The usefulness of microbiological standards for frozen foods is now a controversy in the trade and scientific literature. Most reviewers have given arguments both for and against, and have concluded that they should be applied with great caution. Such standards have the advantage of putting questions of safety on a convenient numerical basis. Canadian workers have reported that promulgation of standards has invariably raised the hygienic level of the products controlled. Bacteriological standards have often been associated with the question of safety to the consumer. Everyone recognizes that food poisoning bacteria are a potential danger in any food. But many have argued that the history of food poisoning outbreaks from frozen foods is excellent and that there is no need for standards; on the other hand, proponents of standards have pointed to the incomplete investigation and reporting of outbreaks, and have argued that there may be more outbreaks than we realize. They have pointed to laboratory studies that have shown grossly mishandled precooked frozen foods to be truly dangerous. Some have proposed that pathogens should be absent from foods; but others have questioned that a microbiological standard can accomplish this end. Some pathogens, such as Salmonella or Staphylococcus have been shown to be so ubiquitous that their presence in some commercial foods is unavoidable. Also, sampling and analytical methods have been described as inadequate to guarantee that pathogens present will be detected. Some have argued that control at the source is a better way—through inspections of the plant operation, by enforcement of handling codes, or by processing procedures such as pasteurization, which would be more certain to result in a pathogen-free food. A most important part of any of the proposed standards is a “total count” of viable aerobic bacteria. English workers have found that foods causing poisoning outbreaks usually had total viable counts above 10 million per gram. On the other hand, these same workers found Salmonella on meats with very low total viable count. The assumption by many that low total count indicates safety has been shown to be not always true. Furthermore, high counts of nonpathogenic organisms, such as psychrophilic saprophytes would have no public health significance. The relation between bacterial level and quality is open to less controversy. Some authorities have pointed to bacterial level as a measure of sanitation, adequacy of refrigeration, or speed of handling. Others have indicated that to determine which of these factors caused a high count would be impossible with only a total count on the product as a guide. Some investigators have said a high count affects flavor adversely before actual spoilage is evident, and this may be a factor in competition on today9s market. It is well established that initial bacterial level will affect the shelf-life of a chilled product. Methods of analysis are more nearly adequate for counts than for pathogens, but they need improvement, and should be clearly specified as part of any bacteriological standard. Foods with high count could sometimes be brought into compliance merely by storing them for a sufficient period frozen, or by heating them slightly. This has been cited by some authors as a disadvantage of bacteriological standards. The enterococci and the coliform group (except Escherichia coli) have been shown to be ubiquitous and therefore should not be used alone to indicate fecal contamination. Although E. coli has greater significance, its source should be determined each time it is found. Various reviewers have expressed the need for caution in the application of standards. The principal precautionary arguments we have found are as follows: 1) A single set of microbiological standards should not be applied to foods as a miscellaneous group, such as “frozen foods” or “precooked foods.” 2) Microbiological standards should be applied first to the more hazardous types of foods on an individual basis, after sufficient data are accumulated on expected bacterial levels, with consideration of variations in composition, processing procedures, and time of frozen storage. 3) When standards are chosen, there should be a definite relation between the standard and the hazard against which it is meant to protect the public. 4) Methods of sampling and analysis should be carefully studied for reliability and reproducibility among laboratories, and chosen methods should be specified in detail as part of the standard. 5) Tolerances should be included in the standard to account for inaccuracies of sampling and analysis. 6) At first, the standard should be applied on a tentative basis to allow for voluntary compliance before becoming a strictly enforced regulation. 7) Microbiological standards will be expensive to enforce. 8) If standards are unwisely chosen they will not stand in courts of law.