The spacecraft Chang’E-4 landed on the far side of the Moon in January; something of a triumph for the Peoples’ Republic of China as it was a first. It was more than a power gesture at a time of strained relations between the PRC and the US, for it carried a rover (Yutu2) that deploys a panoramic camera, ground penetrating radar, means of assessing interaction of the solar wind with the lunar surface, and a Visible and Near-infrared Imaging Spectrometer (VNIS). The lander module itself bristles with instrumentation, but Yutu2 (meaning Jade Rabbit) has relayed the first scientific breakthrough.
The dominance of the Lunar Highlands by feldspar-rich anorthosites, which form when feldspars that crystallise from magmas float because of their lower density, gave rise to the idea that the Moon initially formed as a totally molten mass. That this probably resulted because the early Earth collided with a Mars-sized protoplanet stems from the almost identical chemical composition of the lunar and terrestrial mantles, as worked out from the composition of younger basalts derived from both, together with the vast energy needed to support a large molten planetary body condensing from a plasma cloud orbiting the Earth. Such a giant impact is also implicated in the final stages of core formation within the Earth.
A core formed from molten iron alloyed with nickel would have acted as a chemical attractor for all other elements that have an affinity for metallic iron: the siderophile elements, such as gold and platinum. Yet the chemistry of post-moon formation basaltic melts derived from the Earth’s mantle contain considerably more of these elements than expected, a feature that has led geochemists to wonder whether a large proportion of the mantle arrived – or was accreted – after the giant impact.
A tool that has proved useful in geochemistry on the scale of entire planets – well, just the Earth and Moon so far – is measuring the isotopic composition of tungsten, a lithophile metal that has great affinity for silicates. One isotope is 182W that forms when a radioactive isotope of hafnium (182Hf) decays. The proportion of 182W relative to other tungsten isotopes has been shown to be about the same in Lunar Highland anorthosites as it is in the Earth’s mantle. This feature is believed to reflect Moon formation and its solidification after the parent 182Hf had all decayed away: the decay has a half-life of about 9 Ma and after 60 Ma since the formation of the Solar System (and a nearby supernova that both triggered it and flung unstable isotopes such as 182Hf into what became the Solar nebula) vanishingly small amounts would remain.
Oddly, two papers on tungsten and Earth-Moon evolution, having much the same aims, using similar, newly refined methods and with similar results appeared in the same recent issue of Nature (Touboul, M. et al. 2015. Tungsten isotopic evidence for disproportional late accretion to the Earth and Moon. Nature, v. 520, p. 530-533. Kruijer, T.S. et al. 2015. Lunar tungsten isotopic evidence for the late veneer. Nature, v. 520, p. 534-537). The two of them present analyses of glasses produced by large impacts into the lunar surface and probably the mantle, which flung them all over the place, maintaining the commonality of the ventures that might be explained by there being a limited number of suitable Apollo samples. Both report an excess of 182W in the lunar materials: indeed, almost the same excess given the methodological precisions. And, both conclude that Moon and Earth were identical just after formation, with a disproportional degree of later accretion of Solar nebula material to the Earth and Moon.
So, there we have it: it does look as if Earth continued to grow after it was whacked, and there is confirmation. Both papers conclude, perhaps predictably, that the early Solar System was a violent place about which there is much yet to be learned…
When I learned about the unveiling of Lunar Mission One (LM1) , a few days after the global excitement about ESA’a Rosetta mission following Philae’s 12 November 2014 landing on a far-distant comet and success with its core experiments, it did cross my mind that here was a bit of a let-down in PR terms. There’s an old saying – ‘What can follow the Lord Mayor’s Show?’ – and the thrill of Philae’s landing rivalled any of the events at the 2012 London Olympics, plus the science it and Rosetta promise is likely to be about as leading-edge as it will get for quite some time. So what does LM1 offer that might achieve a similar scoop, and indeed your prospect of virtual immortality?
