A new gap test for the determination of the sensitivity of an explosive to shock initiation has been developed. The tester uses water as the medium in the gap between the donor charge and the explosive to be tested. With this design, the gap width is easily adjustable to within a small fraction of the standard deviation of the mean gap for common high explosives. During trials to establish the gap width, a computer code drives a test sequence based on D-optimality statistics that uses knowledge of all the previous responses in the trial. This procedure gives economical and reliable estimates of the mean gap width and the standard deviation of the mean. Some illustrative examples of results are given and discussed, and the influence of temperature on the results is addressed in an appendix.