Chapter 1: Tales a Child Can Tell

Garniss, lend me your knife for a second, will you," I whispered. Like most geologists, Garniss always carries some kind of implement for prying small rock samples loose when the need unexpectedly arises, as it so often does. Geologists are as inquisitive as kids, poking into everything. In Garniss's case the implement was a small Swiss Army knife, secured on a cord around his neck, together with a small hand lens. Garniss handed over the knife, wondering what I had in mind.

It was a Monday morning in early September of 1992, and we were in a small, hospital-green room at Gadjah Mada University on the Indonesian island of Java. Aside from Garniss and me, there were half a dozen people in the room. These included the eminent Indonesian anthropologist Teuku Jacob. In his late sixties and recently retired as chairman of his department, Jacob nevertheless retained preeminence in Javanese anthropology. Jacob's longtime assistant, Agus Suprijo, was also present, as were Ann Getty, Meg Starr, and Sharie Shute, colleagues of Garniss's and mine. Conversation rippled easily around the room, but with overtones of excitement and an odd reverence.

A plain wooden table occupied the center of the room, supporting an angle-poise lamp, its connection cord hanging free. On the otherwise plain walls hung an abstract painting titled "The Spirit of Mojokerto," incongruous in its modernity. Cushioned on a thick mat of foam rubber on the table rested a collection of dark brown objects, gnarled and cheating easy identification to the untutored eye. They were ancient human fossils -- jaws and teeth, and a skull -- which Jacob had just retrieved from the two refrigerator-sized safes that flanked the armored grille entrance to the room. These fossils, the prized objects of Jacob's collection, are rarely seen, even by professionals in the fossil-hunting business. Scholars with serious research programs have to apply to Jacob for permission even to see them, let alone touch them, for scientific study. And even those few who succeed in obtaining official permission have to wait for Jacob's final OK, for he alone is permitted to remove the fossils from the safes. Even Agus, his assistant, is not given access to the vault without Jacob, and the possibility that Agus might study the fossils on his own is out of the question, despite his credentials as an anatomist. Jacob maintains an assiduous -- some might say obsessive -- protection of the fossils.

Some of the relics arrayed on the table Jacob himself had recovered from the island's ancient sediments; others were found by earlier workers, long ago, going back to the 1930s. Java holds a special place in the annals of anthropology as the source of some of the earliest-discovered fossils of human ancestors. Principal among these is the tiny cranium of the so-called Mojokerto child, who died when he or she was a tender five to six years of age -- an event that, by most estimates at the time of our gathering in the hospital-green room, was a million or so years back in human prehistory.

Discovered some six decades ago, the child's cranium has no face and no jaw; and yet, bereft as it is of any recognizable physiognomy, the small, rounded object still elicits a sense of awe in the observer, a connection with a distant past, not so much forgotten as unknown, perhaps even unknowable. Frustrating though they are in their muteness, relics such as these have drawn anthropologists into the search for human origins for more than a century. Their aim is at once the scholarly pursuit of uncovering our species' evolutionary history and a personal mission of understanding what made us human. Hold a fossil like Mojokerto in your hand, and you have what is at the same time the object of scientific investigation into the byways of human prehistory and a petrified moment of your own ancestry.

Garniss and Jacob, friends and colleagues for more than thirty years, were talking about the visits to various fossil sites that we had all made during the previous week and swapping stories about their earlier experiences in these places. Agus remained quiet, out of Javanese respect for his mentor. Ann and Meg were gingerly handling a distorted fragment of lower jaw, feeling the weight of bone turned to rock. Sharie was taking photographs. I was sitting at the table with the child's cranium in front of me. The fossil bone was a rich brown color, the rock matrix black. I slowly picked up the cranium, turned it over, and scrutinized the topography of the underside, thinking about what I was going to do. It would mean big trouble, but if we wanted to make sense of what we had seen at the Mojokerto site a few days earlier, I was just going to have to risk it.

On our visit to the site of the skull's discovery we had seen white pumiceous rock, and yet the rock matrix infilling the child's skull in front of me was black, at least on the surface. If the matrix was black throughout the cranium, it would hold the awful implication that the skull had not been found at the site where we, and the anthropological world, had been told it was found. We would therefore be wasting our time dating the pumice from the Mojokerto section, and what most anthropologists believed -- that no one really knew where the skull was found -- would turn out to be true. If you do not know where a particular fossil was found, if you don't know which ancient sediments it was buried in, then you cannot know how old it is, because it is from the sediments, not from the fossil itself, that an age can be determined. In that case, no matter how exquisite the fossil, no matter how interesting its anatomy, its value in helping anthropologists piece together the path of human history is much diminished. The child's skull hung in that limbo.

I focused my attention on a large oval lump in the bottom surface of the cranium. Given the cranium's shape, I knew the lump could not be fossilized bone but had to be rock matrix filling the space where the child's brain once was. It would be a good spot.

"Garniss, lend me your knife for a second, will you...?" After Garniss handed it over, I quickly unfolded the blade and started to scrape at the surface of the small raised lump. A fine black dust sprayed out as I worked, moving quickly so that I could get done what I needed to do before the inevitable happened. Just being allowed to see and handle the fossil had been privilege enough. Taking a steel blade to so valuable and valued a relic was sacrilege.

A thundering silence engulfed the room. Seconds passed like hours.

"Garniss," Jacob eventually said, maintaining a tense calm, "come with me. I'd like to talk to you." The two men left the room, and Jacob suggested that they go to his office. It was clear to Garniss, as he told me later, that Jacob was upset by what I had done, but, in a manner that was very typical of Jacob, he spoke politely about various unrelated matters for some five minutes. At last, in a very low-key way, he came to the subject of his concern. "You and I have been friends for many years, Garniss," he said. "And you know that I have to get permission from the committee even to allow you to look at the fossils, let alone do anything with them." Garniss thanked Jacob for his help with the project and for making it possible for us to have access to the fossils.

That was all that was said. Beneath the pleasant exchange, though, there was considerable tension. "I knew that Jacob himself was the committee, that he didn't have to seek permission from anyone," recalls Garniss now. "He didn't mention the knife, and neither did I. But by the way he spoke, and in what he didn't say as much as in what he did, he made it very clear that he thought we had overstepped an important barrier."

It was only when Jacob had left the room to talk with Garniss that I felt free to display my excitement at what I'd seen.

SEPARATE JOURNEY, THE SAME GOAL

We had flown into Yogyakarta a week earlier, on 30 August, after attending the 29th International Geological Congress in Kyoto, Japan, where we participated in a session on geochronology, the science of determining the age of rocks. This branch of geology is extremely important when you are reconstructing Earth history, including the story of human evolution as it was played out in places such as Java. But Java wasn't much in my mind at the conference. In fact, I hadn't given Java much thought at all while planning this trip to the South Pacific, because I had dinosaurs on my mind, not human ancestors.

Ever since the maverick Nobel Prize-winning physicist Luis Alvarez had shaken up the world of paleontology with his suggestion in 1980 that the Age of Dinosaurs was brought to an abrupt end 65 million years ago when a giant asteroid slammed into the Earth, dinosaurs were on the minds of many people whose business it is to find the age of rocks. Alvarez's claim was based on the discovery of unusually high levels of the element iridium, together with the unique ratios of it with the other platinum elements, at the transition between the end of the Age of Dinosaurs and the beginning of the Age of Mammals, the so-called Cretaceous/Tertiary boundary. Iridium is rare in the Earth's continental rock but common in asteroid rock. The humongous explosion generated by the asteroid's impact -- the energy equivalent of a billion nuclear bombs -- would have vaporized the asteroid, created an immense crater, and filled the atmosphere with debris that would have darkened the skies for months. Much of life cannot survive for long in such a world; pretty soon, argued Alvarez and his colleagues, half the world's species succumbed, the dinosaurs being the most prominent victims. Eventually, the dust debris settled, creating a clear chemical signal -- the dust layer rich in iridium -- of a cataclysmic episode in Earth history. Or so Alvarez said.

