We examined the influence of experience with a macronutrient on subsequent macronutrient selection. For 4 days, rats ate chow, or chow and a single macronutrient source. They then had simultaneous access to protein, carbohydrate (CHO) and fat sources, according to a standard macronutrient self-selection paradigm. When selecting among the macronutrient sources, rats pre-exposed to CHO ate more CHO, and those pre-exposed to fat ate more fat, relative to the other groups. Rats pre-exposed to protein ate more protein than did those pre-exposed to CHO or fat but not more than those that received no macronutrient pre-exposure. These selection patterns persisted for at least 12 days, when the test was ended because of the low protein intakes and poor growth of the rats pre-exposed to the CHO and fat sources. After 34 days of recovery with only chow to eat, the rats were again allowed to choose among the three macronutrients, and their patterns of selection were essentially unchanged. Similar results were found in a second experiment in which a 5-day interval was interposed between macronutrient pre-exposure and macronutrient selection. These findings show that experience with the macronutrients typically used in self-selection experiments can have a large, long-lasting, and sometimes detrimental effect on subsequent food selection by rats. Prior experience can be a more powerful influence than nutritional wisdom in determining the rat's food choice.