 Well good evening. I'm Michael Rowan Robinson and I'm delighted to welcome you all to the fifth Institute of Physics Isaac Newton Lecture. The Isaac Newton Medal is the Institute's top award and it's a great pleasure and privilege to introduce the 2012 recipient Martin Rees. Martin is one of the leading figures in astrophysics and cosmology in the world. His 1966 paper on the appearance of relativistically expanding radio sources showed how apparently faster than light phenomena can be generated in special relativity. With various collaborators he developed the modern picture of a massive black hole origin for active galactic nuclei. He's been extremely influential in the development of models for gamma ray bursts. In cosmology he's written key papers on cooling dynamics and fragmentation of massive gas clouds, on how galaxies form through baryonic condensation in cold dark matter halos, and on how quasars may control galaxy formation and evolution. Martin is professor of astrophysics and cosmology at the Institute of Astronomy Cambridge. He was director there for ten years altogether and played a major role in the development of the Institute as one of the major astronomy institutes in the world. He was president of the Royal Astronomical Society from 1992 to 94, has been astronomer royal since 1995 and was president of the Royal Society from 2005 to 2010. In that role he provided leadership on many of the great science and society issues of the day. He won the gold medal of the Royal Astronomical Society in 1987, the Bruce Medal of the Astronomical Society of the Pacific in 1993, the Faraday Award of the Royal Society for Science Communication in 2004 and the Crawford Prize of the Royal Swedish Academy in 2005. He's an honorary fellow of numerous foreign academies. He was made a life peer in 2005 and a member of the Order of Merit in 2007. He was master of Trinity College Cambridge from 2004 to 2012. Martin has been very active in outreach, as of course we all have to be these days. He's the author of several popular science books including Just Six Numbers and Our Final Century and has a new book in preparation. He's been a trustee of the British Museum and of the Science Museum. He was the 2010 Wreath Lecturer. I've known Martin for 50 years and I know that he will give a brilliant lecture this evening. His title is From Mars to the Multiverse. Martin Wreath. Well, thank you very much. It's a great privilege to give this lecture and a great pleasure to be introduced by Michael Rural Robinson who is indeed my oldest astronomical friend. I realize that I'm speaking to a general physics audience, not to astronomers. And in case there are any hardcore theorists here who've never wondered about those things in the night sky, here's a definition of a star for them. I'll explain this later. Michael and I were lucky to start our research in the mid-1960s. Those years saw the birth of relativistic astrophysics. Astronomers discovered the first compelling evidence that our universe had expanded from a big bang. The thermal microwave background and objects like neutron stars and pulsars were Einstein's theory was crucial and where they displayed very extreme and fascinating physics. And Michael and I both studied strong radio sources far beyond our galaxy, which we later learned were powered by jets energized by black holes. It was an exhilarating time for the young because when so much is new, the old guys don't have a big head start over the young. But the good news for students or post-docs in the audience is that today is an equally good time for young researchers. The pace of advance has crescendoed rather than slackened. Instrumentation and computer power have improved hugely. We've discovered a whole menagerie of exotic objects and surveyed millions of galaxies. The volume and precision of astronomical data is unprecedented and the scale and geometry of the observable universe is well pinned down. We can trace cosmic history confidently right back to the first nanosecond. And these advances bring into focus questions that couldn't even have been posed back in the 1960s. We want to understand not just the way things are, but how the cosmic panorama of which we're apart emerged from the universe's hot dense beginning and what the long-term future is too. The 1960s also saw the beginning of the space age. And this lecture commemorates Newton and I mentioned him here because he must have thought about space travel. Indeed, this well-known picture from the English edition of His Principia is still the neatest way to teach the concept of orbital flight. Newton knew that for a cannonball to achieve an orbital trajectory, it had to go at 18,000 miles an hour. And of course, as we know, that speed wasn't achieved until 1957 with the launch of Sputnik 1. And only 11 or 12 years after that, we had this iconic picture taken by astronauts orbiting the moon and we had the moon landings. Here's Jack Schmidt, the last man on the moon. He's a geologist. He spent three days on its surface back in 1972. The Apollo program is a heroic episode. And it was all over 40 years ago. If momentum had been maintained, there'd be footprints on Mars by now. But actually, of course, people have done no more than circle the Earth in low orbit, many of them more recently in the International Space Station. But space technology, of course, has burgeoned for communications, environmental monitoring, sat nav, and so forth, and we depend on it every day. And for astronomers, it's revealed the far infrared, the UV, X-ray, and gamma-ray sky, telling us huge amounts about the high-energy universe. And unmanned probes to other planets have been back pictures of varied and distinctive worlds. A very quick tour. Five million miles away on the way to the moon, looking back at the Earth. You see that with the sun shining from the right. And when you get to Mars, here's a picture of the red planet. And this is the Curiosity probe, which landed last August and will do on Mars what Jack Schmidt did on the moon. It'll trundle around for about 10 years studying the lunar geology. This is a picture, a talk of its own surroundings. You can see the geological layers there. It's landed in the top left where that ellipse is of this 50-mile crater. It's going to trundle around and eventually climb the three-mile-high mountain in the middle. Further afield, here's Jupiter. The four Galilean moons have all been observed