Thirty-five polished stone axes and shaft-hole implements which had been previously assigned by petrographic (partially destructive) study to Group XVIII (assumed to be from the Whin Sill dolerite source in northern England), were characterised non-destructively using a combination of portable X-ray fluorescence analysis (PXRF) and magnetic susceptibility measurements. All the implements were compared with the Whin Sill, and with other petrologically and/or visually similar rock sources in the north of England and in Scotland, using a database of chemical elements constructed from the published literature, and statistical atypicality testing. Twenty-nine of the implements are good matches for the Whin Sill as expected, but three do not match Whin Sill rocks and may therefore need to be re-assessed as members of Group XVIII. A further three implements have ambiguous characteristics, perhaps as a result of surface weathering. One axe, an axe-hammer and an adze that had not previously been assigned to implement groups were also analysed, as a test of the robustness of the non-destructive methods. All three implements could be distinguished from Group XVIII on the basis of chemical and magnetic characteristics. Most of the distribution of Group XVIII implements is best explained by the opportunistic use of glacial erratics rather than trade from the primary source. Possible alternative sources for the non-Group XVIII implements include the Scottish Old Red Sandstone age lavas, which could also have been obtained as erratics. The chemical and magnetic characteristics established in this paper provide a non-destructive alternative to thin section study of potential Group XVIII artefacts.