AbstractTwo mid‐scale laboratory gallery fire tests, the propane gallery and the BELT test, were constructed and evaluated. In both cases the fundamental parameter that can be used for ranking flammability of a belting is the total quantity of heat input into the sample, rather than length burned or flame‐out time. The temperature and velocity of the ventilating air used were found, experimentally and theoretically, to be critical factors in determining the results of the tests. The two tests correlated qualitatively on a pass|fail basis; however, the BELT test was affected more by the thickness of the sample: the thicker, the more probable to pass the test. The BELT test has the advantages of easier control and producing fewer pollutants. The results of BELT, propane gallery and critical oxygen index tests on some beltings are compa