We show that a combinatorial question which has been studied in connection with lower bounds for the knapsack problem by Brimkov and Dantchev (2001) is related to threshold graphs, threshold arrangements, and other well-studied combinatorial objects, and we correct an error in the analysis given in that paper.