Profound skepticism is not too strong a phrase to characterize the scientific world's response to the asteroid-impact theory at the time. After all, geologists were deeply wedded to the notion that the Earth's topography on all scales -- from the smallest curve in the course of a stream to the highest peak in a large mountain range -- was formed by the gradual accumulation of small changes. This perspective goes by the term Uniformitarianism, and was developed by a nineteenth-century Scottish geologist and contemporary of Charles Darwin's, Charles Lyell. In developing the uniformitarian view, Lyell overthrew the long-established notion that the Earth had been periodically roiled by great catastrophes, such as global flood. The intellectual brainchild of the French geologist Baron Georges Cuvier, this view, known for obvious reasons as Catastrophism, came to be viewed with deep suspicion because, not unnaturally for its time, the theory had distinctly religious overtones to it. One of Cuvier's proposed thirty or so catastrophes that mark Earth history as written in deep geological strata was said to have been the Noachian Flood. The triumph of Uniformitarianism over Catastrophism in the mid-nineteenth century was therefore seen as the triumph of modern science over anachronistic fancy.

When Alvarez started talking about rocks falling out of the sky as playing a major role in shaping Earth history, he seemed to be trying to drag geology back to the dark ages of Catastrophism. He was, of course, but in a different guise. Almost a decade was to pass before Alvarez's theory met with acceptance, the result of the inexorable accumulation of many different lines of evidence. These days, asteroid impacts are recognized as important in shaping the flow of life during Earth history, possibly bombarding the Earth every 30 million years or so, each time wiping out a large proportion of living creatures in the process. The potential hazard of asteroid impact is now taken so seriously that U.S. government agencies are planning to spend hundreds of millions of dollars to try to detect the next asteroid with Earth's number on it, attempting to keep humans from going the way of the dinosaurs. (Hollywood responded by releasing two major movies on the topic in 1998, Deep Impact and Armageddon.) One piece of evidence about the putative Cretaceous/Tertiary impact remained elusive even after geologists accepted the theory, however: where was the crater of the right size and the right age?

In the spring of 1992, Luis Alvarez's son, Walter, and his student Alessandro Montanari approached me with a small piece of rock that they had obtained from a well drilled by Pemex Oil in the Yucatán Peninsula of Mexico. The site was within a large circular structure, known as Chicxulub, that some scientists, such as A. R. Hildebrand, had begun to think was the result of an impact from a large extra-terrestrial bolide. Not only was Walter convinced that the rock they held represented the melt rock from this impact, but, given the geology of the site, he argued that the impact was of the right age for the Cretaceous/Tertiary boundary, and might represent the "smoking gun" that they had been looking for.

At that time I was working in eastern Montana, with Lowell Dingus and Bill Clemens, on dating the extinction of the dinosaurs. While I liked the idea of obtaining a date for the melt rock from Chicxulub, I found it difficult to see the association between the dinosaurs of Montana and an impact site in Mexico. Since I had just obtained excellent dates of around 65 million years ago for the last of the dinosaurs in Montana, my curiosity over this impact site led to my collaboration with Walter and Alessandro, if only perhaps to show them that the ages of the two events were different. Much to my surprise and Walter's delight, this turned out not to be the case. Instead, the ages for the two sites were indistinguishable from each other.

When these results were published in the journal Science on August 14, 1992, they caused a tremendous sensation in the popular media, which was no surprise. What was surprising was that they provoked little negative response from researchers who had previously been loud in their criticism of the impact theory of dinosaur extinction. The intellectual tide had shifted to an acceptance of an impact at or near the Cretaceous/Tertiary boundary. But some skeptics continued to say, "OK, so there was an impact...but did it cause the demise of the dinosaurs?" What more of a smoking gun could these people demand than irrefutable evidence of catastrophic impact right on the cusp of the extinction, 64.98 million years ago?

The Kyoto conference was therefore a rewarding time for me. I was in my mid-thirties, confident, even a little brash, and I loved the notoriety the work generated. Much of my work, like that of most scientists, had gone along unnoticed by anyone but a few specialists in the field, with little public fanfare. The glamor of the media attention over the dating of the Chicxulub crater seemed to me to be my proverbial fifteen minutes of fame. Having never stepped into the arena of human origins research, I had no idea of the potential for notoriety that lurked there, too. Of all the disciplines in science, paleoanthropology has an apparently endless capacity for stirring emotions, in its practitioners and in spectators alike. And of all the disciplines in science, paleoanthropology boasts perhaps the largest share of egos, often engaged in intemperate defense of cherished hypotheses and sometimes in fierce public and private battle with other egos. For me, the journey to Java was to be a journey to a different kind of science. Garniss was not so naive.

