by Neil Shubin
The best places to look are those where we can walk for miles over the rock to discover areas where bones are “weathering out.” Fossil bones are often harder than the surrounding rock and so erode at a slightly slower rate and present a raised profile on the rock surface. Consequently, we like to walk over bare bedrock, find a smattering of bones on the surface, then dig in.
So here is the trick to designing a new fossil expedition: find rocks that are of the right age, of the right type (sedimentary), and well exposed, and we are in business. Ideal fossil-hunting sites have little soil cover and little vegetation, and have been subject to few human disturbances. Is it any surprise that a significant fraction of discoveries happen in desert areas? In the Gobi Desert. In the Sahara. In Utah. In Arctic deserts, such as Greenland.
This all sounds very logical, but let’s not forget serendipity. In fact, it was serendipity that put our team onto the trail of our inner fish. Our first important discoveries didn’t happen in a desert, but along a roadside in central Pennsylvania where the exposures could hardly have been worse. To top it off, we were looking there only because we did not have much money.
It takes a lot of money and time to go to Greenland or the Sahara Desert. In contrast, a local project doesn’t require big research grants, only money for gas and turnpike tolls. These are critical variables for a young graduate student or a newly hired college teacher. When I started my first job in Philadelphia, the lure was a group of rocks collectively known as the Catskill Formation of Pennsylvania. This formation has been extensively studied for over 150 years. Its age was well known and spanned the Late Devonian. In addition, its rocks were perfect to preserve early limbed animals and their closest relatives. To understand this, it is best to have an image of what Pennsylvania looked like back in the Devonian. Remove the image of present-day Philadelphia, Pittsburgh, or Harrisburg from your mind and think of the Amazon River delta. There were highlands in the eastern part of the state. A series of streams running east to west drained these mountains, ending in a large sea where Pittsburgh is today.
It is hard to imagine better conditions to find fossils, except that central Pennsylvania is covered in towns, forests, and fields. As for the exposures, they are mostly where the Pennsylvania Department of Transportation (PennDOT) has decided to put big roads. When PennDOT builds a highway, it blasts. When it blasts, it exposes rock. It’s not always the best exposure, but we take what we can get. With cheap science, you get what you pay for.
And then there is also serendipity of a different order: in 1993, Ted Daeschler arrived to study paleontology under my supervision. This partnership was to change both our lives. Our different temperaments are perfectly matched: I have ants in my pants and am always thinking of the next place to look; Ted is patient and knows when to sit on a site to mine it for its riches. Ted and I began a survey of the Devonian rocks of Pennsylvania in hopes of finding new evidence on the origin of limbs. We began by driving to virtually every large roadcut in the eastern part of the state. To our great surprise, shortly after we began the survey, Ted found a marvelous shoulder bone. We named its owner Hynerpeton, a name that translates from Greek as “little creeping animal from Hyner.” Hyner, Pennsylvania, is the nearest town. Hynerpeton had a very robust shoulder, which indicates a creature that likely had very powerful appendages. Unfortunately, we were never able to find the whole skeleton of the animal. The exposures were too limited. By? You guessed it: vegetation, houses, and shopping malls.
Along the roads in Pennsylvania, we were looking at an ancient river delta, much like the Amazon today. The state of Pennsylvania (bottom) with the Devonian topography mapped above it.
After the discovery of Hynerpeton and other fossils from these rocks, Ted and I were champing at the bit for better-exposed rock. If our entire scientific enterprise was going to be based on recovering bits and pieces, then we could address only very limited questions. So we took a “textbook” approach, looking for well-exposed rocks of the right age and the right type in desert regions, meaning that we wouldn’t have made the biggest discovery of our careers if not for an introductory geology textbook.
Originally we were looking at Alaska and the Yukon as potential venues for a new expedition, largely because of relevant discoveries made by other teams. We ended up getting into a bit of an argument/debate about some geological esoterica, and in the heat of the moment, one of us pulled the lucky geology textbook from a desk. While riffling through the pages to find out which one of us was right, we found a diagram. The diagram took our breath away; it showed everything we were looking for.
