We study the dynamical and statistical behavior of the Hamiltonian mean fie
ld (HMF) model in order to investigate the relation between microscopic cha
os and phase transitions. HMF is a simple toy model of N fully coupled rota
tors which shows a second-order phase transition. The solution in the canon
ical ensemble is briefly recalled and its predictions are tested numericall
y at finite N. The Vlasov stationary solution is shown to give the same con
sistency equation of the canonical solution and its predictions for rotator
angle and momenta distribution functions agree very well with numerical si
mulations, A link is established between the behavior of the maximal Lyapun
ov exponent and that of thermodynamical fluctuations, expressed by kinetic
energy fluctuations or specific heat. The extensivity of chaos in the N-->i
nfinity limit is tested through the scaling properties of Lyapunov spectra
and of the Kolmogorov-Sinai entropy, Chaotic dynamics provides the mixing p
roperty in phase space necessary for obtaining equilibration; however, the
relaxation time to equilibrium grows with N, at least near the critical poi
nt. Our results constitute an interesting bridge between Hamiltonian chaos
in many degrees of freedom systems and equilibrium thermodynamics. (C) 1999
Elsevier Science B.V All rights reserved.