The heats measured in isothermal titration calorimetry (ITC) experiments generally need to be adjusted for heats of dilution of the reagents in the syringe (titrant) or cell (titrand) or both, to enable correct analysis of the data with standard fit models. This is often done through "blank" experiments with one of the reactive species missing. For ionic solutes, the resulting dilution heats can be large and can vary significantly with injection number; for nondissociating solutes, they are typically much smaller. When mixing heats vanish, there remains an "instrumental blank" of order -1 μcal but varying with both rate and volume of injection. In protein binding studies the titrant and titrand are normally prepared in a common buffer solution, yielding negligible mixing heat and reducing the blank correction to just the instrumental contribution. With some buffer preparations, even 5% concentration mismatches give negligible mixing heat. These results raise questions about the common use of an ad hoc constant in the least-squares analysis of protein binding data.