close up. And the Cassini probe went to Saturn. This is a lovely picture. This is a picture taken by Cassini, which shows an eclipse of the sun by Saturn. Cassini is beyond Saturn, lined up at a distance such that Saturn blocks out the sun, but the rings of Saturn are in sunlight. And just where that arrow is, there's a little spot which is actually the Earth, the scene from that distance. Cassini carried in its cargo bay a European robotic probe called Huygens, whose task was to land on Saturn's giant moon Titan. And that's what it did. Left and center, pictures taken on the way down. Right-hand side where it landed. They look a rather nice place, rivers and a little lake, but the rivers are liquid ethane. Temperature minus 160 degrees centigrade. Well, I hope that during the coming decades, the entire solar system will be explored and mapped by flotillas of tiny robotic craft and maybe robotic fabricators out there too. But will people follow them? That's a topic for another lecture. I think they will, but as high-risk adventurers, rather than for any practical purpose. And the other question that we're always asked is, is there life out there already before humans or post-humans get there? Well, prospects look pretty bleak in our solar system, but if we widen our horizons beyond the scale of any probe we can now envisage to the other stars, then prospects look very different. Perhaps the hottest topic in astronomy now is the realization that many other stars, perhaps even most, are orbited by retinues of planets just like the Sun is. These planets are not detected directly, but they're inferred by precise measurements of their parent star. There are two main methods, both very simple in principle. Here's the first. If a star is orbited by a planet, then both planets and stars move around the center of mass the barycenter, and a star being more massive, of course, moves in a smaller orbit. But careful Doppler measurements of the stellar spectrum can detect these motions, even if they're only a few meters per second. This is one case where you see the sinusoidal velocity distribution of a circular orbit of a planet. And by this technique, of course, you can infer the mass of the planet and the length of its year, and also whether it's circular or eccentric in its orbit. And this is just a list of planets found by this technique. This is not very up to date, but several hundreds have been found. Some stars have several planets around them. This evidence pertains mainly to giant planets, planets rather like Jupiter or Saturn, the giants of our solar system. Detecting an Earth-like planet by this technique is much harder, because the Earth, for instance, induces a motion in the sun but of only a few centimeters per second, and that's too small, 10 to the minus 10 C, to be detected in the shift of a spectral line. But there's another equally simple technique which can detect Earth-like planets. And this is to, as it were, look for their shadow. The star would dim slightly if a planet moves across in front of it. An Earth-like planet transiting a sun-like star causes a fractional dimming, recurring once per orbit of one part in 10,000. And the Kepler spacecraft has been pointing steadily at an area of sky about 7 degrees across and monitoring, photometrically, the brightness of over 150,000 stars, doing that at least twice every hour for every star with a precision of one part in 100,000. It's already found more than 2,000 planets, many no bigger than the Earth. And, of course, it only detects the transits of those whose orbital plane is nearly aligned with our line of sight. For everyone it finds, you can expect there are a order of 50 more. We're especially interested, of course, in possible twins of our Earth. Planets the same size as ours on orbits with temperatures such that water neither boils nor stays frozen. And the real goal, of course, is to see them directly, like in this simulation. We can't do this yet. To do it is hard. To realize how hard it is to actually image a planet, suppose an alien astronomer with a powerful telescope was viewing the Earth from, say, 50 light-years away. The distance of a nearby star. Our planet would seem, in Carl Sager's nice phrase, a pale blue dot, very close in the sky to its star, our Sun, that outchines it by many billions. You're looking for a firefly next to a searchlight, as it were. But if the aliens could actually detect the planet, our Earth, they could learn quite a bit about it. The shade of blue will be slightly different, depending on whether the Pacific Ocean or the landmass of Asia was facing them. So they could infer the length of the day, that there were oceans and continents, the gross topography, and something about the climate and the seasons. By analyzing the faint lights, they could infer that it had a biosphere. Well, we can't do this yet, but within 20 years, this huge telescope planned to be built by the European Southern Observatory, unimaginatively named the Extremilars Telescope, ELT, which has a mosaic mirror 39 meters across. That will be drawing inferences like those I just mentioned about Earth-like planets orbiting sun-like stars. And that will indeed be fascinating. What surprised people about these planetary systems is their great variety. There are many Jupiter-like planets very close in, whose year is only a few days. There are some planets on very eccentric orbits. There are some planets on orbits counter-rotating relative to the spin of their parent star. And there's a planet orbiting a double star that has two suns in its sky. And that, by an incident, was found by two amateur citizen scientists who accessed the time series from the Kepler spacecraft of a set of stars, and they happened to strike lucky and find this rather odd set of transits indicating a planet orbiting a double star. Well, the nature of some of these planetary systems was surprising, but the existence of planets wasn't surprising because we've learnt that stars form via the contraction of clouds of dusty gas as shown in this cartoon. If the cloud has any angular momentum, then as it contracts, it'll rotate faster and then it'll spin off a dusty disk around a protostar as in this cartoon. And in such a disk, gas condenses and closer in the less volatile dust agglomerate into rocks and planets. And this should be a generic process in many protostars. Here's another flashback to Newton. This is a quote from his