Selenium was determined in infant formulas sold in Canada and compared with the selenium content of Canadian breast milk samples. Formulas were purchased in the Ottawa area during 1992 (prior to selenium addition to infant formulas) and during 1993 and 1994. Samples (1-2 g) were digested with a mixture of nitric, perchloric, and sulfuric acids and selenium in the digestate determined by a diamino-napthalene (DAN) fluorometric method. All samples were analyzed in duplicate or triplicate. Samples of nonfat milk powder reference material (NIST 1549) were analyzed with each batch of infant formula or human milk and coefficients of variation for within-run determination were 3.1% and day-to-day reproducibility was 7.6%. Unsupplemented infant formulas contained 3 to 21 mg Se/liter, compared to supplemented formulas which contained 16 to 35 mg Se/liter. Human breast milk samples from women in Eastern Ontario contained 13 to 25 mg Se/liter (17.7 { 3.3, mean { SD, n Å 10). Assuming an intake of 800 ml of formula or human milk per day, unsupplemented formulas would provide 2 to 17, supplemented formulas 13 to 28, and human milk 11 to 20 mg/Se day, respectively. Thus, we found that several infant formulas sold in Canada can provide approximately one-fifth to onethird of the infant's requirement for selenium (10 mg/day) and that some supplemented formulas can provide twice the requirement.