Susan Turner & Randall F Miller. American Scientist. Volume 93, Issue 3. May/Jun 2005.
The world’s oceans, lakes and rivers harbor about 30,000 species of fish. In addition to a diverse array of bony fishes, scientists have described some 830 species of Chondrichthyes-sharks, skates, rays and other cartilaginous fishes. Many paleobiologists are working to decipher the fossil record left by the bony fish-our ancestors after all, and what we generally find on the end of a fishing line. Others are concerned with the mysterious and closely connected story of how the sharks and their relatives came to be. Chondrichthyans are remnants of an ancient lineage, rare survivors from an explosion of forms that mostly disappeared before the end of the Devonian Period, 359 million years ago.
Thanks to the rarity of intact shark fossils, the transition from primitive fish to sharks and bony fishes is poorly understood. Following the discovery of a relatively complete early shark fossil, paleontologists hope to better understand this important evolutionary progression. Initial examination of the fossil led to surprising implications. First, it became clear that early sharks were more diverse than scientists had imagined. Next, the presence of pectoral spines on the new fossil suggests that relationships among bony and cartilaginous species might be somewhat closer and also more complex than had been thought. Thus, this type of shark could help scientists work out the evolution of bony fishes, the Osteichthyes. This and other fossils being discovered around the world look set to revolutionize our view of early shark and fish evolution.
A Lot More Fish in the Sea
We begin our story with the earliest fish, going far back to a time when the planet was already old, but vertebrate life was just getting started. There were neither sharks nor fishes as we know them today. Instead, creatures that defy simple classification were swimming through the Earth’s seas.
During the Ordovician Period, some 488 to 443 million years ago (mya), the first jawless fishes, or agnathans, arose. During the next geological period, the Silurian (443-416 mya), agnathans diversified. Through an elegant modification of gill arches (specialized gill structures), some evolved into the first jawed fishes, which in turn diversified during the Devonian (416-359 mya) into all the fishes known today. (One branch split off to become land-dwellers and eventually humans.) Among these jawed fishes were shark-like creatures, at the time but one kind of a bizarre array of marine fauna. Also exploring the evolutionary possibilities in the sea at this time were myriad strange vertebrates and invertebrates including placoderms (armored fish), trilobites and ammonoids-all now extinct. The sharks and bony fish succeeded in nature’s great gamble where other creatures fell by the wayside. In large part, their success began with the evolution of jaws.
Jaws enable a predator to grasp live prey while consuming it. Teeth allow the prey’s body parts to be cut or chewed for easy digestion. Such a useful adaptation greatly increased the evolutionary opportunities for early fishes and sharks. Jawless fishes, the forerunners of today’s lampreys and hagfishes, hung on, but in an evolutionary backwater as parasites and scavengers. Without jaws and teeth, it seems, modern land vertebrates would never have arisen. Because of the importance of this transition, paleontologists have longed to have a complete picture of how primitive ancestral species evolved into the sharks whose fictional counterparts terrorize moviegoers today.
How sharks and bony fishes are related is a key part of the puzzle. Complete fossil fishes from this era are exceedingly rare, but sharks’ teeth are common. But from teeth alone it’s difficult to infer taxonomic classifications. So the hunt for good fossil fishes continues. On July 4, 1997, a team led by one of us (Miller) found a fossil fish in Canada that was a very special catch indeed: the most complete shark fossil specimen from the early Devonian. Identified as Doliodus problematicus, a species known previously only from teeth, this specimen has enabled close study of the body shape and dental characteristics of a species thought to be among the earliest sharks-perhaps a missing link between sharks proper and the so-called spiny sharks, or acanthodians. So it was of particular interest that the fossil’s teeth or tooth rows, were found in situ-a unique circumstance for a fossil this old. The fossil shows that 400 million years ago, sharks had already developed the complex battery of teeth that made Jaws so frightening.
Long in the Tooth
It’s not easy to reconstruct fossil animals. And among the most contentious fossils are those belonging to the specimen we found. Its name, Doliodus problematicus, alerts you to its troubled history: It means “problematic deceiver.” These and many other fossil teeth have since confused paleontologists.
Typically millimeter-sized, Doliodus teeth were first found by a British collector known to us only as Mr. Jex. Despite this character’s historical evanescence, his fossils persisted in museums and other collections, and more Doliodus teeth have been uncovered by others. Although they have received plenty of attention from eminent scientists, they puzzle paleontologists to this day.
