58 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 122 



tion in A. lineatopus adequately by itself, nor is there a point inter- 

 mediate between them that will describe it if one considers dominance 

 hierarchy and territoriality to be end points of a continuum as Davis 

 (1959) suggests. It would be possible to do as Greenberg and Noble 

 (1944) did in their paper on A. carolinensis and describe certain aspects 

 of A. lineatopus behavior in terms of territory and others in terms of 

 hierarchy; even this approach is not completely satisfactory. It 

 seems to me less important to try to fit observed data into labeled 

 pigeon-holes than to describe that data usefully. 



An approach to the description of the situation in A. lineatopus is 

 suggested in a paper by Allee (1950), wherein he distinguished be- 

 tween two types of hierarchies ("peck right" and "peck dominance") 

 on the degi'ee of predictabiUty of the outcome of disputes. Extending 

 this approach, one can consider agonistic behavior in terms of the 

 kinds of information necessary to predict dispute outcomes. This 

 seems to me to provide the basis of a more general system of descrip- 

 tion and analysis than does a dichotomy between territory and 

 hierarchy. 



Both territory and hierarchy can be described in these terms: ter- 

 ritory as a system in which the outcome can be predicted on the basis 

 of the place where the encounter occurs, and hierarchy as a system in 

 which the outcome can be predicted on the basis of the results of 

 previous encounters. 



To my knowledge, this approach has never been developed in quite 

 the way I am doing here, though every useful generalization about the 

 social behavior of a species can be paraplirased as a prediction about 

 what an individual wiU do under the specified conditions. Ethologists 

 (e.g., Neil, 1964, in his study of Telapia) have used the predictive ap- 

 proach, but they have been interested largely in predicting the next 

 action of an animal during a dispute rather than in predicting the 

 eventual outcome. The cues the ethologists use seem largely to be 

 evidence that indicate the internal state of the animal, which in turn 

 determines what it will do. The cues discussed below seem to be in- 

 volved in actually determining or affecting the internal states of the 

 lizards. 



From the discussion in the preceding section, two factors suggest 

 themselves as bases for predictions in A. lineatopus: relative size of 

 anole and place of encounter. 



Among vertebrates generally these two characteristics, size and 

 familiarity with the site, are undoubtedly among the most important. 

 Probably the only factors likely to rival these are sex, breeding condi- 

 tion, and, in some cases, age. Brown (1963) found that, in SteUer's 

 jays, where the encounter took place was the most important factor in 

 determining (or predicting) which of the two bu'ds would dominate. 



