The influence of swimming activity on gill paracellular permeability has not been measured previously in fishes. We critically assessed the use of tritium-labeled polyethylene glycol ([ 3 H]PEG-4000) for this purpose, a substance that is also a classic marker for extracellular fluid volume, glomerular filtration rate and drinking rate. Tests (8 h) on resting freshwater trout showed that when measuring [ 3 H]PEG-4000 clearance from the plasma in the efflux direction, correction for a large excretion via glomerular filtration was essential, necessitating urinary catheterization. When measuring [3 H]PEG-4000 clearance from the water in the influx direction, correction for a significant uptake by drinking was essential, necessitating terminal gut removal, whereas glomerular filtration losses were minimal. After correction for these alternate routes of loss and uptake, [ 3 H]PEG-4000 clearance rates by efflux from the plasma and by influx from the water were identical, showing that gill paracellular permeability is not rectified, and can be measured in either direction. The influx technique with terminal gut removal was used to assess gill paracellular permeability in trout without urinary catheters freely swimming at 1.2 body lengths s −1 for 8 h. Branchial [ 3 H]PEG-4000 clearance rate (by influx from the water) increased significantly by ~80% in accord with a similar measured increase in O 2 consumption rate. Thus in trout, gill paracellular permeability does increase during exercise, in accord with the traditional concept of the osmorespiratory compromise.