Unlike NASA or ESA missions, LM1 is to be a crowd-funded private enterprise by Lunar Missions Ltd, and for that the subscribers will want something in exchange. Through Kickstarter anyone can have a punt to help raise the initial £600 thousand goal by midnight on 17 December. Apparently that sum is to fund 3 years full-time work by a professional management team to raise further mission funds from commercial partners to take the project further: it will cost at least £0.5 billion. At this stage you can pledge any sum you wish, but what you get in return depends on your generosity. Highlights are: for £3 to 15 the reward is ‘Our eternal thanks and a place in space history’; >£15 gets you a certificate and a place in an online ‘wall of thanks’; >£30 escalates to your name being included in a digital ‘time capsule’ taken to the Moon and buried, plus membership of the Lunar Missions Club; >£60 entitles you to a voucher to invest in your own digital ‘memory box’ to go in the capsule – one of ‘millions and millions’ – and a vote on key decisions; for >£300 you can ‘Meet the Team’; >£600 gets you annual meetings and a chance to ballot for the landing module’s name; for higher contributions there are invitations to the launch (>£1200), sealing of the digital archive capsule and your name engraved on the lander (>£3000); and – wait for it – you get a place in the viewing gallery at Mission Control if you can stump up more than £5000.
For those contributing £60 or more, what goes in the much vaunted digital ‘Memory Box’ is on a sliding scale, from the equivalent of a text message to a strand of your hair and the DNA in it. One catch, if you are thinking of resurrection, is that it will be at the bottom of a 5 cm diameter hole at least 20 m deep. The buried digital archive will also contain a record of all living species on Earth and the entire history of humankind to date, but a continually updated copy will also be freely available online. Wikipedia seems not to be associated for some reason, but every item in this public archive will be peer-reviewed through an editorial board to whose deliberations schools, colleges and universities can contribute. The buried, multi-Terabyte, digital capsule is said to have a life of perhaps a billion years. Currently the longest lived data storage (~1500 years) is still ink on vellum, whereas the most advanced static and optical digital media are estimated to have a maximum 100 year lifetime, subject to technical obsolescence. On the plus side, privacy is guaranteed, partly by the nature of the storage. So, for £10000 Joe and Josie Soap will figure on a kind of cenotaph but who- or whatever digs up the module will learn absolutely nothing about them and but conceivably could clone them from their anonymous strands of hair.
What are the science goals for an LM1 landing scheduled for 2024 that cannot be achieved by lunar-lander and sample-return missions currently under state-funded development by China, Russia, NASA, Japan and India before LM1 reaches the ‘Go/No Go’ stage? The landing is planned for the Moon’s South Pole, on the rim of a major crater. There, LM1 will drill a hole to between 20-100 m deep, using a maximum of 1 kW of solar power – this ‘will also be a major leap forward for safer and more efficient remote drilling on Earth’: make of that claim what you will. Such a hole is said to enable sampling of pristine lunar rock in 15 cm lengths of 2.5 cm diameter core through the debris of the impact that caused the crater. The core samples are to be chemically analysed in the lander to test the hypothesis that Earth and Moon shared their origins. Future missions may pick up the cores and return them for more detailed analysis on Earth. But consider this: the oldest rocks known from the Apollo programme are approximately 4.4 billion year-old, feldspar-rich anorthosites that are thought to have formed the lunar highlands through fractional crystallisation of an early magma ocean that immediately followed Moon formation. Any unfractionated lunar material is only likely, if at all, at far greater depths than 20 m, and none was found or even suggested among the 0.4 tonnes of samples returned by the Apollo missions, which have been repeatedly analysed using advanced instruments. Indeed, near-surface debris from a crater rim is unlikely to be any more diverse lithologically than the various kinds of lunar surface from which the Apollo samples were collected, and may be contaminated by whatever caused the cratering and by the immense, long-lived heating at the impact site itself.
Compared with the prospect of advancing understanding of the origins of life and the Earth’s oceans, and the early stages of Solar System evolution from data provided by Rosetta and Philae, LM1 might seem less exciting, though the buzz being hyped is that it would be a People’s Mission. Yet those who place their punt on it and the commercial concerns that ultimately earn from it are two different sets of people. The ambitious global education wing will, of course, face competition from the growth of MOOCs in the science, technology, engineering and maths area that have a considerable head start, but it does have a noble ring to it. Whatever, if you make a pledge before midnight on 17 December this year and the ‘pump-priming’ target is not met by then, you pay nothing. If £600 thousand is raised there is no going back and only 10 years to wait. But what a challenge, you may well think… LM1 definitely has the edge over Virgin Galactica, but here on Earth there are probably a great many more vital challenges than either.
