Previous methods for the comparison of growth in groups of exptl. animals subjected to differing treatments consisted of visual estimation of significant differences from a plot of weight against time, or application of the "[image]" test to the mean weights at a particular time. These conventional methods often give an unconvincing or ambiguous analysis of the results. The present paper describes how time-weight data (growth curves) were treated in a comparative study of exposures of rats to vapors of 5 analogous halogenated hydrocarbons. The wt. data were expressed in the form of frequency distributions, the weekly wt. of each rat alive at the time of weighing being considered a point in the distribution. Separate frequency distributions were prepared for [male] and [female] rats; these data were also combined to show the overall effects of the agents. The method of "chi square" reveals probability levels of significance, as does also the ratio of 3 times the standard error of the difference to the difference between the mean weights; however, the former is preferred. Values for the coefficient of skewness calculated for the distributions of the control groups, treating the sexes both combined and separately, justify the use of normal statistics.