optics. Newton understood planetary orbits, but he didn't understand why these orbits were all close to the ecliptic plane unlike those of comets. Of course, the dusty disk origin, which I've just shown you, accounts for this. So we've stepped back in the causal chain from where Newton was. And what I'll do later in my talk is to outline how the causal chain has been pushed back further still to the formation of galaxies, stars and atoms, and right back to the first nanosecond of the Big Bang. Well, first, what about stars and atoms? We see stars forming. The early picture was a cartoon. This is a real picture of the Eagle Nebula, 7,000 light-years away. And we see stars dying. This is what the sun would look like in about six billion years. Another dying star. A star dying rather messily here. And here, the remnant of a more massive star. This is the famous crab nebula, the remnant of an explosion of a supernova witnessed by Chinese astronomers in 1054 AD. Well, supernovae are important because as Fred Hoy was the first person to fully realize, before a massive star explodes, it's built up a sort of onion skin structure where nuclear fusion has transmuted material further up the periodic table closer in, rather like that, and then it all flung out again into interstellar space when it explodes and then it merges with the interstellar medium and then new stars form. So what's going on in our galaxy is a sort of ecological recycling where pristine gas goes into the first stars, is partially transmuted to the higher elements and then new stars form. So all the atoms that we are made of are processed through stars. Indeed, each of us has inside us atoms from many hundreds of different stars which exposure to supernovae all over the Milky Way, all more than four or four and a half billion years ago before our solar system formed. So our galaxy is a sort of ecological recycling system. Let's now enlarge our spatial horizons to the extragalactic realm. If you could get two million light years away from our galaxy and look back, you'd see something like this. This of course is Andromeda, the nearest big galaxy to us. It's a spinning disk viewed obliquely with a hundred billion stars spinning around a central hub. Here's another well-known galaxy, the whirlpool seen face on. And we have huge samples of galaxies to study. This shows the results of a survey of galaxies. You could see they're grouped together in clusters and they thin out here because the fated ones are a long way away from the sea, but they are in clusters which are spread through space. Well, how much can we actually understand about galaxies? Physicists who study particles can of course probe them and crash them together in accelerators. Astronomers can't do experiments. Astronomers can't crash real galaxies together. And galaxies change so slowly that they only offer us a sort of snapshot of each one. But we can do experiments in the virtual world of our computer. And here is such a virtual collision of two galaxies. Tens of 15 times faster than would actually happen. Everything in one galaxy such as gravitation pool and everything else, they merge together in this sort of train wreck. And this will happen, by the way, to our galaxy and Andromeda. They will crash together but not for another four billion years. And when we look at the real sky, we find systems like this. And having done simulations like the ones I just showed you, we can infer that these two galaxies have got dangerously close. Each is raising tidal plumes on the other. And when we came back in say, tens of eight years, then these will probably have merged into one amorphous galaxy. We can do or redo the simulations like the ones I showed you and of other galaxies, making different assumptions about the mass of stars, the mass of gas in each galaxy and so forth, and see which of the simulations matches best what we actually see in the sky. And the most important thing we find when we do these simulations and use other evidence is that all galaxies are held together by the gravity not just of what we see, not just of the gas and the stars, but they're all embedded in a swarm of particles which are invisible, but which collectively contribute about five times as much matter as your neatoms do. This is so-called dark matter, a swarm of particles made in a big bang. We don't know what they are, but we know that they are almost non-interacting. The four lines of evidence I've written down here. There's a time to go into them all, but the second one, looking for gas in a cluster of galaxies, here's a map of the Perseus cluster showing the profile of the X-ray gas, and from data like this you can infer the temperature of the X-ray gas and therefore you can infer the depth of potential well and therefore you can infer the mass of the cluster and the mass you infer is about five times more than the total mass of the gas and the stars that you see in the cluster. And just a word about the fourth on the list, gravitational lensing. This is shown here in this picture. The brighter objects on the left are galaxies in a cluster, about a billion light-years away, and then the fainter objects are galaxies further away still. You can see these little arcs and things, and what's happening is that the gravitational field of the cluster is acting like a rather poorly-figured converging lens, and so it's distorting things in the background, distorting them into a sort of streaky pattern in a way. And so we infer by just doing a simple optical inversion, the mass distribution and the total mass, and again by this argument we again infer that the dark matter is about five times more important in providing the gravity of the cluster and of the individual galaxies as the gas and stars themselves. Non-baryonic material is more important than baryonic material. And just to quantify this in very round numbers, we have a number of lines of evidence, the ones I've mentioned, plus evidence from cosmology, that the baryon density is about 4% of what's called a critical density. The critical density is the density of the simplest cosmological theory, the so-called Einstein-Essiter model, which is flat and has zero cosmological constant. And that's the way we sort of calibrate cosmological densities. So ordinary baryons are about 4% of the critical density, and the dark matter inferred by a variety of arguments is between 25 and 30%. I said 30% there, it's between 25 and 30%. And this is a very important and well-established result. We can test ideas on how galaxies evolve, because we can not only look at the nearby ones and model them, but we can look at galaxies as they were 1 billion, 2 billion, 3 billion, 4 billion years ago by looking far out into space. And this picture shows a tiny patch of sky, just a few arc minutes across, and in it you see hundreds of smudges. These smudges are galaxies, some fully equal of our own. But they're so far away that for some of them, the light set out more than 10 billion years ago, so they're being viewed when they were young. And you see, for instance, if you take their spectra, that the proportion of heavy elements is lower than in a present-day galaxy because there's been less time for the reprocessing by stars and supernovae to build up the heavy elements from pristine hydrogen. This object here is the most distant object, which has a really well-established redshift and distance. This is a paper two years ago by astronomers at Imperial College in Sampson, Cambridge. And this is the spectrum of the light from that particular object. I show this tracing. The main point, which you see, is that the Lamin Alpha line of hydrogen, normally in the far UV, 12, 16 angstroms, is stretched by a factor of more than 8 to about 1 micron. And this has a redshift of 7. The redshift is described as 1 plus the redshift. It's the ratio of the observed to the emitted wavelength. This object, incidentally, isn't a typical galaxy. It's especially bright because the starlight is swamped by reprocessed emission by some central black hole. It's what's called a quasar. This is just a simulation of gas swirling down into a black hole. There's been a lot of studies of this. But in the centers of most galaxies, we think there are black holes. And when they're being fueled, the gas swirling down into them, getting magnetized, produces energy which outshines the stars in the galaxy. That's called a quasar. And these very distant objects are much brighter and easier to see if they're behaving like quasars, not just the stellar model. But what about looking still further away? If we look further out, we look back to a time, perhaps before there are any galaxies. So what do we know about the earlier universe? Well, we have reasons to think that our universe started from a hot, dense, big bang, the microwave background. And when people are told that, they sometimes worry about whether it's consistent with the second law of third dynamics that offers dense, big bang to end up as our structured cosmos with immense intricacy, including us, stars and galaxies, and huge contrast between the blazing surfaces of stars and the dark night sky. Well, the answer to this seeming paradox lies in the force of gravity. Gravity enhances density contrast rather than wiping them out. Any patch that starts off slightly denser than average would decelerate more because it feels extra gravity. Its expansion lags behind until it eventually stops expanding and separating out. This movie shows a simulation of part of a virtual universe. It molds at a main large enough to make thousands of galaxies. The expansion is scaled out, so the picture stays the same size, time-scaled in a giga-years on the bottom, and you can clearly see incipient structure unfolding and evolving. Slightly over-dense regions lag behind to condense out. This picture just shows the dark matter. Here's another picture which shows separately the dark matter and the baryons. And when the structure starts to condense, then the potential wells form, and the baryons fall into the potential wells, but they can dissipate and cool and become more centrally concentrated. So the red here is the baryons, and they're going to turn into stars. So this is a model for the sort of formation of a group of galaxies, showing the dark matter in blue and the baryons in red. Moves of this kind portray how galaxies emerge 16 powers of 10 faster than it actually happened. And each of the galaxies is then an arena within which stars, planets, and perhaps life can emerge. And there's one important point. The initial fluctuations fed into the computer model, the small amplitude fluctuations, they weren't arbitrary. They're derived from observations of the microwave background. The microwave background comes from when the universe was about 300,000 years old, when the fluctuations had small amplitude, and projects like the Wilkerson Microrave Anisotropy Satellite are able to map the whole sky and study the fluctuations. And this shows the regions of above average temperature, below average temperature, and the fluctuations have an amplitude of only about 10 to the minus 5, very small, but if you feed in those fluctuations and their scale dependence into the simulation and then run it forward, then what's very gratifying is that you end up with a universe structured like the present universe. That's shown in this picture due to Max Tagmark. The solid curve shows the density contrast as a function of scale at the present time, which is predicted by putting in the fluctuations we observe in the early universe in the microwave background and just running them forward under gravity as in those simulations. And the points plotted show the actual density contrast as a function of scale derived by different techniques, looking at clusters of galaxies and looking at galaxies, etc. So there's good evidence that there is a link between the early universe with those fluctuations and the present universe and its gravity that gets from one to the other. And this is one of the reasons why we fail on the right lines in following up the Big Bang model. And this indicates the claim that structure emerges by clustering of the gravitational dominant dark matter during cosmic expansion. Another flashback to Newton. This is a letter to Richard Bentley when he was asked how the stars formed. And you can see what he said there. He imagined an infinite universe where things would convene partly in one place and partly in another to make an infinite number of great masses. And thus might the sun and fixed stars be formed supposing the matter was of a lucid nature. Well, he wasn't thinking of a dark matter, of course. He didn't realize what kept the stars shining. Something that especially interests me is what might call the dark age and how it ended. This is a picture I showed earlier, looking back, looking out to greater