Introduction. In late February 2005, the National Reference Center for Salmonella reported an increase in the number of Salmonella serotype Agona isolates from infants in France during January–February 2005. Methods. A case was defined as an infant in whom Salmonella Agona had been isolated since January 1, 2005, in connection with a fever or diarrhea. A descriptive study and a case-control study were quickly initiated. Results. One hundred forty-six cases were identified over a five-month period. The case-control study showed that all cases had consumed infant formula brand A, whereas none of the controls had consumed it. The recall of brand A formula was followed by a rapid decrease in the number of cases who had consumed this formula, but by an increase in the number of cases consuming brand B formula. Five batches of brand B powder had been manufactured on the same production line as brand A powder. Forty-four cases (31%) were linked to brand A powder and 92 (65%) to brand B powder. Discussion. This is the first documented outbreak of S. Agona infections in France. This investigation demonstrated the inadequacy of in-house Salmonella testing in powders for detecting low-level and heterogeneous contamination. It also highlights the importance of investigating all cases to assess the effectiveness of control measures. (R.A.)