OSMOPHOBIC AND OSMOPHILIC EFFECTS**

The osmophobic effect has been an accessible phrase to introduce biochemists to ideas of how osmolytes affect biochemical processes. It was built by analogy with the hydrophobic effect, central to protein folding, and the observation that urea primarily interacts with the backbone of proteins, and not significantly with side chains. We have revisited experiments underpinning the original formulation of the osmophobic effect and found that there are likely several errors in the data used, including errors that would indicate an increased role of side chains in the osmophobic effect. We compare the predictive model based on the solubilities of protein backbone and side chains (group transfer free energy model) with a model based on molecular surface types (preferential interaction model). We show that our new solubility data is more consistent with vapor pressure osmometry data used to construct the preferential interaction model. We conclude that available evidence supports using models for osmolyte effects on biochemical processes that rely on exposed atomic surface types rather than on molecular functional groups. But instead of discarding useful terminology that helps newcomers more quickly grasp the importance of the topic, we propose reframing the osmophobic effect from the idea that osmolytes primarily fear the protein backbone to more broadly applicable osmophobic and osmophilic effects where osmolytes are attracted to and excluded from different atomic surface types