1. Experiments were conducted to examine the role of the spiracles in discontinuous respiration, to define the kinds of behavior that spiracles can show, and to clarify the manner in which tracheal Po₂ and Po₂ interact to provoke various modes of spiracular activity. To accomplish this, records were made of the movements of the spiracular valves of diapausing pupae and developing adults of the Cecropia. Polyphemus and Cynthia silkworms, in air and in gas mixtures.

2. Spiracular valve movements in these silkworms occur in repeated cycles, with periods of from a few minutes to many hours. Each cycle consists of an open period or spiracular burst (which corresponds to the CO₂ burst), a closed or constriction period, and a flutter period which ordinarily occupies most of the cycle. In pupae with long cycles. the respiratory events occur virtually in slow motion when compared with other insects, and this permits careful analysis of complex events in gas exchange and spiracular behavior which are not readily separable in other insects. Evidence is presented that each spiracular act (fluttering, burst, valve closure) is a response to a specific chemical stimulus: the gaseous composition of the tracheal system.

3. In pure O₂, the flutter period is ordinarily eliminated. As ambient Po₂ decreases, fluttering reappears and the flutter period progressively lengthens, until, in Po₂'s below 15%, the spiracular bursts disappear and fluttering is continuous.

4. Ambient Pco₂ ordinarily has no effect until it increases above 5%, where-upon the cycles shorten. This shortening occurs at the expense of the flutter period, which progressively diminishes as Pco₂ increases. When Pco₂ rises above about 15%, the cycles break down completely and the valves flutter continuously.

5. Intubating even one pupal spiracle eliminates the cycles in the remaining 13, and the valves stay permanently constricted, presumably because normal triggering stimuli for spiracular activity are absent.

6. The spiracles of intubated pupae can be caused to open by lowering the Po₂ or raising the Pco₂ The Pco₂ which opens the spiracles varies with the ambient Pco₂: in a typical pupa, in 2.3% O₂, 5% CO₂ opened the spiracles, whereas in 4.7% O₂, 10% CO₂ was required.

7. From these and other data, it is concluded that the cyclical movements of the spiracles result from cyclical changes in tracheal composition. In particular, fluttering is initiated by low Po₂, while spiracular bursts are caused by high Pco₂.

8. Evidence is presented to prove that the pupa possesses an independent O₂-sensitive mechanism, which is quite separate from any CO₂-sensitive mechanism. It is also argued that low Po₂ does not affect spiracular behavior by virtue of anoxia-produced acidity, nor by virtue of acidity releasing bound carbon dioxide.

9. The spiracular behavior of these silkworms is compared in detail with the picture of spiracular behavior in the flea provided by Wigglesworth, and it is concluded that there is no fundamental difference between the two, except for the flutter period which seems peculiar to the pupa. Evidence is presented that the flutter period holds the key to the disparate rates of gas exchange between bursts, which remains the central problem of discontinuous respiration.