Garniss, now approaching eighty-one, has been closely associated with many of the important developments in geochronology, particularly dating volcanic minerals based on the slow conversion of an isotope of potassium to an isotope of argon. The accumulation of the argon isotope acts like a clock, measuring the passage of time: the more of that particular isotope of argon there is in a rock, the older the rock is. Given certain assumptions and calculations, a very reliable age can be worked out, as will be described in a little more detail later in the chapter. Known prosaically as potassium/argon dating, the technique became the pillar of establishing the ages of early human fossils. One of the first occasions on which the technique was applied in African prehistory was to try to determine the age of the first human fossil discovered at the now famous Olduvai Gorge, in Tanzania, in 1959.

For almost three decades Louis and Mary Leakey had scoured the ancient sediments of the gorge, looking for relics of our ancestors. Signs of prehistoric daily life were everywhere, in the form of hundreds and thousands of simple stone tools, and collections of bones that had once been dinner for the gorge's inhabitants. Then, in a story many times told, Mary Leakey spotted massive, humanlike teeth eroding out of a small slope one afternoon in August 1959. When alerted by Mary, Louis, who had been resting back at camp with a high fever, quickly raced to the spot. The teeth proved to be the tip of a paleontological iceberg, which was the complete (but shattered) cranium of what Louis later named Zinjanthropus boisei, or East African Man. The fossil was also nicknamed Nutcracker Man, in recognition of its millstone-like molars.

Louis quickly recruited Garniss (and his colleague Jack Evernden) to find an accurate date for Zinjanthropus. The principle of finding the age of an ancient fossil is fairly simple but indirect, since at present there is no reliable way of discovering the age from the fossil itself. If you know the age of the layer of rock that covers a fossil, then you know that the fossil is at least this old. For instance, imagine that an individual died a million years ago, at which point its bones began to be covered by sediments. Imagine also that at some time afterward a volcano erupted, spewing volcanic ash over the local countryside and further covering the bones. If, in the present day, you take a sample of that ash and subject it to dating analysis, you will get a result that is a little less than one million years of age. You can then say that the individual lived somewhat earlier than that date. If you date ash layers above and below a fossil, you can say that the individual lived between these two dates, making the estimate more precise. And if the fossil was actually buried in the ash layer, in an ancient version of what happened at Pompeii, you can be most specific of all: the age of the ash is the date the individual died. Unless, of course, an individual's bones somehow become reworked into ancient sediments that were deposited long before the individual lived, which, via certain tricks of nature, can happen. The geologist's job is to look for clues to the true history of the bones.

Volcanoes have been extremely active in the history of East Africa, which has given paleoanthropologists the opportunity to create a very good time frame for sorting the events of human evolutionary history as they unfolded there. Garniss's landmark application of potassium/argon dating in 1961 produced a shock to the world of anthropology: Zinjanthropus had lived 1.75 million years ago (now, with a change in the decay constants of the potassium isotope, the age is recalculated to 1.85 million years), more than three times deeper into the past than anyone had imagined. The discovery of the fossil, and its startlingly ancient age, propelled Louis Leakey to world fame. The collaboration between Garniss and Louis seemed set to ripen into a productive partnership, one that would be of great benefit to the science. Within a few years, however, the relationship soured, as Louis repeatedly refused to believe dates that Garniss and Evernden produced for other fossils, when the dates did not jibe with what Louis wanted or believed. Garniss developed a tremendous respect for Louis Leakey the man, but not for Louis Leakey the scientist. Eventually Garniss vowed never to set foot in East Africa while Louis was alive. He kept his vow, although not for as long a time as he was prepared to; Louis Leakey died in October 1972.

About a decade after the breakup with Louis, Garniss was locking horns with another Leakey: this time it was Richard, Louis and Mary's equally famous son. Shortly before Louis's death, Richard found a large-brained human ancestor in the sediments on the eastern shore of Lake Turkana, in northern Kenya. Richard was convinced that the skull, known simply by its museum accession number of 1470, was the earliest known specimen of our own lineage, that is, Homo. Using an advanced version of the potassium/argon dating technique, British geochronologists claimed that 1470 had lived and died 2.6 million years ago, making it more than half a million years older than any previously discovered member of the genus Homo. The skull did for Richard what Zinjanthropus had done for Louis: it propelled him to worldwide recognition.

