Biometrical problems in the study of early mammalian embryos in vitro

During recent years techniques have been evolved which permit the development of embryonic mouse cleavage stages in vitro. These methods are now sufficiently reliable for use in quantitative experiments such as the mapping of concentration‐response surfaces to discover the interaction of effects produced by various components of the medium. In studies on the development of mouse embryos in vitro conducted in our laboratory, a series of specialized biometrical problems has arisen. In the present paper, the theoretical basis for analyzing the responses of two‐cell mouse ova in vitro is discussed. The responses are quantal in nature and are transformed before statistical treatment. The application of the angular and probit transformation to this data is illustrated. In certain cases, allowances must be made for a control portion of embryos failing to develop. The analyses are illustrated by two examples, one concerned with the determination of optimum conditions, and the other with the estimation of the median effective dose of a metabolic inhibitor.