The argument stopped, and planning for a new field expedition began.
On the basis of previous discoveries made in slightly younger rocks, we believed that ancient freshwater streams were the best environment in which to begin our hunt. This diagram showed three areas with Devonian freshwater rocks, each with a river delta system. First, there is the east coast of Greenland. This is home to Jenny Clack’s fossil, a very early creature with limbs and one of the earliest known tetrapods. Then there is eastern North America, where we had already worked, home to Hynerpeton. And there is a third area, large and running east–west across the Canadian Arctic. There are no trees, dirt, or cities in the Arctic. The chances were good that rocks of the right age and type would be extremely well exposed.
The Canadian Arctic exposures were well known, particularly to the Canadian geologists and paleobotanists who had already mapped them. In fact, Ashton Embry, the leader of the teams that did much of this work, had described the geology of the Devonian Canadian rocks as identical in many ways to the geology of Pennsylvania’s. Ted and I were ready to pack our bags the minute we read this phrase. The lessons we had learned on the highways of Pennsylvania could help us in the High Arctic of Canada.
Remarkably, the Arctic rocks are even older than the fossil beds of Greenland and Pennsylvania. So the area perfectly fit all three of our criteria: age, type, and exposure. Even better, it was unknown to vertebrate paleontologists, and therefore un-prospected for fossils.
The map that started it all. This map of North America captures what we look for in a nutshell. The different kinds of shading reflect where Devonian age rocks, whether marine or freshwater, are exposed. Three areas that were once river deltas are labeled. Modified from figure 13.1, R. H. Dott and R. L. Batten, Evolution of the Earth (New York: McGraw-Hill, 1988). Reproduced with the permission of The McGraw-Hill Companies.
Our new challenges were totally different from those we faced in Pennsylvania. Along the highways in Pennsylvania, we risked being hit by the trucks that whizzed by as we looked for fossils. In the Arctic we risked being eaten by polar bears, running out of food, or being marooned by bad weather. No longer could we pack sandwiches in the car and drive to the fossil beds. We now had to spend at least eight days planning for every single day spent in the field, because the rocks were accessible only by air and the nearest supply base was 250 miles away. We could fly in only enough food and supplies for our crew, plus a slender safety margin. And, most important, the plane’s strict weight limits meant that we could take out only a small fraction of the fossils that we found. Couple those limitations with the short window of time during which we can actually work in the Arctic every year, and you can see that the frustrations we faced were completely new and daunting.
Enter my graduate adviser, Dr. Farish A. Jenkins, Jr., from Harvard. Farish had led expeditions to Greenland for years and had the experience necessary to pull this venture off. The team was set. Three academic generations: Ted, my former student; Farish, my graduate adviser; and I were going to march up to the Arctic to try to discover evidence of the shift from fish to land-living animal.
There is no field manual for Arctic paleontology. We received gear recommendations from friends and colleagues, and we read books—only to realize that nothing could prepare us for the experience itself. At no time is this more sharply felt than when the helicopter drops one off for the first t
ime in some godforsaken part of the Arctic totally alone. The first thought is of polar bears. I can’t tell you how many times I’ve scanned the landscape looking for white specks that move. This anxiety can make you see things. In our first week in the Arctic, one of the crew saw a moving white speck. It looked like a polar bear about a quarter mile away. We scrambled like Keystone Kops for our guns, flares, and whistles until we discovered that our bear was a white Arctic hare two hundred feet away. With no trees or houses by which to judge distance, you lose perspective in the Arctic.
The Arctic is a big, empty place. The rocks we were interested in are exposed over an area about 1,500 kilometers wide. The creatures we were looking for were about four feet long. Somehow, we needed to home in on a small patch of rock that had preserved our fossils. Reviewers of grant proposals can be a ferocious lot; they light on this kind of difficulty all the time. A reviewer for one of Farish’s early Arctic grant proposals put it best. As this referee wrote in his review of the proposal (not cordially, I might add), the odds of finding new fossils in the Arctic were “worse than finding the proverbial needle in the haystack.”
It took us four expeditions to Ellesmere Island over six years to find our needle. So much for serendipity.
We found what we were looking for by trying, failing, and learning from our failures. Our first sites, in the 1999 field season, were way out in the western part of the Arctic, on Melville Island. We did not know it, but we had been dropped off on the edge of an ancient ocean. The rocks were loaded with fossils, and we found many different kinds of fish. The problem was that they all seemed to be deep-water creatures, not the kind we would expect to find in the shallow streams or lakes that gave rise to land-living animals. Using Ashton Embry’s geological analysis, in 2000 we decided to move the expedition east to Ellesmere Island, because there the rocks would contain ancient streambeds. It did not take long for us to begin finding pieces of fish bones about the size of a quarter preserved as fossils.
Our camp (top) looks tiny in the vastness of the landscape. My summer home (bottom) is a small tent, usually surrounded by piles of rocks to protect it from fifty-mile-per-hour winds. Photographs by the author.
The real breakthrough came toward the end of the field season in 2000. It was just before dinner, about a week before our scheduled pickup to return home. The crew had come back to camp, and we were involved in our early-evening activities: organizing the day’s collections, preparing field notes, and beginning to assemble dinner. Jason Downs, then a college undergraduate eager to learn paleontology, hadn’t returned to camp on time. This is a cause for worry, as we typically go out in teams; or if we separate, we give each other a definite schedule of when we will make contact again. With polar bears in the area and fierce storms that can roll in unexpectedly, we do not take any chances. I remember sitting in the main tent with the crew, the worry about Jason building with each passing moment. As we began to concoct a search plan, I heard the zipper on the tent open. At first all I saw was Jason’s head. He had a wild-eyed expression on his face and was out of breath. As Jason entered the tent, we knew we were not dealing with a polar bear emergency; his shotgun was still shouldered. The cause of his delay became clear as his still shaking hand pulled out handful after handful of fossil bones that had been stuffed into every pocket: his coat, pants, inner shirt, and daypack. I imagine he would have stuffed his socks and shoes if he could have walked home that way. All of these little fossil bones were on the surface of a small site, no bigger than a parking spot for a compact car, about a mile away from camp. Dinner could wait.
With twenty-four hours of daylight in the Arctic summer, we did not have to worry about the setting sun, so we grabbed chocolate bars and set off for Jason’s site. It was on the side of a hill between two beautiful river valleys and, as Jason had discovered, was covered in a carpet of fossil fish bones. We spent a few hours picking up the fragments, taking photos, and making plans. This site had all the makings of precisely what we were looking for. We returned the next day with a new goal: to find the exact layer of rock that contained the bones.
The trick was to identify the source of Jason’s mess of bone fragments—our only hope of finding intact skeletons. The problem was the Arctic environment. Each winter, the temperature sinks to minus 40 degrees Fahrenheit. In the summer, when the sun never sets, the temperature rises to nearly 50 degrees. The resulting freeze-thaw cycle crumbles the surface rocks and fossils. Each winter they cool and shrink; each summer they heat and expand. As they shrink and swell with each season over thousands of years at the surface, the bones fall apart. Confronted by a jumbled mass of bone spread across the hill, we could not identify any obvious rock layer as their source. We spent several days following the fragment trails, digging test pits, practically using our geological hammers as divining rods to see where in the cliff the bones were emerging. After four days, we exposed the layer and eventually found skeleton upon skeleton of fossil fish, often lying one on top of another. We spent parts of two summers exposing these fish.
This is where we work: southern Ellesmere Island, in Nunavut Territory, Canada, 1,000 miles from the North Pole.
Failure again: all the fish we were finding were well-known species that had been collected in sites of a similar age in Eastern Europe. To top it off, these fish weren’t very closely related to land-living animals. In 2004, we decided to give it one more try. This was a do-or-die situation. The Arctic expeditions were prohibitively expensive and, short of a remarkable discovery, we would have to call it quits.
