Five-hundred popular upholstery fabrics were obtained from commercial fabric shops in North Carolina. The ignitability of each upholstery fabric was characterized using a particular smoldering cigarette in a specified protocol with six replications. If one or more of the smoldering cigarettes resulted in smoldering of the fabric, the fabric was classified as ignitable. Of the 500 fabrics, 355 were found to be non-ignitable and 145 were classified as ignitable. The differences between the properties of ignitable and non-ignitable fabrics were examined. The results show that the upholstery fabrics' contents of sodium and potassium salts, their concentrations of cellulosics, and their basis weights (areal densities) correlate with ignitability. An aqueous rinse of upholstery fabrics rendered them substantially non-ignitable by the smoldering-cigarette used in this study.

Upholstery fabrics for residential use were obtained from various fabric shops and manufacturers' catalogs. Screening was conducted with four experimental cigarettes of varying design. By the National Institute of Standards and Technology (NIST) cotton duck mockup method, two of the cigarettes display "low" ignition propensity and two show "high" ignition propensity. Of the fabrics obtained for this study, 316 smoldered when in contact with at least one of the experimental cigarettes. Further examination within this set of upholstery fabrics showed once again that cigarette ignition propensity ranking are dependent on fabric characteristics. The number of fabrics that showed ignition propensity rankings opposite to the NIST test "ranking" was similar to the number that agreed with it. That is, on the fabrics with rankings opposite to the NIST ranking, cigarettes of "low" ignition propensity by the NIST test showed more ignitions than cigarettes of "high" ignition propensity by the NIST test. These results are consistent with those of a previously published study with a smaller set of ignitable fabrics. Fabric weight was a key factor in determining fabric ignition behavior; two of the cigarettes showed increasing ignition propensity and two showed decreasing ignition propensity with fabric weight. For the majority of fabrics, however, cigarette design was unimportant in determining ignition behavior; that is, no differences in ignition behavior were observed with the different cigarette designs. The physical and chemical properties of the cotton duck fabrics of the NIST test are such that they repre-JOURNAL OF FIRE SCIENCES, VOL.

The NIST method for measuring the ignition propensity (IP) of cigarettes on a test mockup meant to represent the flat surface of upholstered furniture was revised and tested [1]. The revisions included a number of dele tions and additions. The most important addition to the original test was the in clusion of selected, smolderable upholstery fabrics. The deletions were, in general, procedural details known or believed to be unnecessary. Three replicate runs of the method using the same experimental and control cigarette pair, as well as a series of runs using experimental and control cigarette pairs with deliv eries of 5, 11 and 16 mg of tar, were performed. The modified method provides re peatable and statistically valid data. Adding commercial, nonreversal upholstery fabrics to the test did not add to the ability of the test to measure IP behavior and, therefore, they were eliminated in the final proposed method. A reversal-type fabric should be included in the test, but it should be specially pro duced for this purpose [9]. With these further modifications, the test is consider ably simpler to perform than the original test.