Universe by number: can cosmology be as easy as one, two, three?

Natural History, Dec, 2002 by Charles Liu

The cosmos is a cluttered place. Stars dot the night sky in every direction--hardly surprising, given the hundreds of billions of stars in our galaxy alone. If our own solar neighborhood is any guide, each star among those billions could host a handful of major planets, dozens of moons, millions of asteroids and other minor planets, and billions (if not trillions) of icy comets. Now throw into the center of the galaxy a supermassive black hole two and a half million times the mass of the Sun, add a few billion Suns' worth of free-floating interstellar gas and a few trillion Suns' worth of unseen dark matter, and you get the approximate contents of a single galaxy, our own Milky Way. Finally, multiply that by a hundred billion or so--the number of galaxies between us and our cosmic horizon--and you have the number of objects in the visible universe. It truly boggles the mind.

That sheer quantity alone poses a serious dilemma for us astronomers. As eager as we may be to study every single thing out there, assigning ourselves such a task would be like asking an entomologist to get personally acquainted with every insect on Earth. The numbers leave us no choice but to pick and choose our targets, and then to classify and categorize them: planets into gas giants and terrestrials; galaxies into spirals, ellipticals, and irregulars; and so on. Most of what we understand about the cosmos begins with organizing what we observe into families; all of us, scientists or not, classify things in order to transform raw information into knowledge. But how do astronomers deal with the vast majority of objects in astronomical images--the foreground stars, the background galaxies--when we know almost nothing about the objects but their position and apparent brightness? What can we do with them to learn something about the universe?

Here's a start: we can count them.

Simple addition goes a long way in astronomy. That is particularly true in cosmology, the often difficult pursuit of knowledge about the universe as a whole. In the 1920s Edwin Hubble showed that the Milky Way was just one of a vast number of galaxies scattered throughout the universe. At about the same time, inspired by the idea put forth in Einstein's general theory of relativity that space itself is curved, an international community of physicists and astronomers such as Alexander Friedmann, Georges Lemakre, Willem de Sitter, and Albert Einstein proposed various models of the shape of the universe. Because galaxies are the glowing tracers of matter in the universe, it made sense to map the number and brightness of galaxies all across the sky. The maps could then be compared with the pre dictions of the models to see which model was not ruled out.

Eight decades have passed, and we astronomers are still at it. But along the way we've managed to shed some light, not just on cosmic geometry but also on cosmic evolution. These insights aren't exactly a bonus; they're a product of scientific necessity resulting from the realization that counting galaxies has two major complications.

First, we are embedded in the curved space that Einstein first described; we look out at a vast universe along sight lines distorted by the very galaxies we are trying to observe. Imagine drawing a fine grid of dots on a rubber sheet. Now stretch the sheet tightly against a table, then a saddle, then a beach ball. An ant clinging to the middle of the sheet, aware only of the sheet's two-dimensional surface, would see the dots spaced differently in each case. We earthlings are in the same predicament as the ant, except that our "curved sheet" has three dimensions instead of two. Taking the galactic census of the entire universe therefore requires that we take curvature into account, or we'll misinterpret what we see.

Second, when we look outward into space, we also look backward in time. The observable universe is about 13 billion years old, which we know because its edge is 13 billion light-years away. By definition, then, every single galaxy ill the universe must be younger than 13 billion years of age. Pictures of the universe thus combine past with present. Infant galaxies growing rapidly, adolescents forming new stars at a furious rate, adults making stars gently and steadily--they're all in the mix, all at different distances, and all with different histories of birth and growth.

It gets worse. We also have to take account of a number of other processes, such as galaxy clustering and galaxy merging, surface brightness, and the obscuring power of dust. All of these processes can change with time, drastically increasing or decreasing the number of galaxies we see. It all adds up to a huge mess of cosmic change, collectively called galaxy evolution. Ignoring this factor would completely confound any interpretation of galaxy counts.

To deal with the complications of spatial curvature and galaxy evolution, cosmology-by-galaxy-counts proceeds in stages. First we count the galaxies. Then--just as our predecessors did-we make models of the universe by positing various shapes and evolutionary rates for it, predicting the galaxy counts we would get for each model. Finally, we compare the data with the predictions and tweak the models until they match. The more galaxies we count, the more accurate--and complex--the models get.