distances. The microwave background photons have traveled freely since an era when the universe was about 300,000 years old and was a temperature of 3,000 degrees. Thereafter, hydrogen cooled down and there was no electron scattering and the primordial radiation shifted into the infrared and the universe became literally dark. And it stayed dark until the first stars formed and lit it up again. So when did this happen? Well, I showed you a picture of a quasar which sent out its light when the universe was about 1 billion years old. So by that time, clearly a lot had happened. And by that time also, there had been enough heating that the primordial gas had been re-ionized and so it had been very cold. It was heated up. We can infer that also from the spectrum. And the last number of galaxies had been assembled. But how much further back did the action actually start? When did the dark age actually end? When to answer this question, astronomers have to look for very much more distant galaxies even than the one I showed you. And one way they do this, incidentally, is by using nature's telescopes, the clusters of galaxies I showed you. If you look behind one of those clusters of galaxies, it acts as a gravitational lens, then there will be some places where you are seeing a magnified image of distant objects. And some candidate galaxies at huge distances have been found in just that way, further away than the picture I showed you. Another way in which we might go further is to use gamma ray bursts, which as Michael mentioned is something I've worked on. Gamma ray bursts are a very spectacular kind of star death when instead of getting just a supernova, you get a jet which if it's pointed towards you, gives you a brightness for a few seconds which is a million times brighter than a galaxy. And if gamma ray bursts form from an early generation of stars at very high red shifts, even when the universe was only 200 million years old, we would see them. So there are things to look out for. And another thing we can do is to study the gas itself. And before doing that, let me mention that in the next decade we expect a number of new developments. The ALMA radio telescope millimeter and submillimeter is being completed in New Mexico and the ELT is being planned. The X-ray project is not actually approved yet. The James Webb telescope is due for launch in 2018. And there's a project called the Square Kilometre Array which is going to be very useful for studying gas at high red shifts. You can study the 21 set of midline of hydrogen at high red shifts and as you will do tomography of the universe because you've got the red shift and direction you know the distance. So you can actually do three dimensional tomography. Square Kilometre Array is being built half in South Africa and adjoining countries. And the other half is going to be on this conveniently situated underpopulated island in the southern hemisphere there. And when this is built it will have the combination of sensitivity and spectral resolution to be able to do this tomography. I'll just show you a simulation of what it might see if you look at an area of sky and scan through frequency space to different red shifts. Going up from red to seven where the gas is mainly ionized. Going back further. This is a simulation making certain assumptions about the galaxies. And that is the sort of information which we will have from the SKA which will be a clue to how the gas changed from being cold and neutral to being ionized as galaxies formed and then grew hierarchically into the big galaxies that we already can observe. Here's the time chart of cosmic evolution again from the hot dense beginning to today's complex cosmos. We can highlight several essential requirements for the emergence of our present universe from simple amorphous beginnings. And I'm going to give you a list of the requirements. First we need gravity. As I showed in those simulations gravity enhances density contrasts as in the movies. It's a very weak force but then it of course holds together stars. On the atomic scale it's about 40 parts of 10 weaker than the electric force but of course large objects are electrically neutral so gravity always wins. And this is my favorite pedagogical diagram. It shows log mass upwards, log radius along the horizontal. And you can see it got protons and a hydrogen atom here and the black hole line slope one in the log log plot and a black hole the size of a proton has a mass of tens of 38 protons. That's because gravity is so weak. And if you look at solids they're on a slope three in the log log plot going up from a proton. And we go up we have people with sugar lumps, people asteroids etc. But when you get up to planets gravity makes them round. Any object more massive than Jupiter gets squeezed to make a star. And that is three-halves power of this ratio of the strength of gravity to the strength of the electric forces. And incidentally I plotted the quantum uncertainty going like one over mass and the place where those two lines meet on the left is the Planck scale where a black hole is no bigger than its Compton wavelength. That's the scale where quantum gravity can't be avoided. Now from diagrams like this you could predict what stars were like even if you lived on a perpetually cloud bound planet. You could predict the action takes place when you get big enough objects to be crushed by gravity despite the weakness of gravity compared to the micro physical forces. Look at the picture you can see that the scale is so large because gravity is weak. If you imagine the hypothetical universe where gravity was say tens of tens times stronger. 