People have often sought solitude in the starry night sky, and it is an appropriate place for that. The night is dark because, in cosmic terms, our sun and its family of planets are very lonely. Neighboring stars are so far away that they look like mere specks of light, and more distant stars blur together into a feeble glow. Our fastest space probes will take tens of thousands of years to cross the distance to the nearest star. Space isolates us like an ocean around a tiny island.

Yet not all stars are so secluded. About one in 10 belongs to a cluster, a swarm of hundreds to tens of thousands of stars with a diameter of a few light-years. In fact, most stars are born in such groups, which generally disperse over billions of years, their stars blending in with the rest of the galaxy. What about our sun? Might it, too, have come into existence in a star cluster? If so, our location in the galaxy was not always so desolate. It only became so as the cluster dispersed in due time.

A growing body of evidence suggests just that. Although conventional wisdom once held that the sun was an only child, many astronomers now think it was one of about 1,000 siblings all born at nearly the same time. Had we been around at the dawn of the solar system, space would not have seemed nearly so empty. The night sky would have been filled with bright stars, several at least as bright as the full moon. Some would have been visible even by day. Looking up would have hurt our eyes.

The cluster into which the sun was probably born is now long gone. I have pieced together the available data and made an educated guess as to what it might have looked like. From these inferred properties, I have calculated the possible trajectories of former cluster members through the galaxy to figure out where they might have ended up. Although they have scattered and intermixed with millions of unrelated stars, they should be identifiable with the European Space Agency's Global Astrometric Interferometer for Astrophysics (GAIA) satellite, which was launched in 2013. The motions of their orbits and their sunlike chemical compositions should give them away. And reuniting with our long-lost stellar siblings should enable astronomers to reconstruct the conditions under which a shapeless cloud of gas and dust coalesced into our solar system.

Memories of Our Birth
The most compelling evidence that the sun has siblings emerged in 2003, when Shogo Tachibana, now at Hokkaido University, and Gary R. Huss, now at the University of Hawaii at Manoa, analyzed two primitive meteorites that are thought to be almost pristine leftovers of solar system formation. They detected nickel 60, the product of the radioactive decay of iron 60, in chemical compounds where, by rights, iron should be found. It seems a game of chemical bait and switch took place in the meteorite: the compounds originally formed from iron, the iron metamorphosed into nickel, and the nickel was locked in place, forever an interloper.

The iron 60 had to be synthesized and injected into the solar system's meteorites within its radioactive half-life, which recent estimates peg at approximately 2.6 million years. That is a cosmic eyeblink. Therefore, the iron had to come from very nearby—and the likeliest source is a supernova explosion. Based on various isotopic measurements, Leslie Looney of the University of Illinois and his co-workers argued in 2006 that a supernova must have detonated within a distance of five light-years when the sun was only 1.8 million years old. The supernova might have been as close as 0.07 light-year.

If the sun had been as secluded as it is today, the location and timing of the supernova would be quite a coincidence. Was a massive star simply passing by when it decided to blow up? No other supernova has ever gone off at such close range; if it had, it would probably have wiped out life on Earth. A much more plausible explanation is that the newborn sun and the exploding star were fellow members of a cluster. With stars packed so tightly together, a close supernova would not have been so improbable.

The Starry Clusters Bright
The idea that the sun originated in a star cluster is at odds with the classical view of clusters that is still common in textbooks. Astronomers have traditionally classified clusters into two types: so-called galactic, or open, clusters and globular clusters. The former are young, sparsely populated and located primarily in or near the plane of our galaxy. The prototypical example is Praesepe, also known as the Beehive cluster or as M44. It was one of the first objects at which Galileo pointed his telescope 400 years ago, in 1609. What looked like a splotch of light revealed itself as an array of stars—up to 350 of them, all born about 700 million years ago.