Sir Arthur Smith Woodward, doyen of paleontology at the Natural History Museum in London, formally described Doliodus (as Diplodus) teeth in 1892. Ramsey Traquair, his counterpart at the Royal Museum of Scotland in Edinburgh, revised the description the following year and changed the name to Doliodus. Traquair realized the teeth were arranged in tooth rows, the conveyor-belt arrangement of teeth that enables elasmobranchs-modern sharks-to constantly replace lost teeth and agreed with Woodward that Doliodus was a shark. Succeeding generations of paleontologists erroneously changed Woodward’s and Traquair’s classification. Taxonomically, then, Doliodus fossils have been homeless, passed back and forth from the sharks to the acanthodians. It was not until recently that one of us (Turner) finally recognized that Doliodus teeth indeed belonged to a shark. Thus the species took on a new significance in early shark evolution.
Problems with classification aren’t the only kind to have plagued students of the “problematic deceiver.” Paleogeographers were puzzled too. Its fossils are found primarily in rocks in Canada, from the Devonian part of a continental agglomeration called Euramerica. But most early shark fossils, including some that looked like Doliodus, are known from the Gondwana supercontinent, which lay thousands of miles to the southeast. So was Doliodus not a shark, or were sharks a lot more widespread than thought? With so little evidence to go on, it’s hard to decide. So goes the study of early shark evolution: Every tooth counts. In the case of Doliodus, scientists were keen to discover more. Indeed, before the Canadian team set out, one of us (Turner) advised them to “Keep an eye open for shark teeth.” Little did we know how good that advice would turn out to be.
Our group that day consisted of New Brunswick Museum staff and six-year-old Sasha Miller. Sasha turned over slabs of rock, and one of us (Miller) worked on their stratigraphy-analyzing the rock layers to place the associated fossils in geologic time. Meanwhile, Jeff McGovern of the museum staff and Heather Wilson of the University of Manchester excavated fossil-rich mudstone a few yards away. Aside from sharks’ teeth, the exotic fossil fauna included jawless cephalaspid fish, placoderms, acanthodians and weird invertebrates such as giant sea scorpions, or eurypterids. The sediment was deposited a little after 407 million years ago (according to the dates of volcanic rocks underneath) in a calm lagoon or estuary; scattered land-plant fossils indicate proximity to an ancient shoreline.
Before long, McGovern handed over a specimen that we quickly realized was a significant discovery: the oldest known articulated shark fossil. All its teeth were in place, and almost the entire front part of the shark had been beautifully preserved, including the braincase, jaws and pectoral fins, all in their anatomically correct positions. The mudstone had entombed the shark’s cartilage, teeth, scales and a long, thick spine jutting out along the front edge of each pectoral fin, a feature previously unknown in cartilaginous fishes-all this in a specimen of a species identifiable as Doliodus, the deceiver previously known only from teeth. By amazing good fortune, our colleague Richard Cloutier, from the Université du Québec, recovered part of the braincase a month later. But why were the discoveries important? Having a complete articulated fossil (with parts connected) enables paleontologists to fill significant gaps in knowledge.
The most important feature of this fossil is its paired pectoral spines. These suggest that many isolated fossil spines might have belonged to sharks rather than acanthodians as previously believed. The adaptive significance of these spines is uncertain. They were prevalent in acanthodians, placoderms and these new “sharks,” but are present today only in a few sharks (on dorsal fins) and catfish. The spines may have helped ward off giant eurypterids and other predatory arthropods. If they began as additional enlarged scales on the leading edges of the lateral fins, then they might also have been a result of sexual dimorphism, the differentiation of male and female body shape. Until now, most fossil spines have been attributed to acanthodians, which are characterized by having spines on all their paired fins.
So perhaps Doliodus wasn’t a shark at all. Features of the fossil blur the distinction between acanthodians and early chondrichthyans. To put the new fossil into the context of early fish evolution, it is necessary to look in more detail at the origins of sharks from their jawless ancestors.
Where Do They All Come From?
During their early evolution, fishes diversified into dozens of taxonomic groups, which paleontologists recognize by studying the fossils and coining excruciating polysyllabic names. The names of these groups (the genera or species, collectively taxa) are given based on morphology, or physical characteristics, alone; the biological concept of a species (which rests on the ability of individuals to produce viable offspring) cannot be tested directly in extinct animals. The various groups’ origins are generally regarded as monophyletic, each having arisen from a single common ancestor-separate branches on the evolutionary tree arising from a single stem.
Some paleontologists hypothesize that vertebrates originated in Gondwana, because most fossils come from areas that were once part of that paleocontinent. However, tiny cone-shaped fossil teeth from conodonts, eel-like animals with a finned tail and a possible notochord (a supporting rod later incorporated into the vertebral column in true vertebrates), are widely found in strata from a broad span of geologic time-the Cambrian to the late Triassic, some 500 to 100 mya-in areas that once were also part of Euramerica. On this basis, others hold that conodonts are the most primitive vertebrates and that vertebrates arose in Euramerica. Others refute the classification of conodonts as vertebrates and say that the fossils are not even teeth.
