I am reading Harry Markowitz’s “ Portfolio Selection ” ( Journal of Finance , 1952), and I cannot determine whether Figure 4 contains a labeling error or whether I am misunderstanding the paper. On PDF page 11 (journal page 86), Figure 4 shows a three-dimensional coordinate system with axes X1X_1 , X2X_2 , and X3X_3 , together with three lines labeled l1234l_{1234} , l234l_{234} , and l124l_{124} . On the following page, Markowitz explains Figure 4 in footnote 10: The attainable set is represented, in three-space, by the tetrahedron... Let S123S_{123} be the subspace consisting of all points with X4=0X_4=0 ... For each subspace Sa1,,aaS_{a_1,\ldots,a_a} , we can define a critical line la1,,aal_{a_1,\ldots,a_a} ... If a point is in Sa1,,aaS_{a_1,\ldots,a_a} and is efficient, it must be on la1,,aal_{a_1,\ldots,a_a} . My understanding is that each ll -line represents the efficient set within the corresponding subspace, and that its subscripts identify the securities that may have nonzero portfolio weights. If that interpretation is correct, l124l_{124} should lie on the face where X3=0X_3=0 , because only securities 1, 2, and 4 are included. In Figure 4, however, l124l_{124} appears to lie on the “floor” of the tetrahedron, where X1=0X_1=0 and X2X_2 , X3X_3 , and X4X_4 may be nonzero. That face would therefore seem to correspond to S234S_{234} . Likewise, l234l_{234} appears on the left face of the tetrahedron, where X3=0X_3=0 , which would seem to correspond to S124S_{124} . The label l1234l_{1234} makes sense because that line passes through the interior of the tetrahedron, where all four portfolio weights may be nonzero. Does Figure 4 contain a typo, with the following labels accidentally reversed? $$ l_{234}\longleftrightarrow l_{124}