This study was designated to examine the pathogenicity of several strains of Mycoplasma gallisepticum (R, F, S-6, 227 and A5969) and laboratory derived substrains. Preliminary results indicated that the nine M. gallisepticum strains differed markedly in their pathogenicity for chickens. A comparison was made between various in vivo and in vitro methods for quantitative evaluation of pathogenicity. Reproducibility, convenience, and relevance to clinical observations were considered. Two in vivo tests were employed. In one case 2-week-old chickens were infected with M. gallisepticum by aerosol. Air sac lesion score, ability to reisolate M. gallisepticum from trachea and air sac, and serological response to M. gallisepticum were determined 2 weeks post infection. The second test was based on the ability to reisolate M. gallisepticum 3 days after intratracheal inoculation at different dose levels. In this way it was possible to calculate a median tracheal infection dose for each of the strains tested, a parameter which reflected their ability to colonise this organ. Scanning electron microscopy (SEM) was used to examine changes in the surface morphology of the infected trachea and photometric analysis of the SEM lesion was performed. Certain strains multiplied profusely in the trachea of healthy birds without causing detectable pathological changes. Chick tracheal ring (TR) cultures were used for estimating pathogenicity in vitro. In the standardised TR system a method for quantitative evaluation of mycoplasma-TR interaction was devised. The use of the TR system as a model for evaluating in vitro pathogenicity was rapid and less subject to environmental variation than the in vivo tests. The test was economical with respect to time, space and birds.