From 1959 to 1962, I was working as a research assistant to Prof. R. M. Garrels in the geochemistry laboratory at Harvard University. We were measuring the solubilities of various carbonate minerals in solutions saturated with carbon dioxide. Our attempts to measure the solubilities of highly soluble minerals such as hydromagnesite and trona led us to consider that complex ions or ion pairs such as MgCO3 or [NaCO3] - might exist in some of the more concentrated solutions. Venturing into the field of aqueous electrochemistry, using the available electrodes [pH, Na +, K +] and assumptions about solution theory, charge balance, and individual ion activity coefficients, we measured dissociation constants for several carbonate and bicarbonate species. Next we attempted to extend the concept of individual ion activity coefficients and ion pairs toward understanding the factors affecting the solubility of calcite in seawater, which meant working at much higher ionic strength and also having to consider sulfate ion pairs. Finally, we were emboldened to try to calculate the major ion species distributions in seawater, assuming that Cl - was completely ionized. The first version of the manuscript, submitted to the American Journal of Science, was severely criticised by both electrochemists and marine geochemists. The manuscript was rewritten, the solubility of gypsum was discussed, and the title of the paper was changed from something like "Activities and concentrations of major ion species and ion pairs in seawater" to the final simpler and more emphatic one. Both Bob and I were very proud of the fact that the model remained valid as long as twenty-five years after publication.