I am close to completing improved versions of my stochastic frontier functions, but I am trying to get better starting values. This leads me to the following problem. Suppose that a function is given a list X containing the independent variables, which may or may not include the constant (default variable no 0). How do I create a new list X_one which consists of all of the independent variables other than the constant? In effect, this list is the complement of X && const, but I think that I am missing something because I can't see an obvious way of generating it. Gordon