In contrast, globular clusters are very old, densely populated and located all around the galaxy, not just in a plane. The first was discovered in 1746 by Italian astronomer Giovanni Maraldi and is now known as M15. It contains about a million stars with an age of about 12 billion years.

The trouble is that neither category fits the sun. Its advanced age of 4.6 billion years suggests it should have been born in a globular cluster, yet its location points to a galactic cluster. Recently, however, we have realized that not all clusters fall neatly into one of these two classical types [see “The Unexpected Youth of Globular Clusters,” by Stephen E. Zepf and Keith M. Ashman; Scientific American, October 2003].

What changed our minds was the star cluster R136, which is located in one of the Milky Way's small satellite galaxies, the Large Magellanic Cloud. First spotted in 1960, R136 was initially thought to be a single, giant star 2,000 times as massive as the sun and 100 million times as bright. But in 1985 Gerd Weigelt and Gerhard Baier, both then at the University of Erlangen-Nürnberg in Germany, used new high-resolution imaging techniques to show that R136 is actually a cluster of about 10,000 stars a few million years old. It is as dense as a globular but as young as a galactic cluster. With characteristics of both types, R136 was the missing link between them. Since then, observers have found several clusters like R136 in our galaxy. Other galaxies such as the Antennae contain hundreds if not thousands of them.

The discovery that stars continue to form in clusters so dense they could be mistaken for a single star was astonishing. It led to considerable consternation among theorists. On the one hand, we were relieved because we had not been able to explain R136 as a single superstar. On the other hand, we had to reconsider everything we thought we knew about star clusters. We now think that the majority of stars, including the sun, are born in tight clusters such as R136. A cluster forms out of a single gas cloud and, over time, evolves into either a galactic or globular cluster depending on its mass and environment.

Dreams from Our Stellar Fathers
The members of a cluster span a range of masses, with a few heavy stars and a multitude of lightweight ones. The least massive, with a tenth the mass of the sun, are the most common, and for every factor of 10 increase in mass, the abundance of stars drops by about a factor of 20.

Thus, for each star of 15 to 25 solar masses—the size of the one that went supernova near the newborn sun—a cluster contains some 1,500 lesser stars. This number sets the minimum mass of the sun's birth cluster. The maximum mass is set by the fact that the larger a cluster is, the longer it takes for massive stars to settle toward the center, where they have the greatest likelihood of affecting their smaller brethren. Based on my simulations, the cluster probably contained fewer than about 3,500 stars.

A star of 15 to 25 solar masses lives for six million to 12 million years before blowing up, so it must have formed about this long before the sun did. In other clusters, such as the famous Trapezium cluster in the Orion nebula, astronomers have found that massive stars are usually the first to form, with sunlike stars arising several million years later.

A cluster of the inferred mass was too flimsy to evolve into a globular cluster. Instead it dispersed after 100 million to 200 million years. The massive stars at its center shed gas in stellar winds (similar to but much more intense than the solar wind) and eventually exploded, reducing the density of material in the cluster and thereby weakening its gravitational field. Consequently, the cluster expanded and might have fallen apart. Even if it survived this early outgassing, interactions among stars and the tidal forces exerted by the rest of the galaxy drove its slow dissolution.

Before the cluster disintegrated, stars were so densely packed that one could easily have passed through the solar system. A stellar close encounter would have pulled planets, comets and asteroids from their original circular, planar orbits into highly elliptical and inclined orbits. Many comets beyond a distance of 50 astronomical units (AU), past the orbit of Neptune, have highly skewed orbits. The internal dynamics of the solar system seem incapable of accounting for these peculiar orbits; the bodies are beyond the gravitational influence even of Jupiter. The most likely explanation is that they were stirred up by a star passing 1,000 AU away. The planets, though, have very regular orbits, indicating that no stellar intruder ever came within 100 AU of the sun.

