Frequency shift/concentration calibration curves applicable to the use of shift reagents (SRs) for in vivo 23Na MRS studies can be obtained from experiments with whole blood. Here, they are reported for titrations of rat blood with the SRs DyTTHA3- and TmDOTP5-. There are a number of considerations that must be made in order to derive accurate calibration curves from the experimental data. These include the effects of bulk magnetic susceptibility (BMS, since the SRs are paramagnetic), the effects of water flux (since addition of the SR stock solution to blood renders the plasma hyperosmotic), and the consequences of restricted distribution of the SR anion in the erythrocyte suspension. We give in some detail the BMS shift theory that obtains in this case and show also how it applies to excised perfused organ as well as in vivo studies. Also, we report significant effects of adjuvant Ca2+ additions in the TmDOTP5- titrations. These are very important to the successful use of this SR in vivo. Finally, our considerations of BMS lead naturally to an understanding of its manifestations in the shifts of the 1H2O resonance frequencies of cell suspensions and tissues induced by SRs. Since these are being increasingly reported, and often misinterpreted, we devote an experiment and some discussion to this subject. We show that this phenomenon cannot be used to quantitatively discriminate intra- and extracellular 1H2O signals.