“Boffins” is not a term that is familiar to most Americans. So when the science and technology writer Simon Sharwood conferred this distinction on me and my collaborator Matthew Reece in the British journal the Register, I wasn’t initially sure what to think. Was the author criticizing us and our foolish ways, or was “boffins” a term like “pulchritude,” which sounds pretty bad—but is actually a term of flattery?
I was comforted to learn that “boffins” just means scientists or technical experts—albeit possibly overly focused ones. But my initial fear that the word might have meant “loony people” or some such was not entirely unfounded given that the subject Sharwood reported on was our work about dark matter and meteoroids—with a brief shout-out to extinctions. The idea was that dark matter could effectively sling comets out of the Oort cloud so that they periodically catapulted into Earth, possibly even precipitating a mass extinction.
Even for particle physicists like Matthew and me who try to keep broad and open minds, messy phenomena like meteoroid strikes tied to the complicated dynamics of the Solar System as a whole and to dark matter on top of that seemed like an uncertain route. On the other hand, dark matter (!), meteoroids (!), dinosaurs (!). The five-year-olds in us were intrigued. As were the adults who were curious to learn more about the Solar System. Not to mention the scientists we are, who wanted to know if we could actually learn something more about these various pieces and how they might all tie together. After all, although we have yet to measure the presence of a dark disk, a satellite that measures a billion stars in the Solar System could have the sensitivity to decide the issue within the next five years—and test whether our proposal is correct.
Just in case this scenario, the richness of the ideas, or the upcoming satellite measurements weren’t argument enough to pursue this line of inquiry, the day I asked Matt if he wanted to consider the project was the day that the Chelyabinsk meteoroid struck. Although most of the many meteoroids hitting the Earth or its atmosphere are so small that we don’t usually notice them, the particular meteoroid that exploded on February 15, 2013, was 15 to 20 meters wide—large enough to glow brilliantly and release the equivalent of 500 kilotons of TNT. That this meteoroid exploded three days after a question from the University of Arizona audience had started me thinking about periodic meteoroid impacts—and the day on which I had proposed to Matt that we delve deeper into the subject—struck us as remarkable, and pretty funny. We were wondering if we should study why extraterrestrial objects reach the Earth and on that very day something did. How could we not proceed?
I will now describe our research that ties together many of the ideas this book has explored and explain how dark matter could affect the planet on an approximately 30 to 35 million year time scale. If we are correct, not only did a roughly 15-kilometer-wide asteroid hit the Earth 66 million years ago, but the trigger for that impact was the gravitational influence of a dark matter disk in the midplane of the Milky Way.
THE SCENARIO
We now have a picture of the Milky Way galaxy with its bright disk of gas and stars, and inside, perhaps another, denser disk composed of interacting dark matter. The Milky Way came into existence more than 13 billion years ago when dark matter and ordinary matter collapsed into a gravitationally bound structure. Perhaps a billion years after the galaxy halo formed, ordinary matter radiated away energy to begin to form the brightly lit disk we now see. If some of the dark matter interacts and radiated dark photons sufficiently quickly, it too collapsed into a thin planar region we call a disk. This might have all taken a while to complete, but the narrow dark disk would have been established long ago.
Meanwhile, roughly four and a half billion years in the past, the Sun and the Solar System were formed. Planets subsequently emerged from the disk of matter that circled around the Sun. After the planets’ formation, Jupiter moved inward and other giant planets moved outward and, in doing so, scattered material in the disk. Some of that material moved very far away to the distant region of the Oort cloud, where small icy objects are bound to the Sun only by a very weak gravitational pull.
The Solar System then circled around the galaxy every 240 million years. But with a period of perhaps 32 million years, the Solar System, while traveling along its dominantly circular path, also bobbed up and down through the galactic plane. The gravitational pull of the disk acted on the Sun throughout this journey, serving as a restoring force every time the Solar System escaped as far as it could in the vertical direction above or below the plane. Because the galaxy provides so little friction, the Solar System repeated its vertical motion through the galactic plane on a periodic basis, with the restoring force of the plane acting on it every time that it passed through.
Moreover, when the Solar System was in or close to the galactic plane, the distorting gravitational tidal effects of the disk acted on it most strongly. During these particularly strenuous intervals, the tidal influence of a thin dense disk of dark matter might have disrupted the tranquility of some of the weakly bound objects in the Oort cloud, which would otherwise have continued mostly undisturbed along their distant orbits. Once in range of the dark disk, Oort cloud’s icy objects were unlikely to all stay in place in the face of this bumpy ride.
