The planets of our solar system follow nice, predictable orbits, but it was not always so. In the chaotic early days of the solar system, Jupiter and its fellow giant planets seem to have migrated from their birthplaces into the stable orbits that we observe today.

The migration of giant planets has been invoked to explain a number of features of planetary systems, such as the uneven spacing among the objects of the Asteroid Belt in our solar system. Migration would also explain the huge planets in other planetary systems known as "hot Jupiters" that orbit extremely close to their host stars, far closer than where they could have plausibly formed.

Now a new study, published online June 5 in Nature, demonstrates that a peculiar migration of Jupiter—first inward, then outward—could account for Mars's relatively small size. (Scientific American is part of Nature Publishing Group.)

Researchers have built a number of numerical simulations to try to trace how the planets formed in the tens of millions of years following the solar system's birth. But often, Mars has posed a problem. "As things would slowly build up, essentially what always would happen is you'd end up with planets of about the size of Earth and Venus where they should be," says lead study author Kevin Walsh, a planetary scientist at the Southwest Research Institute in Boulder, Colo. "But the object that ended up around the location of Mars was the size of Earth as well." In actuality, compared with Earth, Mars is only about half its diameter, and about one tenth its mass.

The problem would go away if there were simply less raw material available to Mars when the planet developed. To try to account for that paucity, Walsh and his colleagues designed a model in which Jupiter's motion sweeps out many of the planetesimals (small bodies) that collided to form the terrestrial planets. Today, Jupiter orbits the sun at about five times the Earth–sun distance, or five astronomical units (AU). But interactions with the gaseous disk that surrounded the young sun could well have drawn Jupiter inward in the first millions of years following the solar system's birth.

Walsh and his colleagues found that if Jupiter drew all the way in to about 1.5 AU, then retreated outward due to gravitational interactions among Saturn and the gaseous disk, its motion would scatter most of the planetesimals situated beyond Earth's orbit. So while Earth and Venus continued to accrete more protoplanetary material, Mars's growth further out was stunted. The entire inward and outward migration could have taken place within a few hundred thousand years; the planets of the inner solar system would then have finished their development over tens of millions of years.

Jupiter's migration would account for the divergent sizes of the terrestrial planets, and it also seems to gibe with the current state of the Asteroid Belt, which lies between the orbits of Mars and Jupiter. "The real big test for this was the Asteroid Belt, which is right there in the firing line where Jupiter goes back and forth," Walsh says. "We really didn't know if we'd get an asteroid belt that made any sense." But Jupiter's passage in and out appears to scatter the right amount of material to leave today's belt in place, and it could also account for the fact that the inner and outer parts of the Asteroid Belt have somewhat different populations.

In the simulation Walsh and his colleagues designed, one family of asteroids originates among the giant planets of the outer solar system, and another family originates much closer to the sun. The migratory motions of the giant planets mix those two asteroid families somewhat, but not completely, leaving two discernable populations within the belt. "We ended up with a really nice match with the Asteroid Belt," Walsh says. "The total mass worked out pretty nicely, and we were also able to reproduce this dichotomy in the Asteroid Belt."