1: \begin{abstract}

2: \noindent Recently it has been shown that when an equation that allows

3: so-called pulled fronts in the mean-field limit is modelled with a

4: stochastic model with a finite number $N$ of particles per correlation

5: volume, the convergence to the speed $v^*$ for $N \to \infty$ is

6: extremely slow --- going only as $\ln^{-2}N$. However, this convergence

7: is seen only for very high values of $N$, while there can be

8: significant deviations from it when $N$ is not too large. Pulled

9: fronts are fronts that propagate into an unstable state, and the

10: asymptotic front speed is equal to the linear spreading speed $v^*$ of

11: infinitesimal perturbations around the unstable state. In this paper,

12: we consider front propagation in a simple stochastic lattice

13: model. The microscopic picture of the front

14: dynamics shows that for the description of the far tip of the front,

15: one has to abandon the idea of a uniformly translating front

16: solution. The lattice and finite particle effects lead to a

17: ``halt-and-go'' type dynamics at the far tip of the front, while the

18: average front behind it ``crosses over'' to a uniformly translating

19: solution. In this formulation, the effect of stochasticity on the

20: asymptotic front speed is coded in the probability distribution of the

21: times required for the advancement of the ``foremost occupied lattice

22: site''. These probability distributions are obtained by matching the

23: solution of the far tip with the uniformly translating solution behind

24: in a mean-field type approximation, and the results for the

25: probability distributions compare well to the results of stochastic

26: numerical simulations. This approach allows one to deal with much

27: smaller values of $N$ than it is required to have the $\ln^{-2}N$

28: asymptotics to be valid.

29: \end{abstract}

30: