In pharmacodynamic studies using focal ischaemia models, the size of the infarct measured by quantitative histology is the most important outcome measure. Precise, unbiased and reproducible assessment of infarct volume is of foremost importance. A frequent problem in interventional stroke models is the evaluation of infarcts in animals found dead, where instant post-mortem fixation of the brain cannot be performed. The purpose of this study was to investigate possible bias from perfusion, immediate and 3-h post-mortem delayed immersion fixation on the measured volumes of cerebral infarction, oedema and hemispheres in a rat embolic stroke model. Thirty-six male Sprague-Dawley rats were thromboembolized into the internal carotid artery. After survival for 24 h, the animals were divided into three groups: group 1 - immediate perfusion fixation; group 2 - immediate immersion fixation of the brain; and group 3 - animals left dead for 3 h at room temperature before removal of the brain for immersion fixation. Following histological preparation and evaluation, the volumes of the hemispheres and infarction were measured by quantitative histology and planimetry. Brains fixed by immersion were 7% larger than the perfusion-fixed brains. Delaying the immersion fixation for 3 h may increase hemisphere volume by a further 12%. Independent of the fixation procedure, the size of infarction was approximately 40% of the ipsilateral hemisphere, and the oedema was approximately 11% of the size of the infarct. The used planimetric technique was accurate with measured values within +/- 2% of the factual value. In conclusion, sizes of hemispheres, infarction and oedema in absolute volume measures are influenced by the effect of unwanted variation of brain size caused by biological factors and artificial shrinkage caused by fixation, dehydration and heat treatment of the specimens. Infarction and oedema expressed relatively in per cent of hemisphere and infarct, respectively, are robust measures independent of the investigated fixation procedures.