Even as the Schopf–Brasier battle raged, a small cadre of less publicized researchers labored to craft a convincing case for fossils even more ancient than Apex. This new breed of paleontologist doesn’t depend on questionable black blobs. They probe rocks for fossils far smaller than microscopic cell-like spheres or segmented filaments. Remarkably, the fossils they seek consist of the very atoms and molecules of once-living organisms.

When a cell dies, its vital chemical structures quickly fragment and decay. Almost always the essential atoms of biochemistry—carbon, hydrogen, oxygen, nitrogen, and more—disperse and return to the environment. Earth’s vast but nevertheless finite reservoirs of life-sustaining atoms play their parts over and over and over again. Most of the atoms in your body were once part of mastodons, dinosaurs, trilobites, even the earliest living cells. Take a moment to look at the palm of your hand and imagine the fantastic yet unknowable histories of its countless trillions of atoms. Earth’s biosphere is the ultimate recycling machine.

Atoms almost always recycle, but once in a great while, under an unusual concatenation of geological circumstances, a dying organism will find itself encased in an impermeable rock tomb. If a worm is

swept away and buried in a sudden mudslide, if a colony of deep-sea microbes solidifies in chert, if a winged insect dies ensnared in sticky tree sap, then it’s just possible that some of the organism’s original atoms and molecules will become trapped as well. Such a trapped fossil animal or microbe may persist through eons in its original form, or it may decay to a shapeless dark splotch. Nevertheless, its hermetically sealed atoms and molecules are the remains of past life, so they qualify as fossils just as legitimate as the most elegant coiled ammonite or massive dinosaur.

It’s an amazing feeling to hold a 3-billion-year-old rock that once teemed with living organisms—a sample that contains the very atoms and molecules of cells from the dawn of life. Such rare and precious samples demand a new approach to the study of fossils; traditional descriptive paleontology must morph into analytical chemistry.

A casual conversation during the summer of 1997 with longtime friend Andrew Knoll, professor of paleontology at Harvard University, led me into this fascinating field. Andy and I were attending a Gordon Research Conference on the origin of life, held at New England College in Henniker, New Hampshire. He’s an engaging, articulate, and friendly speaker, and the author of richly illustrated articles and lectures on the diversity of microbial fossils in Earth’s oldest rocks—presentations that opened a new world to me.

When most people hear the word “fossil,” they think of the bones of a dagger-toothed Tyrannosaurus rex or the spiny shell of a trilobite—hard parts that survive the rigors of decay and burial. By contrast, the soft cellular tissues of animals, plants, and microbes almost always rot away without a trace. Only occasionally will an organism die and be buried in rock fast enough to preserve cellular detail. For a micropaleontologist like Andy Knoll, whose specialty is ancient microbes, those rare cellular fossils provide the raw material for a career in science.

The very earliest fossil cells are nondescript objects and difficult to identify, but geologists have documented dozens of localities with numerous clearly identifiable microfossils dating from about 2.9 billion years on. Distinctive bumpy rods and symmetrically spiky spheres, chainlike filaments of repeated rectangles, and curious corkscrew spi-

rals form a panorama of primitive life. Andy’s work surveys the saga of life’s evolution, culminating in the first enigmatic multicellular organisms about a billion years ago.

Throughout our conversations at the Gordon conference, I was struck by the fact that many ancient microscopic fossil forms are preserved in black chert or shale—impermeable rocks that have the potential to preserve chemical traces of the original bacteria. Over a beer, I asked Andy if paleontologists ever analyzed their microfossils with the kind of machines that we mineralogists routinely employed to characterize the atoms and isotopes of our samples. He shook his head and admitted that, while there had been a few pioneering studies, most paleontologists worried almost exclusively about the sizes and shapes, not the chemistry, of their bugs. Then came his deceptively innocent question: “Do you want to collaborate? I’ve got a couple of students with really interesting samples….”