Large features on the near side of the Moon give us the illusion of the Man-in-the-Moon gazing down benevolently once a month. The lightest parts are the ancient lunar highlands made from feldspar-rich anorthosite, hence their high albedo. The dark components, originally thought to be seas or maria, are now known to be large areas of flood basalt formed about half a billion years after the Moon’s origin. Some show signs of a circular structure and have been assigned to the magmatic aftermath of truly gigantic impacts during the 4.1-3.8 Ga Late Heavy Bombardment. The largest mare feature, with a diameter of 3200 km, is Oceanus Procellarum, which has a more irregular shape, though it envelopes some smaller maria with partially circular outlines.
A key line of investigation to improve knowledge of the lunar maria is the structure of the Moon’s gravitational field above them. Obviously, this can only be achieved by an orbiting experiment, and in early 2012 NASA launched one to provide detailed gravitational information: the Gravity Recovery and Interior Laboratory (GRAIL) whose early results were summarised by EPN in December 2012. GRAIL used two satellites orbiting in a tandem configuration similar to the US-German Gravity Recovery and Climate Experiment (GRACE) launched in 2002 to measure variations over time in the Earth’s gravity field. The Grail orbiters flew in a low orbit and eventually crashed into the Moon in December 2012, after producing lots of data whose processing continues.
The latest finding from GRAIL concerns the gravity structure of the Procellarum region (Andrews-Hanna, J.C. and 13 others 2014. Structure and evolution of the lunar Procellarum region as revealed by GRAIL gravity data. Nature, v. 514, p. 68-71) have yielded a major surprise. Instead of a system of anomalies combining circular arcs, as might be expected from a product of major impacts, the basaltic basin has a border made up of many linear segments that define an unusually angular structure.
The features only become apparent from the gravity data after they have been converted to the first derivative of the Bouguer anomaly (its gradient). Interpreting the features has to explain the angularity, which looks far more like an outcome of tectonics than bombardments. The features have been explained as rift structures through which basaltic magma oozed to the surface, perhaps feeding the vast outpourings of mare basalts, unusually rich in potassium (K), rare-earth elements (REE) and phosphorus (P) know as KREEP basalts. The Procellarum polygonal structure encompasses those parts of the lunar surface that are richest in the radioactive isotopes of potassium, thorium and uranium (measured from orbit by a gamma-ray spectrometer) – thorium concentration is shown in the figure.
Tectonics there may be on the Moon, but the authors are not suggesting plate tectonics but rather structures formed as a huge mass of radioactively heated lunar lithosphere cooled down at a faster rate than the rest of the outer Moon. Nor are they casting doubt on the Late Heavy Bombardment, for there is no escaping the presence of both topographic and gravity-defined circular features, just that the biggest expanse of basaltic surface on the Moon may have erupted for other reasons than a huge impact.
Mercury is quite different from the other three Terrestrial Planets, having a significantly higher density. So it must have a considerably larger metallic core than the others – estimated to make up about 70% of Mercury’s mass – and therefore has a far thinner silicate mantle. The other large body in the Inner Solar System, our Moon, is the opposite, having the greatest proportion of silicate mantle and a small core.
The presently favoured explanation for the Moon’s anomalous mass distribution is that it resulted from a giant collision between the proto-Earth and a Mars-sized planetary body. Moreover, planetary theorists have been postulating around 20 planetary ‘embryos’ in the most of which accreted to form Venus and Earth, the final terrestrial event being the Moon-forming collision, with smaller Mars and Mercury having been derived from the two remaining such bodies. For Mercury to have such an anomalously large metallic core has invited mega-collision as a possible cause, but with such a high energy that much of its original complement of silicate mantle failed to fall back after the event. Two planetary scientists from the Universities of Arizona, USA, and Berne, Switzerland, have modelled various scenarios for such an origin of the Sun’s closest companion (Asphaug, E. & Reuffer, A. 2014. Mercury and other iron-rich planetary bodies as relics of inefficient accretion. Nature Geoscience, published online, doi: 10.1038/NGEO2189).
Their favoured mechanism is what they term ‘hit-and-run’ collisions in the early Inner Solar System. In the case of Mercury, that may have been with a larger target planet that survived intact while proto-Mercury was blasted apart to lose much of it mantle on re-accretion. To survive eventual accretion into a larger planet the left-overs had to have ended up in an orbit that avoided further collisions. Maybe Mars had the same kind of lucky escape but one that left it with a greater proportion of silicates.
