This thesis aims to clarify the formation mechanisms of damage in steel bearings associated with white etching matter (WEM). WEM is nano-crystalline, equiaxed ferrite and is found bordering isolated 'butterfly' defects and extensive, networked white etching cracks (WECs). Recently, WEM formation has been attributed to severe plastic deformation (SPD), induced by various rubbing interactions which arise during rolling contact. However, it has not been assessed whether the morphologies of butterflies and WECs is consistent with this thinking. In this work, selected WECs and butterflies are explored, in 3D, via ion beam serial sectioning; with the positions of the cracks and WEM recorded using various electron microscopy (EM) techniques. The data was used to create 3D renderings, which provide insight into the formation of WEM. It was observed that WEM often lies predominantly along a single side of a WEC; which appears to contradict crack-face rubbing mechanisms. It is proposed that WEM can instead form due to SPD resulting from the concentration of in-service stresses, in the vicinity of crack-tips. Unevenly decorated WECs may then result from cracking along the boundary of the generated WEM. Conversely, WEM was observed to be thicker along mid-sections of the crack network, suggesting on-going WEM growth behind the advancing crack-tip. It is therefore proposed that WEM associated with cracks forms according to multiple SPD-driven mechanisms. For the investigated butterflies, a variety of crack and WEM morphologies were observed. It was not possible to ascribe the formation of the observed WEM solely to rubbing between the initiating inclusion and matrix; given the 3D shapes of the butterflies and an observation of untransformed material between the butterfly inclusion and its WEM 'wing'. Overall the observations appeared to be more consistent with a formation mechanism based on SPD driven by the concentration of in-service stresses around the butterfly inclusion. Selected 2D areas were also investigated with various techniques. In the case of WECs, the focus was cracks exhibiting an uneven decoration of WEM. The aim was to assess whether pre-existing microstructural differences were responsible for the observed WEM asymmetry. No such differences were found however, raising questions regarding the validity of crack-rubbing mechanisms. Wavelength dispersive spectroscopy (WDS) revealed that, contrary to expectation, WEM around WECs may be depleted or enriched in carbon, relative to the matrix material, despite dissolution of the matrix carbides. Conversely, butterfly-WEM was revealed to be consistently depleted. Carbon levels below the (carbide-free) matrix appear to imply expulsion of solute carbon from the WEM. It is suggested that WEM forms according to SPD-driven recrystallisation, leading to a relatively defect-free microstructure which initially expels carbon. It is also proposed that the varying levels of carbon observed are a consequence of variation in the accumulation of carbon-stabilising lattice defects, under continued rolling. Nano-indentation confirmed WEM is harder than the parent material, but no correlation was found between the hardness and carbon concentration; indicating the structural changes of the WEM are the main cause of its increased hardness, rather than supersaturation with carbon, as reported previously. Broadly, the results of this thesis reveal a high degree of variability in WEM, which warrants further investigation and needs to be accounted for in any future mechanisms which are proposed.