The water-TiO2 interaction is of paramount importance for many processes occurring on TiO2, and the rutile TiO2(110)-(1×1) surface has often been considered as a test case. Yet, no consensus has been reached whether the well-studied surface O vacancies on the terraces are the only active sites for water dissociation on rutile TiO2(110)-(1×1), or whether another channel for the creation of H adatoms exists. Here we use high-resolution scanning tunneling microscopy and density functional theory calculations to tackle this long-standing question. Evidence is presented that a second water dissociation channel exists on the surfaces of vacuum-annealed TiO2(110) crystals that is associated with the ⟨11¯1⟩ step edges. This second water dissociation channel can be suppressed by blocking of the ⟨11¯1⟩ step edges using ethanol.