One possible scenario is that proto-Mercury was indeed the body that started the clock of the Earth-Moon system through a giant impact. Yet no-one will be satisfied with a simulation and some statistics. Only detailed geochemistry of returned samples can take us any further. The supposed Martian meteorites seem not to be compatible with such a model; at least one would expect there to have been a considerable stir in planetary-science circles if they were. For Mercury, it will be a long wait for a resolution by geochemists, probably yet to be conceived.
Charting the variation in gravitational potential across a planet provides a measure of the distribution of mass beneath its surface. That depends on both the planet’s actual shape and on internal variations in rock density. The Earth’s gravity has been mapped with varying degrees of precision, depending on sample spacing, by surface measurements using gravimeters. Doing gravity surveys from space cannot be so direct, however. One ingenious approach for the gravitational field over the oceans is to measure the mean height of the ocean surface using radar beams from a satellite. Since this is affected by variations in the gravitational field, partly due to bathymetry and partly because of varying density beneath the ocean floor, removing the calculable bathymetric effect leaves a gravitational signal from the underling lithosphere and deeper mantle. The first satellite to illuminate the Earth with radar microwaves, Seasat, gradually built up such a gravitational map of the deep Earth over a period of 105 days in 1978, which was followed up by other satellites such as the ERS series and Topex-Poseidon.
It is not so easy to map gravity precisely above a solid planetary surface, but through the GRACE experiment this can be done by measuring very precisely the distance between a pair of satellites that follow the same orbit. As the gravitational field changes so too does the separation between the tandem of satellites; an increase in gravity pulls the satellites closer together and vive versa. GRACE has provided some fascinating data, such as estimates of the withdrawal of groundwater from large sedimentary basins and shrinkage of ice caps. However, GRACE is limited in its resolution of gravitational anomalies by the fact that Earth has an atmosphere above which such tandems must be parked in orbit to avoid burning up. The higher the orbit, the more degraded is the resolution. This effect is much less for Mars and non-existent for the Moon.
A sister experiment to GRACE has been orbiting the Moon since September 2011: the Gravity Recovery and Interior Laboratory (GRAIL). First the tandem orbited at 55 km, then 22 and for a brief period 11 km, before running out of thruster fuel on 17 December 2012 and crashing into the lunar surface. Results from the highest orbit resolve lunar gravity to 13 km cells, recently reported on-line in three papers (Zuber, M.T. and 16 others 2012. Gravity field of the Moon from the Gravity Recovery and Interior Laboratory (GRAIL) Mission. Science, doi 10.1126/science.1231507; Wieczorek, M.A. and 15 others 2012. The crust of the Moon as seen by GRAIL. Science, doi 10.1126/science.1231530; Andrews-Hanna, J.C. and 18 others 2012. Ancient igneous intrusions and early expansion of the Moon revealed by GRAIL gravity gradiometry. Science, doi 10.1126/science.1231753). From crater gravitational signatures due to variations in surface topography it seems that the early bombardment of the lunar surface far exceeded previous assumptions. Impact effects dominate the GRAIL data at this resolution, but 2% of the information relates to structures hidden at depth.
There are linear gravity anomalies extending over hundreds of kilometres, which may be huge igneous intrusions in the form of dykes; perhaps reflections of early influences of early extensional tectonics in the Moons lithosphere. Estimates point to this having been due to an up to 5 km increase in the lunar radius, probably as a result of thermal changes. The dominant feature of the lunar surface is not the near-side flat basaltic maria, visually prominent as they are, but the far more rugged lunar highlands which stand far higher because of the lower density of their constituent feldspar-rich anorthosites. GRAIL permitted a bulk estimate of the density of highland crust that turned out to be substantially lower, at 2550 kg m-3 – compared with 2600-2700 for granite and 2800-3000 for basalt – than originally estimated from samples returned by the Apollo mission. This forces a reassessment of the thickness of highland crust from 50-60 km to between 34 and 43 km, with a near-surface layer that has a porosity of around 12%, probably resulting from its awful battering. A thinner highland crust than previously assumed presents a bulk geochemical picture that need not be more enriched in ‘refractory’ elements, such as aluminium and calcium, than is the Earth.
