The power of a Doppler signal is theoretically proportional to the volume of blood within the sample volume of an ultrasound (US) beam and, hence, may provide a means of detecting in vivo changes in the cross-sectional area of cerebral vessels, such as the middle cerebral artery. The purpose of this study was to examine the relationship between power and vessel size for signals recorded from a wall-less flow phantom. The results demonstrate the importance for the in vitro case of maximising the received signal power for each channel to obtain the true relationship between power and size, and show that a nonproportional relationship observed between the two parameters is primarily caused by high-pass filtering and nonuniform insonation. In addition, an investigation of the reproducibility of power values after transducer repositioning has shown that variation occurs even when extreme care is taken to maximise the received signal intensity. The implications of these results for the in vivo use of the Doppler signal power method are discussed.