We present numerical results on the properties of young binary and multiple stellar systems .
Our analysis is based on a series of SPH + N -body simulations of the fragmentation of small molecular clouds , that fully resolve the opacity limit for fragmentation .
These simulations demonstrate that multiple star formation is a major channel for star formation in turbulent flows .
We have produced a statistically significant number of stable multiple systems , with components separations in the range \sim 1 - 10 ^ { 3 } AU .
At the end of the hydrodynamic stage ( 0.5 Myr ) we find that \approx 60 \% of stars and brown dwarfs are members of multiples systems , with about a third of these being low mass , weakly bound outliers in wide eccentric orbits .
Our results imply that in the stellar regime most stars are in multiples ( \approx 80 \% ) and that this fraction is an increasing function of primary mass .
After N -body integration to 10.5 Myr , the percentage of bound objects has dropped to about 40 % , this decrease arising mostly from very low mass stars and brown dwarfs that have been released into the field .
Brown dwarfs are never found to be very close companions to stars ( the brown dwarf desert at very small separations ) , but one case exists of a brown dwarf companion at intermediate separations ( 10 AU ) .
Our simulations can accommodate the existence of brown dwarf companions at large separations , but only if the primaries of these systems are themselves multiples .

We have compared the outcome of our simulations with the properties of real stellar systems as deduced from the IR colour-magnitude diagram of the Praesepe cluster and from spectroscopic and high-resolution imaging surveys of young clusters and the field .
We find that the spread of the observed main sequence of Praesepe in the 0.4 - 1 M _ { \odot } range appears to require that stars are indeed commonly assembled into high-order multiple systems .
Similarly , observational results from Taurus and \rho Ophiuchus , or moving groups such a TW Hydrae and MBM 12 , suggest that companion frequencies in young systems can be indeed as high as we predict .
The comparison with observational data also illustrates two problems with the simulation results .
Firstly , low mass ratio ( q < 0.2 ) binaries are not produced by our models , in conflict with both the Praesepe colour magnitude diagram and independent evidence from field binary surveys .
Secondly , very low mass stars and brown dwarf binaries appear to be considerably under-produced by our simulations .