A strange pattern has emerged in the orbits of planets around other stars. It turns out, planets known to have companions orbiting the same star have very circular paths. However, when a planet is the only one we've detected in its system, the orbit is often very long and thin.

As Johannes Kepler demonstrated, and Newton explained, objects orbit the Sun in ellipses, although for our system's eight planets these are not far off from circular. For generations, astronomers have wondered whether this is a typical or an unusual feature of our family of worlds. The question has huge implications for the prospects of extraterrestrial life – the more eccentric (that is, far from circular) an orbit is, the harder for life to flourish between being frozen and boiled. Answers may also shed light on planetary formation.

Although we have discovered thousands of planets around other stars, all we have learned about their orbits has been the average distance. Dr Ji-Wei Xie of Nanjing University in China has changed that by measuring the eccentricities of 698 planets discovered by the space telescope named in Kepler's honor. The study is published in the Proceedings of the National Academy of Sciences.

Xie found an odd pattern to the results. Of those studied, 368 represented singles – the only planet in a system that Kepler detected. The other 330, known as multiples, are part of systems where more than one planet was observed crossing the face of their star. The singles may not be the only planets orbiting those stars, but any others either have orbits angled such that they don't pass in front of the star from our location or are too far out to do it often enough for Kepler to have seen the pattern.

A perfectly circular orbit is considered to have zero eccentricity, while a value of one indicates a parabola. The singles Xie observed had an average eccentricity of 0.32, but the typical multiple was just 0.04. For comparison, Earth has an eccentricity of 0.017. Mars, at 0.09, is the most eccentric planet in the Solar System since Pluto's demotion.

The multiples look very much like the Solar System, with planets traveling in near-circular orbits in almost the same plane. The singles, however, appear to be operating very differently. Nevertheless, the eccentricity of most singles is in keeping with the previous data we had for exoplanet eccentricity, which was mostly based on planets of Jupiter's mass or larger.

If it were not for the other planets in our system, Earth might also have an elongated orbit, which might well have prevented our evolution.

Eccentricities were calculated by measuring how long it takes for the planet to transit the star. Using Kepler's third law, this gives the distance between the two at the time. This can be compared with the average distance, calculated using the time between transits.

These are the same for circular orbits and for eccentric orbits where the transit happens to occur at the average distance. For other orbits, the difference gives an indication of eccentricity. An individual planet's eccentricity cannot be found this way, but Xie collected enough data to allow statistical modeling.