100 THE PHYSIOLOGY OF INSECT SENSES 



in the anterior trunk. Some species respond by a complete cessation of 

 movement, a 'freezing'; some respond by 'freezing* followed by 

 thrashing. The response probably serves as a protective device. It re- 

 mained to Minnich (1925, 1936, 1937) to attempt a physiological 

 investigation of the response. He studied seven species of butterfly 

 larvae and eight species of moth larvae, bringing to a total of twenty- 

 eight the number of lepidopterous larvae known up to that time to be 

 sensitive to sound. It is likely that all caterpillars detect sound. The 

 range of frequency sensitivity extends from 32 to 1,000 c/s. By cutting 

 larvae into fragments and noting that the fragments responded, 

 Minnich showed that the receptors for sound were diffusely distributed 

 throughout the body. 



Early experiments were confined to relatively hairy species, and in 

 these the responses to sound were abolished when the hairs were 

 loaded with flour dust or droplets of water. Upon removal of the dust 

 or water, responsiveness returned. A steady stream of air directed 

 against the hairs inhibited response. These experiments support 

 strongly the idea that the hairs are the sound receptors. On the other 

 hand, similar behavioural responses to sound were obtained with 

 'hairless' species. Although even these species possess minute setae, 

 Minnich (1936) entertained doubts that sound reception here was in- 

 deed mediated by hairs. The question is still open. 



In the belief that the sensory hairs of the hirsute caterpillars were 

 resonant structures, that is, that different lengths and sizes would reso- 

 nate to different frequencies of stimulus, Minnich (1936) attempted to 

 fatigue the response to one frequency and test for a response to 

 another. He found, in fact, that fatiguing at low frequency inhibited 

 responses to all higher frequencies, but that the reverse was not true. 

 As Pumphrey (1940) pointed out, these results are inconsistent with a 

 hypothesis of reaction by resonance. He pointed out further that the 

 results obtained in the experiments on fatiguing are explicable on the 

 grounds that low-frequency stimuU will be much louder to the animal 

 than those of equal intensity but higher frequency and would therefore 

 probably have a greater fatiguing effect. This interpretation is based on 

 a study of action potentials recorded from cereal hairs of cricket and 

 cockroaches (Pumphrey and Rawdon-Smith, 1936 a, 1936 b, 1936 c). 



The cerci of cockroaches and Gryllus and many other Orthoptera 

 are clothed with extremely delicate, long, lightly hinged hairs (Sihler, 

 1924). They are typical sensilla trichodea. In Periplaneta americana 

 there are several hundred on each cercus. Each hair is about 0-5 mm. 

 long and i 0-005 mm. in diameter. They move visibly to light puff's of 



