It is now over 40 years since Chen and his associates1 published a classic report on the treatment of cyanide poisoning using dogs as the animal model. Until the current report of Graham et al in this issue (see p 1051), this therapeutic regimen was virtually unchallenged as the definitive treatment for cyanide poisoning in humans. The presentation by Graham and his colleagues certainly must be considered an assault upon the sacred cow. They raise a number of provocative issues and have the temerity to challenge the application of the therapeutic combination of sodium nitrate and sodium thiosulfate. The basic lesion in cyanide poisoning is the product of a very tight combination between the cyanide ion and the ferric ion of cytochrome oxidase. The affinity between cytochrome oxidase and cyanide is very tight,2 and inactivates this key respiratory enzyme. The rationale behind the use of sodium nitrate was