Methodology was developed for measuring the gastrointestinal endogenous phosphorus (P) outputs and true P digestibility values in studies with piglets. Four barrows, average initial body weight 6.8 kg, were fitted with a simple T-cannula at the distal ileum and fed four diets according to a 4 x 4 Latin square design. Four cornstarch-based diets containing four levels of P (1.1, 2.1, 3.2 and 4.3 g/kg diet) on a dry matter (DM) basis were formulated from soybean meal (SBM). Each experimental period comprised 8 d with a 4-d adaptation and 4-d collection of ileal digesta and feces. The apparent ileal and fecal P digestibility values in SBM were affected (P < 0.05) by P levels in the assay diets. The ileal and fecal P digestibility values increased from -24.8 to 37.1% and from 18.8 to 42.5%, respectively, as P contents increased from 1.1 to 4.3 g/kg DM diet. Linear relationships (P < 0.05), expressed as g/kg DM diet intake, between ileal and fecal outputs and dietary inputs of P, suggested that the endogenous P outputs can be determined by linear regression analysis. The endogenous P output was higher (P < 0.05) in ileal digesta than in feces (0.86 +/- 0.09 vs. 0.31 +/- 0.06 g/kg DM diet intake). There was no difference (P > 0.05) between the true ileal (50.7 +/- 7.1%) and fecal (48.5 +/- 5.4%) P digestibility values in SBM. These results suggest that differences in P contents between assay diets are primarily responsible for the large variability in apparent P digestibility values reported within the same ingredient. Apparent digestibility values underestimate the true digestive utilization of P by approximately 25%. True rather than apparent P digestibility values should be determined and used in diet formulation for pigs. In addition, this study shows that the gastrointestinal endogenous P output is important in whole-body P requirement and homeostasis.