30 powers of 10 dot 40 weaker than the Microsoft forces. This picture would look basically the same shape. Stars would still exist as gravitation biofusion reactors but tens of minus 15 the mass. You get a small scale speeded up universe. And that is the explanation for the definition I gave you at the beginning of why stars have that mass. Well the first prerequisite therefore is gravity and the weaker the better because that gives you a bigger range of scales between the micro world and the cosmos. But there are other requirements. There could be no complexity if the universe stayed in thermal equilibrium as our universe was for the first 300,000 years. So that's a requirement. There must be an excess of matter over antimatter. And another requirement is that chemistry must be non-trivial. If hydrogen were the only elements chemistry would be an easy subject but there would be no complexity. And so this requires a balance between the nuclear force that binds nuclear together and the electromagnetic force that separates them off. So we get the binding energy which is familiar to everyone as a function of atomic number. There must also be stars to transmute the pristine simple elements into the rest of periodic table. And probably more than one generation of stars. And there must be as it were a tuned cosmic expansion rate. The universe mustn't collapse too soon otherwise we know time for complexity to evolve and it may stay in thermal equilibrium. On the other hand it can't expand so fast that the expansion energy overwhelms gravity and structures can never pull themselves together. So there must be a sort of tuning there. And there must also be non-zero fluctuations in the early universe. I showed that small fluctuations grew and became non-linear but if there were no fluctuations at all then after 10 billion years our universe would still be cold neutral hydrogen. No stars, no planets and no people. Well, time to art once again. We can trace back to one second. That's when hydrogen helium were made. Indeed we can probably be confident back to a nanosecond. The reason I take that time is that that's when each particle had about 50 GeV of energy. About as much as can be achieved in the LHC. And incidentally the entire visible universe would be squeezed to the size of our solar system at that epoch. But questions like where did the fluctuations come from and why did the early universe contain the actual mix we observe of protons, photons and dark matter. Take us back still further to the brief instance where our universe was hugely more compressed still where the physics is very uncertain. Indeed back perhaps to the time when the energies were about 10 to 16 GeV when experiments offer no direct guide to the relevant physics. At this point let me put in a health warning. I'm going to be a bit speculative from now on. This magazine cover shows the universe when it was a trillionth of a trillionth of a trillionth of a second old actual size. And if we go back to 10 to 16 GeV then everything that we now see in our universe is compressed not to the size of our solar system which it is at 50 GeV but down to literally that size. And according to a popular theory the entire volume we can see with our telescopes inflated at this stage 10 to 16 GeV from a hyper dense blob no bigger than this. Well I wouldn't have had the opportunity to mention this but two previous Newton lecturers Alan Goothe and Ed Whitten both spent almost their entire lecture talking about this extreme era and far be it from me to second guess such great pundits. So we do suspect although we don't know the physics the crucial features of the universe were imprinted back then. A few years ago astronomers had a big surprise. It doesn't change anything much I've said about cosmic history but it does alter our perspective on the far future. They found that the expansion of our universe instead of slowing down as you would expect because everything turns to gravitation pull and everything else it was actually speeding up. And this was famously found by looking at the Hubble diagram at distant supernovae treating them with standard candles where you can of course compare the Hubble expansion rate back in the past when you see the distant supernovae with the expansion rate now. And we found that the expansion rate was accelerating now to everyone's surprise implying that the gravitational attraction that was slowing it down was overwhelmed on the cosmic scale by some mysterious new force laden to empty space which pushed things apart. Independent evidence that there was something extra came in a quite different direction. When one looks at the microwave background fluctuations we understand enough about the physics of the early universe when the radiation matter was coupled to know there will be sort of sound waves. And we can calculate there's a particular wavelength where the sound waves will have maximum amplitude and that's just a rigid rod a particular length scale which we can calculate for the early universe. And then we can look at the W map data and see where these fluctuations are maximal and see what angle of scale that's on. And this picture the red is what you would expect in a flat universe and as a function of scale and the so called Doppler peak is a high peak there. And the point I want to mention is that the microwave background peak is on the scale you expect if you have a flat universe. But if we had a universe of 0.3 and nothing else in it then we would get something like the yellow which is quite which is way outside the error bars. So in effect what we've got is we've got a very narrow triangle looking along to this wavelength and the angles of that narrow triangle exactly 180 degrees when you add them up and in that sense the universe is flat. Well this also told us that there was not just the matter we see and so we have in my view this network of convincing arguments. If it was just a supernova acceleration I wouldn't have believed this but there's a separate argument not just the Hubble diorama supernovae. Knowing that the Irving and Dart matters 0.3 and knowing that we're in a flat universe with Doppler peak that says that 70% of the mass energy of the universe must be in something which is unclustered. And something which is unclustered and it must be something which is less important in the past than now because it was equally important in the past it would have stopped the growth of galaxies. So it's something which therefore must have a negative pressure so it is less important in the past. So we could infer that 70% of the mass energy universe was in something with negative pressure and in Einstein's equations it's rho plus 3P over C squared which determines gravity and so we could have predicted an acceleration. So even without the supernovae we could have predicted the acceleration. And I think it's that network of arguments which together is compelling even if the supernovae themselves might not be. Well another interesting question. How extensive is what we might call physical reality? The part that we can talk about in science. Well we can see out to 10 or 15 billion light years the very distant galaxies around us. But that limit is essentially