The biexponential relaxation behavior of the sodium nucleus affects the accuracy of quantitative measurement of in vivo tissue sodium concentration (TSC). Theoretical analysis and in vivo experimental results are used to demonstrate the extent of the large bias in the measured TSC that arises when the relaxation behavior in vivo differs significantly from that of the calibration standards which is when a significant fraction of the total sodium signal decays with a relaxation time much shorter than the echo time (TE) used for imaging. This bias can be as large as 20% for measurements of TSC in a normal rat brain with TE = 2 ms. Our findings indicate that shortening the echo time (TE < 0.5 ms) by projection imaging is a reliable means of obtaining accurate in vivo estimates for TSC using MR.