1: \begin{abstract}

2: The stochastic Euler scheme is known to converge to the exact solution of

3: a stochastic differential equation

4: with globally Lipschitz coefficients and 

5: even with coefficients which grow at most linearly.

6: For super-linearly growing coefficients convergence

7: in the strong and numerically weak sense remained an open question.

8: In this article we prove for many stochastic

9: differential equations with super-linearly growing

10: coefficients that

11: Euler's approximation does not

12: converge neither

13: in the strong $L^p$-sense 

14: nor in the numerically weak sense

15: to the exact solution.

16: Even worse,

17: the difference

18: of the exact solution

19: and of the numerical

20: approximation diverges to infinity

21: in the strong $L^p$-sense 

22: and in the numerically weak sense.

23: %%%This divergence is also established for other numerical schemes such as

24: %%%the Milstein scheme.

25: \end{abstract}

26: