the disc.
So could such a jet, produced by activity in the core of IC2497,
have excited the glowing gas in the Voorwerp? Such a scenario
seemed more plausible after a close look at the neighbouring gal-
axy itself. It seemed to be warped, its disc twisted and marked by
thick dust lanes—features which suggest a recent interaction
with a neighbour. Such an interaction, or even a merger, might
have funnelled material down into the galaxy’s central region
where the black hole lurks, piling on material and making the
presence of a jet much more likely, and if such a jet existed at
right angles to the galaxy’s disc, it looked plausible that it would
hit the Voorwerp directly.
196 Serendipity
Mystery solved. Yet the details didn’t add up. If there was
enough material falling into the black hole to produce a jet able
to excite the Voorwerp, then the gas and the dust that immedi-
ately surrounds the black hole should also be warm, glowing
brightly in the infrared. The Voorwerp’s neighbour is bright in
the infrared—it’s what astronomers call an LIRG, or luminous
infrared galaxy—but it wasn’t bright enough to allow for the
necessary activity. Despite the neatness of the explanation, it
was clear that we couldn’t hide a powerful enough source of
activity at the core of the galaxy to account for the Voorwerp’s
brightness.
This was confusing, but slowly, in discussions and emails and
phone calls, those of us on the Galaxy Zoo team who were
increasingly consumed with trying to answer Hanny’s simple
question realized that we had to consider the possibility that the
Voorwerp was an evolving object, capable of change. The dis-
tance between the Voorwerp and the galaxy is about 50,000
light years.* This means that light travelling from the centre of
the galaxy to the Voorwerp, and then heading to Earth, would
arrive 50,000 years after light taking the direct route. Put another
way, the Voorwerp is a light echo—it reveals to us the state of its
neighbour galaxy as it was 50,000 years ago.
The fact that the Voorwerp is highly excited now tells us that,
back in the stone age here on Earth, the black hole at the centre of
IC2497 would have been feeding heartily, even though today it is
relatively quiet. Had our Cro-Magnon ancestors had stone age
binoculars, they would have found the galaxy—now too faint to
be seen with small telescopes—easily visible, the nearest example
of what we call a quasar. In the time that’s passed since then—
* That’s the projected distance on the sky—the distance you get if you assume that the lines joining us to the Voorwerp and the Voorwerp to the galaxy make a right angle. If the Voorwerp is in front of or behind the galaxy, that distance might really be only 25,000 light years or as much as 75,000 light years.
Serendipity 197
not long in the 13.8 billion year long history of the Universe—the
galaxy must have quietened down. Perhaps the black hole ran
out of fuel, or the activity around the black hole might have
heated the surrounding gas, preventing further collapse in a pro-
cess known as feedback. In either case, as long as something dra-
matic had changed in the galaxy in the last few tens of thousands
of years, it would explain what we saw.
This is an exciting idea. We can monitor what the black holes
at the centres of galaxies are up to on human timescales, return-
ing year after year to the same objects to see what, if anything,
has changed. We can see too how the population of active galax-
ies changes over cosmic time, by comparing the population of
galaxies that we see at different epochs of the cosmos’s history.
But being able to trace changes on a timescale of thousands of
years, in-between these two more accessible timescales, is unique.
Not that having had the idea was enough. Getting from there
to a proper result was hard work, and a particularly hard slog for
me; my PhD had made me an expert in using radio telescopes,
not in handling data from anything with a mirror, and so I was
learning what to look for in spectra and in the other data we
gathered as we went along. With much help from Bill and others,
eventually a short paper was ready, and we sent it off to our nor-
mal research journal.
We didn’t have to wait long for the referee’s report to appear in
my inbox. Unfortunately, it was as critical as anything I’d seen.
My analysis had, it said, basic problems, and the interpretation
I’ve just spent paragraphs convincing you of was overblown. Our
conclusion that the black hole couldn’t currently be active
enough to account for the Voorwerp’s observed brightness was
wrong, we were told.
In some fields, such a bad report would have meant the end of
the paper’s chances of publication. In astrophysics, there’s a
more collaborative approach and it’s possible to go back and
198 Serendipity
forth with the referee until all is well, or until mutual exhaustion
ends the process, at which point the editor as umpire steps in and
ends the fight. This tangle with the referee was one of the most
exhausting of my career, and we went back and forth more than
a dozen times; what had started as a four page paper weighed in
at thirty pages by the time we had satisfied every niggling doubt
in their mind (or, I suppose, won through sheer exhaustion).
This effort mattered. One reason for publishing in a journal is
to show publicly that a result stands up, that it has at least passed the minimum standard required for peer review. A second might
be to spread the word, communicating a result to one’s col-
leagues. Another goal is to build up a public track record of
work—even though there are well-documented problems with
the approach, we use someone’s publication record when con-
sidering them for jobs or for promotions. The fourth reason,
though, is to stake a claim for posterity. Despite all the wrangling
and the back and forth, we needed the paper out so that Hanny—
as a co-author alongside the rest of us—could get the proper sci-
entific credit for her discovery.
The trouble was, the Voorwerp was big news. Hanny had been
on Dutch television, who were reporting on the mysterious blob,
and we’d been blogging to share our progress and our excite-
ment with the Galaxy Zoo volunteers. As a result, a Dutch team
of astronomers had decided to take a look at the new object
themselves, using a network of radio telescopes to get a very dif-
ferent view of the system, one more detailed than had been
obtained at similar wavelengths before. Before our referee was
happy and our paper could be accepted, this rival group pub-
lished their own paper which told a different story.*
* This Dutch group were kind enough to add Hanny and a few of the Galaxy Zoo team to the author list for the paper, but it was still annoying not to get there first.
Serendipity 199
A quick glance at their data showed that the Voorwerp was
bigger th
an we thought. It turns out that only part of it has been
excited enough to glow brightly when viewed in visible light;
there exists a much longer streamer of cold gas wrapped round
the galaxy, exactly as one might expect in the aftermath of a
merger. The Milky Way’s satellites, the large and small Magellanic
Clouds, have been disrupted by the encounter with our larger
galaxy, and trail gas and stars behind them as a result. The conse-
quences of a more dramatic merger should be even more spec-
tacular, producing long ‘tidal tails’, streamers of gas that here
happen to be in the right place to be excited. Was the Voorwerp
just part of such a tidal tail, illuminated by a still-active jet?
The cause of that excitation seemed to be visible for the first
time in the Dutch radio data. Deep in the heart of IC2497, it turns
out there is indeed a jet of material, moving fast and heading
straight for the Voorwerp. This is exactly what we’d originally
expected, and completely consistent with a currently active gal-
axy. By the time our paper was published we had to argue that
there were two possibilities—we were either right, and the galaxy
was now quiet, or our calculations were wrong and the jet discovered
by the Dutch team was evidence that we’d made a mistake.
Luckily, there was a clear test available. Kevin—he of the original
classifications that inspired Galaxy Zoo—and friends applied for
and won time to use two telescopes in space. XMM-Newton and
Suzaku are European and Japanese telescopes that look for x-rays, high-energy radiation emitted most commonly by very hot gas
such as that swirling around an active black hole which is still
growing. Using x-rays also allows us to peer through the dust
clouds that surround the centre of a galaxy like IC2497, getting
directly to the heart of the action.
When we used these telescopes to look at the centre of the
Voorwerp’s neighbour we saw precisely nothing. There were a
200 Serendipity
few stray photons recorded by the detectors, but nothing worth
writing about. In other words, the galaxy was dead, just as we
had predicted. And the implications are startling. The black hole
at the centre of IC2497 probably weighs in at a few million solar
masses, and yet a system dominated by an object that large has
managed to change its behaviour completely in the small matter
of a few tens of thousands of years.
In the past, the Voorwerp must have been the brightest quasar
in the sky, and as it would be too much of a coincidence to have
the nearest such object behave oddly, such behaviour must be
relatively common. Much better to assume that such things hap-
pen all the time than to argue that the nearest such galaxy just
happened to misbehave shortly before we came along to watch.
The lesson of the Voorwerp is that galaxies switch from active to
passive—and presumably back the other way—all the time.
Classifying a galaxy as either active or quiescent becomes not a
property of the galaxy, like its mass, but a statement about what
stage of its life a galaxy is in.
Even the story of our own Milky Way changes when you start
thinking like this. The centre of our galaxy is a quiet place today,
containing a supermassive black hole but one that is quiet and
devoid of any substantial accretion disc. There was great excite-
ment a few years ago when what appeared to be a gas cloud of
about the same mass as Jupiter, named G2, appeared likely to fall
in, and coffin-chasing astronomers were keen to watch. As it
turned out, G2 survived its close passage around the black hole
(it didn’t come close enough to pass the event horizon, the point
of no return) and may well be something more substantial than
just a cloud of gas.* Just the fact that everyone got so excited
* The most favoured hypothesis seems to be a pair of stars embedded in a cloud of gas, for a variety of complicated reasons.
Serendipity 201
about it, though, tells you that not much happens in the centre of
the Milky Way.
Look away from the disc of the galaxy with the right eyes and
the picture changes. NASA’s satellite Fermi has been mapping the sky in light that is even more energetic than the x-rays that
betrayed the Voorwerp’s secrets. The gamma-ray sky is marked
by two large bubbles of hot gas, extending symmetrically for tens
of thousands of light years and centred on the galaxy’s heart. A
good explanation for these structures, now known as the ‘Fermi
bubbles’, is that they’re a pair of shock waves exciting the thin gas that exists beyond our disc, the echoes of a time not that long ago
when the centre of our galaxy was home to an active black hole.
As material fell on the Milky Way’s central engine, it shone so
brightly that as its radiation travels out into space it can still
excite its surroundings.
So even the Milky Way can be active, and the Voorwerp helps
us understand this behaviour. We also managed to get Hubble
Space Telescope images and data, which made the Voorwerp yet
more famous. (It made a brief appearance in one of David
Letterman’s monologues, being revealed at the end as nothing
more than a smear on a camera lens, easily wiped away.) In the
meantime, Galaxy Zoo volunteers in search of their own discov-
eries quickly assembled a set of ‘Voorwerpjes’ (Figure 24).* Each
of them is a glowing blob of hot gas, excited by activity around a
galaxy’s central black hole.
The variety of shapes seen in the Voorwerpje sample, many of
which have now been imaged by Hubble too, is remarkable. There are long braids of gas, which seem to twist round each other.
* Voorwerpen would be the Dutch plural, but these are smaller versions, so we use Voorwerpjes—the diminutive. It’s amazing what you end up having to know in this job.
202 Serendipity
Figure 24 Example Voorwerpjes imaged with the Hubble Space Telescope.
The emission is from excited oxygen and shows a remarkable range of
shapes, which are difficult to explain.
There are dense clouds, in front of, behind, and surrounding
their host galaxies. There are a surprising number of rings, and
features that look rather like the still-mysterious ‘hole’ in the
Voorwerp. The complexities of these shapes are either caused by
Serendipity 203
an uneven distribution of gas or rapid changes in the luminosity
of material falling down into the black hole, and they don’t make
understanding these systems easy. Nonetheless, after a lot of
work it seems clear that about a third of galaxies with visible
Voorwerpen* have faded dramatically, just like the original.
Hanny’s find is still inspiring new scientific discoveries a dec-
ade after it was found, but it can seem like an sideshow compared
to the main Galaxy Zoo story. Galaxy Zoo wasn’t set up to find
giant glowing gas clouds, nor to reward volunteers who got inter-
ested in curiosities they came across while classifying. The vast
majority of the scientific papers which have been published as a
re
sult of the project use the results produced by volunteers click-
ing away on the main interface, few of whom get the recognition
and dose of fame meted out to Hanny, and pay no attention to
random discoveries made and discussed on the forums.
Its importance is rather that it turned out to be the first of
many similarly serendipitous discoveries, and in many cases the
volunteers went far beyond just pointing at the presence of an
object. The idea that citizen scientists could do more than just
spotting odd stuff first became obvious when we came across
the Green Pea galaxies—a good few months after our volunteers
had started thinking about them.
As the name suggests, these are small, round, and green
objects which appear in the background of some of the images
from the Sloan Digital Sky Survey that populated the original
Galaxy Zoo. A small group of volunteers, calling themselves the
Peas Corps,† set out to collect them but also, importantly, to find
out what they were. First they noticed that the peas were all the
* Got the plural in too!
† It took me months to realize this was a joke. The revelation finally came when I was on stage, giving a lecture, and said the name out loud.
204 Serendipity
Figure 25 A ‘pea’ as found by Galaxy Zoo volunteers. This tiny galaxy is undergoing a dramatic burst of star formation; most of the light we see is due to emission from excited oxygen.
same colour, a blueish-green; and then that the only other things
this colour were small, irregular galaxies. Remarkably, the Peas
Corp organized their own version of Galaxy Zoo to sort through
the 15,000 or so objects that shared this hue, emerging with a set
of a few thousand peas (Figure 25). Sloan also provided spectra
for these objects, and the volunteers went and grabbed this extra
data. When they inspected their haul, they noticed that their
light is dominated by emission from oxygen.
The significance of that discovery wasn’t immediately appar-
ent, but they quickly decided that having a strong oxygen line
was a good test of whether something was a ‘pea’ or not. Having
tested their sample, they sent an email announcing to the Galaxy
Zoo team—their professional counterparts—that they’d discovered
not an odd object or a new galaxy, but a new type of galaxy.
The Crowd and the Cosmos: Adventures in the Zooniverse Page 24