RELATION BETWEEN PREY AND PREDATOR — ROEDER 30I 



concerned in prey location and capture will undoubtedly be a com- 

 plex undertaking. However, the behavioral studies mentioned above 

 lead to a number of definite postulates which Mittelstaedt is attempt- 

 ing to test by electrophysiological methods, A parallel morphological 

 study of neuron relations in the prothoracic and head ganglia of the 

 mantis is also essential to a coherent picture of the mechanisms of this 

 interesting segment of behavior. 



DISCUSSION 



The achievement of rapid response times at the expense of much 

 central nervous space occupied by a few giant nerve fibers is com- 

 mon throughout the invertebrates. The well-known giant motor axons 

 of cephalopods provide an example. Giant internuncial fibers mediat- 

 ing the withdrawal reflex in annelids have also been intensively in- 

 vestigated (Bullock, 1948, 1953) and a similar situation exists in 

 the Crustacea (discussed by Wiersma, in press). Thus, some of the 

 neurophysiological experiments related above have been performed on 

 other organisms, and the general significance of giant fiber systems 

 in the avoidance of predators is widely recognized. 



In spite of this there have been few attempts to relate neurophysio- 

 logical information of this kind to the actual performance of adaptively 

 significant behavior in the intact animal. Some large fraction of neuro- 

 physiological endeavor is rightly directed to establishing a physico- 

 chemical basis for nerve function, but it seems to me that we now 

 have sufficient information on the behavior of nerve cells (see Eccles, 

 1957) to justify a more intensive effort to find a neural basis for the 

 behavior of animals. This was probably the original object of nerve 

 studies, and it seems to be still valid today. 



One of the difficulties is due to the fact that many sensory inputs 

 are available to direct the behavior of the intact animal. While only 

 one, or a few of these, direct behavior at any given moment, others 

 may take over in short order and in any case play an indirect role by 

 determining the excitatory state of central neurons. A response 

 mechanism with many potential inputs presents a formidable tech- 

 nical problem to the neurophysiologist. For this reason the startle 

 response offers a better chance of analysis since one sensory input 

 overrides all others during its performance. 



Another complication lies in the instability of most behavior. This 

 is apparent even in the startle response of the cockroach. If a puff 

 of air is applied a few times in quick succession to a previously undis- 

 turbed group of cockroaches the startle response wanes noticeably 



