None

Dear Bob: Aurelio's contention is quite correct. In fact, the error was corrected in the final version that has been accepted for the new journal. In fact, the only limits that are needed are products (which certainly exist; they are disjoint unions) and the equalizer of a group of automorphisms (which can be shown to exist). The latter is needed to do the dual of the symmetric algebra construction; the group is the symmetric group acting on tensor powers. Michael ------------------------------------------------------------------------------