1: \begin{abstract}

2: 

3: We simultaneously study the  dynamics of the growth of errors and the question of the

4: faithfulness of simulations of $N$-body systems.  The errors are quantified through the numerical 

5: reversibility of small-$N$ spherical systems, 

6:  and by comparing fixed-timestep runs with different stepsizes. 

7: The errors add randomly, before exponential divergence sets in,    

8: with exponentiation rate virtually independent of $N$,  but scale saturating as $\sim 1/\sqrt{N}$, 

9: in line with theoretical estimates presented.

10: In a third phase, the growth rate is initially driven by 

11: multiplicative enhancement of errors, as in the exponential stage. It is then 

12: qualitatively different for the phase space variables and 

13: mean field conserved quantities (energy and momentum); for the former, 

14: the errors grow systematically through phase mixing,  

15: for the latter they grow diffusively. For energy, the $N$-variation of the `relaxation time' of error growth follows 

16: the $N$-scaling of two-body relaxation.  This is also true for angular momentum 

17: in the fixed stepsize runs, although the associated error threshold is higher and the relaxation time smaller.

18: Due to shrinking saturation scales,  the information loss associated 

19: with the exponential instability decreases with $N$ and the 

20: dynamical entropy vanishes at any finite resolution as $N \rightarrow \infty$. 

21:  A distribution function depending 

22: on the integrals of motion in the smooth potential is decreasingly affected. 

23: In this sense there is convergence to the collisionless limit, 

24: despite the persistence of  exponential instability on infinitesimal scales. 

25: Nevertheless, the slow $N$-variation in its saturation points to the slowness of the convergence.  

26: 

27: 

28: 

29: \end{abstract}

30: