It may be of interest that the divers subjective evaluation of fish 

 responses led them to the conclusion that the model's speed was highly 

 important, but that shape, size, and color were marginally important 

 if they had any effect at all. This is probably because the fish 

 reacted more synchronously, and swam faster when a fast model approached. 

 The great variability of response distance masked the more subtle 

 effects of color and size. Speed of the model affected parameters of 

 the response which were clearly seen by divers, but did not affect 

 the distance at which the response was initiated. 



The possibility that Chromis recognize some shapes and respond differ- 

 entially to different ones is not ruled out by our shape series. In 

 this series, no natural shapes were used. The novelty of the square, 

 circular, and round models could have been their overriding character- 

 istic. We did run a model with the area of the medium series, and 

 with the basic shape and color of a bar jack. But most of our films 

 of this experimental comparison of natural amd unnatural shapes were 

 not possible to analyze, or produced extremely meager and variable 

 data. There is need for more careful assays of the role of shape and 

 pattern in evoking the escape response. 



There is little question that we could use models to study parameters 

 affecting an escape response in Chromis in a quantitative way. There 

 is little reason to doubt that the response we studied was the same 

 type that is evoked by some natural occurrences. There was reason- 

 able correspondence between the sizes and speeds of fish and models 

 which evoked response. Our models were most often run at speeds 

 exceeding those of all but a few species of local fish, but the 

 independence of reaction distance and model speed over such a range 

 as we found would have led us to predict an extension of the principle 

 to lower speeds. In addition, some fish did not provoke responses 

 which we would have predicted, and some fish which did not approach 

 sufficiently close to evoke responses (according to our experimental 

 data) did cause escape responses. We suspect that more complex 

 visual and auditory cues associated with such intruders are recognized 

 by the Chromis and influence their escape responses. 



VI-208 



