Random noise and MR signal uniformity were analyzed for a period of 9 months using two head coils on a 1.5 T commercial imager. Signal response from a cylindrical phantom filled with a 0.1 mmol/L CuSO4 solution was incorporated in a correction scheme and the effect of correction on uniformity and accuracy of tissue volume determination was measured. There was little change in random noise (CV < 5%) and image uniformity (CV < 15%) over the time of the study. The uniformity in a 130 cm2 region was 4% for a double saddle coil and 5.3% for a mirror coil. After correction, uniformity was improved to 1.1% for the double saddle coil and 1.2% for the mirror coil. In a 40 cm2 central region the uniformity was approximately three times better than in the 130 cm2 region. A second phantom, mimicking the relaxation times of CSF and the brain, consisted of saline encapsulated in vials of volume ranging from 35.6 ml to 249.6 ml and placed in a 0.1 mmol/L solution of MnCl2. Uniformity correction reduced the average error in volume measurement from 8.6% to 6.1%.