Too often we hear statements concerning ‘group housing’ and ‘stalls’ that imply that all such systems are the same. Group systems in particular are extremely different in how they control feed intake and manage aggression. It is important that we understand which particular group and stall systems are being compared. Gestation housing for sows is more complex than is often implied by generalizations. There are numerous group housing systems being used for gestating sows, but they generally fall within four general types. Floor feeding systems involve spreading a limited amount of feed over a large floor area and there is relatively poor control over individual feed intake, and high levels of aggression during feeding. The animals are often kept in small groups to minimize both re-grouping and feeding related aggression. Another type of group housing provides a feeding site for each sow, and feed is dropped into these sites at a speed slightly slower than the animals can eat it. Each sow stands at a feeding site waiting for the feed, rather than attempting to take feed from another sow. These ‘trickle’ feeding systems control individual feed intake and feeding associated aggression fairly well, but animals must still be re-grouped and group size must be kept small. A third type of group housing provides individual feeding stalls, and an open area for the non-feeding period. Feed intake is very well controlled, as is feeding-associated aggression. However, such a system still involves re-grouping aggression, requires a great deal of space, and still requires individual stalls. A final type of group housing system involves the use of electronic sow feeders. In such a system each sow in the group wears an electronic identification tag and a computer and an electronic stall control feeding. When the sow enters the feeding stall she is identified by means of her tag, her records are checked on the computer, and her daily allotment of feed is provided to her. This system provides very good control over individual feed intake and prevents much of the feeding associated aggression, but still involves re-grouping fighting. The feeding and control equipment is more costly than in most other systems, but the computer can be used in other aspects of management. Not all group-housing systems are the same. Each achieves the goals of controlling individual feed intake and aggression to a different degree. If the farm has a stall housing system it is usually one size usually 22 or 24 inches (56 or 61 cm) wide. The Code of Practice makes the logical recommendation that the size of stall should depend upon the size of the sow. This study used the suggested sizes in the Code to develop a formula for width based on sow weight. The demographics were then estimated for the herd once a stable population was reached. Based on this data, stalls ranging from 22 to 28 inches (56 to 72 cm) in width were included, in proportions that should match the size of sows in the mature herd. For each weekly breeding group there are four narrow stalls for small gilts, several stalls for second and third parity sows, a few less for 4th and 5th parity animals, and two 28 inch (72 cm) stalls for 6th parity animals. Hopefully by allotting animals to stalls based on their size it will reduce injuries and improve longevity.