Spinal skates (Raja spp.), or specimens under Dial, Ciba (0.4 cc. per kg.), were employed. When such means of circulatory arrest were used as exsanguination, ligation of the conus arteriosus, and large doses of pilocarpin (8 mg. per kg.), respiration continued for about forty-two minutes after the blood-flow had ceased. Experimental investigation of the effect of water-flow indicated it to be a primary factor in respiratory regulation as the movements disappeared when water-flow was stopped (they were diminished very much by too great water-flow). Special movements, "gasping", would then appear. Normal rhythm returned as soon as water-flow was resumed. This picture was profoundly altered by perfusion of cocaine solution (0.25% in sea-water) through the mouth and gills. If the water-flow be stopped for a longer period, in about five minutes would appear regular, though infrequent, and often weaker contractions of the branchial arches. The period of time during which these contractions would continue would depend upon the state of excitability of the respiratory centres. Certain experiments carried out on the central nervous system tend to confirm the idea of a segmental arrangement of the respiratory centres. A double mechanism for respiratory control is indicated; a weakly developed automaticity of the centres, and a finely elaborated reflex mechanism, the reaction of which varies with the stimuli received.