“…First, the monolayer partitioning model is fully predictive, has no free parameters, and is only constrained by measurements not related to the droplet experiments, in contrast to the 2D van der Waals model. Second, the Gibbs model tends to overestimate partitioning, as discussed in greater detail in previous modeling studies (17,43,44), and has compared poorly against experimental and field studies in some cases (12,20), providing unphysically large estimations of surface layer thicknesses (20). Here, we compare the experimental data for the glutaric acid-Triton X-100 droplets to a Gibbs model (22) and, similar to these previous studies (12,17,20,43,44), we observe that the extent of surface partitioning is overpredicted relative to the measurements (i.e., the Gibbs model overestimates N surface and underestimates N bulk ), thereby predicting much higher surface tensions than experimentally measured (SI Appendix, Fig.…”