because there's a horizon. The shell around us delineating the distance that light can have traveled since the Big Bang. But that shell around us has no more physical significance than the circle which delineates your horizon if you're in the middle of the ocean. If you're in the ocean you don't think that the ocean necessarily ends just beyond your horizon although you can't be sure. And so we would suspect that there are many more galaxies beyond our horizon. Indeed you can quantify that because the micro background temperature if you look as far as you can in that direction and that direction doesn't differ by more than one part in 10th to the 5th. So that suggests if we're in some huge finite structure the gradient across is very gentle and it probably goes thousands of times further in overall extent than our visible horizon. And maybe far further still than that. Maybe even so far that all combinatorial possibilities are repeated. So far beyond the horizon we could all have our avatars. Indeed there could be a replica of our entire Hubble volume if we went on far enough. Well be that as it may even conservative astronomers are confident that the volume of space time within range of our telescopes. What astronomers have traditionally called the universe is only a tiny fraction of the aftermath of our big bang. And there's nothing novel about this. Indeed if you'd asked astronomers in the 1980s about that they would have expected that because their favorite model then was the Einstein and the Sitter model which went on with infinite. And then we were seeing only a 5th bit of it. But in that model we'd have predicted that eventually as the universe went on expanding light from those distant galaxies we can't now see would eventually have time to reach us because of the deceleration. But it's now different for the acceleration because galaxies that we can now see will eventually disappear. They cross the horizon like things falling into a black hole. And galaxies that are now beyond our horizon will never come within our horizon. They're never in principle even be observed. So most astronomers now are happy to believe that there are galaxies which we can never ever even in principle observe and they're part of what we would regard as physical reality. But there's more to it than that because as Alan Gooth discussed when he gave his Newton lecture, course of a model for the physics of tensor 16 GeV suggests a model called so-called eternal inflation. According to which our big bang is just one island of space-time in a vast cosmic archipelago. And that illustrates symbolically here, there's our region we can see, the galaxies beyond our horizon, but that whole thing is just one part, one island of space-time among many others. Well this is speculative physics, but this is physics not metaphysics. When the multiverse is mentioned and pictures like this shown, people say well these domains aren't observable so they aren't part of science. But I'd like to contest this by sort of a version therapy, you know what that is, that's when you, if you're scared of spiders, you first are shown a little spider a long way away and you end up with a tarantulas quarrel over you. So I mentioned already that there were galaxies beyond our horizon. We're relaxed about that. And I mentioned that in the accelerating universe there were galaxies beyond our horizon which will forever be beyond our horizon. Most people are now relaxed about that. They're in the aftermath of our Big Bang. But why is their reality status necessarily higher than that of the aftermath of other Big Bangs? If there are other Big Bangs. And of course we don't know if there are. We'll only take the existence of the other Big Bang seriously when and if we have a theory which describes physics at 10 to 16 GeV which can be tested in other ways. You won't observe them but if we have a theory of the Big Bang at 10 to 16 GeV and if it predicts something like this then we'll take that prediction seriously. So again as Ed Witton said in his lecture at his challenge for 21st century physics is to see which branch of this decision tree is correct. Are there many Big Bangs rather than just one and if there are many are they all governed by the same physics or not? Witton didn't think so. He thought there could be a huge number of different vacuum states. Different values of Einstein's lambda with different microphysics. And if Witton's right what we call laws of nature may in this grander perspective be local bylaws governing our cosmic patch. If that's the case many patches could be stillborn or sterile because the laws prevailing in them might not allow any kind of complexity. They might not fulfill the list of eight requirements I showed. We therefore wouldn't expect to find ourselves in a typical universe. We'd be the typical member of the subset where complexity could evolve. And this is what's called anthropic selection. Some people film with the mouth at this idea but I think it would be inevitable if Witton were right. Well I earlier went through a list of some requirements that must be fulfilled if a complex cosmos like ours is to emerge. And it's interesting to explore what range of parameters would allow this. Those who are allergic to multiverses can regard this just as an exercise in counterfactual history. What if it were different? Just like historians speculate on what might have happened to America if the Brits had fought more competently in 1776. And biologists speculate on how a biosphere might have evolved differently if the dinosaurs hadn't been wiped out. But I'll illustrate this style of reasoning with two parameters which I mentioned earlier. The first is lambda, the vacuum energy, which string theorists suspect could span a whole range. We know that if it's too large and positive, if it's overwhelmed gravity before galaxies had formed then we wouldn't be here. If it was negative then the universe would have collapsed too soon. But I want to mention another fundamental number which is not very well explained. It's what I call Q. This is the amplitude of the fluctuations. This is a number which measures how rough the universe is. And Q is the temperature fluctuations observed by W MAP. It's about tens of minus five. And that value determines the amplitude of fluctuations. It determines the scale of nonlinearities in the structure of the universe today. It determines how big clusters are, etc. And