1: \begin{abstract}

2: We study the  performance of nonparametric Bayes procedures

3: for  one-dimensional diffusions with periodic drift. We improve 

4: existing convergence rate results for Gaussian process (GP) priors with fixed hyper parameters.

5: Moreover, we exhibit several possibilities to achieve adaptation 

6: to smoothness. We achieve this by considering hierarchical procedures 

7: that  involve  either a prior on a  multiplicative scaling parameter, or 

8: a prior on the regularity parameter of the GP. 

9: \end{abstract}