The French Institute for Public Health Surveillance and the National Reference Center for Yersinia conducted a mail survey of a random sample of French medical analysis laboratories (MALs). The objectives were to: determine the number of stool cultures tested for Yersinia performed by these MALs; estimate the frequency of Yersinia isolation in these stool cultures; compile an inventory of the various methods used to isolate Yersinia; and determine the rate of strain identification and characterization. Of the 953 community and hospital MLs contacted, 483 (51%) participated in the survey. The vast majority of these laboratories test stool cultures for Yersinia, but not systematically. This testing was performed on only half (53%) of the 256,871 stool cultures conducted in 2003 and resulted in the isolation of Yersinia in 333 cases (isolation rate of 0.25%). The most commonly used method for isolating Yersinia from stool samples is adequate. It relies on the use of CIN medium incubated at 28–30°C for 24 to 48 hours, sometimes preceded by enrichment. Identification of the genus and species is most often performed using the API20E identification gallery, which, without additional testing, does not allow for a fully reliable species diagnosis. More in-depth characterization of strains (biotype, serotype) is rarely performed, even though it is the only method capable of distinguishing pathogenic strains from non-pathogenic ones. Antibiograms are performed according to appropriate procedures (Mueller-Hinton medium incubated at 37°C) and include antibiotics that may be part of the treatment regimen for yersiniosis. Overall, this study shows that testing for Yersinia was far from systematic within the LAMs in 2003 and that there was significant variation in their ability to isolate Yersinia. The number of Yersinia infections is likely significantly underestimated in France, and in the absence of tests to distinguish pathogenic from non-pathogenic strains, unjustified treatments are likely being prescribed. (R.A.)