General anesthesia has been shown to induce a cephalad shift of the end-expiratory position of the diaphragm in recumbent human subjects. The authors used the technique of sonomicrometry in chronically instrumented dogs to measure the length changes occurring in the costal and crural diaphragm during anesthesia. Seven dogs were studied in lateral decubitus; first awake, and then during pentobarbital anesthesia. The end-expiratory length (LFRC) of the crural segment increased gradually and reached a plateau after 30 min of anesthesia. Costal LFRC did not change. The results were similar when the hemidiaphragm under study was placed in a gravity-dependent or in a non-dependent position. In the awake state, variable levels of post-inspiratory or tonic diaphragmatic EMG activity were observed, which disappeared during anesthesia. The authors conclude that anesthesia induces a 7-8% increase in end-expiratory length of the crural, but not of the costal, diaphragm. This selective adjustment is not due to a pressure gradient effect, but is compatible with a loss of tone in the crural diaphragm.