To determine whether incomplete carbohydrate absorption plays a role in mediating the effects of formula change on babies' crying, breath hydrogen excretion and behavior were measured in 17 normal formula-fed infants who entered a feeding trial at 28 days of life. The trial permitted two comparisons between (1) lactose and reduced lactose soy-based formulae, and (2) the infant's usual pretrial formula and the subsequent soy-based variable-carbohydrate formulae. Reduced lactose formula was associated with a small reduction in H2excretion (from a mean of 15 to 7 ppm,p= .07) but no difference in crying or fussing. However, compared with their usual pretrial formula, the change to soy-based variable-carbohydrate formulae was associated with a substantial and sustained reduction in H2excretion (mean 32 to 11 ppm,p< .03) and a modest 21% decline in fussing (90.4 to 71.5 min/24 hr,p< .08). By 8 days after the formula change, there was a 40% decline (90.4 to 53.9 min/24 hr) in fussing. These results suggest that, although behavioral changes due to differences in carbohydrate content are unlikely in normal infants, formula changes involving protein and carbohydrate can reduce colonic gas production and may have some effect on crying. Such effects may be implicated when feeding changes occur in normal infants, but their potential role in treatment of crying problems (colic) is yet to be demonstrated.