From these facts, I have estimated the dimensions of the cluster. For another star in the cluster to pass 1,000 AU away with reasonable probability over the cluster lifetime, the cluster had to be less then 10 light-years in diameter. Conversely, for a star not to come within 100 AU, the cluster had to be greater than three light-years in diameter. In short, the sun's birth cluster looked like R136 but much less dense, so that stars were far enough apart not to interfere with planet formation.

Solar Genealogy
Theorists can go further and ask where exactly in the galaxy the birth cluster was located. The solar system revolves around the galactic center in an almost circular orbit, more or less in the disk. At the moment, we are located about 30,000 light-years from the center and about 15 light-years above the plane of the disk, orbiting at a speed of 234 kilometers per second. At this rate, the sun has done 27 circuits since its formation. Its orbit is not a closed loop but a somewhat more complicated shape determined by the gravitational field of the galaxy, which astronomers infer from the motion of stars and interstellar gas clouds.

Assuming, provisionally, that the gravitational field has not changed over the past 4.6 billion years, I have projected the orbit backward in time and deduced that the sun was born 33,000 light-years from the center and 200 light-years above the galactic plane. What makes this position puzzling is that the outer reaches of the galaxy are poorer in heavy elements than the inner parts. The most distant regions may lack enough material to make planets, let alone life [see “Refuges for Life in a Hostile Universe,” by Guillermo Gonzalez, Donald Brownlee and Peter D. Ward; Scientific American, October 2001]. Although the sun's putative birthplace is not quite so impoverished, it is still poorer in heavy elements than the sun is. Based purely on the sun's heavy element composition, astronomers would have expected it to form 9,000 light-years closer to the center.

Maybe the supernova that seeded meteorites with iron 60 also enriched the sun with heavy elements. Or maybe my orbital calculation went astray because the gravitational field of the galaxy has changed or because the sun's orbital path was diverted slightly by the gravity of nearby stars or gas clouds. In that case, the sun was born closer to the center than I estimated, and its composition is not so anomalous.

The sun's ex-family members, too, should be orbiting around the galactic center at more than 200 kilometers per second. Yet their relative velocity, which is determined by their mutual gravitational forces in the original cluster, is only a few kilometers per second. Like clumps of cars on a highway, they stick together even though they are no longer bound to one another gravitationally. The original swarm has spread into an arc only very gradually. After 27 orbits, it should stretch about halfway around the galaxy.

My calculations suggest that about 50 of the sun's brothers and sisters should still be within 300 light-years of our current location and that about 400 stars are within 3,000 light-years. Depending on the stars' original relative velocity and the timing of their departure from the cluster, the sun either follows in their orbital footsteps or they in ours.

The best place to look for them is in the plane of the galaxy in the direction the solar system is moving or in exactly the opposite direction. We have been looking for them in a catalogue of stars assembled by ESA's Hipparcos satellite [see “The Star Mapper,” by Philip Morrison; Scientific American, February 1998]. But the Hipparcos data are not precise enough to identify any siblings. A breakthrough came earlier in 2014, when Ivan Ramirez of the University of Texas at Austin and his co-workers measured the chemical composition of the yellow dwarf star HD162826. They found it to be very similar to our sun's, as if it had been exposed to the same supernova that enriched our solar system. HD162826 could be one of our lost siblings, although more detailed information will provide more proof. Such proof could come from the ESA satellite GAIA, which has two telescopes measuring the position and velocity of some one billion stars over five years, creating an essentially complete census of stars brighter than 22nd magnitude within some 30,000 light-years of the sun. The solar siblings will stand out in these data because they lie nearly along the sun's past and future path. Identifying even one sibling of the sun, or confirming HD162826's genealogy, will provide crucial information about the very early days of the solar system. Theorists will be able to compute the birthplace of the sun with greater certainty and determine, for example, whether the gravitational field of the galaxy has changed substantially or not. Solar siblings will also be excellent places to look for habitable planets. Although we seem very alone in the galaxy, it was not always this way. Many of the sun's seeming idiosyncrasies—particularly its nurturing of life—might make sense in the context of its family.