A lifetime is very long relative to the picosecond it takes for two atoms to form a molecule, but it is the blink of an eye compared to many natural phenomena, from the rise of mountain chains to the collisions of galaxies. To answer questions that take more than a lifetime to resolve, scientists hand their efforts down from one generation to the next. In medical science, for example, longitudinal studies often follow subjects well after the original researchers have passed; some studies that are still ongoing started as far back as the 1920s. The record for the most extensive sequence of uninterrupted data gathering in history may belong to the ancient Babylonians' Astronomical Diaries, which contain at least six centuries' worth of observations from the first millennium B.C.; those records have revealed recurring patterns in such events as solar and lunar eclipses.

In most fields of scientific research, however, some of the most interesting and fundamental questions remain open because scientists simply have not had enough time to pursue them. But what if time were no object? I recently spoke with leading researchers in various fields about the problems they would attack if they had 1,000 years—or 10,000 or even a million—to make observations or perform experiments. (To keep the focus on the science rather than on futurology, I asked them to assume they could use only technology that is state of the art today.) Condensed versions of their intriguing replies follow.


Robert Hazen, earth scientist at George Mason University

In the early 1950s Stanley Miller and Harold Urey of the University of Chicago famously showed that some basic building blocks of life, such as amino acids, form spontaneously given the right conditions. It seemed that solving the mystery of the origin of life could be just a matter of combining the right chemicals and waiting long enough. It has not turned out to be that simple, but over 10,000 years or so a modern version of the Urey-Miller experiment might yield some rudimentary self-replicating molecule able to evolve through natural selection—in short, life.

An experiment to simulate the origin of life has to take place in a geochemically plausible environment and start from scratch. The primordial soup may have contained millions of different kinds of small molecules, which could combine and react in an astronomical number of possible ways. In the ocean, though, they would have been so diluted that the chances of any two molecules running into each other, much less reacting chemically, were very low. The most plausible explanation is that self-replicating molecules first assembled on the surface of rocks. The wet surfaces of primordial Earth would have constituted a vast natural laboratory, running perhaps 1030 little experiments at any one time, over a period of maybe 100 million to 500 million years.

A 10,000-year laboratory effort could attempt to re-create this situation by running huge numbers of tiny experiments simultaneously. These molecular nurseries would look from the outside like rooms filled with racks of computer servers, but inside there would be chemical “labs-on-chips” containing hundreds of microscopic wells, each with different combinations of compounds reacting on a variety of mineral surfaces. The chips would constantly and autonomously monitor the reactions to check for signs that a molecule had gone into runaway self-replication.

Experimenters could cut down the time needed from millions to thousands of years by focusing on combinations of chemicals that are most likely do something interesting. With luck, eventually we will learn enough about how nature works to trim this time down to a few decades.


Gerald Gabrielse, physicist at Harvard University

The basic laws of physics appear to be universal and eternal: so far as we know, all protons have the same amount of electrostatic charge, light always travels at the same speed, and so on. Yet certain proposed models of reality allow for variations, and some astronomical studies have claimed, controversially, to have seen small changes. Meanwhile all laboratory data have held steady. My lab, for instance, has measured the strength of the electron's magnetism—the most precise measurement, to my knowledge, of any property of a fundamental particle. If repeated for thousands of years, such an experiment might see a shift.

To measure the electron's magnetism or, more precisely, its “magnetic moment”—the subatomic analogue of a bar magnet's strength—we confine a single electron to a plane with an electrostatic field and use a magnetic field to force the electron to move in circles. We keep our apparatus at less than a tenth of a degree above absolute zero so that the electron's motion is in its state of lowest possible energy. With radio-frequency waves, we then force the electron's magnet to flip. The particle's response and, in particular, the rates at which we can make it flip depend on its magnetic moment, which we can then determine to three parts in 1013.

If the magnetic moment had changed by one part in 1,000 over the entire history of the universe and if the change had gone on at a constant pace all along, our experiment would have already detected it. Of course, science can never prove that something is exactly constant, only that its rate of change is extremely small. Moreover the rate of change could be much slower now than it was in the early universe, making it difficult to spot in the lab. But if we repeated our experiment over 10,000 years and saw no change, that stability would place stringent constraints on any theoretical predictions of changing constants. (It would also cast doubt on assertions that experimental observations of light from distant quasars have detected slight changes in the strength of the electromagnetic interaction since the early moments of the universe.)

Naturally, our techniques and those of other labs are certain to improve. I suspect that increasingly clever methods will enable us to make more progress in far less time than 10,000 years.


