A force-free surface (FFS) ${\cal S}$ is a sharp boundary separating a void from a region occupied by a charge-separated force-free plasma. It is proven here under very general assumptions that there is on ${\cal S}$ a simple relation between the charge density $\mu$ on the plasma side and the derivative of $\delta=\E\cdot\B$ along $\B$ on the vacuum side (with $\E$ denoting the electric field and $\B$ the magnetic field). Combined with the condition $\delta=0$ on ${\cal S}$, this relation implies that a FFS has a general stability property, already conjectured by Michel (1979, ApJ 227, 579): ${\cal S}$ turns out to attract charges placed on the vacuum side if they are of the same sign as $\mu$. In the particular case of a FFS existing in the axisymmetric stationary magnetosphere of a "pulsar", the relation is given a most convenient form by using magnetic coordinates, and is shown to imply an interesting property of a gap. Also, a simple proof is given of the impossibility of a vacuum gap forming in a field $\B$ which is either uniform or radial (monopolar).

We consider a three-dimensional bipolar force-free magnetic field with non zero magnetic helicity, occupying a half-space, and study the problem of its evolution driven by an imposed photospheric flux decrease. For this specific setting of the Flux Cancellation Model describing coronal mass ejections occuring in active regions, we address the issues of the physical meaning of flux decrease, of the influence on field evolution of the size of the domain over which this decrease is imposed, and of the existence of an energetic criterion characterizing the possible onset of disruption of the configuration. We show that: (1) The imposed flux disappearance can be interpreted in terms of transport of positive and negative fluxes towards the inversion line, where they get annihilated. (2) For the particular case actually computed, in which the initial state is quite sheared, the formation of a twisted flux rope and the subsequent global disruption of the configuration are obtained when the flux has decreased by only a modest amount over a limited part of the whole active region. (3) The disruption is produced when the magnetic energy becomes of the order of the decreasing energy of a semi-open field, and then before reaching the energy of the associated fully open field. This suggests that the mechanism leading to the disruption is nonequilibrium as in the case where flux is imposed to decrease over the whole region.

The so-called ``symplectic method'' is used for studying the linear stability of a self-gravitating collisionless stellar system, in which the particles are also submitted to an external potential. The system is steady and spherically symmetric, and its distribution function $f_0$ thus depends only on the energy $E$ and the squarred angular momentum $L^2$ of a particle. Assuming that $\partial f_0 / \partial E < 0$, it is first shown that stability holds with respect to all the spherical perturbations -- a statement which turns out to be also valid for a rotating spherical system. Thus it is proven that the energy of an arbitrary aspherical perturbation associated to a ``preserving generator" $\delta g_1$ [i.e., one satisfying $\partial f_0 / \partial L^2 \{ \delta g_1, L^2 \} = 0$] is always positive if $\partial f_0 / \partial L^2 \leq 0$ and the external mass density is a decreasing function of the distance $r$ to the center. This implies in particular (under the latter condition) the stability of an isotropic system with respect to all the perturbations. Some new remarks on the relation between the symmetry of the system and the form of $f_0$ are also reported. It is argued in particular that a system with a distribution function of the form $f_0 = f_0 (E,L^2)$ is necessarily spherically symmetric.

We have performed a series of high resolution N-body experiments on a Connection Machine CM-5 in order to study the stability of collisionless self-gravitating spherical systems. We interpret our results in the framework of symplectic mechanics, which provides the definition of a new class of particular perturbations: The preserving perturbations, which are a generalization of the radial ones. Using models defined by the Ossipkov-Merritt algorithm, we show that the stability of a spherical anisotropic system is directly related to the preserving or non-preserving nature of the perturbations acting on the system. We then generalize our results to all spherical systems. Since the ``isotropic component'' of the linear variation of the distribution function cannot be used to predict the stability or instability of a spherical system, we propose a more useful stability parameter which is derived from the ``anisotropic'' component of the linear variation.