. J. Chcm. 61, 473 (1983).Partial molar volumes V" of the Fe", ~i " , Fe3', ~r " , ~h " , and M(NH,),0HZ3' (M = Cr, Co, Rh, Ir) aqueous ions have been obtained by extrapolation from density measurements at 298.15 K on the perchlorate salt solutions containing added HCIO, where ncccssary to suppress hydrolysis or oxidation. The values for thcsc high-spin aqua-ions and others in the literature can be represented satisfactorily for ionic chargcs z = I through 4 by whcre ?:,, is thc absolute partial molar volume in cm3 mol-' of MT'(aq) rclativc to ?:,, = -5.4 cm' mol-' for H'(aq), and r-is the Shannonradius in pm of M" with coordination number n,. Volumcs of activation AV,,:Vor watcr exchange correlate inversely with V" within the series ~" ( a q ) and M ( N H~) ~O H~" (~~) at least, allowing predictions of AV,,:!: to be made from V" or the above equation. The physical origin of these correlations is considcrcd. Our interest in the precision measurement of volume properties of aqueous metal ions stems from the observation (1, 2) that volumes of activation AV,," for the exchange of coordinated water on M2+(aq), M(NH3),OH:+(aq), and possibly M3+(aq) seem to correlate inversely with the apparent molar volumes 4, or partial molal volumes at infinite dilution V0 of the respective aqueous ions within each of these series. If this observation has any general validity, the transition states volumes must be much less sensitive to the nature of M than are either AV,,* or vO, i.e., the transition states must resemble each other relatively closely within a series despite some differences in reaction mechanism which AV,," is thought to reveal (1). Furthermore, the correlations themselves would permit prediction of unknown AV,," values from known V0 data (or vice-versa). Finally, if a physically meaningful model is available for the prediction of VO, we have the prospect of predicting AV,,*, and so arguably the mechanistic predisposition of a metal aqua-ion, from basic principles and quantities.The establishment and subsequent application of such correlation and models. however. require values of V0 that are . * accurate to within 2 dm3 mol-I or better. Good modern data are available for aqueous Mn2+, Co2+, and ~i " (3-7), but for Fe2+(aq) there exist only older V0 values which range from -17.7 to -26.4 cm3 mol-' (8-10) (relative to v0 = 0 for H+(aq)), possibly because of the rapidity of aerial oxidation of Fe2+(aq) in the absence of added acid (1 1). For Fe3+(aq), the V0 value cited in Millero's comprehensive review (10) can be traced back to density data of precision no better than 1 part in 10" which implies uncertainties in the derived molar volume of at least + 2 cm3 mol-I at 0.05 mol L-I; solution densities reproducible to about 2 parts in lo6 are now routinely measurable. In any case, addition of substantial amounts of acid is necessary to suppress hydrolysis and polymerization of iron(111) (pK, -2).[I] Fe3+(aq) + H 2 0 S H+ + Feo~"(aq) S polymers Suppression of hydrolysis is also important for Rh3+(aq), -3), for which n...