My own research on life’s emergence had to that point focused on bottom-up chemical experiments, trying to synthesize life’s molecular building blocks, but the top-down approach also has great appeal. I’ve always loved fossils and was more than happy to associate myself with a real paleontological pro, albeit in a modest support capacity. Agreeing to the offer, I immediately envisioned an arsenal of microanalytical tools that might be brought to bear on the problem. Our conversation soon turned to technical details: the number of samples, their size, the degree of chemical alteration, and more.

The first samples arrived at Carnegie’s Geophysical Laboratory from Harvard within a few months, and many more followed. A suite of 400-million-year-old plants from Canada, slices of ancient black soils from Australia, 3-billion-year-old microbial mats from South Africa, bizarre spiky spores from a billion-year-old Chinese formation—wonderful fossils holding some of the secrets of life’s past. Our Carnegie team quickly confirmed that ancient fossils have the potential to provide three important types of microanalytical data: chemical elements, isotopes, and molecules. Of the three, the composition of chemical elements is arguably the easiest to measure.

The mineralogist’s tool of choice for analyzing chemical elements is the electron microprobe, a costly but indispensable piece of hardware in geology departments around the world. The machine works by firing a narrowly focused beam of electrons at a highly polished piece of rock, typically an inch or so across. The energetic electron beam

excites the rock’s atoms, which in turn emit a spray of X-rays. It turns out that every element of the periodic table produces its own slightly different suite of X-rays of different wavelengths. The task, then, is to capture these X-rays and measure their diagnostic wavelengths.

Microprobe analysis of most elements has become routine and automated. The machine is a workhorse, operating 24 hours a day for analyses of silicon, magnesium, iron, and other rock-forming elements. But the lightest elements, including the one of greatest interest to us—the key biochemical element carbon—pose a severe analytical challenge. Lighter elements tend to be rather inefficient at producing X-rays, while the relatively few X-rays that are produced have rather low energies. Both of these factors complicate carbon analysis. What’s more, we routinely use carbon to coat rock samples prior to probing, in order to make them electrically conductive. The coat is essential to prevent the sample from building up an electric charge while being bombarded by electrons, but it can mask any carbon in our samples. In short, the Lab’s usual microprobe procedures wouldn’t work. We’d have to devise new protocols.

The Geophysical Lab’s deviser of protocols is Christos Hadidiacos, microprobe jockey extraordinaire. For more than 30 years, Chris has maintained and upgraded the Lab’s electron microprobe. He knows every trick in the book and constantly invents new ones to push the limits of analysis. Complex circuit diagrams, cryptic numerical tables, and other papers decorate his office, while technical manuals and electronics catalogs fill his bookcases and rows of volleyball trophies line the shelves above.

I explained the carbon analytical problem to him—one he had never faced before. The Geophysical Laboratory probe had been used almost constantly for decades but almost always to study rocks and experimental run products—samples with mostly heavier elements. But it took Chris less than a minute to figure out a possible fix. “We can try raising the current,” he said. “That might work.” He began asking detailed questions about the expected amount of carbon and what other elements might be present. Then a frown. “We’re not going to be able to use a carbon coat, are we?”

I shook my head, knowing that this was a potential deal killer. We needed an electrically conductive coating, and that meant a metal. But a carbon coat would mask our fossils, and most metals absorb X-rays

so efficiently that we’d never see the low-energy carbon X-ray signal. “Any ideas?” I asked.

Again, it took him less than a minute. “Maybe we could use aluminum. Just vaporize a bit of aluminum foil.” He was smiling again, pleased at the simplicity of his solution. Aluminum, element 15, should be light enough itself to allow most of the carbon signal through. It was definitely worth a try.

I handed over the first of my fossil specimens, a pair of 2 × 3-inch rectangular thin sections of 400-million-year-old plant fossils from Rhynie, Scotland, a classic chert locality. The samples had been collected many years ago as isolated flinty boulders in old stone walls; no rock outcrop has ever been found. The precious thin sections arrived courtesy of Kevin Boyce, a bright-eyed, soft-spoken grad student in Andy Knoll’s group. Kevin had trolled through Harvard’s somewhat neglected paleobotany collection as part of his thesis work on the evolution of leaves. He hoped that the Rhynie samples, which preserve cellular structures of some of the oldest known land plants, might reveal clues about the chemical evolution of plants.

The first step was to apply the thin aluminum coating. Our antiquated but serviceable vacuum coating system consists of a well-worn metal housing about the size of a washing machine with vacuum pumps and hoses arranged inside. A 4-inch-square platform with wire electrodes sits on top, while vacuum gauges and control valves project from the side.

Chris snipped a 1-centimeter-square piece of aluminum foil, crumpled it up and placed it in a small wire basket attached to the electrodes. He arranged the two Rhynie chert thin sections on the metal platform, and then lowered a 2-foot-tall dome-topped bell jar so that its rubber-lined base made an airtight seal. The old pumping system labored for a quarter of an hour to achieve the desired vacuum, but then it took only a fraction of a second to apply an electric current and vaporize the aluminum foil. Aluminum atoms flew off in all directions, coating everything inside the bell jar, including our samples. Chris released the vacuum, raised the bell jar, and the fossils were ready for the probe.

The Geophysical Lab electron microprobe is an awkward-looking tabletop machine that sits in its own small room. [Plate 3] Two chairs flank the workbench, which is dominated by a 4-foot-high cylindrical

tower that looks like a model of some futuristic fortification. The tower houses the electron gun, the heart of the probe. At the tower’s top, a coiled tungsten filament generates electrons; a series of ring-shaped electromagnets focus the electrons into a narrow beam as they accelerate downward onto the sample. The base of the tower is cluttered with five boxlike attachments, called spectrometers, that measure X-rays, plus various vacuum lines, power cables, and viewing ports.

A curious combination of instrument panels controls the electron gun hardware and X-ray detectors. To the right, a computer monitor displays all the machine’s vital statistics—beam current, spectrometer settings, sample position, and more. In sharp contrast, a 1980s-vintage slant-front console, sporting two antiquated 6-inch black-and-white video screens and more than a dozen plastic knobs reminiscent of a classic Star Trek set, dominates the central table. Like an old house that has undergone decades of renovation, the Geophysical Lab probe has been through a lot of upgrades.

I carefully secured one of the thin fossil sections into a shiny metal sample holder, closed the sample port, and waited a couple of minutes for the machine to achieve the high vacuum necessary to stabilize the electron beam. Meanwhile, Chris fiddled with the computer controls, raising the electron current to about ten times its normal settings. It took him a few moments to center and focus the intense beam onto a carbon-rich portion of our sample. We were about to discover whether or not we could detect fossil carbon atoms.

It worked! A carbon signal of 600 X-ray counts per second stood out sharply from the 30-count-per-second background. We were in business. It was a simple matter to select half a dozen areas, each about a fiftieth of an inch square, to map. Slowly but surely, the microprobe beam scanned across the sample, measuring the carbon concentration point by point. It took about 3 hours to produce one map. We put the probe on automatic, happy with our rapid progress, and headed outside to the Lab’s sand court for our afternoon game of volleyball.

The analytical procedures took a bit of tweaking. The initial aluminum coatings were too thick, the beam settings not quite optimal. But within a few weeks, we were producing a steady stream of colorful maps; regions rich in carbon atoms from ancient life-forms stood out boldly against the carbon-poor fossil matrix. Cellular features less than a ten-thousandth of an inch across were clearly visible. Armed with

these maps, Kevin Boyce, with his sophisticated botanical eye, was able to describe and interpret cellular detail never previously seen. [Plate 3]

Making these carbon maps, watching the fine details emerge, is great fun. Each map is formed from a two-dimensional array of point analyses, just like the pixels on your computer screen. We typically employ a quick 400 × 400-point array for reconnaissance, while slower 500 × 500-point arrays yield beautifully detailed maps with colors representing the concentration of carbon—red for the highest carbon content, followed by orange, yellow, and the other spectral colors. We play with map colors like a high-tech video game to heighten the contrast and highlight features of special interest.

