A graphic method employing "normal equivalent deviate" plots, was used to visualize the results of antituberculosis vaccine assays in groups of test mice. This method identifies relatively normal distributions of data by making them fall closely about straight lines. Three relatively normal post-challenge death time distributions were found. Two of these were expected, but the third or "accelerated" distribution came as a surprise and may be a new finding. It represents infections that are sufficiently overwhelming to mask the effects of active immunization in vaccinated mice, and it can only be identified and bounded by graphic means. Graphic analysis was used to spot the most likely post-challenge day on which to terminate each experimental mouse group before calculating results. This method, which yielded end points that varied from group to group, was compared with the Youmans fixed 30-day post-challenge method of ending experiments. Both methods are simple and practical, but graphic analysis was adjudged preferable for theoretical reasons.