Poma can thus arcuatus) or had shapes similar to that of the jacks 

 (parrotf ishes, Scaridae; goatfishes, Mullidae). These fish moved 

 relatively slowly, and this may have accounted for the absence of 

 observed responses. On one occasion when a grey angelfish made a 

 rapid dart toward the reef, the Chromis did give a distinct escape 

 response. 



Sound or water-borne vibration caused escape responses on some 

 occasions. Objects dropped on the metal barge overhead or on the 

 bottom of a few meters from the school provoked distinct escape 

 responses, but the fish paid no attention to the continuous sounds 

 of passing motorboats. There were some escape responses whose 

 causes were not apparent. They may have been correlated with 

 passing mackeral or barracuda at a distance of 3 to 7 meters, but 

 not all such events caused responses. On the other hand, they 

 could have been caused by sounds not audible to the divers. 



CONCLUSIONS 



The absence of a difference between reaction distance for the clear vs 

 the light opaque models could be interpreted to mean that sound plays 

 a major role in evoking the responses. But it could also mean that 

 Chromis easily saw the clear model (as did the divers) which was thus 

 not a good control for sound. The results of the strings model point 

 toward the latter explanation. Several other results demonstrate that 

 vision does play a role in evoking the escape responses. One is the 

 independence of reaction distance and speed of the model. If noise 

 generated by the model triggered the escape, a faster model which 

 generates more noise, would cause a reaction at greater distance. 

 We have not measured noise as a function of model speed, and we can- 

 not set limits on the extent to which sound might contribute to the 

 response. The significant difference between reaction distance for 

 dark and light models can only be interpreted to implicate visual 

 involvement. The models of different color were identical with the 

 exception of color. Further experiments might elucidate the 

 importance of sound and vibration in triggering escape. 



To a diver, the contrast between the light models and the background 

 (either grey coral sand, light surface, or slightly darker water) was 

 less than the contrast between the dark models and the background. 

 We cannot tell whether it is the darkness of dark models, or their 

 greater contrast which made them more effective at evoking escape. 

 A more complete grey series would have to be evaluated to allow this 

 distinction. 



The trend toward earlier response with increasing model size could 

 have been an area effect, but probably not a maximum diameter effect. 

 This is because the small ellipse was of the same maximum dimension 

 as the medium circle, but the latter was more effective (this was not 

 a highly significant result, but was suggestive). 



VI-207 



