1: \begin{abstract}

2: Most realistic, off-lattice surface simulations are done 

3: canonically---conserving particles. 

4: For some applications, however,

5: such as studying the thermal behavior of rare gas solid 

6: surfaces, these constitute bad working conditions. 

7: Surface layer occupancies are believed to change with temperature, 

8: particularly at preroughening, and naturally call for a grand canonical 

9: approach, where particle number is controlled by a chemical potential. 

10: We report preliminary results of novel realistic 

11: grand canonical Monte Carlo simulations of the 

12: Lennard-Jones (LJ) fcc$(111)$ surface, believed to represent a quantitative 

13: model of {\it e.g.}\ Ar$(111)$.

14: The results are successful and highly informative for temperatures 

15: up to roughly $0.8\,T_m$, where clear precursor signals of 

16: preroughening are found. At higher temperatures, convergence to

17: equilibrium is hampered by large fluctuations.

18: \end{abstract}

19: