An examination of the results of a recent international exchange of tungsten strip lamps reveals that an addition can be made to its analysis. An application of the analysis of variance to the reported data demonstrates, with high probability, the existence of differences between laboratories. It also demonstrates, with the same high probability, the existence of a scale change error in some laboratories, but not in others, at the place where one changes from the H range to the XH range on a Leeds and Northrup pyrometer. Since both sorts of differences are, thus, demonstrably greater than the precision that is available in any one laboratory, the sources of them can, presumably, be found. Procedures for doing so are suggested.