We present a scenario for the variability of the microquasar GRS 1915+105. This starts from previous works, leading to the tentative identification of the Accretion-Ejection Instability as the source of the low-frequency Quasi-Periodic Oscillation of microquasars and other accreting sources. We follow the physics of this instability: its conditions (the magnetic field and geometry adapted to MHD jet models), its instability criterion, and its consequences (cooling down of the disk, heating and excitation of the corona). Comparing them to the observed properties of the source, in particular the detailed properties of its spectral states, we first derive a model for the $\sim$ 30 minutes cycles often exhibited by GRS 1915+105. In our model this is a limit cycle determined by the advection of poloidal magnetic flux to the inner region of the disk, and its destruction by reconnection (leading to relativistic ejections) with the magnetic flux trapped in the vicinity of the central source. We show how it leads to natural explanations for observed behaviors of GRS 1915+105, including the three basic states of Belloni et al (2000). We then discuss how this could be extrapolated further to understand the longer-term variability of this and other microquasars.

We present an instability occurring in the inner part of disks threaded by a moderately strong vertical (poloidal) magnetic field. Its mechanism is such that a spiral density wave in the disk, driven by magnetic stresses (rather than self-gravity as in galactic spirals), becomes unstable by exchanging angular momentum with a Rossby vortex it generates at its corotation radius. This angular momentum can then ``leak'' as Alfven waves emitted toward the corona of the disk thus providing, as an element of the accretion process, an energetic source for a wind or a jet. As galactic spirals, this instability forms low azimuthal wavenumber, standing spiral patterns which might provide an explanation for low-frequency QPOs in low-mass X-ray binaries.

Fragmentation of the planetesimals is found to play a role whose importance is greater than previously believed and depends on the average characteristic size of the primordial planetesimals but not on their mass distribution. In the small size range the mass spectrum is strongly modified with the formation of a small bodies tail but, on the other hand, the growth of the most massive objects is significantly slowed down. The formation of planetary embryos much more massive than the rest of the swarm is delayed; nevertheless the time-scale problem for the growth of the outer planets is not revived.