In the meantime, while all these inanimate objects were going about their business, life on Earth began to form about three and a half billion years ago and complex life proliferated about three billion years later—540 million years in the past. Life has had its ups and downs since then, as diversification competed with extinctions. Five major extinctions punctuated this time frame known as the Phanerozoic era. The last of these occurred 66 million years ago, when a meteoroid strike devastated the Earth.
Up to the time right before impact, the dinosaurs were oblivious to any distant Solar System snafus. Icy bodies orbited through the Oort cloud, with only the occasional small change to their orbits from the distant tug of the Milky Way disk, which acted with varying strength according to the Sun’s distance from the midplane. The orbits of a few of these icy bodies became so distorted that their paths took them to the inner Solar System, where some broke away from their initial trajectories under gravity’s distorting effect. At least one of these icy bodies might have turned into a comet on a collision course with the Earth.
From the perspective of the Oort cloud, it was a relatively minor disturbance. One or at most a few of its icy bodies got dislodged. But from the perspective of 75 percent of life on Earth, including the venerable dinosaur, the meteoroid that was set to strike was apocalyptic. Yet even if the dinosaurs had been sentient, conscious beings, they wouldn’t have noticed anything extraordinary was about to happen when the comet first appeared. Though the comet’s nucleus was bright enough to see during the day and its long tail was visible throughout the night, the comet would have betrayed no notable sign of the imminent devastation it was poised to deliver. This impression might have changed as the comet descended, when fire and debris lit up the sky. But whatever the doomed creatures might have seen or envisioned, once gravitational influences had altered the comet’s course, the fate of those animals was irrevocably sealed.
The comet would shortly slam into the Yucatán, pulverize its target, and end a journey that would culminate in massive global destruction. When the meteoroid whose impact created the Chicxulub crater struck, the impact vaporized the comet as well as the ground near where it hit—sending up plumes of dust that scattered all around the globe. Fires burned the Earth’s surface, tsunamis flooded coastlines both near the impact and on the opposite side of the planet, and poisonous materials rained down even more dangers. The food supply was decimated, so any land-dwelling creatures that did manage to survive the immediate aftermath probably starved to death in the weeks and months afterward. Most life didn’t stand a chance when confronted with such sudden and drastic changes to the Earth’s climate and its various habitats. Only ground-burrowing mammals and airborne birds remained when conditions eventually improved to perpetuate advanced life into the uncertain future of a very different age.
It’s a dramatic picture, and the basic facts of the meteoroid impact are by now well established. The many observations of geologists and paleontologists have confirmed that a big object hit 66 million years ago and that at least 75 percent of life on the Earth then died. Shortly I will describe how a dark disk might have been the trigger responsible for dislodging the comet that was responsible for all this devastation. But first I’ll explain the genesis of the idea.
INCEPTION
Many benefits accrue from sharing physics with the general public through books and lectures. But because time invested in these activities can detract from ongoing research, I frequently have to prioritize and choose which talk requests I should accept. However, on some happy occasions, my research profits from what I had wrongly feared would be a distraction by taking me to visit people I would not have ordinarily encountered or by introducing me to an idea I might otherwise not have considered.
In February 2013, I reaped such a reward from the astrophysicist Paul Davies’s invitation to give the Annual Lecture he hosts at Arizona State University’s BEYOND Center. Despite my hesitations about too much travel, ASU has a very strong cosmology research group so I was happy to agree not only to give the public lecture, but also to present a seminar the following day to the experts in the department there. The seminar would be more focused on my recent research, which was the double-disk dark matter idea I have already described.
The physicists in attendance asked several excellent questions about the model—its detectability and its implications for the cosmic microwave background radiation, for example. But I was thrown for a loop when Paul asked me if the dark matter disk was responsible for the dinosaurs’ demise. I confess that at the time, I hadn’t thought much—really ever—about dinosaurs in my scientific research, which had been focused on elementary particles and elements of the cosmos. But Paul informed me of potential evidence for periodic meteoroid strikes and the absence of a good explanation for them. He asked whether a dark matter disk might fit the bill—and in the process reminded me about the meteoroid that had extinguished the terrestrial dinosaur.
