The comparative studies undertaken by 7 laboratories in 6 countries show that the calculation I.U.s did not as anticipated minimize but actually enhanced the variability of results of Rabies antibody estaminations in the sera of HDCS vaccines. The high biological variance in the method(s) may not have been considered by individual laboratories and any neglect of fundamental biostatistical laws, unfortunately, diminishes the theoretical advantage of using the "International Standard (I.S.)" as a "tertium comparationis". Perhaps the intrinsic variability of the I.S. should be re-evaluated and it is conceivable that a pure IgG fraction of Rabies antiserum would show less variability. Intralaboratory variation might be reduced by agreeing that only a geometric mean of the I.S., and not a single value obtained in an individual test, should be used for calculation of I.U.s. Application of the principles of biochemical and pharmacological methods, such as test-to-test control of the I.S. and its analytical variances might well enhance the reproducibility of the results. MNT, RFFIT, PRT and CFT were unable to detect antibodies in HDCS vaccinees until 7 days after the first vaccination. The establishment of methods for detecting early antibody requires further investigation.