1: \begin{abstract}

2: Conditional stability estimates allow us to characterize the degree of

3: ill-posedness of many inverse problems, but without further

4: assumptions they are not sufficient for the stable solution in the

5: presence of data perturbations. We here consider the stable solution

6: of nonlinear inverse problems satisfying a conditional stability

7: estimate by Tikhonov regularization in Hilbert scales. Order optimal

8: convergence rates are established for a-priori and a-posteriori

9: parameter choice strategies. The role of a hidden source condition is

10: investigated and the relation to previous results for regularization

11: in Hilbert scales is elaborated. The applicability of the results is

12: discussed for some model problems, and the theoretical results are

13: illustrated by numerical tests.

14: \end{abstract}