Is Cosmology in Shambles?

[O]ur results may indicate that any direct DM [dark matter] detection experiment is doomed to fail….
–Moni Bidin, Carraro, Méndez, and Smith

[T]he missing satellites problem becomes a catastrophic failure of the standard cosmological model.
–Pawlowski, Pflamm-Altenburg, and Kroupa

Dark matter is the source of incredible frustration for astronomers and cosmologists. On the one hand, the data supporting its existence looks very strong from a wide variety of observations, but…on the other hand, after decades of work, we still lack direct evidence for it. We haven’t identified what kind of particle it is, and enough results are ambiguous that some people even argue it doesn’t exist. Nonexistence poses its own problems, though, or else I think most of us would be happy to see it go.

Now to make things fun, two different groups of researchers are publishing studies saying they find no evidence for dark matter in and around the Milky Way, respectively. The first study (by  C. Moni Bidin, G. Carraro, R. A. Méndez, and R. Smith) looked for the gravitational effects of dark matter in the galactic neighborhood near the Solar System, and failed to find any. The second study (by M. S. Pawlowski, J. Pflamm-Altenburg, and P. Kroupa) examined the distribution of satellite galaxies orbiting the Milky Way, and found they lie nearly in a plane, where dark matter models predict the satellites should be distributed in a more-or-less spherical pattern. I’ll fill in more about each of these studies shortly, but note in both cases, the authors make very strong statements about the very existence of dark matter, including the quotations that begin this post. In fact, the National Geographic coverage of the second article states things even more strongly than the paper. Pavel Kroupa, the third author of the study, is quoted as saying, “It means that we have to completely and utterly rethink cosmology…. Cosmology is basically in a shambles now.”

You can probably guess already that I don’t see things quite that way. I’m no expert on galactic dynamics (the purview of the first paper) or on structure formation (which the second focuses on), so if there are any problems with the data or methodologies, I can’t spot them. (If any readers are experts and can provide more detailed analysis, please let me know.) However, there are a couple of general points I want to make before summarizing the cases for and against dark matter’s existence.

One concern I have with the strong statements in the papers and press coverage has to do with the nature of science itself. While science is largely the art of falsification—ruling out ideas using experiment and observation, an idea we owe to philosopher Karl Popper—it’s also rare to have a single definitive experiment that Changes Everything. The most radical results are generally recognized as such only after a significant amount of time has passed, and experiments are typically more ambiguous. Dark matter has always been a slippery concept, with plenty of evidence for it, yet enough ambiguity to leave many open questions; it seems to me that there are so many “definitive” observations that they just confuse matters. I’ll summarize a few of those in the following section.

This isn’t to say we should dismiss the observations: based on my understanding of galactic dynamics and structure formation, these studies should be taken very seriously, and may indeed show dark matter doesn’t exist…but it’s arrogant to say they’ve shown that just yet.

The Evidence For and Against Dark Matter

For those who need a quick refresher, dark matter is the name we give to our ignorance: it’s a convenient and relatively simple explanation for why a lot of the mass in the Universe is invisible.(In fact, I think “invisible matter” is a better term than “dark matter”—”dark” to me implies the color black, which absorbs all wavelengths of light, while all light seems to pass through dark matter without interacting. It’s more transparent than anything in our ordinary experience.

The reason we infer its existence is through gravity. A wide variety of effects are most simply explained if there is a significant amount of mass that doesn’t register in our telescopes. Let’s look at a few of these in turn; other examples exist as well, so this list isn’t exhaustive.

Case Study 1: The Rotation of Spiral Galaxies

The observed rotation curve of spiral galaxy NGC 3198 (top line), along with the expected curve (labeled “disk”) if the visible matter – stars, gas, dust – were the only mass present in the galaxy. The third curve represents the inferred shape of the dark matter halo that surrounds us and binds us and holds the galaxy together.

In the late 1960s and early ’70s, astronomer Vera Rubin measured the velocities of stars near the edges of spiral galaxies. She found that all these stars rotated nearly at the same speed, right out to the visible limit of galaxies. This result is not expected from our understanding of gravity; instead, based on the amount of stars, gas, and dust in the galaxies, the outermost stars should revolve much more slowly around the center of the galaxy than the stars closer in. Since that time, Rubin’s results have held up for a huge number of spiral galaxies: the effect is real, and contrary to our understanding of how gravity works.

The standard explanation: spiral galaxies contain a large amount of mass that isn’t made of stars, gas, dust, or anything ordinary. This missing mass is dark matter. Matter made of atoms—the stuff of everyday experience for us—absorbs and reemits light to varying degrees, so it doesn’t hide completely, even if it’s cold and spread out through space. Over time, astronomers were able to rule out neutrinos as a possible candidate for dark matter as well: neutrinos have a very tiny mass (if they were massless, they wouldn’t affect the rotation of galaxies strongly), but it seems to be too small. Rather than stick around galaxies, they seem to flow away at high speeds. That pretty much leaves dark matter as a new, unknown type of particle. (Again, I’m summarizing here—there’s enough history and science of dark matter to fill books.)