Such unanticipated results from the low-resolution mode of the GRAIL experiment have its science team almost salivating at prospects from the sharper ‘pictures’ that will arise from the lower altitude orbits.
- Violent past revealed by map of moon’s interior (sciencenews.org)
- NASA moon-mapping mission to come to a crashing end (thehimalayantimes.com)
Although a few would-be space faring countries have ambitions, a post-Apollo crewed mission to the Moon is unlikely for quite a while. Yet moon-struck curiosity goes on: currently there is a surge in re-examining the lunar samples brought back more than 40 years ago. The Lunar Sample Laboratory Facility in Houston holds about a third of a ton of rock and regolith. I suppose part of the reason why lunar rocks are being re-analysed – in fact some for the first time – is because new or improved methods are available, but frustration among a growing community of planetary geochemists having little more than meteorites to peer at probably plays a role as well. Since Hartman and Davis first suggested it, the giant impact theory for the Moon’s origin has dominated geochemical ideas. Most tangible is that of a magma ocean, floated plagioclase crystals from its fractional crystallisation probably having formed the glaring white lunar highlands composed of anorthosite. More subtle are ideas about what happened to the Mars-sized planet that did the damage to Earth and flung vaporised rock into orbit to accrete into the new Moon, and the effects of the stupendous energy on the geochemistry of all three bodies. Directed at all that is new research on isotopes of zinc (Paniello, R.C. et al. 2012. Zinc isotope evidence for the origin of the Moon. Nature, v. 490, p. 376-379).
The focus on zinc is because it is easily vaporised compared with more refractory materials, such as calcium an titanium, and as well as being ‘volatile’ it has five naturally occurring isotopes with relative atomic masses of 64 (the most abundant), 66, 67, 68 and 70. In general, isotopes of an element behave in slightly different ways during geological and cosmological processes, which changes their proportions in the products; a process known as ‘mass-fractionation’. Paniello and colleagues from Washington University, Missouri and the Scripps Institution of Oceanography, California USA found that Moon rocks are enriched in the heavier isotopes of zinc yet depleted in total zinc compared with terrestrial rocks and meteorites supposed to have come from Mars. Unlike those two planets the Moon’s zinc deviates from its abundance relative to other elements recorded by chondritic meteorites. This zinc depletion tallies with volatile loss from incandescent vapour blurted from the colliding planets. But it doesn’t help with the detailed predictions from the giant-impact model. A variety of scenarios suggest that the Moon should be made from remnants of the inbound impactor’s mantle, yet studies of other elements’ isotopes indicate that the Moon is rather Earth-like. But not those of zinc, so it looks like they have to be explained by a complete rethink of the whole hypothesis (Elliott, T. 2012. Galvanized lunacy. Nature, v. 490, p. 346-7).
The South Pole and the farside of the Moon contain, at 2500 km across and 13 km deep, the largest impact structure in the Solar System: the South Pole-Aitken (SPA) basin. Being partly camouflaged by many later craters up to several 100 km across, typical of the lunar far side and the lunar highlands in general, the SPA basin formed early in the Moon’s cratering history, and is unlike the mare basins of the near side that are filled with basalt lavas. The light colour of the lunar highlands into which the SPA basin was excavated signifies that they are dominated by almost pure feldspar in the form of anorthosite rock. These anorthosites are prime evidence for the former melting of much if not all of the Moon at the time of its formation: low-density feldspar with a very high melting point could only have accumulated with the degree of purity of anorthosite if early-formed crystals floated to the top of the magma ocean.
The other feature of feldspars is that they are among the least magnetic of minerals, so it came as a surprise that the northern rim of the SPA basin is studded with positive magnetic anomalies (Wieczorek, M.A. et al. 2012. An impactor origin for lunar magnetic anomalies. Science, v. 335, p. 1212-1215). Lunar samples returned by the Apollo Programme are consistently lacking in all but the weakest remanent magnetism, suggesting that the Moon either never had a magnetic field or if it did the field was extremely weak. Even if it did once have a magnetic field, the anomaly patterns are small with high amplitude and reminiscent of a target hit by a shotgun blast. Similar anomalies are scattered on the near side.