we live in the universe where Q is tens of minus five. And no theories determine exactly why Q has that value. The Planck satellites may give new dates that helps to pin it down, but at the moment we don't know. So let's just be counterfactual and ask what would happen if Q were different. If Q were tens of minus three we'd have a much lumpier universe. And then Q's black hole, 20-18 solar masses, would have formed soon after recombination and we'd probably get no stars and no galaxies. If it was tens of minus four, just a factor of ten bigger than in our actual universe, this might be quite an interesting place to live in. Because what would happen then is that galaxies would form early and be much bigger. You would have disk galaxies like ours, but the size of a cluster, tens of 15 solar masses. And this might be quite an interesting universe. We could live there. The only problem is that everything would be rather denser and it might be hard to have a solar system that remained unperturbed by a passing star for long enough for life to evolve. What about a smooth universe than ours? This wouldn't be so good. It's called an anemic universe because their structures take longer to form and the potential wells are shallower, maybe hard for stars to form and supernodes will blow all the gas out again and it might be hard to get second generation stars. So Q isn't pinned down, but we could say roughly that the range tens of minus four, tens of minus five is the best sort of range. And we can then plot a diagram like this. We can say if we don't know about these two parameters, lambda and that, we can say that we could exist in this shaded domain. Q could be in the range I've shown along there and for any value of Q, the lambda must be low enough to have allowed galaxies to form before the cosmic proportion takes over. So if Q was tens of minus four or tens of minus five, then galaxies would have formed ten times earlier so we could tolerate the value of lambda a thousand times higher. So what I would hope in the long run is to have a theory which tells us whether these numbers could be different and would put a measure on it so we would know whether our universe there is a typical member of the habitable subset. We can't do that but that's the kind of thing in principle we might do. Well I started this talk by describing newly discovered planets orbiting other stars. I'd like to end with a flashback to planetary science 400 years ago, even before Newton. This is Kepler's picture of course. At that time Kepler thought that the Earth was unique and it all was a circle related to other planets by beautiful mathematics. We now realize that there are zillions of stars each with planets around them and the Earth's orbit is special only in so far as this is in the range of radii and eccentricities compatible with life. So maybe we're due for a comparable conceptual shift on a far grander scale. We've had the shift on the top about planets and so now we are realizing that our big bang may not be unique any more than planetary systems are. Its parameters may be environmental accidents like the details of the Earth's orbit and the hope for neat explanations in cosmology may be as vain as Kepler's numerological quest. So if there's a multiverse it'll take our Copernican demotion one stage further. Our solar system is one of billions of planetary systems in our galaxy. Our galaxy is one of billions in the observable universe but this entire panorama may be a tiny aftermath of our big bang which itself may be just one among billions of big bangs. It may disappoint some physicists if some key numbers they're trying to explain turn out to be mere environmental contingencies no more fundamental than the parameters of the Earth's orbit around the sun. But in compensation we'd realize that space and time were richly textured but on scales so vast that we're not directly aware of it even when we make astronomical observations. At a conference in Stanford a few years ago there was a panel discussion where the panelists were asked how strongly they'd bet on the multiverse concept. I said that on the scale would you bet your goldfish, your dog or yourself I was about at the dog level. Andrei Linde the inventor of eternal inflation he said he was far more confident he spent 25 years working on eternal inflation. And Stephen Weinberg later said he'd happily bet Martin Reese's dog and Andrei Linde's life. But all that is speculative. Finally let me say a word about the far future. Long range forecasts are never reliable and the biggest uncertainty is about the nature of empty space the vacuum energy. Is it really something which is constant like Einstein's lambda or is it going to change in the far future? Some people think that space may be destined for a phase transition where lambda decreases and maybe changes sign. If that's the case we could end up with a big crunch where everything collapses. On the other hand some people although this is a minority view think that the repulsive force may get stronger. And then we have what's called a big rip where eventually planets get torn apart and then everything else ends up by being disrupted by this ever-growing force. But these are both rather unlikely and the best and most conservative bet is that this force lambda is unchanging. It is just Einstein's lambda in the modern guise. And if that's so we have almost an eternity ahead with an ever colder and ever emptier cosmos. Galaxies will accelerate away and disappear over an event horizon rather like the inside of a black hole. And all that's left for far future astronomers will be the remnants of our own galaxy andromeda and smaller neighbors. Protons may decay, dark matter particles anionate, occasional flashes when black holes evaporate and then silence. And that's perhaps a good note on which to finish this lecture except once again I'd like to emphasize that the last 10 minutes have been speculative and future progress will depend 95% on better data and better instruments less than 5% on armchair theory. Thank you very much. Congratulations, what a pleasure. Thank you very much. Thank you. So what we've got to do is to actually exhibit this. Yes. Alright. January. January. October. We've got to do worse than that. I hope you've been able to do that. Thank you very much. Congratulations, Martin. Thank you very much. Thank you. Thank you very much. Thank you very much. Thank you very much.