A highly elegant cartoon of the structure of a spiral galaxy. The dark matter halo is the black circle (generally considered to be spherical), which is much larger than what we usually think of as the “galaxy” – the luminous portion containing the highest concentration of stars. (The disc contains the spiral arms, which despite their prominence in the appearance of a spiral galaxy aren’t important for the current discussion.)

However, it’s not enough to say there’s dark matter—we also have to know what it is and where it is in spiral galaxies. The simplest model involves a spherical halo of dark matter surrounding and permeating the visible parts of a galaxy, but extending much farther beyond. This picture works very well for the rotation of stars in the outermost regions of spiral galaxies. Based on the inferred density of dark matter in the region of our Solar System, a variety of experiments are attempting to detect it (including at least one I’ll be describing in my book). An interesting recent paper discusses the likelihood that if dark matter does exist, particles will occasionally strike atoms in our bodies.

Things get stickier close to the centers of galaxies: the dark matter density in our simplest physical models doesn’t match observations. Briefly, models show too much dark matter near the centers of galaxies, a sharp uptick known as a cusp. That’s not a secret: astronomers are well aware of this problem, and a number of fixes have been proposed over the years. Part of the problem is that the closer we look toward the center of any galaxy, the more messy things are: stars move in chaotic orbits and there’s a lot more stuff in general to keep track of.

Enter the paper of Moni Bidin, Carraro, Méndez, and Smith. These researchers studied the motion of roughly 400 red giant stars within the Milky Way’s disc (the region of the galaxy that contains the spiral arms, as well as our Solar System). Using data from three different surveys, they reconstructed a fully three-dimensional view of the position and velocity of each star. They then used a detailed theoretical model to obtain the amount of dark matter that must be present to generate these star motions, assuming a spherical halo and a flat rotation curve, as above.

Even with generous error estimates, they found that none of the usual dark matter distribution models work with their data. Again assuming their dynamics are OK (and I stress that I’m no expert), this means either the density of dark matter is much lower than expected from halo models, or…there is no dark matter at all. While the authors briefly consider the possibility that other, possibly non-spherical dark matter halos may be responsible for the odd results, it boils down to the problem of describing all non-rabbit animals—once you break away from the simplest case, there are too many alternatives, and no particular reason to select one over another. To their credit, the researchers don’t claim to have easy answers, and acknowledge they’ve raised more questions than they’ve answered. For example, how is it the Sun has such a large velocity, unless there’s a lot more mass than is visible?

Case Study 2: Galaxy Clusters

The Coma Cluster, which Fritz Zwicky used to argue that much of the mass of the universe is invisible.

While Vera Rubin’s galactic rotation result was certainly important, the very first hint about dark matter dates back several decades. In 1933, astronomer (and  notoriously prickly character) Fritz Zwicky measured the motion of galaxies in the Coma Cluster—a large, relatively nearby collection of over 1000 galaxies bound together by their mutual gravitation. Zwicky found that seven galaxies within the cluster were moving far too quickly relative to each other: some hidden mass was making them move faster. Though his calculations were pretty rough and seven galaxies don’t make a large sample, subsequent work has borne out Zwicky’s analysis: about 90% of the mass of the Coma Cluster is invisible.

We should note that the dark matter in a cluster is in addition to the dark matter in individual galaxies. I spoke about dark matter in galaxy clusters in a podcast I recorded for the 365 Days of Astronomy site last year (transcript here), and in my opinion, the evidence is stronger for dark matter in clusters than it is in individual galaxies. Part of the reason for that is because galaxy clusters are simpler in many ways than galaxies. Since hot gas and dark matter within a cluster are both more massive than the galaxies themselves, dorky cosmologists like me often leave the galaxies out entirely when talking about cluster properties!

The Bullet Cluster is a galaxy cluster in the process of merging. The pink regions are the hot gas as imaged using X-rays; the blue is the dark matter as found using gravitational lensing. (For more on lensing, see the transcript of another podcast I recorded last year.)

Of course, no discussion about dark matter in galaxies can ignore the Bullet Cluster, shown at left. The pink in the image is X-ray imaging of the hot gas in the cluster, which shows the violent shockwave from which the name “bullet” originates. The blue is the location of the mass of the cluster, as seen by gravitational lensing. According to the most widely accepted explanation, the Bullet Cluster is actually two galaxy clusters in the process of colliding. During the collision, the gas shocked and heated up, but most of the mass—in the form of dark matter—was unaffected by the collision.