Paul’s question was too good to ignore. The answer wasn’t straightforward and I had a lot to learn before more definitively responding, but dark matter and the dinosaurs certainly seemed like a connection that could teach me—and potentially scientists more broadly—quite a lot. I asked Matthew Reece whether he would be interested in studying the possibility that meteoroid strikes were triggered by our proposed dark disk, which sounded more connected to physics than a question about the dinosaurs.
Matt was the obvious choice of collaborator. He played a crucial role in the initial DDDM research, has a cool technical head, and is scientifically open-minded when it comes to new ideas—more so than you would anticipate from his decidedly conservative demeanor. He doesn’t make the common fallacy of assuming anyone—even overly confident more senior colleagues—has correctly guessed everything.
But most important, Matt is an excellent physicist with very high scientific standards. When he does something, you can be sure it is on solid footing. Still, I wasn’t sure how he would react to such an apparently crackpot suggestion. I was very pleased when Matt found the idea intriguing and recognized its potential scientific merit. Paul Davies was interested too, but he already had many ongoing research projects and graciously chose to be in contact but not to participate.
So after listening with amazement to news about Chelyabinsk on the very day we started discussing the idea, Matt and I pressed on to see what we could learn. Our goal was to turn this crazy notion of a dark disk causing meteoroid strikes into testable science. As model builders and particle physicists, Matt and I try to be receptive to new ideas and interpretations. But we also are fully aware of the importance of remaining unbiased and careful. These qualities were essential in the research I’ll now describe.
THE DARK DISK AND THE SOLAR SYSTEM
As explained in Chapter 14, in order to be realistic in our goals, Matt and I decided first to pare down our investigation. Despite our curiosity about dinosaurs, we initially left aside the additional challenges endemic to extinctions and focused solely on meteoroids and Solar System dynamics and a possible periodicity in the physical crater record. With the extinction issue on the back burner, we could directly investigate the dark disk’s potential influence on comets and whether it might be responsible for periodic meteoroid strikes. We could decide afterward how well our meteoroid predictions explained any particular known impact, including the one responsible for the K-Pg extinction.
We then made sure that none of the previous proposals for periodic triggers that might dislodge Oort cloud objects could successfully explain a periodic signal. If a more conventional mechanism sufficed, then no one, ourselves included, would bother evaluating the consequences for the crater record of a more exotic scenario—no matter how cool and seductive it might be.
However, as explained in Chapter 15, conventional triggers don’t work. With only the standard Milky Way disk, the tidal effects from the galaxy are too smooth and perturbations from stars are too infrequent. Neither the usual tidal effects, Nemesis, Planet X, nor the Milky Way’s spiral arms suffice to trigger frequent or notable enough comet showers. These earlier proposals didn’t yield either the correct time between galactic plane crossings or sufficiently sudden strikes to match the crater record. For example, with only normal matter in the disk to influence the motion, the Sun’s vertical oscillation period would be more like 50 or 60 million years—too long to match the available data.
This left two possible conclusions: either the periodicity wasn’t real, as might well turn out to be the case, or the more interesting logical alternative is correct and the trigger is unconventional. With the previous suggestions ruled out, it made sense for us to ask whether our proposed dark disk could succeed where ordinary matter alone failed and give rise to the required periodicity and change in rate. Indeed, the dark disk has just the necessary properties to address the inadequacies of the normal matter disk. With a thin disk of dense dark matter, the disk tidal force can successfully account for both the period and time dependence of perturbations to the Oort cloud.
Recall that throughout their existence, Oort cloud objects are subject to the disk tidal force from ordinary matter and occasionally to the more intermittent but nonetheless important influence of passing stars. These effects serve to move around the distant, relatively weakly gravitationally bound bodies of the Oort cloud and nudge them closer to the Sun. The tidal effect of the Milky Way plane can then give one final tug that might bring these icy bodies into precarious, very eccentric orbits that jut inward to about 10 times the Earth-Sun distance, where the gravitational pull of the large planets will very likely remove them from the Oort cloud. This pull will either fling such comets out of the Solar System or pull them in so that they enter tightly bound orbits in the inner Solar System. These disruptions account for the generation of long-period comets, with several new ones entering the Solar System every year. Occasionally, however, perturbed objects get deflected out of their orbits altogether, and at those times deviant comets might strike.
