The reliability of the capture probability equilibrium model developed by Su and Lee (2008) for population estimate was tested in three-directional extended foraging arenas connecting to large Plexiglas cubes (96 by 96 by 96 cm) containing ≈100,000–400,000 workers of the Formosan subterranean termite, Coptotermes formosanus Shiraki. After the release of marked termites in the arenas, the capture probability was averaged for three directions at equal distance from the release point. The daily data of directionally averaged capture probability were subject to a linear regression with distance as the independent variable to identify the capture probability equilibrium. When the daily data produced significant regressions with regression slope |b| ≤ 0.05 or |b| ≈ 0.05, the directionally averaged capture probability was considered to have reached equilibrium, and the regression intercept was used in the Lincoln index to derive the population estimate. Of the four laboratory colonies tested, three met the criteria, and the equilibrium models yielded population estimates that were not significantly different from the known numbers of workers in the arenas.