1: \begin{abstract}

2: \noindent

3: This paper 

4: shows how the theory of

5: % develops the use of

6: % demonstrates that an approach based on 

7: Dirichlet forms 

8: can be used

9: to deliver proofs of

10: % allows the proof of 

11: optimal scaling results for Markov chain Monte Carlo algorithms

12: (specifically, Metropolis-Hastings random walk samplers)

13: under regularity conditions

14: which are substantially weaker 

15: % than have been 

16: than those

17: required by the original approach (based on the use of infinitesimal generators).

18: The Dirichlet form methods have

19: the added advantage of providing an explicit construction of the underlying infinite-dimensional context.

20: In particular, this enables us directly to establish weak convergence to the relevant infinite-dimensional distributions.

21: \end{abstract}