Broiler chickens were used to alter the partitioning of ME between maintenance and production. They were fed amounts of feed that ranged between ad libitum and 75 to 80% ad libitum each day from a BW of 1.1 to 2.2 kg. The experiment was done with 2 strains of broilers, both sexes, and 2 forms of feed (mash or pellets). Data were collected to determine live performance, digestibility of the feed, and carcass composition. The daily amount of feed did not affect the ME of the feed, but broilers fed limited amounts of feed gained less BW per day and had a larger feed:gain ratio. For most measurements, strain of broiler, sex, and form of feed also were significant factors. Because broilers that gained BW more slowly required more days to gain 1.1 kg and more feed during that time, a larger proportion of the energy was used for maintenance. The net energy theory proposes that heat increment, net energy for maintenance, and net energy for production are constants for a feed ingredient or feed. Results from the present research did not support the net energy theory. A different model was proposed that used ME as the basis for energy partitioning. Amounts of feed or energy per day had no effect on the ME content of the feed; however, amounts of energy consumed per day had dramatic effects on the proportion of the ME from each gram of feed that was used for maintenance, product, or heat increment. A model was developed that showed these effects and the feed:gain ratio over a wide range of daily energy intake.