Very soon, however, a swell of evidence began to imply that 1470 was in fact less than 2 million years old, not 2.6 million. Richard would hear none of it, declaring that the science used to date the skull was the most modern available, and therefore the age must be right. Besides, being more than half a million years younger than the declared age would diminish 1470's importance in the world of anthropology. A tremendous fracas boiled up; accusations and counteraccusations of perfidy -- and worse -- were hurled around. Garniss, having been sucked into the fight, eventually put the nail in 1470's coffin: he showed that the skull was a little less than 1.9 million years old. (An Australian geochronologist, Ian McDougall, produced a similar age for 1470 at about the same time as Garniss's work.) Only recently has Richard quelled the anger he felt over the episode.

So, Garniss was not naive about what it means to be a journeyman geochronologist in the world of paleoanthropology. He had no illusions about what going to Java might mean, and, in any case, he had had a preview: our August 1992 trip to Java was not his first. He had gone there in 1969, fueled by National Science Foundation money and with the hope of applying potassium/argon dating to the Mojokerto child. The age of the Javan fossils is important to anthropologists because it relates to when humans first expanded their territory beyond the African continent. The human family originated in Africa, just as Charles Darwin predicted over a century ago; no self-respecting anthropologist doubts that, and we now know our prehistory stretches back about 5 or 6 million years. But when did human feet first touch Eurasian soil? There was (and still is, in some people's minds) a lot of doubt about that. The Mojokerto child might hold the answer -- if an accurate age could be established, that is.

The answer Garniss got in 1969 when, back in Berkeley, he analyzed some volcanic rock samples he had taken from the Mojokerto fossil site was 1.9 million years. This was almost twice as old as prevailing anthropological theory allowed, so the result was sensational. Or rather, unbelievable. "Everyone said I must be crazy," Garniss now recalls. "They said, 'You must have got it wrong,' and that was that." There is nothing more intransigent than a group of scientists clinging to a cherished theory in the face of counterevidence. There was room for some doubt about the age Garniss had obtained, however, because the volcanic rock he had used for the analysis was less than ideal. "It contained very low levels of potassium, and that compromises the accuracy of the date," Garniss concedes. "Low levels of potassium means that low levels of argon are produced in the rock, through radioactive decay. And with low levels of the two elements to measure, there is a significant margin of error in the result, with the technology available at the time." In this case, Garniss calculated the age of the Mojokerto child to be 1.9 million years, but with a margin of error of plus or minus 25 percent. This meant that the child might have lived as long ago as 2.5 million years, or as recently as 1.5 million. "Even the young age is still a lot older than everybody believed," Garniss points out. "But still people refused to listen. They had their idea of what the date should be -- that is, less than one million years -- and that was that. No discussion."

It has to be admitted, however, that the way the date came into the public arena didn't exactly inspire confidence in Garniss's claim. For instance, the initial announcement of the work was contained in two short paragraphs reporting on a symposium at the University of California in June 1970. The date of 1.9 million years was erroneously said to apply to an entirely different fossil specimen from the Mojokerto child, a mistake that was corrected in the pages of the journal Science not long after. Even with the error corrected, Garniss admitted in several publications that the low level of potassium in the mineral and the high level of atmospheric argon "greatly reduce [the date's] precision and accuracy."1 Now Garniss simply says, "It was the best I could do at the time."

Two decades later, Garniss was prepared to try again. The reason? As so often happens in science, new technology made possible a greater precision. The potassium/argon dating technique had developed to the point at which it was possible to work with extremely small samples, rather than on the samples of several grams that were required previously, and with minerals low in potassium, as in Java. This allowed for much more precise dating of rock minerals there.

