1: \begin{abstract}

2: The possibility of having a delocalization transition in the 1D de

3: Moura-Lyra class of models (having a power-spectrum $\propto q^{-\alpha}$)

4: has been the object of a long standing discussion in the literature.

5: In this paper, we report the first numerical evidences that such a

6: transition happens at $\alpha=1$, where the localization length (measured

7: from the scaling of the conductance) is shown to diverge as $(1-\alpha)^{-1}$.

8: The persistent finite-size scaling of the data is shown to be caused

9: by a very slow convergence of the nearest-neighbor correlator to its

10: infinite-size limit, and controlled by the choice of a proper scaling

11: parameter. Our results for these models are consistent with a localization

12: of eigensta\textcolor{black}{tes that is driven by a persistent small-scale

13: noise, which vanishes as $\alpha\to1^{-}$. This} interpretation in

14: confirmed by analytical perturbative calculations which are built

15: on previous work. Finally, the nature of the delocalization transition

16: is discussed and the conclusions are illustrated by numerical work

17: done in the $\alpha>1$ regime.

18: \end{abstract}

19: