By a process of accelerated training, in which 5-aminoacridine is added gradually to an actively growing culture of Bact. lactis aerogenes, strains showing a considerable degree of adaptation to the drug have been obtained in a time interval too short for any appreciable selection of pre-existing mutants to occur. Since 5-aminoacridine is antagonized by acids, the pH changes during the experiments have been carefully followed and it is shown that the adaptation can take place in media which remain neutral, become slightly acid or even become slightly alkaline. The adaptation is usually, but not invariably, unstable and in general is largely lost after one subculture in drug-free medium. If, however, the sequence of accelerated training followed by subculture in drug-free medium is repeated several times the adaptation is gradually stabilized. The results are discussed in terms of theories of induced adaptation and of mutation followed by selection and it is concluded that the latter does not apply. In addition the objections of Yudkin and his associates (1957, 1959) to some earlier accelerated training experiments carried out by Baskett (1952) are shown to be without foundation and certainly do not apply to the present set.