But this type of perturbation does not in itself suffice to explain periodic impacts. For periodic strikes to occur, a rapid change in the rate of disruptions to the Oort cloud must occur at regular intervals. Furthermore, to match the available evidence, the period must be in the range of 30 to 35 million years. If even one of these criteria fails, a proposed explanation for periodic meteoroid strikes won’t do. And neither criterion is satisfied for any conventional suggestion.
However, the addition of the denser and narrower dark disk addresses both these issues exceedingly well. Once you accept the possible reality of periodic meteoroid strikes, a dark disk is indeed a very promising idea. The dark disk exerts an influence that is both more intense and more rapidly varying with time than the conventional disk of the galactic plane—the two essential requirements to create spikes in comet intensity.
With the dark disk included in the Milky Way plane, the Sun’s vertical oscillation period would be shorter than the period induced by the conventional Milky Way disk alone because the gravitational pull with the addition of the dark disk’s matter is stronger. On top of that, according to current matter density determinations, the Solar System oscillates only about seventy parsecs above and below the galactic plane—a much more limited range than the thickness of the full ordinary matter disk. The narrow dark disk, which therefore would encompass the Solar System throughout a reasonably large fraction of its trajectory, can have a disproportionately large influence on the Solar System’s motion as it bobs up and down through the plane.
The other merit to a thin dark disk is that, even so, the Solar System can pass through it quickly enough to induce a spike in the comet rate that lasts a million or so years. Because of its big time-dependent influence, the dark disk triggers further perturbations each time the Solar System crosses the galactic plane, creating comet showers on a regular basis—during each plane crossing—that otherwise would be triggered only very infrequently by closely approaching stars. The enhanced tidal effect takes place when the Solar System crosses the narrow region occupied by the dark disk. Only during this passage and perhaps for one or two million years following will comet strikes be enhanced.
When the Solar System passes through the disk on this time scale and is subject to an enhanced tidal force—a spike in the force if it happens sufficiently quickly—icy bodies in the Oort cloud might get dislodged and a few might even come hurtling into our planet at about 50 km/sec. Once set off track in this manner, the trip is quick—perhaps a few thousand years. But the perturbation that set it off in the first place takes place more slowly—generally taking one to a few orbital periods to get going. This means that in a time period between about one hundred thousand and a million years, the fate of comets that have come too close to the Sun will be determined, and some of those could account for comet showers that hit the Earth’s atmosphere or the Earth itself.
Matt and I worked out the predicted trajectory and the scenario was a success—at least within the confines imposed by the limited and somewhat shaky data. But there was one final check that we hadn’t completed, as was pointed out to us by the referee who reviewed our paper for the prestigious physics journal Physical Review Letters. In addition to determining the motion of the Solar System in the presence of the dark disk, we calculated the rise and fall in density of the Solar System’s environment as it passed through. We needed to know the density since we assumed that whatever disturbed the Oort cloud would be proportional to this degree of matter concentration. After all, more mass means more tidal influence, which should mean more disturbances. We therefore assumed that density would serve as a useful proxy for the rate of meteoroid strikes, as indeed turned out to be the case.
But we hadn’t yet explicitly confirmed that the tidal distortion of the Oort cloud by which the dark disk acts on the icy bodies in the cloud was sufficient to rain down comets at the correct rate. Fortunately for us, Scott Tremaine and Julia Heisler had already done a lot of the heavy lifting about a decade earlier, so we could simply apply their results. And indeed our assumption had been right. The enhanced density creates just the sort of pull needed to dislodge comets on the appropriate time scale.
I was actually impressed by the referee’s useful suggestion. These days, referee reports, in which colleagues who should be experts review papers before approving them for publication, are often either rubber stamps or vehicles for aggrieved authors looking for citations. This referee’s suggestion actually taught us some physics. The tone had been dismissive, but we learned something from following it up. We had to deal with some misguided criticisms too—but since we had been careful to check papers and experts beforehand, we could readily pinpoint the flaws in those critiques.
In the end, Matt and I calculated the preferred density and thickness values for a dark disk that matched the crater record, and found that they were in line with our previously existing DDDM model, which by that time we knew was consistent with existing galactic measurements. The new, even better wrinkle that Matt and I found in our research was that not only was the dark disk allowed, but it was actually favored if you take the disk seriously as the instigator of comet strikes.
