Abstract. The starting point of this article is an old question asked by Feferman in his paper on Hancock's conjecture [6] about the strength of ID*. This theory is obtained from the well-known theory ID] by restricting fixed point induction to formulas that contain fixed point constants only positively. The techniques used to perform the proof-theoretic analysis of ID* also permit to analyze its transfinitely iterated variants ID*. Thus, we eventually know that 11 D" | = 11D* |. §1. Introduction. The theories ID a of iterated inductive definitions formalize hierarchies of least (definable) fixed points. In the past years, these theories have been exhaustively studied and their proof-theoretic analysis has been carried out a long time ago, (cf. Buchholz et al. [3]). Also their metapredicative relatives ID a , that speak about hierarchies of (not necessary least) fixed points are well understood by now. The proof-theoretic ordinal of IDi is due to Aczel [1], who used a recursion theoretic argument, nowadays known as Aczel's trick, to embed IDi into 2}-AC. The theories ID" of n-times iterated inductive definitions have been analyzed by Feferman in connection with Handcock's conjecture in [6]. The proof-theoretic analysis of ID tt has been carried out in all details by Jager, Kahle, Setzer and Strahm [9]. _ Some problems however, have remained unsolved: In the theories ID a , induction on fixed points is dropped completely. It is natural to study theories, where fixed point induction is only restricted. Kreisel pointed out in [11], that "an inductive definition tells you what is in P^ not what is not in 9 s *". As mentioned in Feferman [6], this motivated to consider restricted versions of I Di such as IDj, a theory credited to H. Friedman where the scheme for proof by induction on fixed points is restricted to formulas that contain fixed point constants only positively. The question for a sharp upper bound is raised loc. cit. No answer to this question has yet been published, although partial results have been attained: If the fixed point axioms of ID, are restricted to so-called accessibility inductive definitions, then the resulting theory \D*{stf%W) can be embedded in S'-DC as sketched by Feferman in [6]. There, it is also stated that Friedman [8] introduced the theory IDj and showed that its ordinal is bounded by a\, where ao := so and a"+i := y>a"0. Further, upper bounds for