Water displacement is often cited as the ‘gold standard’ technique for limb volume measurements (Rabe et al., 2010). The usual implementation results in only a single volume measurement for the entire submersed limb (Damstra et al., 2006). Various optical techniques exist which produce cross-sectional profiles of limbs (e.g. perometry or laser scanning imaging) but they are relatively expensive and are typically considered to have lower accuracy than water displacement. Leg profiles can be obtained cheaply using multiple circumferential measurements but usually only at relatively low resolution with inaccuracies inherent within frustum approximations. We have attempted to modify water displacement volumetry to enable high resolution leg profile measurements to be made at low cost. Our technique involves submerging the leg in a custom-made container (the current prototype is a home-made plywood boot, volume ~12 L) filled with water (~27°C) which is then emptied at a constant rate (~59 ml.s-1) with a diaphragm pump (PLD-1205, Shysky Tech; 12V 25 W) whilst the height of the water is measured (20 Hz) using multiple pressure transducers (MPX5010DP, Freescale Semiconductors). During pumping the rate at which the height declines allows the leg cross-sectional area to be calculated (cross sectional area = flow / rate of change of height).Fig. 1A shows two traces of cross-sectional area calculated from the rate of decline of height plotted against height. The top line shows the empty boot profile and the lower line the profile when the boot contains a leg. The difference between the two lines gives the leg cross-sectional area. We have plotted the scan for the leg in Fig 1B as a symmetrical plot scaled for linear dimensions (√area) assuming the leg to be circular in cross-section (which it is obviously not) and rotated so that the foot is at the bottom. The bulges associated with the calf, thigh and foot are clearly visible, although the cylindrical assumption produces an odd ‘foot’ shape. The accuracy of the technique depends greatly upon the noise within the pressure transducers and the reliability/reproducibility of the pump. We have employed multiple pressure transducers and modified then to allow the removal of air since early experiments showed that air locks and the associated moving columns of water within the pressure transducer produced large artefacts. We also found that emptying the boot produced quieter traces since the velocity of water within the boot is low minimizing Bernoulli artefacts. Removing air bubbles and ensuring the boot is stable are also critical for producing accurate results. The result of various optimizations is a system that can report cross-sectional area (or regional volume) changes with high accuracy and repeatability at very low cost. To illustrate, we have made a profile of a water bottle on a metal block – a photograph of the bottle and the scan are show in Fig. 1B. The intricate profile of the bottle is easily resolved by the system. Further optimizations are possible, in particular measuring the flow rate (which may vary if air bubbles transit the diaphragm pump), reducing the excess cross-sectional area of the boot and the addition of auto-calibrators (sharp volume constrictions) within the boot should allow for improve accuracy and increased scanning speed. There are inherent problems associated with such measurements. First, the water exerts a pressure on the leg which will reduce venous volume. Second the measurement requires the subject to remain still (although multiple consecutive scans can reduce this noise). Third the measurement takes time during which the muscle pump is not active. Finally, it does involve getting wet which we have found reduces the willingness for the elderly to act as participants. Nevertheless, the system allows for a more detailed investigation of limb volumes than finances might otherwise allow.