1: \begin{abstract}

2: We consider one-step methods for integrating stochastic

3:     differential equations and prove pathwise convergence using

4:     ideas from rough path theory. In contrast to alternative

5:     theories of pathwise convergence, no knowledge is required of

6:     convergence in $p$th mean and the analysis starts from a

7:     pathwise bound on the sum of the truncation errors. We show

8:     how the theory is applied to the Euler--Maruyama method with

9:     fixed and adaptive time-stepping strategies. The assumption

10:     on the truncation errors suggests an error-control strategy

11:     and we implement this as an adaptive time-stepping

12:     Euler--Maruyama method using bounded diffusions. We prove the

13:     adaptive method converges and show some computational

14:     experiments.

15: %Include keywords and mathematical subject classification numbers as needed.

16: % \keywords{stochastic differential equations \and numerical

17: %   methods \and rough

18: %     path theory}

19: % \PACS{PACS code1 \and PACS code2 \and more}

20: %\subclass{ 	65C30 }

21: \end{abstract}

22: