Atlantic City Paper - Discussion of the paper of Prof. Howe on the Use of the Tri-Axial Diagram and Triangular Pyramid for Graphical Illustration (see p. 346)

R. H. THURSTON, Cornell University, Ithaca, N. Y. (communication to the Secretary): The " tri-axial diagram" was, I think, first employed by me in the work of the " U. 8. Board Appointed to Test Iron, Steel and Other Metals." A model was exhibited to the American Association for the Advancement of Science, and the diagram was first published, and the model described, in a paper presented to the same society at its Nashville meeting in 1877." The problem then presented to its inventor was to find a method of planning a research involving proportions of ternary compositions in such manner that, a limited number of such compositions being tested, the results should give the law of variation in the physical properties of all possible combinations of the three elements, and the actual numerical values of such as were important to the investigator. The investigation of the binary alloys, copper and tin, copper and zinc, and tin and zinc, had been completed and satisfactorily discussed, employing the usual bi-axial diagram, of which the ordinates represented the one, and the abscissæ the other, of the characteristic properties to be studied, the composition being commonly indicated by the abscissæ. In these cases it was easy to ascertain, by the study of a limited number of compounds, the law of variation in strength, ductility, etc., with composition; but, for a long time, the ternary alloys were not attacked, because no simple and satisfactory method presented itself of securing results significant for the whole field by the investigation of a comparatively small number of alloys. The tri-axial diagram was finally hit upon as a means of representing and interpreting results, and researches were at once begun, and in due time satisfactorily completed.