Oedema is caused by impaired lymphatic drainage and/or increased microvascular filtration. To assess a postulated role for the latter in postmastectomy oedema, filtration was studied in the forearms of 14 healthy subjects and 22 patients with chronic, unilateral arm oedema caused by surgical and radiological treatment for breast cancer. A new non-contact optical device (the Perometer) and a conventional mercury strain gauge were used simultaneously to record forearm swelling rates caused by microvascular filtration during applied venous congestion. Filtration rate (FR) per 100 ml tissue was measured over 10-15 min at a venous pressure of 30 cmH2O, a pressure reached in the dependent forearm (FR30), and then at 60 cmH2O (FR60). Apparent filtration capacity of 100 ml soft tissue (CFCa) was calculated from FR60 - FR30/30, after adjustment for bone volume. The Perometer and strain gauge gave similar results in normal and oedematous arms. Mean CFCa in healthy subjects was (3·8 ± 0·4) × 10-3 ml (100 ml)-1 cmH2O-1 min-1, close to literature values. In the patients, FR30 was 47 % lower in the oedematous forearm than in the opposite, unaffected forearm (P = 0·04). FR60 showed a similar trend but did not reach significance (P = 0·15). The values of CFCa of (2.2 ± 0.5) × 10-3 ml (100 ml)-1 cmH2O-1 min-1 in the oedematous arm and (2.8 ± 0.5) × 10-3 ml (100 ml)-1 cmH2O-1 min-1 in the unaffected arm were not significantly different (P = 0.47). When differences in arm volume on the two sides were taken into account, the total fluid load on the lymphatic system of the oedematous forearm was (411.0 ± 82.2) × 10-3 ml min-1 at 30 cmH2O and (1168 ± 235.6) × 10-3 ml min-1 at 60 cmH2O, similar to the normal side, namely (503.7 ± 109.3) × 10-3 ml min-1 and (1063 ± 152.0) × 10-3 ml min-1, respectively (P >= 0·50). The filtration capacity of the entire oedematous forearm (CFCa scaled up by total soft tissue volume), (25.4 ± 6.2) × 10-3 ml cmH2O-1 min-1, was not significantly greater than that of the normal forearm, (18.3 ± 2.6) × 10-3 ml cmH2O-1 min-1 (P = 0.40). The results indicate that no major change occurs in the microvascular hydraulic permeability-area product of the forearm, or in the total filtration load on the lymph drainage system during dependency, in the arm with postmastectomy oedema compared with the normal arm. This argues against a significant haemodynamic contribution to postmastectomy oedema.