A brief response to the latest word from BLS about workplace fatalities: 1. BLS attempts to tabulate fatal workplace injuries primarily; most fatal diseases are not recorded. The rule of thumb in this business is that there are 10 occupational disease fatalities for each fatal injury. 2. BLS greatly undercounts the rate of fatal occupational injury. The best source is NIOSH's Census of Fatal Occupational Injuries (CFOI), which appears after a longer delay than BLS. Journalists and others should look to this source for their numbers. No one, to my knowledge, has ever investigated the extent to which the BLS fatality rate is correlated over time with better measures of serious risk on the job. So each year, when the new BLS numbers are released, a spate of news stories appears which make pronouncements on occupational safety and health that have no real factual basis. The best single discussion of occupational fatality rates, traumatic and nontraumatic, is the study done for NIOSH by J. Paul Leigh et al. There is a summary in the Jan. 1997 Annals of Internal Medicine; the full report (which may be available from NIOSH) will appear in book form later this year or sometime next year, published by U of Michigan Press. Peter