A NEW ERA OF GEOCHRONOLOGY

The modern approach for dating rocks goes back to the beginning of this century, when Ernest Rutherford suggested that the natural radiation in them might be exploited as built-in clocks. With information about the rate of decay of certain isotopes and a way of measuring the decay products, rocks may be dated, he said. There are different kinds of clocks, based on different isotopes. Since the 1950s three principal clocks have been exploited for rocks: uranium series, rubidium/strontium, and potassium/argon. This last system, or at least a version of it, is the most useful technique for anthropologists interested in early human prehistory because it works in much of the relevant time range and is applied to volcanic rocks and minerals.

Many rocks contain traces of potassium, particularly some minerals in volcanic rocks, such as potassium feldspar, amphiboles, and various micas, common minerals of volcanic eruptions. Naturally occurring potassium contains a small quantity of potassium-40, an isotope that decays slowly and at a known rate to produce argon-40, a noble gas. During volcanic eruptions the hot molten lava, rising from a great depth in the Earth, loses whatever argon it contains as the pressure lowers and the lava erupts onto the surface of the earth. This effectively sets the argon clock to zero. As time passes after being expelled by an eruption, a potassium-containing rock will accumulate higher and higher levels of radiogenic argon-40, the product of decay of potassium-40 in the rock. Devised in 1948 by Tom Aldrich and Alfred Nier of the University of Minnesota and developed by the Berkeley geophysicist John Reynolds in the 1950s, the potassium/argon clock operates on the simple principle that the more radiogenic argon-40 a rock contains, the older it is. The amount of potassium in the rock or mineral sample has to be measured, of course, because the clock is based on the quantity of argon-40 that would accumulate over particular periods of time from a potassium source of a certain amount. Then, with appropriate technical adjustments, the potassium/argon clock can give reliable dates, particularly for relatively young rocks. By "young" we geologists mean a million years or less rather than hundreds of millions of years. (The technique is also suitable for rocks as old as several billion years.)

To date a rock using the conventional potassium/argon technique therefore requires two separate measurements: the amount of potassium in the rock or mineral, and the amount of radiogenic argon-40 that has accumulated in it. The two measurements involve two separate experiments on two separate samples of the rock, using different methods, since argon is a gas and potassium is a solid. It is a cumbersome but workable approach, one that has been the workhorse of geochronology for many years. But in 1965 Craig Merrihue, a graduate student of Reynolds and Garniss at Berkeley, hit upon an idea that would allow the two measurements -- of potassium-40 and argon-40 -- to be made simultaneously on one sample. Naturally occurring potassium contains a large and constant proportion of a second isotope, potassium-39, which under normal circumstances is stable. Blast it with neutrons in a reactor, however, and it becomes argon-39, cousin of argon-40. Merrihue realized that argon-39 could therefore serve as a vicarious measure of potassium-40 in a rock. When the irradiated minerals are fused at high temperature within vacuum, both forms of argon gas are released and can be determined almost simultaneously in a mass spectrometer: the argon-39 provides a measure of the potassium contained in the sample, and the argon-40 represents the cumulative ticks of the radiogenic clock. All the information required for dating the sample is therefore obtained from the same sample, at the same time, using the same method.

Merrihue developed this so-called argon-40/argon-39 technique with Grenville Turner, who until recently was at Sheffield University, but Merrihue never saw it applied, as he lost his life in a climbing accident in 1966. Various labs were involved in refining the new technique over the next decade, including those of Turner in Sheffield, Jack Miller in Cambridge, Brent Dalrymple and Marvin Lanphere in Menlo Park, and Derek York in Toronto.

Fusing a rock sample in the argon-40/argon-39 technique usually requires raising its temperature to about 1600°C, which was achieved in the early years through radio-frequency induction. The sample was enclosed in a crucible at high vacuum and heated for some 45 minutes. The released gases were first passed through scrubbers to remove reactive ingredients and then expanded into a mass spectrometer, a machine for measuring the amounts of argon. It was a time-consuming business; one sample might occupy a machine for an entire day. The recent revolution in geochronology -- taking the argon-40/argon-39 technique to a new level -- has focused on the means of heating the sample and on the possibility of using small samples.