The dark disk’s surface density should be about one-sixth of that of matter in the ordinary disk. That’s enough to be interesting but not so much as to overwhelm any currently understood phenomena. It’s a sizable chunk of the dark matter—not one-millionth, for example, but instead a few percent. Should this dark component exist, it would be sizable enough to have measurable influences and therefore would be worthy of attention. Furthermore the disk thickness might be less than one-tenth the thickness of the ordinary matter disk—less than a few hundred light-years in width compared to ordinary matter with a roughly 2,000 light-year thickness. Again, it is this narrowness of the dark disk that explains why it can conceivably trigger dramatic effects on a periodic basis.
We found the dark disk with the right density was favored by a factor of three. A key contributor to this new conclusion with better statistical support was the look-elsewhere effect that I mentioned in Chapter 15. With a definite model for what might trigger periodic strikes, we could not only better predict the period, but we could also do so more reliably. In fact, our intention in the paper went beyond demonstrating that a dark disk could succeed in explaining periodic comet showers in a way that ordinary galactic components can’t. We wanted to make a second point, which had to do with statistics, and how to evaluate the significance of these or any other results.
As discussed in Chapter 14, most existing searches for periodicity tried to match a periodic function for the up-and-down motion of the Solar System—a sine wave, for example—to the data. This can be interesting, but it doesn’t capture the full story. We don’t have to guess the motion of the Solar System. If we knew everything about the galaxy and the Sun’s initial position, velocity, and acceleration, we could use Newton’s laws of gravity to compute the Sun’s motion and predict the period we should expect. After all, the motion of the Solar System isn’t random, but has to be consistent with its underlying dynamics. Even with imperfect knowledge of the density distribution and the parameters of the Sun, the range of possible trajectories—and hence possible periods—is restricted.
Matt and I incorporated what we know about the densities of known matter in the galactic disk—allowing for the full range of possible values supported by current measurements—and added the matter contribution of a dark disk. Our goal was to check whether there was evidence for periodic cratering rates that matched the Solar System’s motion once we took into account everything we know about the measured disk components—stars, gas, etc.—plus the dark disk component we had introduced.
The measured contributions of ordinary matter restrict the range of possible trajectories that the Solar System can take since the gravity of the matter in the disks—both ordinary and dark—acts on the Sun and influences its motion, thereby reducing the influence of the look-elsewhere effect. Matt and I used the measured densities to predict the periodic motion of the Solar System and compared the times of galactic plane crossings to the reported crater creation times to check how well they match. Although predictions without an underlying model don’t discriminate sufficiently, we found that with existing measurements accounted for, the statistics do favor periodic meteoroid onslaughts with a period of about 35 million years. Recent data improvements indicate the period is likely to be a bit shorter even—perhaps 32 million years.
The dark disk was critical to making the scenario work and generating the favored impact rate. Turning the story around, with the better match of crater data to Solar system motion, a dark disk is actually preferred. Future data should be analyzed with this sort of model in mind in order to yield the best statistical significance. The results will then either strengthen our result—or rule it out.
AND THE DINOSAURS . . .
After Matt and I had sorted everything out and our research had been accepted by Physical Review Letters, we posted our results to the online journal repository that provides immediate Internet access to the research papers known as preprints that are yet to be published. Matt did the actual submission. We had conservatively titled our paper “Dark Matter as a Trigger for Periodic Comet Impacts.” But to my surprise, Matt had edited the comments section—generally used to describe format or revisions to the submission—to read “4 figures, no dinosaurs.” I thought this was pretty funny since we had studiously avoided any explicit mention of dinosaurs in our paper, which focused on the crater record and its more direct contact to physics. But of course we had had this connection in mind all along and even jokingly referred to our work as “the dinosaur paper.” I suppose that had I paid more attention, I wouldn’t have been surprised the next day by the degree of online interest in our work, which was reported in many blogs and journal websites—including the “boffins” piece—almost always accompanied by some rather entertaining graphics.
But this does bring me back to the dinosaurs. Having established at least a first attempt to put data together with models to predict comet impacts, and knowing that this is not the final word but will be improved with future measurements, we did then look to see how well our model agreed with the timing of the Chicxulub event. Our calculations showed that, depending on the improving measurements of ordinary matter in the Milky Way disk, meteoroid strikes should occur about every 30 to 35 million years. Since we passed through the galactic plane within the last two million years, a comet dislodged from the Oort cloud one complete oscillation (two disk crossings) in the past might indeed have come hurtling down to Earth 66 million years ago, at the time of the K-Pg extinction, to wreak its enormous destruction. As an aside, if we passed within the disk less than a million years ago, we could even be on the tail end of an enhanced comet flux and have the potential to see heightened impacts today. But much more likely is that, aside from a truly random and extremely unlikely event, the Earth passed through a little further back and we won’t witness another Chicxulub for about another thirty million years.
