Since insect blood usually lacks oxygen-carrying pigments it has always been assumed that respiratory needs are met by diffusion in the gas-filled lumen of their tracheal systems. Outside air enters the tracheal system through segmentally arranged spiracles, diffuses along tubes of cuticle secreted by tracheal epithelia and then to tissues through tracheoles, thin walled cuticle tubes that penetrate between cells. The only recognized exceptions have been blood cells (hemocytes), which are not tracheated because they float in the hemolymph. In caterpillars, anoxia has an effect on the structure of the hemocytes and causes them to be released from tissues and to accumulate on thin walled tracheal tufts near the 8th (last) pair of abdominal spiracles. Residence in the tufts restores normal structure. Hemocytes also adhere to thin-walled tracheae in the tokus compartment at the tip of the abdomen. The specialized tracheal system of the 8th segment and tokus may therefore be a lung for hemocytes, a novel concept in insect physiology. Thus, although as a rule insect tracheae go to tissues, this work shows that hemocytes go to tracheae.