To the Editor: In the October 16 issue, Levine and Silvis reported on absorption of whole peanuts, peanut oil, and peanut butter and analyzed their data with Student's t-test. The multiple application of the t-test in this experimental design is inappropriate. Such multiple applications produce additive effects through Type I errors. Therefore, even though the authors cite significant differences at the 0.05 level, in actuality those figures are much higher and are thus beyond the level that could be considered significant. The exact level of α could be calculated if their data were available. The proper analysis to use would. . .