Wishing to extrapolate in vitro observations of granulocyte function and pharmacology made with human cells to animal models of diseases in which we believe granulocyte stimulation to play a major role, we examined techniques for preparation of canine granulocytes and conducted a survey of the function and pharmacology of those cells. Isotonic density gradients of PercollTMproved a simple and highly satisfactory method of preparation. Canine granulocytes in most respects paralleled human cells in function and pharmacology, except that canine cells lacked receptors for formylated oligopeptides and resisted them as stimuli; canine plasma contained a heat-labile inhibitor of canine PMN aggregation, oxidative metabolism, and myeloperoxidase release; canine PMNs were not inhibited in aggregation by protease inhibitors such as aprotinin; canine response to ibuprofen and steroids was more variable than that of human cells, and synergy between those agents was less readily demonstrated; heterologous stimulation (canine cells by human C5a or vice versa) led to a different time course and maximum response from those observed in the homologous systems. Canine granulocytes were readily marked with indium-111, and functioned normally in vitro and survived well in vivo after marking. We conclude that the dog is a suitable animal for studying the role of stimulated PMNs in disease, as long as the observed differences are taken into account in experimental design and data interpretation.