The existence of ventricular fibrillation can be recognized with certainty only by direct inspection of the heart or electrographically. Ventricular fibrillation in the normal dog''s heart is a progressive condition from which spontaneous recovery does not occur. Restoration of a coordinated rhythm can occur only after every vestige of fibrillary activity has disappeared. The normal duration of fibrillation is generally more than 15 min.; often it exceeds 30 min. As the central nervous system does not survive for so long, revival can not be expected if fibrillation is allowed to run its natural course. Fibrillation of the ventricles can be abolished artificially by use of KCl, perfused through the coronary system via a carotid artery and the aorta, but only provided injections are given under high pressure. Fluid injected under low pressures has a tendency first to fill the arterial dead space and then to drain from the arterial system by the large branches and imperfectly closed semilunar valves, with the result that the animal becomes water-logged before an efficient coronary circulation is established. After cessation of fibrillation a coordinated beat can be restored by similar per-fusion of an isotonic solution containing an excess of Ca. Since coronary perfusion demands the sacrifice and liga-tion of a carotid artery under aseptic conditions[long dash]a performance requiring rather elaborate perfusion apparatus, etc.[long dash]it is frequently not adaptable to practical needs. Ventricular fibrillation in the dog can also be abolished promptly by injecting 1 cc. per kgm. of a 5% KCl solution as divided doses into cavities of each ventricle. Massage is not required to produce these effects. Injections into both ventricles, however, appear to be essential to produce a prompt arrest. Effective doses introduced into the right ventricle delay the arrest of fibrillation so much as to make their practical use in resuscitation of doubtful value. The large doses of KC1 employed appear to have no permanent toxic effect when followed by equivalent doses of Ca, as in our procedure. Subsequent injection into the ventricular cavities of a solution containing 5% CaCl2 and 0.1% hepa-rin in doses equal to 1 cc. per kgm. causes a resumption of a coordinated beat provided the fluid is pumped through the coronary vessels by massage. Without massage it is ineffective. No chemical substitute for massage has been found as a result of extensive experimentation. While the methods offer some promise of success when applied promptly to patients after electric shock, hope of resuscitation in man by methods so far devised is necessarily limited to cases in which the surgical possibilities and circumstances approximate those obtainable in laboratory animals. The fact that cardiac massage is required is a serious drawback to the practical use of the intracardiac method in the field.