The SPA basin is elliptical, suggesting that the projectile responsible for it struck at an oblique angle. The far=side magnetic anomalies cluster exactly where impact modelling would suggest for debris displaced by impact from a northward travelling body. The interpretation arrived at by Mark Wieczorek of the Parisian Institut de Physique du Globe and colleagues from MIT and Harvard University in the US is that the anomalies mark landing sites for large fragments of an easily magnetised, iron-rich asteroid that excavated the basin. Moreover, the same impact might explain magnetic anomalies much further from the basin, on the lunar near side. The remaining mystery is how fragments of the impactor came to be magnetised. The impact would have ensured their being heated well above the temperature of the Curie point at which even the most magnetically susceptible materials lose their magnetisation. The most likely possibility is that the fragments attained their magnetised state at a time when the moon did have a core-generated magnetic field, albeit weak.
- Moon’s magnetic material may have come from an asteroid (latimes.com)
- NASA video guide to the Moon – http://lunarscience.nasa.gov/articles/video-a-tour-of-the-moon/
- NASA video of lunar evolution – http://lunarscience.nasa.gov/articles/video-evolution-of-the-moon/
The most significant discovery from the Apollo lunar landings is that the Earth and Moon shared a fiery early history, when a planetary body around the size of Mars slammed into the Earth to fling off vaporised rock that condensed to create the Moon. Such a catastrophic event reset the geochemistry of the Earth, and both it and the Moon likely had an early phase dominated by a deep ocean of magma. The evidence for a magma ocean comes mainly from the lunar highlands which are dominated by almost pure calcium plagioclase feldspar (the rock anorthosite), suggesting that this high-temperature, low-density silicate mineral crystallised and then floated to the surface of the Moon. Yet there is a great deal of evidence about the Moon that did not depend on people setting foot on its surface. For instance, detailed photographic records of the surface and extremely precise measurements of the surface elevation stem from cheaper orbital missions, including coverage of the unvisited far side of the Moon.
The face of the Moon never seen from Earth has long been known to have one of the largest impact basins in the solar system, the South Pole – Aitken basin. Analysis of the far side’s surface elevation data from the Lunar Orbiter Laser Altimeter (LOLA) also shows that it is significantly higher than the near side. It is also far more heavily cratered than the near side. Now there is a plausible explanation for the dichotomy: the Moon received another stupendous blow (Jutzi, M & Asphaug, E. 2011. Forming the lunar farside highlands by accretion of a companion moon. Nature, v. 476, p. 69-72). But how come that didn’t blast the Moon apart or re-melt it and allow it to re-shape to a near perfect sphere? The modelling study suggests that if the culprit slowly collided – around 2-3 km s-1 – it would have wrapped around the early Moon to plaster the surface with debris, nicely shown by the paper’s graphics. Such a ‘slow’ impact is only possible from a co-orbital companion moon, objects from outside the Earth-Moon system inevitably being accelerated by gravity to at least the equivalent of its escape velocity (about 11-12 km s-1). That exceeds the speed of sound through rock, leading at least to a very large hole, shock metamorphism and, with a massive body, to extensive melting (the energy would be ½ mv2) rather than the observed lunar far-side bulge. Jutzi and Asphaugs’s modelling comes up with a companion moon around 1200 km across, that may have formed from the same massive event that created the Moon itself. It could have accreted from the impact-induced vapour disc at a Trojan point in the lunar orbit, where gravitational forces balance to keep orbital objects apart. The gradual expansion of the lunar orbit in response to tidal forces – large in the early history of the Earth-Moon system – could have destabilised the balance so that the companion moon slowly drifted towards the Moon and eventual collision.
One such modelling becomes closer to known reality, i.e. the far-side bulge, it gets more tempting to look for secondary possibilities. One of these the effect of such a ‘slow’ impact on the remaining magma ocean on the Moon. It may have blurted that by then deep molten layer to the side opposite the impact. That, the authors suggest, may be responsible for the geochemical peculiarities of the flood basalts that filled the much later lunar maria on the near side. There are no signs of these KREEP basalt floors to large later craters on the far side, such as the Aitken basin, formed around 4.0 to 3.8 Ga ago at the same time as the near-side maria. A variety of new instruments orbit the Moon and more are planned, so this model presents a nice hypothesis for them to test: what is the betting that a robotic lander might eventually be sent to return samples from the enigmatic far side?
- Did a Slo-Mo Crash Create the Two-Sided Moon? (news.sciencemag.org)
- You: Second moon may have collided with our moon, say scientists (guardian.co.uk)