A team led by Degan Shu from the Early Life Institute at Northwest University, Xi’an, China, claims the earliest vertebrates based on fossils from the lower Cambrian Qiongzhusi Formation near Chengjiang. These specimens give a glimpse into the first fishlike designs, but they bear no sign of the hard bone and dentine structures regarded as fully vertebrate features.
The Chinese fossils do highlight the close connection between protovertebrates and early sharks or fishes. However, it’s still not clear at what point sharks can be said to exist for certain. When do the first sharks unequivocally appear in the fossil record? Textbooks still parrot the conventional thinking that no fossil sharks are found before the late Devonian, but this dogma ignores work from the last three decades. The oldest microfossils definitely attributable to sharks are scales in Silurian strata (440 mya) of Siberian and Arctic Russia. (Modern sharks’ skin is covered with minute scales, or denticles, that give the skin a sandpapery texture.)
A whole new suite of forms dating from just a little later, about 430 mya, has been found in Cornwallis Island and other islands in the Canadian Arctic. These scales bear features, such as a relatively wide and short base compared to the crown, that distinguish them from other scales and have led to their being assigned to a new shark family dubbed the Kannathalepididae. Comprising a series of odontodes-mineralized structures that are the basic unit of vertebrate teeth-these scales are also found in early Silurian deposits in Canada, possibly pushing the date of the earliest sharks even farther back. Early Silurian deposits in the Tarim Basin of western China have also yielded fin spines associated with sharklike scales.
Are these fossils true sharks? If so, the lineage was apparently toothless for millions of years. The first indisputable shark teeth do not turn up until about 50 million years after the appearance of these first putative shark scales in the late Ordovician. And they are quite recognizable.
Among the most significant recent contributions to a modern classification of fossil shark teeth was that of Herman Mader, working during the 1980s at the University of Göttingen in Germany. Based on his 1986 description of scales and teeth, several modern groups of sharks had their origins pushed back to the very beginning of the Devonian. He was also the first to present an evolutionary tree, or cladogram, for sharks based on teeth. Sadly, Mader did not continue his work for health reasons.
The oldest shark teeth described by Mader belong to Leonodus carlsi. This shark lived around 418 mya, on either side of the Silurian-Devonian boundary, in what is today southwestern Europe. Like Doliodus, Leonodus fossils are now being studied by paleontologists in Spain and Germany as articulated specimens, which provides some assurance on their taxonomic status as chondrichthyans. Other teeth belonging to putative sharks are found earlier, but their kinships remain unresolved. Fossil teeth display internal as well as external structure-some have typical dentine nutrient canals, for example-and these structures, along with other features, are under study in an attempt to elucidate the evolutionary affinities among the various groupings. Especially with the discovery of Doliodus’s special internal scales on the gill arches, of particular interest is how sharks are related to a contemporaneous group of scale-covered jawless fishes, the thelodonts.
Sharks and Thelodonts
Since sharks must have evolved from an earlier vertebrate, their origins might be easier to establish if this ancestor can be determined. However, the fossil record during the era preceding the earliest evidence of sharks is exceedingly spotty: Virtually all the information comes from isolated scales from other kinds of fishes. Among the most abundant types of these belong to thelodonts. Because of their potential importance as shark ancestors, one of us (Turner) has been investigating thelodont fossils for more than 30 years. Much of this work has been geared to garnering evidence to support the early realization that thelodonts and sharks share key features. A thelodont origin for sharks was put forward by early paleontologists such as Traquair, but the idea was not extensively explored until recently.
Resolving the issue depends on the microsquamose exoskeleton-a skin made up of tiny scales-that is characteristic of thelodonts, and whether this is a feature of primitive vertebrates only or is also seen in advanced fishes.
The nature of the scaly thelodont skin has implications for our understanding of where thelodonts fit into the evolutionary scheme of things. The pattern of scale distribution in thelodonts is complex, comprising varied zones and including both internal scales that line the mouth down to the pharynx and specialized scales around the branchial (gill) region. Branchial scales have been found in several thelodont genera, and thus might be characteristic of thelodonts in general.
Some authors claim that thelodont branchial scales are primitive tooth whorls, putative tooth-rows as seen in the typical pattern of sharks and other jawed fishes. Walter Gross of the University of Tubingen first characterized scales from Loganellia scotica, but he failed to recognize their significance as branchial scales. One of us (Turner) anticipated that all thelodonts would possess these branchial scales, along with other types lining the upper pharynx. This prediction was based on comparing the features of modern sharks with those of L. scotica, from examination of a collection made by Alex Ritchie, now at the Australian Museum.
