Colliding galaxies

Some galaxies trail long, thin streamers of stars, like the Mice, about 300 million light-years away, and the Antennae, somewhere between 45 and 80 million depending on how it is measured. For decades they were a puzzle. In 1972 the brothers Alar and Juri Toomre showed they are tides: when two galaxies swing past each other, each pulls harder on the near side of the other than on its far side, and the stars on both sides get flung out, one stream toward the other galaxy (a bridge), one away from it (a tail).

Their model was deliberately simple, and it is the one running here. Each galaxy is a heavy core, and its stars are a disc of weightless test particles circling it. The cores swing past each other on an open orbit, and every star simply follows the two cores' gravity. A million stars cost almost nothing, because each feels only two masses.

Try Prograde and then Retrograde. The only difference is which way the main disc spins. When its stars orbit the same way the other galaxy passes, they keep pace with the passer-by for a while and feel its pull for much longer, enough to be torn away. Spin the disc the other way and the companion rushes past too quickly to do much. That asymmetry, plain in the Toomres' figures, is why long tails need a prograde encounter.

The Mice and The Antennae use the orbits and disc tilts the Toomres published for those two pairs. Drag to look at them from the angle we see them on the sky.

Everything is in model units: the closest approach is one unit of distance and each core's mass is one. For two galaxies of a hundred billion suns each passing 25 kiloparsecs apart, one time unit is about 190 million years, so a whole encounter takes a billion years or more. The clock shows both.

• The stars do not pull on each other. Real discs are self-gravitating and grow spiral arms and bars of their own.
• There is no dark matter halo and no friction between the galaxies,so the cores swing past and separate. Real pairs lose orbital energy to their halos and eventually merge, which takes a full self-gravitating simulation (Barnes and Hernquist's work is the classic).
• No gas. The bright knots of new stars in the Antennae come from gas clouds colliding, which this does not model.

• Toomre & Toomre 1972 · A. Toomre and J. Toomre, Astrophys. J. 178, 623 (1972), "Galactic bridges and tails": the model and the encounter geometries used here
• Barnes & Hernquist 1992 · J. E. Barnes and L. Hernquist, Annu. Rev. Astron. Astrophys. 30, 705 (1992): dynamics of interacting galaxies
• Barnes 2004 · J. E. Barnes, Mon. Not. R. Astron. Soc. 350, 798 (2004): a modern model of the Mice
• Plummer 1911 · H. C. Plummer, Mon. Not. R. Astron. Soc. 71, 460 (1911): the softened point mass used for each core
• Schweizer et al. 2008 · F. Schweizer et al., Astron. J. 136, 1482 (2008): the Antennae's distance
• Hibbard & van Gorkom 1996 · J. E. Hibbard and J. H. van Gorkom, Astron. J. 111, 655 (1996): the Mice and other mergers in neutral hydrogen