Seven experiments with 324 chicks tested their ability to select a nutritionally adequate diet from separate sources of purified casein and various supplements, including gelatin (a source of two amino acids), a gelatin-creatine mixture (a source of three amino acids), and fiber (nonnutritive bulk). Nonselecting controls consumed the basal purified-casein diet or a supplemented purified-casein diet. Chicks in all selection conditions composed diets that yielded normal intake, normal body temperature, normal activity, and the maximum growth possible for their intake. They also selected components in the same percentages as in premixed diets. In all instances, their selection was nonrandom and regulated. What chicks included in their diet depended on what else was available. Although the specific percentage taken from each dietary component varied across different selection alternatives, these differences affected neither intake nor growth. Selection, per se, incurred a caloric cost. Chicks selecting from fractions of a corn-and-soy diet offset this cost by increasing intake compensatorily, but chicks with a purified-casein fraction did not, suggesting that some unspecified property of casein placed a ceiling on its intake. These findings unequivocally demonstrate that immature chicks not only can self-select nutritionally adequate diets, but can do so with unexpected precision by exploiting different but equally successful strategies.