It is well known that the skin of modern sharks is covered with tiny tooth-like scales, but it was not until 1970 that Gary Nelson, then at the American Museum of Natural History, made an elegant study of their specialized internal scales: intricate banks of scales that line the branchial bars inside the throat. The similarity of these scales to those of thelodonts alerted Turner to the possibility that thelodonts were more closely related to sharks than had been thought. In 1991 two Dutch paleontologists, Wim van der Brugghen and Jo Vergoossen, made beautiful scanning electron microscope images that highlight the complexity of L. scotica branchial scales. Similar scales were found in the Canadian Arctic in 1998 by Mark Wilson and Michael Caldwell of the University of Alberta. Another early Silurian thelodont from Wisconsin has scales that are made up of multiple odontodes. The zonation and growth pattern of thelodont branchial scales now appears more complex than paleontologists first thought and pushes back in time the formation of complex “pre-teeth.”
These details of scale morphology and evolution together imply a need to rethink previous notions of how teeth became evolutionarily associated with jaws. The accepted model is that teeth evolved from the modification of specialized skin scales that became attached to the jaw cartilages (parts of the gill branches that are believed to have evolved into the lower and upper jaws), and the structures evolved in tandem. Imagine a primitive jawless fish whose foreparts (lips) are covered with sharp, mineralized scales, and these toothlike scales attaching to proto-jaws. But there is an alternative scenario. In 2000, based on their claim that conodont fossils are in fact vertebrate teeth, Moya Smith at King’s College London and Michael Coates at the University of Chicago hypothesized that dermal odontodes (or toothlike skin scales) and teeth arose as separate structures. This idea uncouples the synchronous evolution of early teeth and jaws. Fitting this hypothesis with the thelodont observations suggests that these fish were the first to evolve a recognizable “dentition” (the complex branchial denticles)-but within their throat, not on the jaw cartilages.
Significantly, the specialized, whorl-like branchial scales in thelodonts are different in size and shape from the buccal scales, which line the inside of the mouth. The buccal scales more closely resemble the typical external head scales and so could instead have been teeth precursors, consistent with the traditional model. The road to a full understanding of early fish evolution is beset with twists and turns. It might also help to stop and consider what a shark is. As so often in paleontology, clearing up the various questions will require more specimens. This prospect is looking increasingly likely with the spate of new discoveries in recent years.
Fossil Frenzy
Fossils from the first 50 million years of cartilaginous fish evolution were very rare only 15 or 20 years ago but since then a bewildering diversity of specimens has come to light. The Devonian, the “Age of Fishes,” was a time when nature was experimenting with multitudes of different forms, and species were exploring new ecological niches. Clearly this was a crucial time period. Certain specimens have been identified as chondrichthyans, or even neoselachians (modern sharks) such as the early Devonian Mcmurdodus whitei from western Queensland, the teeth of which closely resemble those of the cow sharks. Other studies based on Late Devonian to Carboniferous sharklike fishes typically called “cladodonts” do not actually help our understanding of the early evolution of sharks.
Such analyses in themselves may not be enough. The presence of putative sharks in the Ordovician-the period between the Cambrian and the Silurian-is suggested by scales that are at least as complex as those of accepted sharks. But as mentioned, it’s uncertain whether or not sharks originated as early as the Ordovician. Part of the problem is that modern analyses seem to work back to front. Rather than looking at a fossil and trying to decide if it’s a shark, we need to look closely at the early record and base our definition of “shark” on those fossils. This is necessary to determine whether the modern and the fossil fishes now called “sharks” are indeed closely related. During the first 50 million years we already seem to have several early different types of “sharks.”
This initial period of shark evolution will continue to be an exciting if confusing area of study for paleontologists. Most of the fossils are tiny scales and teeth, found as isolated specimens out of context from the original animal. Imagine you are an alien trying to figure out what a human looks like from a fingernail or molar, and you get a sense of the difficulties. This is why complete fossils are so important. All the thousands of other fragmentary fossils can be put in perspective. Before the new Doliodus fossil, only one other articulated fossil shark from the early Devonian was known, Pucapampella from South Africa. Since Doliodus, an early Devonian Leonodus carlsi has been found. Articulated fossils from the middle Devonian are more abundant. Typical species include Pucapampella from Bolivia, Antarctilamna prisca from Antarctica and Australia and, possibly, Gladbachus adentatus from Germany. What is needed now is a good Silurian example.
The pieces are slowly but surely coming together. The “problematic deceiver” is living up to the name assigned to it more than a century ago, posing as many new questions as it answers. The task ahead is daunting but rich with potential rewards. The first step is to assess the new finds in terms of existing classification schemes. Next, theorists must explore in more detail how existing specimens might demonstrate new aspects of relatedness between the various groups of early fishes. Finally, more fieldwork is needed. Climate and terrain in the areas in which the latest fossils have been found are demanding and even hostile, but we expect that continued exploration will pay off handsomely.