The recent revival of interest in the use of 5% or 7% carbogen in the treatment of carbon-monoxide poisoning has prompted the description of four unusual cases in which gross hyperventilation occurred. The investiga tion of these cases was not very thorough, but such cases are seen so seldom that it is felt that even this incomplete report may be of value in stimulating further research. Hyperventilation in carbon-monoxide poisoning was noted by Haldane (1895) in experiments performed upon himself, but Drinker (1938) concluded that " increased breathing is not a prominent feature." He felt that those patients in whom it occurred would receive especial benefit from inhalation of carbon-dioxide mixtures because Haggard and Henderson (1921) had demonstrated a falling carbon-dioxide content of arterial blood in dogs poisoned with carbon monoxide. A low blood carbon-dioxide content may, however, be due to either respiratory alkalosis (as suggested by Drinker) or to non-respiratory acidosis, and the latter condition would be less likely to benefit from inhalation of carbon dioxide. The pH of the blood would distinguish between these two conditions, but this does not appear to have been measured either in animals or in man.