I think I may have found a bug in the graph1 CD. In the commented property of "digraph", it has map(lambda(...)) which is not a complete application of "map", shouldn't it be map(lambda(...), Z) or something like it, which would make more sense? Andrew Robbins