Ammonia, methane, and odor from a fully slatted pig housing facility equipped with a pit-recharge manure system were modeled with a view to quantify the sources of ammonia, methane, and odor Emissions of ammonia, methane, and odor were measured during three weeks in two environmental chambers containing ten pigs. The chambers had fully slatted floors, and the pit was precharged with 120 L of water After one week; the pigs were removed from the chambers, while measurements of emissions were continued for an additional two weeks. The trial was conducted in duplicate with pigs weighing 25 and 45 kg, respectively. Ammonia and methane were modeled to estimate concentrations in the presence (sigmoidal model) and upon removal of pigs (single pool exponential decay model). The combined model accounted for 89% of the variation in ammonia. The pit contributed 50% to 60% of the ammonia emission, while the slats contributed the remaining 40% to 50%. The mean ammonia emissions with pigs present were 26 and 17 g AU(-1) d(-1) with pigs weighing 25 and 45 kg, respectively. The model explained 98% of the observed variation in methane concentrations. The pigs and freshly excreted feces were the principal sources of methane, while the recharge pit did not emit significant amounts of methane. With pigs in the chambers, the mean methane emissions amounted to 14 g AU(-1) d(-1) irrespective of weight. Geometric mean odor emissions with pigs present were 60 OUAU(-1) s(-1) (95% confidence interval 4.2-8.6 OUAU(-1) s(-1)). Odor emissions were unaffected by the presence of pigs (p > 0.10). This indicates that manure, whether in the pit or on contaminated slats, is the main source of odor in pig houses, whereas the pigs themselves are of minor importance.