Maps of the distribution of fossil carbon atoms can be dramatic and surprising as well as beautiful, revealing subtle cellular-scale details not previously recognized. In a sense, though, this analytical effort is little more than an extension of past morphological studies of fossil size and shape—a slightly more elaborate way to image the specimens. These carbon-rich fossils preserve far more information than just the chemical elements that make them up. That’s why, for more than three decades, geologists have examined ancient life for fossil isotopes.

The fascinating discipline of atomic-scale paleontology has blossomed primarily because all living cells perform a wonderful repertoire of distinctive chemical tricks. Life transforms any collection of its constituent atoms in subtle and surprising ways. Carbon atoms, for example, come in two common varieties—isotopes dubbed carbon-12 and carbon-13. Every carbon atom has exactly six massive positively charged particles called protons in its nucleus; that atomic number, 6, is the chemical definition of “carbon.” The distinction between the two common carbon isotopes lies in the number of neutrons, a second kind of massive particle that also resides in the atomic nucleus. Carbon-12 has six neutrons, while carbon-13 has seven.

The number of neutrons has no bearing whatsoever on carbon’s chemical behavior. You could live equally well on a pure carbon-12 diet or a pure carbon-13 diet. But there is one important physical difference: Carbon-13, with its extra neutron, is about 8 percent more massive than carbon-12. (An average-sized person whose cells were

made entirely with carbon-13 atoms would weigh about 2 pounds more than the same person made entirely with carbon-12 atoms.) As a consequence of this small mass difference, carbon-13 atoms are also a little more sluggish than carbon-12 atoms when taking part in some chemical reactions. So when living cells process carbon-bearing food, they become slightly enriched in the lighter isotope, carbon-12. That characteristic isotopic signature of life can be preserved for billions of years in rock.

Analytical studies of countless carbon-bearing rocks reveal a sharp dichotomy. Most of Earth’s carbon is locked into mineral deposits, notably the abundant carbonates that adorn the landscape with bold limestone cliffs and dissolve to open sublime limestone caverns. Worldwide, this mineral-locked carbon has a well-defined uniform ratio of carbon-12 to carbon-13 of about 99:1—the standard reference value, which is designated as 0. By contrast, living cells are invariably isotopically lighter, with a higher proportion of carbon-12 than in limestone. This difference between limestone and life arises from chemical reactions in cells, which more readily incorporate the lighter carbon isotope.

On the geochemist’s peculiar scale, a 1 percent deficiency of carbon-13 relative to standard limestone is called “−10 per mil,” a 2 percent drop “−20 per mil,” and so on. Such “light” carbon in a rock sample thus carries a negative number value and with it a strong presumption of biological activity. Thousands of ancient fossil specimens, from mammoth bones (about −21) to the 3.1-billion-year-old microbial mats in South African sandstones and the fossils in Western Australian chert (between −25 and −27), bear out this simple relationship. So does fossil coal, the transformed remains of 300-million-year-old swamp life, which typically ranges from −24 to −25. Carbon isotope studies of soft-bodied fossils from the Burgess Shale, a 540-million-year-old British Columbian locality, made famous by Stephen Jay Gould’s Wonderful Life, display a similarly narrow range of values between −25 and −27. The conclusion: If a rock holds an ancient inventory of carbon atoms from once-living cells, then the carbon invariably will be light, even if all morphological signs of life are gone.

