Objective: To examine the plausibility of health claims for Eclipse, a novel smoking article being marketed by the RJ Reynolds Tobacco Co. (RJR) as potentially reducing the risk of cancer and other diseases compared to conventional cigarettes. Data sources: A company product website (www.eclipse.rjrt.com) summarising scientific studies of various versions of Eclipse, and the published review of these studies by an expert panel convened by RJR, an independent study comparing the smoke yields of major carcinogens from Eclipse and two low yield "ultralight" brands (Now and Canton), and an analysis of the levels of these compounds in Eclipse and Premier (its predecessor) over time. Analysis: The overall doses and effects of toxins in the aerosol from Eclipse are smaller than those from most conventional cigarettes on a per mg basis. However, two tests that compared Eclipse on a per cigarette basis revealed that Eclipse was as or more toxic than an ultralight cigarette. Studies show that consumers smoke Eclipse (like they do cigarettes) at puff volumes and puff frequencies far higher than those used for the Federal Trade Commission (FTC) test. RJR's test results, which are based on aerosols generated under FTC conditions, may not reflect actual human dosing, since the operating temperature of Eclipse is highly dependent on these puffing parameters. Even under FTC/International Organization for Standardization (ISO) standard measures, Eclipse smoke carcinogen yields were higher than Now, but similar to Canton. The yields of carcinogens from Premier and different versions of Eclipse have increased overtime. Furthermore, the human studies reviewed by the RJR expert panel do not offer compelling evidence of reduced harm, as they have not been conducted in smokers who have adopted Eclipse. Conclusion: There is as yet unsatisfactory evidence that Eclipse is less harmful than conventional cigarettes. [ABSTRACT FROM AUTHOR]