杏吧原创

SINCE prehistory, humans have looked to the stars and seen gods and monsters. But now there are new constellations in the night sky. Their invisible lines criss-cross the cosmos, linking stars of different ages, colours and brightness in a tangled network that seems at first to have no clear purpose. But unlike the familiar classical constellations, these are more than aesthetically pleasing patterns: they mark out the first family tree of the stars (see 鈥�Family portrait鈥�).

For now this is just a tentative sketch of stellar heredity, but it is already providing intriguing insights. As the tree grows, it could reveal the origin of the sun and tell us how the galaxy grew.

The idea was sparked a few years ago at a seminar at King鈥檚 College, Cambridge, where astronomer Paula Jofr茅 heard anthropologist Robert Foley talking about phylogenetic trees, a way of classifying organisms that lies at the heart of evolutionary science. That piqued her interest. 鈥淭he tree is a tool in many fields of evolution, easy to visualise and well-developed. My field is the evolution of the galaxy, but we never use trees,鈥� she says.

So Jofr茅 approached Foley and asked him whether the same mathematical tools could be used to make a phylogenetic tree for stars. His initial reaction was: no. 鈥淵ou can build a tree out of any data 鈥� you could build one for a set of coins or light bulbs,鈥� says Foley, but the result won鈥檛 mean anything unless the things reproduce and some traits are passed on from one object to the next. Coins and light bulbs don鈥檛 have children, so their trees would be meaningless. Foley thought the same would be true for stars.

But as the two discussed the question, it became clear that, in a manner of speaking, stars do reproduce. Deep within a star, nuclear fusion reactions forge light elements into heavier ones, which are then returned to space as the star reaches the end of its life. The most massive stars burn through their stock of fuel in a few million years and then explode as supernovae to scatter their elemental DNA. They are not around to see any descendants grow. Slightly less massive stars may go through a pulsating red giant stage that throws off some of their outer layers and then burn on for billions of years. In both cases, the ejected matter mixes with existing interstellar gas, fertilising it with heavy elements.

This enriched gas may drift uneventfully for many millions of years before it reaches the next stage of the life cycle, gathering into a cloud large and dense enough to collapse and form new stars. The rather impersonal mechanics of stellar procreation mean that each baby star will have many star-parents, all contributing to the cloud that ultimately decides their baby鈥檚 chemical make-up.

To make sense of this messy process, Jofr茅 and Foley joined forces with astronomer Payel Das at the University of Oxford and biologist Jaume Bertranpetit at Pompeu Fabra University in Barcelona. The set of stars they chose included our own sun and 21 others within 200 light years of us, all broadly similar in size and temperature, and with the spectrum of light they emit betraying their surface composition.

Where a population of related organisms might have similar DNA, stars born from the same cloud of gas should be more similar in chemical make-up than completely unrelated stars. So as their parallel to genetic code, the team took the ratios of 17 different elements with a range of masses, from carbon to barium.

Written in the stars

To create a tree from the data, they had to choose an appropriate tool. 鈥淭here are a lot of techniques available to reconstruct evolutionary history,鈥� says Foley. The most powerful ones, he says, are tailored to analyse the specific four-base code of DNA, and couldn鈥檛 be readily adapted to deal with the fuzzy chemical proportions of stars. So instead, the team reached for a cruder tool known as the neighbour-joining method, which works for any kind of numerical trait, be it height, shoe size or chemical make-up.

Start by identifying the two most chemically similar stars, and declare that they are descendants of one common ancestor 鈥� not a single star, but likely to be a gas cloud enriched by several giant stars. This cloudy ancestor serves as the first node in the tree, with a chemical composition dictated by the average of its two offspring. Then you can look for the star most similar to that ancestor. This star and its doppelg盲nger gas cloud will have a common ancestor of their own, which forms another, higher node. And so on, back through the generations.

with three branches (See 鈥�Family portrait鈥�). The largest, the red branch, holds the sun and eight other stars, all descended from the same common ancestor. You could call these cousins of the sun 鈥� although some of them, such as the chemically closest HD2071, seem to be nieces and nephews, formed from a gas cloud only one generation removed.