The most ancient microbial samples, which often consist of black, carbon-rich splotches in limestone, shale, or other sedimentary rock, have received special attention in recent years. Dozens of studies on billion-year-old rocks from Africa, Australia, Europe, and North

America reveal consistently negative carbon isotope values, but they also point to a significant scatter among the dozens of known microbial fossils more than a billion years old. Photosynthetic microbes, which live on sunlight, tend to lie in the −20 to −30 range. These organisms have dominated the fossil record since about 2 billion years ago, when Earth’s atmosphere became oxygen-rich. However, many types of more primitive single-celled organisms that live off the Earth’s chemical energy are much more efficient at concentrating the light carbon isotope, carbon-12. Values as low as −50 have been found in 3.8-billion-year-old sediments. While these differences help paleontologists interpret the varied lifestyles of ancient microbes, all unambiguous cellular fossils contain some proportion of light carbon.

It’s amazing how nerve-wracking waiting for a machine to produce a single isotope value can be. I was given my most memorable carbon-rich sample in the summer of 2002, during a lecture tour to Australia. A side trip to Sydney’s northwestern suburbs brought me to the campus of Macquarie University. There, in beautiful green landscaped grounds, is the home of the Australian Centre for Astrobiology, whose director is paleontologist Malcolm Walter.

Walter’s work on ancient Australian microfossils was well known to me, and it was a delight to meet him. He welcomed me with a strong handshake, keen eyes, and ready smile and gave me a quick tour of the facility, a tweed jacket his only protection from the mild Sydney winter. My research was unknown to him, though he was well aware of the well-funded NASA Astrobiology Institute, of which I was part. He listened as I described our various research projects, but his ears really perked up when I recounted my collaborations at Carnegie with Andy Knoll and his students. I outlined our procedures for mapping carbon atoms and described some recent carbon-isotope results.

“You might be interested in this,” he said, handing me a small cloth bag containing several black rock fragments. “It’s Strelley Pool Chert, a new find from Trendall in Western Australia. It’s almost as old as the Apex.”

The chance to hold, much less study, one of Earth’s oldest rocks is a rare privilege. My response was rather pointed and less than subtle: “We’d be happy to do the carbon work, if you’d like. I could do it next week when I get back.” Few laboratories in Australia had the facilities to analyze the ancient rock, while Carnegie was set up and ready to go.

A brief cloud of concern seemed to pass across Walter’s face, but

he hesitated only a moment before extracting a thumbnail-sized sample, a tiny fraction of the valuable hoard. “Perhaps you could have a look at this,” he said, and told me that he was eager to find out as soon as possible whether or not the rock’s carbon was light. I slipped it into my pocket, hardly believing my good fortune in having acquired a piece of Earth’s earliest history. The sample would be a top priority on my return, I assured him, and any data would be his to announce or publish as he wished. Our conversation shifted to less scientific matters: the constant stress of raising money for the Institute, and the glories of his sheep farm in the countryside, where he spends his weekends.

For the rest of that Australian trip, the tiny, 3-ounce sample weighed heavily on my mind, and it rose right to the top of a long “to do” list on my return. When I need a carbon-isotope analysis at the Carnegie Institution, I turn to Marilyn Fogel, a biologist who has amassed an impressive arsenal of analytical hardware. Marilyn and her group study the cycling of elements through ecosystems, and there’s no better way to track an element than with isotopes. Carbon and nitrogen isotopes, both of which get progressively lighter as you move up through the food chain, are her specialty. Fogel’s field areas tend to be exotic: the crocodile-infested mangrove swamps of Belize, the parched outback of Australia, the boiling springs of Yellowstone Park. Her growing scientific reputation and thoughtful mentoring style attract a steady stream of postdocs and visitors.

I grabbed the stone and headed down the hallway mid-morning. After the requisite niceties I got directly to the point: “I met Malcolm Walter in Sydney.” I reached into my pocket and handed her the piece of black chert. “Take a look at this.”

“Apex?” she asked. We all had been following the Schopf controversy, and she knew that Apex Chert samples were pretty hard to come by.

“Nope, this is apparently new. Same area but different. He’d like a carbon-isotope value.”