Everything changed over a period of four days in early July 2004. I was flipping rock at the bottom of the quarry, cracking ice more often than rock. I cracked the ice and saw something that I will never forget: a patch of scales unlike anything else we had yet seen in the quarry. This patch led to another blob covered by ice. It looked like a set of jaws. They were, however, unlike the jaws of any fish I had ever seen. They looked as if they might have connected to a flat head.
One day later, my colleague Steve Gatesy was flipping rocks at the top of the quarry. Steve removed a fist-size rock to reveal the snout of an animal looking right out at him. Like my ice-covered fish at the bottom of the pit, it had a flat head. It was new and important. But unlike my fish, Steve’s had real potential. We were looking at the front end, and with luck the rest of the skeleton might be safely sitting in the cliff. Steve spent the rest of the summer removing rock from it bit by bit so that we could bring the entire skeleton back to the lab and clean it up. Steve’s masterful work with this specimen led to the recovery of one of the finest fossils discovered to date at the water–land transition.
The specimens we brought back to the lab at home were little more than boulders with fossils inside. Over the course of two months, the rock was removed piece by piece, often manually with dental tools or small picks by the preparators in the lab. Every day a new piece of the fossil creature’s anatomy was revealed. Almost every time a large section was exposed, we learned something new about the origin of land-living animals.
What we saw gradually emerge from these rocks during the fall of 2004 was a beautiful intermediate between fish and land-living animals. Fish and land-living animals differ in many respects. Fish have conical heads, whereas the earliest land-living animals have almost crocodile-like heads—flat, with the eyes on top. Fish do not have necks: their shoulders are attached to their heads by a series of bony plates. Early land-living animals, like all their descendants, do have necks, meaning their heads can bend independently of their shoulders.
There are other big differences. Fish have scales all over their bodies; land-living animals do not. Also, importantly, fish have fins, whereas land-living animals have limbs with fingers, toes, wrists, and ankles. We can continue these comparisons and make a very long list of the ways that fish differ from land-living animals.
The process of finding fossils begins with a mass in a rock that is gradually removed over time. Here I show a fossil as it travels fro
m the field to the lab and is carefully prepared as a specimen: the skeleton of the new animal. Photograph in upper left by author; other photographs courtesy of Ted Daeschler, Academy of Natural Sciences of Philadelphia.
But our new creature broke down the distinction between these two different kinds of animal. Like a fish, it has scales on its back and fins with fin webbing. But, like early land-living animals, it has a flat head and a neck. And, when we look inside the fin, we see bones that correspond to the upper arm, the forearm, even parts of the wrist. The joints are there, too: this is a fish with shoulder, elbow, and wrist joints. All inside a fin with webbing.
Virtually all of the features that this creature shares with land-living creatures look very primitive. For example, the shape and various ridges on the fish’s upper “arm” bone, the humerus, look part fish and part amphibian. The same is true of the shape of the skull and the shoulder.
It took us six years to find it, but this fossil confirmed a prediction of paleontology: not only was the new fish an intermediate between two different kinds of animal, but we had found it also in the right time period in earth’s history and in the right ancient environment. The answer came from 375-million-year-old rocks, formed in ancient streams.
This figure says it all. Tiktaalik is intermediate between fish and primitive land-living animal.
As the discoverers of the creature, Ted, Farish, and I had the privilege of giving it a formal scientific name. We wanted the name to reflect the fish’s provenance in the Nunavut Territory of the Arctic and the debt we owed to the Inuit people for permission to work there. We engaged the Nunavut Council of Elders, formally known as the Inuit Qaujimajatuqangit Katimajiit, to come up with a name in the Inuktitut language. My obvious concern was that a committee named Inuit Qaujimajatuqangit Katimajiit might not propose a scientific name we could pronounce. I sent them a picture of the fossil, and the elders came up with two suggestions, Siksagiaq and Tiktaalik. We went with Tiktaalik for its relative ease of pronunciation for the non-Inuktitut-speaking tongue and because of its meaning in Inuktitut: “large freshwater fish.”