Because of the uncertainty in the Sun’s position and the lack of knowledge of the precise period, we can only approximately predict the disk crossing times. If the Earth crossed the galactic midplane about two million years ago, an oscillation period of about 32 million years would be optimal for generating an event that occurred 66 million years in the past. Our original crude analysis produced a period of 35 million years, which is a little too big to match the Chicxulub timing—though the uncertainties in the model and in the length of time of enhanced comet strikes still permitted reasonable agreement. Our updated model for the Milky Way disk, which takes into account the more recent measurements of the galaxy’s components, has since brought the period down—leading to a better match to the K-Pg extinction time. But even with the crude model we used for our initial prediction, there is a reasonable probability that the dark disk prediction corresponded to the Chicxulub event.
The primary reason that our results were not yet sufficiently precise is that the matter measurements in the Milky Way changed since we did our initial analysis. We also still haven’t fully modeled the time-dependent galactic environment, such as galactic arms, which are also only poorly known. The density variation from these effects would not suffice to trigger meteoroid strikes. But they might well be sufficient to change by a few million years the precise prediction of the model for when those strikes would occur.
Other factors too contribute to the uncertainty in the exact times predicted for enhanced comet showers. It takes about a million years for the Solar System to cross the galactic plane—longer if the disk is thicker. Furthermore, a time period of up to a few million years might separate the initial triggering event from the actual meteoroid hit on Earth. Thirdly, the crater record and the dating precision are poor. Finding more craters or dating them more precisely would help—though new crater discoveries emerge only infrequently. Not just craters, but dust that gets trapped in rocks, might also help generate a more precise record of when comets have struck.
Evidence for 30 to 35 million year periodicity in the vertical motion of the Sun away and toward the galactic plane might come from unexpected directions too. After Matt and I had written our paper, a particle physics colleague who was aware of my fascination with astronomy, geology, and climate—but who, at the time, didn’t know about the “dinosaur paper”—fortuitously told me about the work of Nir Shaviv, from the Hebrew University of Jerusalem, and his collaborators, who studied climate over the entire 540-million-year Phanerozoic era. Remarkably, they had found a variation of climate with a period of 32 million years—strikingly similar to the period that we had identified. If Shaviv’s result holds up and this periodicity in climate is indeed determined by the Sun’s motion through the galactic plane, the 32-million-year period too would be evidence of a dark disk since ordinary matter alone wouldn’t suffice to yield this relative short interval between disk crossings.
Of course, we don’t need to delve into the past to see the influence of dark matter. If dark matter does indeed have an interacting component that changes the structure of the matter distribution in the Universe, we will learn about that soon—perhaps even before any of the other dark matter searches reach fruition. Only a limited range of dark disk densities can account for the crater data. Future measurements will almost certainly narrow the range of possible predictions, validating or ruling out our proposal.
The analysis my student Eric and I have already done shows that the dark disk with the necessary density and thickness is permitted by observations to date. And the better data to come from GAIA will pin down a disk’s presence, density, and thickness still further. Once this satellite completes its 3-D map of stars in the nearest region of the Milky Way, the dark disk—or absence thereof—will be much better determined. By this indirect route, we might learn much more—not only about the galaxy and dark matter, but about the Solar System’s past as well. If GAIA data establish the existence of a disk with the right thickness and density, it will be powerful evidence of the crater proposal’s viability.
A better punch line would of course be that we precisely calculated the exact date of the dinosaurs’ demise with sufficiently small uncertainty to be confident in the result. But this is a complicated subject involving many challenging measurements. Even so, the amount of progress scientists have made in the last 50 years is nothing short of astounding. Dark matter has been more elusive in many respects than the more readily apparent Earth, the Solar System, and the many other visible elements of the Universe. But through the research I’ve described, physicists are finding new ways to track it down. Whatever the outcome, we can be sure that the galaxy and the Universe, and the inner workings of matter itself, have some fascinating surprises in store.