Hubble Ultra Deep FieldMost people think of astronomy as the science of big things—"vastly, hugely, mind-bogglingly big," as Douglas Adams put it. But judging from last week's American Astronomical Society meeting in Washington, D.C., the most interesting things in astronomy these days are the small(ish) ones. Planet-hunters, having racked up hundreds of Jupiter-size worlds, now seek Earth-size ones. Astrophysicists studying supermassive black holes are seeing the charms of the merely massive ones. Those picking apart the history of galaxies think the clues may lie in galactic dwarfs.

In one of my favorite talks of the meeting, Marta Volonteri of the University of Michigan discussed how supermassive black holes formed—a mystery on every astronomers' top-10 list. Holes with the mass of a billion suns were in place by a cosmic age of about a billion years, which is vexingly fast.

They presumably formed in a two-stage process: a seed black hole nucleated and then sucked in matter. But sucking is slow, and to bulk up to the observed size in time, the seed must have started off with hundreds of solar-masses, at least. Whatever could have begat such a beast? Sufficiently large stars collapse to a black hole when they die, but stars are inherently limited in mass. The gas clouds whence they come tend to fragment, leading to multiple middling stars rather than a single heavyweight one. In today's universe, the best you might hope for is a star that leaves behind a hole of perhaps 10 solar-masses, a woefully inadequate seed.

But the ancient universe had a major advantage: it lacked heavy elements, since stars had not yet had a chance to synthesize them. Those elements are more efficient at radiating heat than hydrogen or helium is, so they help a gas cloud to cool and fragment. Lacking them, gas in ancient times underwent less fragmentation. From this idea, Volonteri mapped out two general strategies for creating seed black holes. In the first, the seeds were the corpses of the very first generation of stars ever to form in the universe, which were a race of Titans. In the second approach, gas clouds of a thousand solar-masses or more became gravitationally unstable and collapsed directly into black holes. Neither process operates anymore, because the universe has gotten too enriched in heavy elements.

Which process accounted for the seeds? It's hard to tell. Regardless of how the seeds formed, all ended up with a mass limited by the available material—like children who have early or late growth spurts, but wind up equally tall. Their origins are lost in antiquity. Or are they? Volonteri pointed out that the universe contains relatively small black holes—seeds that evidently grew hardly at all. Their properties should still reflect how they formed.

Another question on the top-10 list is how galaxies formed. At a widely reported press conference, Garth Illingworth of the University of California, Santa Cruz, presented images from the Hubble Ultra Deep Field, the most sensitive image ever captured by human beings. He announced galaxies at redshifts of 8.5 and 8.7, corresponding to a cosmic age of about a mere 600 million years old. These galaxies were runts, scarcely 5 percent the size and 1 percent the mass of the Milky Way. Taking into account the redshifting, they were unusually bluish, implying a lack of dust and therefore of heavy elements.

Stars within these galaxies appear to be 300 million years old, implying that the galaxies formed at a cosmic age of 300 million years, making them among the first in the universe. Illingworth's team inferred that the star-formation rate back then was a tenth of its present rate. Thus this is the first time astronomers have observed an epoch when the star formation rate was lower than it is today. The star formation rate quickly ramped up, peaked at a redshift of about 3, and has been on the downhill slope ever since.

One caveat: the galaxies are too dim to measure their real spectra, so the team made do with a poor man's substitute: the relative amounts of light in several different wavebands. The galaxies are invisible in visible light, but show up at wavelengths longer than 1 micron; from this, the researchers deduced their redshifts. Given the uncertainties of this procedure, the level of precision implied by the value 8.7 strikes me as dubious. It might be better to say "about 8." As for the stellar ages, the researchers deduced them by mixing and matching stellar types to try to reproduce the observed galaxy colors. The galaxies are significantly brighter at 3 microns than at 1 micron, suggesting they contain relatively red (hence longer-lived) stars. Again, though, a number of 300 million years perhaps overstates the precision of this estimate.

The present-day universe contains similarly runty galaxies, which match the ancient dwarfs both in their size and paucity of heavy elements. Several researchers at the meeting spoke of a curious fact about these little guys: the supernova explosions that go off in them are downright weird.

Andrew Drake of Caltech described the results of the Catalina transient survey, whose main task is to scan for threatening asteroids, but incidentally picks up all manners of blinks and flashes in the heavens, including supernova explosions. Among its discoveries was SN 2008fz, the most energetic ever seen, 10 times brighter than usual. It went off in a smallish galaxy about the size of one of the Milky Way's satellites, the Large Magellanic Cloud. Another was SN 2008iy, which took 400 days to reach its peak brightness, 20 times longer than usual. It occurred in an even tinier galaxy.

Peter Garnavich of Notre Dame announced results from another supernova survey, ESSENCE, including another strange supernova, Y-155. Not only was it 10 times brighter than usual, it got hotter with time rather than cooler. Garnavich argued it was a so-called pair-instability supernova, which occurs when radiation is so intense that it creates new particles of matter. In December, another team concluded that SN 2007bi, a bright and slowly peaking supernova, was also a pair-instability. Both of these went off in low-mass galaxies.

Again, the reason may be the deficit of heavy elements. Pair-instability explosions require a star of gargantuan mass, 150 to 250 times the sun's, which arise only when the heavy element abundance is low. Dwarf galaxies undergo less reprocessing of material over the eons and therefore accumulate fewer of these elements.

They also tend to lose whatever material they have, as two other studies revealed. Stacy McGaugh of the University of Maryland looked at a sample of galaxies of varying sizes and estimated their mass in two ways: first, he added up all the stars; second, he measured the velocity of those stars and inferred how strong gravity must be to contain them. This kind of analysis has been done before, and it demonstrates that galaxies are much heavier than their stellar count would suggest; the difference is made up by dark matter. (An alternative, which McGaugh has argued in the past but did not push at this meeting, is that the discrepancy indicates a failure of the laws of physics.)

McGaugh showed that the smaller a galaxy is, the greater the discrepancy—indicating a proportionately greater amount of dark matter or, more likely, a lesser amount of ordinary matter. With their weak gravity, dwarf galaxies are less able to contain the debris of supernova explosions, so they lose it to intergalactic space.

Niv Drory of the Max Planck Institute for Extraterrestrial Physics reached a similar conclusion through very different means. His team's COSMOS survey examined 300,000 galaxies in the relatively nearby (hence recent) universe and compared the number of galaxies of different masses to simulations of how many there should be, if galaxies consisted solely of dark matter. They found fewer galaxies than the simulations predict, and again the discrepancy worsens as you go down in size—suggesting that ordinary matter gets ejected more readily from smaller galaxies.

It may be a long time before astronomers fully grasp how black holes and galaxies form, but that doesn't seem to concern them. A striking thing about astronomy is that it is self-consciously a multigenerational effort. Many of the keynote speakers said they wouldn’t find out the answers to big questions in their lifetimes, but that younger members of the audience would.

Image: Deepest image ever taken in near-infrared light, made by the Hubble Space Telescope last August. Credit: NASA, ESA, G. Illingworth and R. Bouwens (University of California, Santa Cruz), and the HUDF09 Team