Thorne Lay, seismologist at the University of California, Santa Cruz

The magnitude 9.0 Tohoku-Oki earthquake and tsunami that devastated northeastern Japan in March 2011 took the seismology community by surprise: almost no one thought the responsible fault could release so much energy in one event. We can reconstruct the history of seismic activity indirectly by inspecting the local geology, but this can never fully substitute for direct detection. Modern seismographs have been around for only slightly more than a century, too short a time to give a clear idea of the largest quakes that might strike a certain area every few centuries or more. If we could let these instruments run for thousands of years, however, we could map seismic risk much more accurately—including specifying which regions are capable of magnitude 9.0 even though they have not seen more than magnitude 8.0 in recorded history.

Multimillennial records would also answer another riddle: Do megaquakes—by which I mean tremors of magnitude 8.5 or greater—come in worldwide clusters? Records of the past 100 years or so suggest that they might: six of them occurred in the past decade, for instance, and none in the three preceding decades. Measurements over a longer period would tell us if this clustering involves physical interaction or is just a statistical fluke.

How smart can they get?

“If I evolved chimps or some other nonhuman primate toward greater cognitive abilities, how far would they go?”

Bruce Lahn, geneticist at the University of Chicago

Will we evolve to resist major diseases?

“Humans' diet keeps changing, causing new scourges such as the diabetes epidemic. Over tens of thousands of years, will our bodies adapt?”

Sarah Tishkoff, human geneticist at the University of Pennsylvania


Cole Miller, astronomer at the University of Maryland

Supernovae are rare, occurring perhaps once every several decades in a large spiral galaxy such as ours. The last time one was seen here was A.D. 1604: Johannes Kepler described it as outshining everything in the night sky but Venus. All supernovae recorded in more recent times took place in other galaxies that are millions, if not billions, of light-years away. When we finally see a supernova up close, we will be able to study it not only with ordinary telescopes but also with two new kinds of observatories—one detecting neutrinos and the other, gravitational waves—which will tell us what actually goes on inside the exploding star. If you could wait 10,000 years, you would be virtually guaranteed to get 100 or 200 of these events—enough to distinguish their subtle variations.

The explosion of a star could happen in our galaxy at any time. When it starts, the screens of computers at a handful of gravitational-wave observatories around the world will begin to flash, signaling the passage of ripples in the fabric of space. These so-called gravitational waves are a key prediction of Einstein's general theory of relativity but have so far eluded direct detection. The waves will signal that the star's core has begun to collapse under its own gravitational pull. The compressed matter turns into neutrons and releases neutrinos—particles that can zip through matter and thus escape through the star's outer layers and into space (and reach observatories on Earth). The energy released by the collapse, mostly carried by neutrinos, could blow off the outer layers of the star, making it stupendously bright. In some cases, however, the shock wave might fizzle, yielding gravitational waves but no light. We do not know for sure, because so far we have only seen the final, visible stage (with the exception of a handful of neutrinos from a supernova in 1987). Having thousands of years to observe would make all the difference. The new tools could also let us solve another open question—namely, in what conditions a dying star leaves behind a black hole or a neutron star.


Kristin Persson, theoretical physicist and materials scientist at Lawrence Berkeley National Laboratory

We build things all the time, but how do we know how long they will last? If we are going to build storage for nuclear waste, we need to be sure that the containers will last until the material inside is no longer dangerous. And if we are not going to fill the planet up with trash, it would help to know how much time it takes plastics and other materials to degrade.

The only way to be sure is to put these materials under stress tests for 100,000 years or so and see how they hold up. Then we could learn to build things that truly last—or that degrade in a “green” way.

We could, for example, test such materials as the copper-based alloys and glasses typically used for encasing nuclear waste. (Repositories are supposed to go deep underground in carefully chosen locations. But geologic conditions may change in unpredictable ways within a few thousand years.) Such experiments would expose the materials to accelerated wear and tear and to chemical abuse—say, varying pH. They would dial temperature up and down to simulate the cycles of day and night and of the seasons.

Even materials that seem to be impervious to the harshest conditions over scales of years may actually be degrading in subtle ways: our characterization methods are just not good enough to see whether you have lost a few atoms here and there. Yet over many thousands of years the damage could start to show, letting us know which sorts of materials are best.

Long-term testing would be tremendously helpful for other technological applications as well. Current laboratory and simulation techniques, for instance, cannot predict with confidence how the battery of a new electric car will perform over the next 15 years. Eventually computer simulations may become sophisticated enough to substitute for long-term experiments. In the meantime, though, we need to exercise extra caution when building things that need to last.

Will we eventually wage endless local wars?

“If in a few centuries we run out of cheap fossil fuels and cannot find a replacement, our societies will return from global to local. Will we relapse to tribalism and to endless small wars?”

