Calculations are given illustrating appropriate and inappropriate methods of statistical evaluation of data involving effects on offspring of experimental treatments given to pregnant rats. Failure to recognize the cluster effect in experiments of this nature can give rise to erroneous utilization of statistical procedure.Despite an increasing sophistication in the use of statistics by biologists, in several recently published papers, involving prenatal or early postnatal manipulation, the authors have not presented sufficient information as to their statistical procedures to permit the reader to make an unambiguous evaluation of their findings. In studies of the effectiveness of a treatment, the statistical significance of the results is customarily evaluated by comparing the mean measurement of the treated group with the mean measurement of the control group by the use of a t-test of the form 81 -j t 2 st/l/nl t l/nz t = where El and K2 are the means of the 2 groups, s is the appropriate standard deviation, and nl and n2 are the appropriate sizes of the 2 groups. In some published studies, the appropriate s, n l , and n2 do not appear to have been used, with the result that the

No abstract

No abstract

When tested by the cotton‐pellet method in rats or the tuberculin‐hypersensitivity test in guinea pigs, griseofulvin had a marked antiinflammatory action. The activity was, weight for weight, less than that of the corticosteroids tested concurrently. The anti‐inflammatory action of griseofulvin appears to be independent of any effect upon the pituitary‐adrenal axis.