Two types of experiments were conducted to test predictions of the behavior of coiled springs when they are exposed to elevated temperatures. Test burnings of actual mattresses and other experiments were performed under controlled laboratory conditions that simulated possible fire situations. Findings show that the percentage and rate of "collapse" and the hardness of coiled furniture/bedding springs subjected to high temperatures are dependent not only on temperature of exposure, but also on the time of such exposure, load supported, amount of prior cold work (from fabrication), temperature dependence of the elastic modulus, carbon content of the steel comprising the spring, and even the rate of cooling to room temperature. Results also indicate that the same extent of collapse of coiled springs can be attained by a relatively short exposure to a high temperature or by a longer exposure to a lower temperature, conditions which occur in accelerated and smoldering fires, respectively. Because of the high temperatures attained in short periods by fires involving normal building materials and because of the complex interaction of numerous metallurgical parameters that cannot be deduced from post-fire spring examinations, observation of the "collapsed" state of coiled furniture/bedding springs is not a reliable indicator of whether a fire was initiated by a smoldering cigarette or accelerated by the presence of a hydrocarbon. 2 tables, 5 figures, 21 references.