P ROBABLY no single phase of research concerned with liquid ammonia solutions has contributed more to fundamental chemical theory than studies on the nature of solutions of metals.Over a period of years, Professor C. A. Kraus and his students have provided extensive experimental data relative to the nature of such solutions. Largely from these studies there have been developed the basic theories which are now generally used to interpret the physical and chemical properties exhibited by these solutions.Briefly, it is believed that a solution of an alkali metal, M, in liquid ammonia, contains positively charged metal ions and negative electrons. Depending upon the concentra&on, the electrons may be relatively free, or associated with molecules of the solvent ammonia M 4 M t + e -, e-+ nNHa + e-(NHJ.Since the evidence in support of this view has been adequately treated in reviews appearing in THrs JOURNAL (1, 2) and elsewhere (3, 4), it is sufficient for present that both the sodium ion and the ammonium group are ammonated. Aside from the relative merits of differing interpretations of a given set of experimental data, it should be kept in mind that the ultimate clarification of any controversial question is often advanced by the appearance of views not entirely in accord with otherwise generally accepted theories. Particularly is this true when it causes a revival of interest and thereby results in new evidence being brought forth, or in stimulating a more complete utilization of available data.A method for the determination of the density of solutions of salts in liquid ammonia has been described by Johnson and Martens (8). They have measured the densities of solutions of the chlorides, bromides, and iodides of lithium, sodium, and potassium at temperatures ranging from -32' to -60°C., and over as wide a range of concentration as permitted by the solubility of the salts. In each case, the density was found to be a linear function of the temperature, and a t lower concentrations, apparently a linear function of the mole fraction of dissolved salt. From the density data for ~otassium iodide it has been shown that a near purposes to again emphasize the fact that a wide variety linear reiationship exists between the apparent molal of experimentally observed phenomena may be satis-volume of the solute and the square root of the concenfactorily explained on this basis. tration expressed in moles per liter of solution. This An alternative explanation of the nature of systems empirical generalization has previously been shown to consisting of alkali metals in liquid ammonia has re-apply to many aqueous salt solutions. ~h~ molar e n f l y been proposed (5). The idea has been advanced heats of solution of ammonium chloride, ammonium that the alkali metals exist in ammonia, not as ions, but bromide, lead iodide, mercuric iodide, lithium bromide, rather as colloidal particles, a view which was 6rst ad-lithium iodide, rubidium bromide, absolute ethyl alvanced by Ostwald (6). Hence, this theory also denies cohol, and pyridine i...