In the late 1960s George Megrue of the Smithsonian Institution in Washington, D.C., was experimenting with ways of analyzing the gas content of minerals. He decided to use a pulsed laser for heating small mineral samples, enabling him to drive off the noble gases and analyze them in a mass spectrometer. Later he heard about Merrihue's argon-40/argon-39 dating method and applied it to his system in dating lunar samples in 1973, effectively substituting the pulsed-laser heat source for the radio-frequency heating. Oliver Schaeffer, of the State University of New York at Stony Brook, then applied the pulsed-laser dating technique in a more extensive lunar dating study to samples from the early Apollo missions, but had hardly gotten started before he died suddenly. Derek York, inspired by Shaeffer, brought the technique into the realm of terrestrial rock dating in the early 1980s, using a continuous laser. Soon after York and his colleagues had their single-crystal system up and running, in the mid-1980s, Garniss visited the Toronto lab and was very impressed by what he saw.

Yet to be achieved, however, was the ability to work with young single crystals (that is, just a few million years old). When you have one crystal rather than a thousand, you have only one thousandth of the amount of gas coming off. With older samples, there is still a measurable amount of gas involved. Young crystals, however, need a very sensitive mass spectrometer to measure the tiny amount.

Garniss had stuck very successfully with conventional potassium/argon dating during his long career at the University of California at Berkeley, not venturing into argon-40/argon-39 territory. "But when I saw them measure individual components of gas from a single crystal," Garniss remembers of his visit to York's lab, "I was converted." York's mass spectrometer was not sensitive enough to date young crystals, but new, high-resolution ones were being developed (York had one on order), and it was clear to Garniss that a new era in geochronology was at hand.

At the time Garniss was nearing retirement from Berkeley, although not from geochronology. With his well-equipped, world-renowned lab and a coterie of bright young geochronologists, he was looking for someplace other than the university to set up shop. Donald Johanson, famous for discovering Lucy, the 3-million-year-old human skeleton, in Ethiopia in 1974, had recently established his Institute of Human Origins just up the street in Berkeley. "The institute's board thought it would be a good idea to broaden the institute's scientific base to include geochronology, which is an important part of human origins research, and so they invited us to be part of the institute," Garniss remembers. "It seemed like a perfect marriage, to join Don in his venture."

And so Garniss and his people packed up our lab and moved into the basement of the Church Divinity School of the Pacific, which housed Don's institute. Very soon the lab was making important contributions to the development of single-crystal laser-fusion dating. Primarily through the programming genius of Al Deino and the determination of Brent Turrin, Mac McCrory, and Tim Becker, we were the first to automate the entire process. Within a couple of years we had a system operating that could produce the ages of over 200 crystals in one fully programmed run. Age determinations could now be extremely accurate, even on minerals with low levels of potassium in them, as in Java.

PLANS FOR A RETURN TO JAVA

"I realized that the argon-40/argon-39 technique could give us a way to settle the dating of the Java fossils accurately, once and for all," Garniss recalls. So in the spring of 1990 we unearthed some of the mineral samples he had collected in Java in 1969, and ran them. The preliminary dates: 1.7 million years, with very little margin of error. That wasn't 1.9, as Garniss had obtained previously, but neither was it the one million years that everyone knew "must" be true. In fact, the slightly younger age we obtained was not surprising, given that the new argon-40/argon-39 system allowed us to date much smaller amounts of material. By hand-picking crystals from the pumice rather than processing bulk samples, we were able in the 1990 test to exclude slightly rounded crystals that might have been reworked from older deposits, thus avoiding parts of the original sample that might have contributed to an erroneously old date. In any case, the 1.7-million-year result was intriguing enough to encourage us to seek funds for collecting more rock samples from Java for a whole new set of dating analyses, which would be done on small populations of single crystals carefully selected to avoid any contamination.

The budget of the Institute of Human Origins, with which our geochronology group was now more formally allied, had a so-called contingency fund, which was meant to support low-budget, pilot projects of the sort we had in mind. A modest figure of $6,000 was agreed upon at a meeting in the fall of 1990 for a new venture to Java. But it wasn't until the late summer of 1992 that the Kyoto geochronology congress provided the opportunity to make use of it. Visiting Java on the way back from Japan would be an inexpensive diversion for us. It would also be eye-opening in our quest to produce accurate ages of these important human relics from Java at last, and, in the process, perhaps dramatically alter our understanding of human prehistory: namely, when did humans first wander out of Africa, and how did we come to be the species we are today?