Many astronomers and cosmologists, including myself, consider this to be one of the best pieces of direct evidence for the existence of dark matter. Since its 2004 discovery, astronomers have found several similar systems, including most recently the “Musket Ball Cluster“.

Case Study 3: Structure Formation

From a wide variety of observations, we see that big galaxies form by merging: smaller galaxies collide and lump together. Similarly, galaxies form individually, then come together right now under gravity to make galaxy groups and clusters. In the standard models, it’s dark matter that drives creation of objects on all of these levels; the big-picture view is called structure formation.

6 Degree Field galaxy survey (6dFGS) map of 125,071 galaxies in the nearby universe, showing how they are clustered together.

It’s simple to see how structure formation works. Small halos provide an initial gravitational attraction for gas, which collects and forms stars and small galaxies. The smaller galaxies are attracted to each other, and merge into larger galaxies. The larger galaxies in turn grow by merging with similarly sized objects, or devour smaller galaxies, then throng together as clusters. We observe the large-scale structure (LSS) of the Universe through galaxy surveys, such as the 6 Degree Field survey shown at right.

The simplest model of structure formation has some difficulties, though: it predicts a large number of low-mass galaxies, far more than we actually see. Some observations may have spotted “dark galaxies” (which I covered for Ars Technica), which contain few or no stars, but the results are ambiguous—and if they hold up, we haven’t found enough dark galaxies yet to make up the deficit. This isn’t a problem for the existence of dark matter per se, but it’s a headache for understanding structure formation.

Simulation of the formation of a galaxy and its satellites, using the simplest dark matter model.

Additionally, structure formation models predict that big galaxies will be surrounded by smaller satellite galaxies. The satellites should be distributed more or less equally around their larger companion, as shown in the simulation at the left. This should be true even apart from the “missing satellite” problem in the previous paragraph.

Now we reach the study by Pawlowski, Pflamm-Altenburg, and P. Kroupa: they studied the distribution of satellite galaxies and other objects (such as globular clusters) orbiting the Milky Way. They found a significant deviation from the expected results: instead of equal distribution, they found two largish clumps of satellites, with streams of matter (gas and stars) appearing to connect them. Based on their analysis, the researchers conclude the dark matter structure formation scenario cannot produce a distribution of satellites like that.

Again, I’m not a structure formation expert, so I will not quibble with their results. I do think these results together with the missing satellites pose a significant problem for the simplest dark matter structure formation models, though—but since the lack of small galaxies is a known issue, I don’t know if the distribution of the Milky Way’s satellites makes things that much worse. I invite any experts in the crowd to chime in!

Case Study 4: Cosmic Microwave Background

We started relatively small (galaxies), and we’ll end as big as we can go: the whole Universe! Another way to assess the existence and amount of dark matter out there is to look at the cosmic microwave background (CMB), which is light left over from when the Universe was about 380,000 years old. I wrote an entire post on the CMB, which itself was an expanded version of a talk I gave at ThirstDC, so I’ll briefly summarize the relevant part of that post and let you read the whole thing if you’re interested in the details.

Small fluctuations in the temperature of the CMB are what gave rise to the structures—galaxies and clusters—we see today. By examining how large the fluctuations are in temperature and in physical size, we can figure out how much ordinary matter, dark matter, and so forth comprise the Universe today. This information is known as the CMB power spectrum, and it includes a pretty strong piece of evidence in favor of dark matter’s existence.

The CMB power spectrum, with the three “peaks” highlighted. The first peak indicates the total amount of “stuff” in the Universe – everything together, and the second peak is the contribution from ordinary matter. The third peak is all matter together: that’s what reveals how much dark matter there is in the Universe. For more details, please see my previous post.

The solid line in the power spectrum is the theoretical model that fits the data best. The way theoretical cosmologists work is by changing a bunch of parameters simultaneously and comparing to the real CMB power spectrum. In other words, it’s possible there may be another explanation that fits the CMB data, but doesn’t include dark matter—but I have yet to see a convincing one.

Dark Matter (Probably) Isn’t Dead (Yet)

Hopefully you can see from the four case studies outlined above how complex the issue is. If you just look at galaxy clusters and the CMB, the case for dark matter’s existence looks very strong. If you look just at structure formation and spiral galaxies, the evidence still looks OK (at least to me), but shaky enough to give a thoughtful scientist indigestion. After all, if we conclude that dark matter is present in galaxy clusters but absent in galaxies, it makes little sense! (Similarly, alternative theories of gravity like modified Newtonian dynamics (MOND) seem still to require at least some dark matter to explain objects like the Bullet Cluster, as well as to fit the CMB power spectrum.)

In other words, I am not convinced that dark matter is dead—the models including dark matter have troubles, and those troubles may yet be fatal. As always, science proceeds on evidence, but as long as the evidence pulls us in two directions, it’s very difficult to decide between the existence or nonexistence of dark matter.


