In the past, tool life tests have been performed using a conventional Taylor testing technique. This methodology is expensive and time-consuming. It requires wearing a number of tools until the tool failure criterion has been reached. A number of short-term tests designed to replace the Taylor test have been proposed but they suffer from a number of drawbacks. Many of these tests are performed under non- standard cutting conditions or require special workpiece preparation or equipment. As a result, tool life models developed from these tests are of limited usefulness in predicting tool failure times for conventional machining operations. A methodology is required which combines the time and cost advantages of non-conventional tests with statistical validity and robustness. In this research, two short-term tests are presented which are based on the Taylor test. Response surface models are used to develop the parameters of Taylor's tool life equation. The tests are shortened by using regression equations of flank wear data to predict the tool failure time without wearing the tool to failure. The two methods, abbreviated conventional testing and sequential composite testing, are statistically validated and compared with the Hill Taylor test. The results show that these tests can accurately predict tool life and the resulting Taylor models are not significantly different from those estimated by conventional means. / Master of Engineering