Marilyn doesn’t generally betray excitement, but she immediately knew the significance. This black fragment was one of the oldest rocks on Earth. “Wow, that’s pretty neat!” She turned the object over in her hands. “Yes, I think the machine is available this afternoon.” She paused. “Why don’t you come by around two.”

I passed the time by breaking off a small piece of the chert, crushing it in a mortar, immersing the powder in oil, and peering at the tiny

glassy shards through a powerful microscope. How curious it was—unlike any rock I’d seen before. Myriad tiny black specks, each a few ten-thousandths of an inch across, clouded the otherwise clear, colorless chert matrix. Unlike typical fossil microbes, which tend to occur in clumps and filaments, these dots were uniformly dispersed. They certainly looked like carbon, but were they cells? Would they show a negative isotopic signature?

At two o’clock I showed up at Marilyn Fogel’s lab, as arranged. Isotope experts rely on mass spectrometry, the experimental technique of choice for measuring a sample’s isotopic ratio. Marilyn’s mass spectrometer for carbon-isotope analyses sits in one corner of a 20 × 40-foot room crowded with scientific hardware. Little space is wasted, and you have to exercise care not to bump into sensitive hardware when squeezing between the various experimental stations. These days, mass spectrometers tend to be highly automated and incredibly precise machines, though they still require meticulous maintenance and rigorous standardization procedures to yield reliable results. With a machine like Fogel’s, carbon-isotope analyses are relatively straightforward. [Plate 4]

The mass spectrometer is a conceptually elegant analytical tool grounded on two of the great physical laws of nature. Newton’s second law of motion, F = ma (force equals mass times acceleration), enables the separation of two atoms of different mass. As noted, carbon-12 and carbon-13 differ by about 8 percent in their mass, so if the two isotopes are subjected to an identical force, then the carbon-12 atom will accelerate about 8 percent faster than the carbon-13 atom. Mass spectrometers accomplish this acceleration by applying a second fundamental law, related to electricity and magnetism: Magnetic fields exert forces on electrically charged particles. The analytical technique is to ionize the carbon atoms: Strip an electron from each—by zapping them with a laser, for example—to yield carbon atoms with a positive electric charge, then subject the ionized atoms to a powerful magnetic field. In some mass spectrometers, a massive horseshoe-shaped magnet bends the stream of carbon atoms (bending is a kind of acceleration, as you discover when you ride a twisting roller coaster). Carbon-12 atoms curve in a tighter arc than carbon-13, so the beam of carbon atoms separates into two. Two detectors placed side-by-side measure the relative amounts of the two isotopes. Alternatively, sensitive electronic detectors measure the time of each ion’s flight:

carbon-12 atoms arrive at a target a fraction of a second before carbon-13 atoms.

Fogel’s mass spectrometer works best with a small powdered sample, so I had crushed and ground a chip of chert to specifications. Rocks are generally a lot easier to prepare than plant and animal tissues, which must be freeze-dried first. In duplicate, I carefully weighed about a milligram of rock powder and tightly wrapped the powder into a tiny ball of inert tin metal foil. Marilyn inserted a series of well-documented carbon-isotope reference standards along with my samples, each a crumpled metal sphere about the size of a BB, into ports of the mass spectrometer’s automated sample holder. Then a computer control system took over and we had to sit back and wait for that one tantalizing number.

It takes only a few minutes per sample, but it seemed longer. The standards always come first, of course; we have to make sure everything is working properly. Finally, the machine spewed out a single printed sheet of white paper, crammed with columns of numbers. One number at the bottom was the key: −25.7 ± 0.5. The carbon was light—just what we’ve learned to expect from ancient microbes! The analysis also indicated that about a tenth of a percent of the chert’s mass was carbon. The duplicate run soon followed: −25.9—satisfyingly consistent results.

