Rapid bacteriological study of food-poisoning outbreaks is essential for the prevention of recurrences. Direct application of the coagulase test, Chapman (1944a), to suspected foods and handlers has proved useful for rapid presumptive field investigation, Chapman and Domingo (1946), but the results must be confirmed by cultural methods. The in vivo demonstration of gastroenterotoxigenicity is desirable to add weight to the scientific evidence but it does not lend itself to routine testing of the numerous complaints of "food poisoning" received by health authorities and, because so many strains not associated with outbreaks also produce gastroenterotoxin, the results are not as conclusive as is generally thought [e.g., Fulton (1943) 1. For epidemiological purposes, therefore, a rapid, simple, highly reliable bacteriological procedure would be of inestimable value to the busy sanitarian. The development of such a method will be described here. No claim is made for identification of food-poisoning staphylococci "with certitude,'' indeed absolute accuracy is a biological rarity.Stone (1935) proposed a culture-medium technic by which food-poisoning staphylococci give a clear zone of gelatinolysis around the colonies on a special gelatin-agar medium. As originally applied the method proved to be nonspecific because many strains from outbreaks gave a negative reaction and many pathogenic and nonpathogenic strains from other sources produced considerable zoning. Later, Stone (1938) maintained that, if the test is applied promptly after isolation, nearly all implicated staphylococci show this zoning. Chapman, Lieb, and Cnrcio (1937) pointed out that the reaction is applicable only to pigmented strains. They found that, in addition to the Stone reaction, typical f ood-poisoning staphylococci produce pigment, coagulate plasma, and ferment mannitol and have other properties that are characteristic of pathogenic staphylococci. Other investigators claimed that many f ood-poisoning staphylococci do not give these reactions but a survey of the literature, Chapman (1942), showed that about 95 per cent of the published results of these tests were positive. Most of the organisms that were reported as negative were shown either to have had no established relation to the outbreak or t o have been tested by methods other than those specified. The more recent literature shows a similar high proportion of positive reactions to these tests, Chapman (1947a). I n a series of studies in which the Stone reaction was applied directly to an improved isolation medium [Chapman (1946a) and Chapman and Domingo (1946) ] all strains incriminated in local outbreaks showed considerable zoning in addition to producing pigment, coagulating blood, and 100