Laurence Smith, geographer at the University of California, Los Angeles


Jerry Coyne, evolutionary biologist at the University of Chicago

Most new species in nature appear when a population becomes geographically isolated from other populations. It then adapts to the local environment and, sooner or later, acquires traits that prevent it from successfully mating with the original species or that would make the resulting offspring sterile, or both. The great open question of evolutionary biology is, Which of these two types of reproductive barriers tends to arise first—those that make crossbreeding difficult or those that lead to nonviable offspring?

Speciation occurs over geologic timescales. Thus, although we can see evidence of it in the fossil record or in DNA, we would have to wait a million years or more to see it reach completion. (Much faster routes to speciation have been documented that do not require geographic separation, but they are the exception rather than the norm.) But if we had, say, 100,000 years, we should be able to reproduce it in the laboratory.

The trick would be to work with an organism that produces new generations quickly, such as Drosophila (fruit flies). Researchers would isolate two or more populations in the lab and expose them to different diets and other conditions. You would then need to periodically test each population for genetic mutations and for changes in its anatomy, physiology, and behavior and once in a while have members of different populations meet to see what happens.

In special cases, my collaborators and I have been able to understand reproductive barriers indirectly by looking at many closely related species at different stages of evolutionary divergence. For geographically separated species of Drosophila, we found that the two types of barriers—mating problems and sterile offspring—evolve at about the same rate. But for species cohabiting the same area, interbreeding barriers seem to evolve quicker. It is not clear, however, whether such results apply to all groups of organisms.

To obtain a new species much faster—perhaps in as little as 100 years—you could beef up the selection pressures to be far stronger than they would normally be in nature. In a landmark experiment in the 1980s researchers bred populations of fruit flies to adapt to different environments—as well as to prefer mating with individuals that shared their habitat preferences—in just 25 generations. Yet the conditions in that experiment were artificial, and it is doubtful whether the two populations produced could be regarded as different species. A very long experiment could be much more definitive.


Glenn Starkman, physicist at Case Western Reserve University

The heat of the big bang left behind radiation that has permeated the universe ever since. Space probes have mapped this cosmic microwave background, or CMB, over the entire sky and found it to be extraordinarily uniform save for small, random fluctuations, just as big bang theory had predicted. Such smoothness implies that the early universe was itself uniform. Yet some analyses, including those by my collaborators and me, saw an excess of symmetry between opposite sides of the sky and other anomalies, including a lack of the largest fluctuations, those that should span more than 60 degrees in the firmament.

To find out if these are real features or statistical flukes, we just need to keep observing. The CMB picture we see today is an accident of our place in space and time. The CMB has traveled to us from all directions for 13.7 billion years. Surveying it thus means mapping a spherical surface that surrounds us and has a radius of 13.7 billion light-years—the distance light has traveled in this time. If we wait long enough, the sphere will get bigger and bigger and thus cross new regions of the early universe. The anomalies are so large that it may take a billion years for the CMB sphere to get past them—when the sphere's radius would reach 14.7 billion light-years. If we could wait “just” one million years, most of the anomalies should be still there but slightly changed. By then, we would be able to see if they were on their way to disappearing—suggesting that they are flukes—or if their persistence reveals the presence of larger cosmic structures.

Will our heads get bigger?

“The narrowness of the human birth canal is a major bottleneck on the size of our heads. Will our use of C-sections, continued for hundreds of thousands of years, lead us to evolve larger brains?”

Katerina Harvati, paleoanthropologist at the University of Tübingen in Germany

How will giving birth at later ages change our biology?

“People are having children at an older age, when mutation rates in sperm are higher and the style of parenting is different. After tens of thousands of years, could these cultural changes affect our biology?”

Marcus Feldman, mathematical biologist at Stanford University


Sean M. Carroll, theoretical physicist at the California Institute of Technology

The universe's ordinary matter consists, for the most part, of protons—particles that have been around since the big bang. Whereas other subatomic particles, including neutrons, can spontaneously decay, protons appear to be exceptionally stable. Yet some grand unified theories, or GUTs—attempts to reinterpret all of particle physics as different facets of a single force—predict that protons should break down, too, with average life spans of up to 1043 years, depending on the theory. If we wait long enough, though, could we finally see it happen?

To see the proton decay, all you have to do is fill a large underground tank with water and monitor it for little flashes of light that would go off as the protons in the water's atoms finally died. The more protons you monitor, the higher the chance that you will see one decay. Studies done with existing detectors show that protons last at least 1034 years, values that have already ruled out numerous GUTs. To have the final word, these detectors might need to run for 100 million years. But if we built detectors 100 times larger—making them about the size of a professional football stadium, voluminous enough to hold five million tons of water—just one million years should do. Unifying particle physics might be worth the wait.