Most of the other stars settled into two distinct branches, marked in yellow and pink. But there are also six green suns whose exact relationship to each other or to the other branches remains unknown.

To help make sense of what these branches might mean, the team also gathered data on the ages and motions of each star. All the red branch stars were relatively youthful, none older than our sun鈥檚 4.5 billion years. Taken together, their orbits are orderly, following a circular path around the galaxy very close to the galactic plane. The pinks, by contrast, are over 9 billion years old, with scattered, elongated and slanted orbits. This confirms existing pictures of the galaxy鈥檚 evolution: an older thick disc of stars formed suddenly, and later on a thinner disc developed in which star formation is still going on.

The yellow group is more puzzling. A separate chemical population, these stars are about 8 billion years old, orbiting between the thin and thick discs. One explanation is that they are relics of a cosmic family feud. If about 8 billion years ago another galaxy collided with our own, then gas clouds from the two galaxies would have collided, compressing the gas and triggering a frantic bout of star formation known as a starburst.

This is only one possibility. 鈥淲e can鈥檛 think of a better explanation than a merger,鈥� says Das. 鈥淏ut it is weird you would have as many as four stars from it out of a sample of just 22. The merger should have contributed only a tiny amount to the Milky Way.鈥� Another option the team suggests is that the stars have migrated from further in or further out in the galaxy, where their gas had a different history of enrichment. But then you would expect all the elemental ratios to be different from those of long-term residents like the sun, says Das, whereas some are identical. 鈥淭his is difficult to understand,鈥� she says.

The cluster of stars in green is another puzzle. It is possible that these are all representatives of separate populations, perhaps caused by another galactic collision. Jofr茅, now at the Diego Portales University in Chile, likens this to a natural disaster on Earth. 鈥淵ou have a settled population coexisting, and then a volcano explodes. Families move apart and evolve independently.鈥�

鈥淭here鈥檚 bad news where the sun鈥檚 family is concerned 鈥� its parents have died鈥�

But for now that is very speculative. Deeper insights into the origins of both green- and yellow-branch stars should come as the tree grows. The team is hoping to graft another 10,000 stars onto its branches soon. If this is successful, Das expects the unassigned stars to find their rightful places within the family. And as this larger set of stars stretches across thousands of light years, it could include groups with an entirely separate heritage, their local gas supply evolving a little differently from that of our solar neighbourhood.

Such a mature and majestic tree could give us unprecedented insight into how gas and stars have woven our galactic history. For example, it might shed light on the origin of the thick disc. And it should uncover many more cousins of the sun. There鈥檚 bad news where its immediate family is concerned, though; 鈥淭he parents of the sun have died,鈥� says Jofr茅. Those stars that did most to shape the gas cloud that formed the sun would have perished in the process. But with luck we might find some uncles and aunts, small survivors of its parental generation whose larger short-lived members enriched the protostellar gas cloud.

So far the tree has proved reasonably robust to the team鈥檚 mathematical tests, but that doesn鈥檛 yet mean it is axe-proof. Just as convergent evolution can lead unrelated animals to independently develop similar features, similar chemistry might not necessarily mean shared ancestry. To make sure their conclusions are really solid, the team will try to build a similar tree in different ways. 鈥淭hat鈥檚 a bit technical and boring, but necessary,鈥� says Jofr茅. 鈥淭his is new territory.鈥�

As we set off in search of the sun鈥檚 family, we are expanding our own cosmic connections as well. Every living thing on Earth, from humans to fungi, birds to bacteria, is formed from the same cloud of gas and dust that gave birth to our sun 4.5 billion years ago. The chemical elements vital to our biology were forged in the parents and grandparents of the sun. So in a sense, these newly identified stellar cousins are our family too.

Like all its new-found relations, the sun will eventually go though a red-giant phase and send out some of itself in a strong stellar wind. By that time, five billion years hence, most gas in the Milky Way will have been locked up in all-but-immortal red dwarfs, and the output of a few middleweights like the sun may not be enough to form new stars. But if the timing is right and the sun鈥檚 end coincides with a galactic merger 鈥� perhaps the collision of our galaxy with Andromeda 鈥� then our star could go on to have children of its own.

This article appeared in print under the headline 鈥淭he ultimate family tree鈥