasing was seen very clearly and consistently in all colony members 

 of the present study. 



135 



