Recent experiments have shown that transport properties of molecular-scale devices can be reversibly altered by the surrounding solvent. Here, we use a combination of first-principles calculations and experiment to explain this change in transport properties through a shift in the local electrostatic potential at the junction caused by nearby conducting and solvent molecules chemically bound to the electrodes. This effect is found to alter the conductance of 4,4'-bipyridine-gold junctions by more than 50%. Moreover, we develop a general electrostatic model that quantitatively predicts the relationship between conductance and the binding energies and dipoles of the solvent and conducting molecules. Our work shows that solvent-induced effects are a viable route for controlling charge and energy transport at molecular-scale interfaces.PACS numbers: 31.15. A-,73.30.+y,73.63.-b,85.65.+h Single-molecule junctions, individual molecules contacted with macroscopic electrodes, provide unique insight into the nanoscale physics of charge, spin, and energy transport [1][2][3][4]. To date, the most robust and reproducible approach to assemble single-molecule junctions is the scanning tunneling microscope-based break junction (STM-BJ) technique [5,6], allowing statistically significant measurements of molecular junction conductance [5-9], thermopower [10][11][12], mechanical properties [13][14][15], and binding mechanisms [15]. Previouslydeveloped theoretical approaches have led to quantitative agreement with experiment for molecular junctions, given a good approximation to the junction geometry and a good estimate of the differences in energy ∆E between the junction Fermi energy, E F and the orbital energy of the frontier orbital, either the highest occupied or lowest unoccupied molecular orbital (HOMO or LUMO, respectively) [16]. Theoretical works focusing on this level alignment [17,18] have led to increased understanding and control of molecular junction conductance and thermopower in terms of junction level alignment, with significant impact on experiments [10][11][12]19].Commonly, these experiments take place at room temperature in a non-conductive solvent [5][6][7][8], which has recently been shown to influence both junction formation probability and, in some cases, to alter the conductance [20]. Despite its practical importance, the impact of these solvents on conductance has not yet been fully understood or explained by theory, in part due to the large computation cost [21], and, therefore, continues to be elusive to control. Previous theoretical works have focused on the effect of solvent on the average [22,23] and dynamical [24] molecular junction geometries, and how they affect level alignment and modify the conductance [25]. Another focus has been the coupling of transmission channels due to intermolecular hopping between conducting molecules [26], in the case solvent would influence the formation of multiple simultaneous junctions. However a detailed physical picture and quantitative framework for understandin...