This paper is concerned with the statistical analysis of experiments in which more than one response (or characteristic) is observed (or measured) on each experimental unit. From the results of the multivariate general linear model and tests of significance. for the general linear hypothesis, the multivariate analysis of variance is extended to investigate multiresponse experiments with a one-way classification, the randomized block design, and a two-way classification. Moreover; the 100(1－α) percent simultaneous confidence intervals for all linear functions of the difference of two treatment-effect vectors are derived from the general results of multiple comparisons in the multivariate . analysis of variance. The complicated matrix algebra used in tests of significance and confidence interval estimation is expressed in t he comprehensive forms by the summation notation. Finally, the multivariate analysis of variance is illustrated with the data. obtained from the . perennial experiments with pasture varieties at different locations in Taiwan.