The method employed by Butler and Symons to obtain individual ion shifts for salts in methanol was incorrect since it took no account of the induced changes in the concentrations of free (OH) and free (lone-pair) groups. The shifts caused by such groups are deduced from an infrared–n.m.r. correlation based on a series of unambiguous results. The resulting individual-ion molal shifts are the weighted average of absolute shifts caused by protons in solvent molecules directly bound to the ions, and those in bulk solvent. The link between molal and absolute shifts is the solvation number of the ion. On the assumption that these data should also obey the established i.r.–n.m.r. correlation, solvation numbers have been deduced, apparently with considerable precision.

Some examples of solvation numbers deduced in this way are ClO^–₄(2), Cl^–, Br^–, I^–(4), Li⁺(4), Na⁺(5) and Mg²⁺(6). The last of these is known to be correct from slow-exchange studies at low temperatures, thereby adding confidence in this procedure.

The significance of the absolute shifts obtained is discussed.