  • “Kinematical and chemical vertical structure of the Galactic thick disk II. A lack of dark matter in the solar neighborhood”, C. Moni Bidin, G. Carraro, R. A. Méndez, R. Smith. Accepted to Astrophysical Journal, available at
  • “The VPOS: a vast polar structure of satellite galaxies, globular clusters and streams around the Milky Way”, M. S. Pawlowski, J. Pflamm-Altenburg, P. Kroupa. Accepted to Monthly Notices of the Royal Astronomical Society, available at
  • I also referred extensively to Galactic Dynamics by James Binney and Scott Tremaine (Princeton University Press, 1987) and Cosmology (2nd edition) by Peter Coles and Frencesco Lucchin (Wiley, 2002). See also the additional links within the text of this post.

13 responses to “Is Cosmology in Shambles?”

  1. I think this is a well written and very balanced article, nicely putting the new papers into the larger perspective. However, please allow me two comments:

    1) The Bullet Cluster is most often cited as evidence for DM only. But if you look at it in detail, it also becomes a problem. The relative velocity which the two clusters must have had is not reproduced in cosmological simulations. And then there is the galaxy cluster Abell 520, which seems to be an ‘inverted’ Bullet Cluster as the dark matter there is where the gas is, and not around the galaxies.

    2) The recent paper by Karachentsev might be of interest in the context of this blog post. It is titled: “Missing Dark Matter in the Local Universe”. (arXiv:1204.3377)

    Best regards, Marcel

  2. This is an excellent summary of dark matter.

  3. […] We’re so mixed up about dark matter that Matthew Francis wonders whether cosmology is in shambles. […]

  4. […] We’re so mixed up about dark matter that Matthew Francis wonders whether cosmology is in shambles. […]

  5. […] We’re so mixed up about dark matter that Matthew Francis wonders whether cosmology is in shambles. […]

  6. “After all, if we conclude that dark matter is present in galaxy clusters but absent in galaxies, it makes little sense!” … I dont understand this sentence: what exactly makes “little sense”? After all, Nature does what it does, and if one wants to speak about elegance here, the lagrangian of the standard model of particle physics is perhaps the best counterexample. I’m not saying it is actually the case, it would be premature and indeed arrogant to say so, but if dark matter is indeed present in galaxy clusters and not in galaxies, well, it just means it does not cluster on small scales, which is precisely what dark matter does when it’s not cold but hot (something around the eV-scale or tenish eV-scale). Of course that would not be sufficient to explain galaxy formation and galaxy rotation curves, but that part could be due to modified gravity indeed. Of course, I fully agree that, a priori, an explanation based on dark matter alone (or on modified gravity alone, but that seems to have failed already) to explain the missing mass in clusters and galaxies is of course more elegant than a solution mixing both solutions. But if data are pointing to that direction indeed, it would be dogmatic (or even more) to refuse the possibility that Nature could have chosen to fool us all by doing just that (and if this modification of gravity would explain dark energy too, that would be the cherry on the cake… and the boomerang of elegance). Again, I’m not saying that is actually the case. Just that it makes no sense to say that it makes little sense.

  7. […] doesn’t damage us too much: we know the SM is not all there is, simply from the existence of dark matter, the fact that the SM doesn’t cover gravity, and so […]

  8. […] dark matter candidate shows up soon—if for no other reason than to help me degrump over the dark matter deniers. While we wait, I have a book to write. Share […]

  9. […] The second is CDMS (Cryogenic Dark Matter Search), one of a handful of detectors trying to find dark matter particles as they pass through Earth. I drove to the Soudan site today to take photos, since […]

  10. Hello Matthew,

    Really nice article on the new Dark Matter research!

    I also wrote a blog article about the same Dark Matter research results but yours adds some good information, explanations (the cartoon is wonderful) and perspective. It has such a nice tone (from your years of handling questions at the Planetarium?) readers should enjoy reading it even if they don’t understand cosmology :-)

    And congratulations on getting nominated for the 3-Quarks Daily contest.
    Even though my article (Cosmic Microwave Angular Resolution Surprise) is competing with yours – please let me wish you best of luck.

    You’ve done well !


  11. […] after that while I finish my book proposal. However, if I can make a blatant appeal, my post “Is Cosmology in Shambles?” (spoiler: no) is up for the 3 Quarks Daily Science Prize for best online science writing. […]

  12. […] Siegel chided me for skimping in the discussion of LSS in my earlier post on the subject, “Is Cosmology in Shambles?“. In fact, I evidently equivocated too much for many people’s tastes, leading some to […]

  13. […] riddles: dark matter and dark energy. I’ve written a lot about both: see my posts “Is Cosmology in Shambles?” (answer: no) and “Observation of a Matter Bridge” for dark matter, and […]

%d bloggers like this: