We are investigating whether there exist instances that truly attain the worst case in the number of subtour elimination constraints (SECs) that the Dantzig-Fulkerson-Johnson (DFJ) formulation requires to prove optimality.