It is difficult to demonstrate the presence of tubercle bacilli in pulmonary tuberculosis in children who are so small that they swallow the expectoration. Positive results are rarely obtained when attempts are made to secure some suitable clots of the expectoration by means of a cotton swab. Thus, the report presented by one of us before the congress in Helsingfors, in 1927, on sixty-two. cases of pulmonary tuberculosis in children less than 3 years old did not include a single instance of the successful demonstration of tubercle bacilli by this method.
In 1898, Meunier1 recommended lavage of the stomach of children during fasting, centrifugation of the water and search for bacilli in the sediment. This method, however, was never employed extensively. In 1927, however, Armand-Delille and Vibert2 reported their results with Meunier's method, which they had improved by homogenizing the sediment before staining it after Ziehl-Neelsen's method. They