Until recently it seemed that 1958 would be more notable for excellent work on certain special systems than for powerful contributions to the general theory of solutions. But a paper by Buff & Schindler (1) has done something to redress the balance. Starting with the expansion of the excess chemical potential of one component of a binary solution in powers of the concentration of the other [Buff (2) and Buff & Brent (3)], the radial distri bution functions which enter into the coefficients in this series are now treated by perturbation theory (which is carried out to the first and second orders for simple systems). More explicitly, the interatomic potentials which determine these distribution functions are treated as perturbed forms of that appropriate to the solvent species. First order theory leads to the first order conformal solution theory of Longuet-Higgins (4), involving terms which may be regarded as negligibly small. Ignoring these and comparable terms, Buff and Schindler show that to deal with the second order terms it is neces sary either to approximate to triplet distribution-functions or to make a physically plausible assumption about the solute-solvent distribution func tion in the limit of zero concentration. Two such assumptions are considered, one of which leads to the cell-model formulae of Salsburg & Kirkwood (5), and the other to those of the two-fluid model of Prigogine & co-workers (6) and Scott (7). Evidence is given for preferring the second of these approxi mations.Rather similar work, although carried less far, has simultaneously been published by Mazo (8). This starts from the more cumbersome expressions of Kirkwood & Buff (9). For classical solutions Mazo goes only as far as the first order terms. But for isotopic mixtures he proceeds to the second order, obtaining results closely related to those of Chester (10). Both papers are useful in tying up previously rather unrelated approaches to approximate solution theory.But the reviewer would like to pin his most personal remarks to a paper by Rubin & Sundheim (11). This somewhat obscure paper deals with radial distribution functions in binary gaseous mixtures and discusses the problem of condensation in what is generally known as Born-Green theory. Since the present author was concerned with this problem when last actively work ing on solution theory, a brief comment on it may be permitted.As is well known, the Born-Green theory for a simple fluid produces a nonlinear integral equation for the radial distribution function g(r), which, after being linearised, can be solved by a Fourier-transform procedure. In more detail, if g(r) =exp [fer) -