The effects of the proportion of pigs removed from an established group and subsequent floor space on growth performance during the final 19 d of the finishing period were evaluated using 28 pens of mixed-sex crossbred pigs (mean initial BW = 113.4 +/- 0.57 kg; n = 1,456; approximately 52 pigs per pen). A randomized block design was used with four pig-removal treatments: 1) 0% of pigs removed [Control], 2) approximately 25% of pigs removed, 3) approximately 50% of pigs removed, and 4) approximately 50% of pigs removed and floor and feeder spaces/pig decreased to equal those of Control. A block consisted of four pens with the same number of pigs and sex ratio per pen and with similar initial BW. Pens within blocks were randomly allocated to treatment, and the heaviest animals were removed from Treatments 2, 3, and 4 at the start of the study. Group size and floor space/pig for Treatments 1, 2, 3, and 4 were 52 and 0.65 m(2), 39 and 0.87 m(2), 26 and 1.30 m(2), and 26 and 0.65 m(2), respectively. Each pen contained a six-place feeder that provided 212 cm of total trough space; however, only three-places were accessible to pigs on Treatment 4. Compared with Controls, removing 25 or 50% of pigs resulted in increased (P < 0.001) ADG by 20.6 and 21.0%, ADFI by 10.8 and 7.9%, and G:F by 7.7 and 14.3%, respectively. Average daily gain by pigs on Treatment 4 (50% removal rate and decreased floor and feeder spaces) was greater (P < 0.05) than that of the Controls, but lower (P < 0.05) than that of Treatment 3 pigs (50% removal rate, no adjustment in floor or feeder spaces). No differences were observed among treatments for either morbidity or mortality. These results indicate that removing 25 or 50% of the heaviest pigs from groups of finishing pigs increased growth rate of the remaining pigs, and that the improved performance was only partly due to increased floor and feeder spaces.