290 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 1 37 



escape from, predators it is obvious that in tliese cases a short reaction 

 time has much greater sur\-ival value than detailed information. A 

 millisecond or so within the nervous system must often mark tlie 

 difference between the quick and the dead. 



The adaptive value of speed at the expense of detail in such alarm 

 and escape systems cannot be questioned, but one may well ask why 

 large fibers, such as those in the cockroach ner\-e cord, occupy such 

 a disproportionate amoimt of central nervous space when compared 

 with analogous alarm systems in vertebrates. All vertebrate nerve 

 fibers capable of high-speed impulse conduction are surrounded by 

 thick segments of myehn interrupted at regtdar intervals by nodes of 

 Ranvier. In insects the lipid material surroimding the largest axons 

 is so thin that it can be detected only by special staining and optical 

 techniques. Myelination, for reasons which will not be debated here, 

 appears to improve conduction velocit}' about tenfold when fibers of 

 similar diameter are compared. For instance, at room temperature 

 locust motor fibers lo to 13 microns in diameter conduct at 2.2 meters 

 per second (Hoyle, 1955a), cockroach giant fibers of 30 to 40 microns 

 conduct at 6 to 7 meters per second (Roeder, 1948), while myeUnated 

 frog axons only 7.5 microns in diameter conduct at 25 meters per 

 second (Taylor, 1942). For some reason myelination has never ap- 

 peared in insects. Hence, their simple but %'ital alarm systems must 

 occupy a large amount of space in their nervous systems if reaction 

 times comparable to those of vertebrate predators are to be achieved. 

 Quality of information is sacrificed to speed, which implies a limita- 

 tion or economy in the number of nerve units unless the central 

 nervous system is to occupy a disproportionate amount of space in an 

 organism which for other reasons appears to be limited in total body 

 size. 



This parsimony of neurons makes insects excellent subjects for 

 that branch of neurophysiolog}- which seeks to pro\-ide a neural basis 

 for behavior. The presence of a smaller number of units imphes that 

 the central nervous system must be a simpler communications net- 

 work compared with that of vertebrates and therefore more suscep- 

 tible to eventual analysis, even though at the same time it is obviously 

 an effective system for integrating ner\-e activity- into behavior of 

 some complexity and variet}-. Other technical ad\'antages of insects 

 for this sort of work are the accessibility of the ner\-ous system, its 

 division into anatomically distinct and partially autonomous ganglia, 

 and the ease of maintaining ganglion function without special per- 

 fusion. But let us turn to the question of the transmission of informa- 

 tion in systems containing a minimum of separate pathways. 



