COCKROACHES, ETC. 123 



just to the middle of the abdomen, and consequently 

 these specimens had the thoracic spiracles submerged, 

 but the terminal abdominal spiracles exposed ; the 

 others had the whole abdomen covered with water, 

 but the thoracic spiracles exposed. It was most in- 

 structive to observe that the latter series all died in 

 eight to twelve hours, whilst the reversed specimens 

 were quite lively and well after twenty-four hours, in 

 spite of their constrained position. In the reversed 

 specimens the respiratory movements went on quite 

 regularly, the abdomen rhythmically moving up and 

 down, and air-bubbles issuing at intervals from the 

 thoracic spiracles ; but in the other series of specimens 

 the respiratory movements of the abdomen soon ceased. 



The experiments showed very clearly, first that 

 normally the Water-Cockroaches inspire air by the 

 terminal abdominal spiracles and expire it from the 

 the thoracic spiracles; secondly, that a certain amount 

 of air can be taken in by the thoracic spiracles so 

 that pinioned specimens with these orifices exposed to 

 air live much longer than specimens whose air-supply 

 is entirely cut off. 



I hoped that similar experiments with terrestrial 

 insects would produce similar results, but I was doomed 

 to disappointment. When Stick-Insects of the genus 

 Lonchodes, terrestrial Cockroaches, e.g. Panesthia javanica 

 and Periplaneta australasice, and large Passalid Beetles 

 were experimented with, there was no uniformity 

 in the results obtained. Sometimes the specimens 

 with the abdomen exposed would die before those 

 with the abdomen submerged, and examples of the 

 same species would behave in quite different ways. 

 All the insects struggled violently- at first, and air 



