Squats on open tracks, as opposed to track inside tunnels, have been reported by railway networks across the world. Theoretical research hypothesised that the hydraulic entrapment of water is critical to crack propagation associated with squat growth. The requirement for a fluid to enable crack propagation provides a rational to why squats vanish as soon as the railway enters a tunnel. However, the metallurgical evidence from the ex-service rails damaged by the presence of water at wheel and rail contact interface is lacking. In the current work, a localised section of rail was found to suffer severe surface damage associated with water dropping from an air-conditioning system located in a tunnel roof. This section of rail represented an ideal controlled location for detailed metallurgical investigation to examine the formation of white etching layers (WELs) under both dry and wet conditions. The location allowed to determine the relationship between WELs and squat initiation; and also to investigate mechanics of water on crack propagation and squat growth of ex-service rail. It was found that the presence of water on the rail surface is not an important condition for WEL formation. White etching layer in dry rails was observed to be associated with a distinguished sub-layer; called brown etching layer (BEL). Brown etching layer is thought to be a development stage of white etching layer. Crack initiation on rails inside the tunnel was confirmed to be caused by ratchetting (severe plastic deformation of the surface layer), with the requirement of the rail top surface to contain WEL. Crack propagation was manly influenced by the presence of water. The metallurgical analysis revealed significant amount of oxides existed along the crack faces. The presence of oxides associated with an absence of shear deformation on the cracks faces revealed that crack growth was driven by Mode I loading (tensile opening) associated with the mechanism of hydraulic entrapment of water. Furthermore, both fracture behaviour of some cracks, and confirmation of the presence of corrosive products inside the cracks by EDS analysis, represent strong evidence that stress corrosion cracking was involved in both crack growth and the observed formation of secondary cracks.