But even as we saw these enticing numbers, a nagging doubt remained about the biological origin of the carbon. The oddly uniform distribution of black specks in the chert looked nothing like fossil forms. Indeed, the uniform spacing suggested a more chemical process—a segregation of carbon from chert as oil drops separate from water. Might there be nonbiological pathways to such a light carbon signature? All known life-forms have a negative isotope signature, but does a negative isotope value provide unambiguous proof of life? We were convinced that the Australian chert had a fascinating story to tell.

I e-mailed Malcolm Walters right away with the exciting data and a quick analysis of their possible significance. It would take a bit more work, I thought, but these were certainly publishable results. His reply came slowly, and with a disappointing surprise. Please stop working on the samples, he asked. Evidently, Walter’s Australian colleague, paleontologist Roger Summons, had been promised the chance to analyze the new find. Summons, a pioneer at extracting biomolecules from

old rocks, had recently accepted a professorship at MIT and had already lined up a graduate student to do the work. It wouldn’t do for our efforts to undermine that thesis project.

A deal’s a deal. I put aside the chert and went back to the seemingly endless list of other projects. But it sure was fun while it lasted.

What does a negative carbon-isotope value tell us about an ancient rock? This question came into focus following a surprising announcement in the November 7, 1996, issue of Nature of the discovery of Earth’s most ancient fossil carbon. The Earth’s oldest known rocks, 3.85-billion-year-old banded-iron outcrops from the remote island of Akilia off the southwest coast of Greenland, reveal not the slightest trace of anything that looks like a fossil. Nevertheless, these rocks may contain a modest store of carbon. Even though the rocks have experienced severe alteration through the ravages of temperature, pressure, and time, some of that carbon is encased in the protective mineral apatite. When Scripps Institution of Oceanography geochemist Stephen J. Mojzsis (now at the University of Colorado) and his colleagues collected those rocks and performed the first carbon-isotope analysis at UCLA in 1996, they were delighted to find light carbon, on average a dramatic 3.7 percent lighter than reference limestone. No known abiotic process produces that kind of value. That simple number, −37, was enough to convince many geologists that life had achieved a firm foothold by that ancient date.

Such a result did much more than establish a world record for ancient life. The work of Mojzsis and his colleagues seemed to narrow the window for life’s origin, which presumably couldn’t have emerged until after the last global sterilizing asteroid impact, roughly 4 billion years ago. If signs of life persist in 3.85-billion-year-old rocks, then life arose very quickly indeed.

But, as it turned out, the Akilia rocks posed problems. Earth’s oldest rocks have been through a lot: heated and squeezed and contorted beyond belief. Billions of years inevitably alter the fabric of a rock. Mojzsis interpreted the Akilia formation, with its appearance of intensely folded layers, as metamorphosed oceanic sediments—a perfectly reasonable residence for early cellular life. But when geologists

Christopher Fedo of George Washington University and Martin Whitehouse of the Swedish Museum of Natural History performed a more detailed geological analysis of the carbon-bearing outcrop, the rocks proved to be an ancient molten igneous mass that gradually solidified deep underground from temperatures approaching 1,000°C. The carbon deposits must have formed under extreme metamorphic conditions deep in the crust. Under no circumstance could those rocks have contained life at the time of their formation.

Scientists quickly came up with a range of plausible explanations for the light carbon. Heating experiments, which preferentially release recent organic contaminants, revealed that some of the rocks’ carbon is modern. It’s also possible that some natural nonbiological processes also generate light carbon. Today much of the carbon cycle is regulated by life, and all carbon compounds derived from living organisms are isotopically light. But before the first microbial life, there could have been equally vigorous geochemical processes that separated carbon-12 from carbon-13. If so, then isotopes alone can provide scant help in recognizing life in Earth’s most ancient formations—or from rocks on other worlds, for that matter.

At best, the isotopic evidence from Greenland is ambiguous. And so, in their search for unambiguous proof of ancient life, paleontologists have had to turn to even more elusive fossils—fragments of life’s oldest biomolecules.