Missing: Thousands of planets


This chart was assembled from data on the incomparable Extrasolar Planets Encyclopaedia , maintained and curated by the astronomer Jean Schneider. It depicts the 520 exoplanets detected between 1992 and 2010, divided up by detection technique. The accompanying data sheet includes a few notes and caveats about the assumptions I used to generate the chart. There's also an interactive version available.

Finding 520 planets in less than 20 years is an impressive testament to the skill of modern-day planet-hunters. And the rapid, recent acceleration of detections suggests that the next 20 years will see several thousands of additional planets added to our catalog. But given that there are hundreds of billions of stars in our galaxy alone, shouldn't we actually be finding more planets?

Let's assume that planets are an almost inevitable side effect of a star's formation from a collapsing cloud of gas and dust—a hypothesis that is less controversial by the day.

Now, there are still several ways for a star to be bereft of planets. A star can form simultaneously with one or more companion stars, and the group will be collectively bound together by its own gravity; in this case, the gravitational interactions of the stars may disrupt planet formation. Similarly, the gravitational nudge of a passing star could scatter planets from around their sun, casting them away forever into the inky void. Gravitational nudges could also make planets fall into their host stars, the planets' constituent atoms lending a dirty sheen to the surfaces of their suns, like oil slicks on oceans. These and other planet-stripping scenarios are unavoidable occurrences in the universe, so not all stars have planets, but it's still a good bet that most do.

Standing beneath a pristine night sky in the northern hemisphere, an unaided human can, at best, see perhaps 2,500 stars. Another 2,500 or so can be seen in the skies above the southern hemisphere. These ~5,000 stars are visible by virtue of their being very bright or very close to us, so that they flood our eyes with photons, particles of light.

Photons are like currency for astronomers: In general, the more you have, the more you can do. Consequently, the stars we can see with the naked eye on a dark night are, on balance, the easiest for astronomers to study. So if most of these stars have planets, and the stars are so easy to observe, why are there only 500-odd exoplanets known today, and less than 100 known around naked-eye stars, rather than 5,000 or more?

The answer comes in three parts: First, just as there are more small pebbles than giant boulders in the world, small planets are probably far more common than the large ones that are easier to detect.

Second, we haven't been looking long enough or hard enough to detect all these smaller planets around nearby stars; the first exoplanets were only discovered in 1992, and only in the past couple of years have we gained the capability to reliably detect tinier, more prevalent worlds.

Finally, and most importantly, each detection method we use has its own unique observational biases that can blind it to the presence of exoplanets large and small.

It's been my experience that most misunderstandings of exoplanetary discoveries are caused by a lack of familiarity with the capabilities and limitations of each detection method. If you want to be savvy about the search for life beyond our solar system, if you want to be immune against exoplanetary hype and flim-flam, then you need to know the basics of how astronomers find planets in the first place. Tune in tomorrow for a beginner's guide to planet hunting!


  1. [Given that we think there are thousands of extrasolar planets out there in the universe,] “shouldn’t we actually be finding more planets?”

    The limitation isn’t the number of planets out there to be found, its the time our current methods take to find them, and the number of telescopes and other resources we’re using to do it. If you think you know a faster way to find them, maybe you should be an astronomy professor instead of a journalist.

    1. People who comment without reading past the jump, and hence misinterpret it completely, are delightful in a schadenfreudetastic kind of a way.

      1. But if he’d stopped to actually read the article, then he wouldn’t have been able to post so early. Getting that coveted “First” post requires dedication that slacker posters like us just can’t understand.

  2. If you assume that planets are an inevitable byproduct of stars, then of course we should be finding more planets. That’s a pretty significant assumption to just wave your hands past.

  3. Hey Lee– I enjoyed your article on Gl 581g. I’m an astronomer– I work on Kepler and in particular on the habitability of planets around M dwarfs, and your article was a very intelligent piece of coverage. I look forward to your posts this week– good timing given the upcoming Kepler data release!

    1. Thanks, Lucianne. I hope you’ll keep me honest/accurate, particularly regarding the new Kepler stuff… ;)

  4. I take part in the hunt for planets using the website planethunters.org. They are looking for regular people to help analyze the NASA’s Kepler spacecraft data, if anyone is interested. And I completely agree that once you see the difficulty and the amount of data that has to analyzed you gain a greater respect and understanding of where we are at in respect to finding planets and where we are about to go.

  5. Interesting article. Is it possible to see the “dirty sheen” on a star that pulled in a planet?

    Also, anyone can help find potential planetary transits by playing the “game” over at http://www.planethunters.org. They will be getting the upcoming data release that Lucianne mentioned, and need all the help they can get classifying stars and detecting transits!

  6. I thought most stars were part of binary or multiple systems, forming with companions, and so probably lack planets. Is this no longer thought to be the case?

  7. The method used to find these planets finds planets many times bigger than Jupiter. We still can’t detect smaller sized planets.

    1. It’s *easier* to detect larger planets, but we *can* detect smaller ones, it just takes more work, more time, more resources (mores money).

      For example, the smallest exosolar planet so far (that Wikipedia’s list of extrasolar planets is aware of) is Gliese 581e (yes, the same system as the planet from the recent stories), at just over twice the earth’s mass. Then again, it is apparently also very close to it’s star, which makes detection easier.

    2. The method used to find these planets finds planets many times bigger than Jupiter. We still can’t detect smaller sized planets.

      We have a lot more data about moons of giant planets than we have about planets orbiting stars. All four of the gas giants in our solar system have usable moons. Three of them have giant, planet sized moons. There is arguably more usable real estate orbiting the gas giants in this solar system than there is freely orbiting the sun.

      Titan, for example, is the most habitable place we know off Earth.

  8. I agree with Daniel, the method used is being pushed to its maximum planet finding ability.
    I think the Kepler telescope will probably find quite a few of these “missing” planets.

  9. And further on the examination of Kepler data and Planet Hunters — the satellite uses the transit method, which means looking for dips in the light of a star when a planet passes in front of it. We (by which I mean the planethunters.org user base — I say “we” because I’ve personally looked at over 6000 light curves and was the first to note several candidates for planethood) have been doing pretty well.

    One of our users, Goryus, has been carefully weeding through what people have found and removed the eclipsing binary stars, false positives, and transits already found by the official Kepler team to come up with the following collection of could-be planets:


    Just think, out of the 264 listed in there many of them are going to be planets — and one of them could have been found by you. New data is going up Wednesday afternoon, the 2nd.

    Personally, I find it quite a rush.

  10. If there was a cash bounty for discovering planets, I guarantee more would be found.

    Do you even get to name the planets you find?

    1. At the very least, a lot more would be reported.

      Planets right now are named in order of discovery with lowercase letters; for instance Upsilon Andromedae has planets Upsilon Andromedae b, c, d, and maybe e, in order of discovery. This matches how companion stars are named in multiple systems, except those get capital letters. The very first extrasolar planet, 51 Pegasi b, was called Bellerophon but they’ve shown up too quickly for proper names to be too useful.

  11. Anon: “The very first extrasolar planet, 51 Pegasi b, was called *Bellerophon* […]”


Comments are closed.