The fossil record is one of the most obvious pieces of evidence for evolution. Fossils have been known since human history began, and Aristotle first noted the similarity between fossils and living animals, leading him to conclude that fossils represented deceased creatures, a view supported by Leonardo da Vinci. By the 19th century, people were increasingly beginning to appreciate that some fossils represented extinct animals, and that their positioning in the rocks appeared to represent that passage of time. Now there is also ample evidence from molecular genetics and radiometric dating that the fossil record does represent the evolutionary history of life on Earth. Despite this, creationists still continue to argue otherwise. Some have suggested that the fossil record represents animals killed during one or more biblical floods, however this is incongruent with the evidence available.
The word fossil comes from the Latin fossus, meaning “having been dug up”, and refers to the remains or traces of any living thing, older than 10,000 years. Fossils usually only preserve certain parts such as bones, seeds, teeth and shells; parts that were mineralised prior to death. Trace fossils, by contrast, are fossilised marks that were created while the animal (or plant) was alive, such as a footprint. Fossilisation is extremely rare because it requires very specific conditions. Usually, animal remains will very quickly decay, and be broken apart by scavengers. Even bones can eventually be worn down by erosion before fossilisation can occur. Fossilisation is most likely if an animal is buried very soon after death, preferably in sediment such as sand or mud. Over time, more sediment is laid down on top of the remains, causing compression. Compression can eventually turn sand into rock, encasing the remains inside. Minerals and chemicals cycle through the rock, and the bone, slowly changing it into rock. In other cases the bone may recrystalise and disappear completely, leaving only a hollow outline of it’s original shape.
Even once fossilised, the remains aren’t certain to be included as part of our ‘fossil record’. They must be discovered. But fossils are also being destroyed all the time, as geological events such as tectonic movement and subduction of the Earth’s crust back into the mantle. Sometimes rock can be eroded away, and while this provides an opportunity for it’s discovery, it also opens the fossil back up to the elements. If left undiscovered the fossil will eventually erode as well.
A Slight Lack of Impartiality
The up-shot of all this is that the fossil record is biased in a number of ways, and these biases are inherent in the way the fossilisation occurs. The fossil record is biased towards hard-bodied animals, and also to aquatic animals, as there is more sediment available in a sea or river bed; an ideal substrate for fossilisation. Long-lived, geographically restricted organisms are less likely to be represented in the fossil record because they are generally fewer in number, and tend to have longer generation times. Larger creatures are more likely to be preserved because they can survive coarser-grained sediments, and are more likely to be found purely as a function of their size. It is thought that 99.99% of all life that has ever existed is not represented in the fossil record. Nevertheless, there is still a lot we can deduce from the 0.01% that we have discovered. It is not necessary to study every species, in order to deduce the key characteristics of the group to which it belongs.
Prior to the Cambrian Explosion (542 million years ago), most life on Earth was entirely soft-bodied. Despite the heavy bias against the fossilisation of soft-bodied creatures, we have still been able to learn about the earliest life from the fossil record. Trace fossils are known from around 560 mya onwards, primarily in the form of tubular borrows in made by presumably worm-like creatures in the sediment. These trace fossils increased in complexity slowly in the build up to the explosion of new forms that appeared in the Cambrian. Further back still, there are biochemical ways in which we can detect the presence of life from biochemical signatures left behind; evidence of the metabolism of our earliest ancestors. These signals and evidence for very early life, combined with the molecular clock, date the appearance of life to around 3.5 billion years ago.
When it was observed that certain fossils were associated with certain rock strata, an obvious explanation was that different rock strata were of a different age, and thus the geological timescale was recognised. The pattern of fossils in the rocks is very distinctive; in different rock strata there are different collections of species. Wherever in the world you look, these patterns are present, and they are consistent across different regions. Index species are used to define certain regions of the rock strata. Abundant, short-lived (in evolutionary time) species, are most useful for dating rock strata because they are numerous and they can eliminate all but a small window of time in which that rock must have formed, in order to harbour that species.
The Geological Timescale
Not only is the geological timescale internally consistent, it is also consistent with the geographical positioning of continents in the past. Furthermore, when radiometric dating was invented in the 20th century, these dates further corroborated the sequential nature of rocks. Radiometric dating is based on the decay radioactively unstable isotopes, which are quite abundant in nature. Over time, radioactive isotopes decay into a more stable form by losing electrons, and the length of time it takes for 50% of these atoms to decay is known as the half-life of a radioactive isotope. Different isotopes have different half-lives, and by comparing the ratio of radioactive to stable isotopes in rock or organic material with modern conditions, we can calculate its age.
Carbon-14 has a half-life of 5,700 years, becoming Nitrogen-14 (a stable isotope) through beta decay. This means it can be useful for dating any organic material, such as bone or wood up to about 60,000 years old. Radiocarbon dating can be used to directly date remains because animals incorporate carbon (in all available isotopes; 11, 12, 13 and 14) into their bodies during life. When they die, no more carbon enters into the body, and as the radioactive isotopes decay the ratio of radio- to stable-isotopes changes. Uranium-thorium dating can be used to date rocks, cave deposits, soil, bone, coral and wood up to 500,000 years old, and is based on the decay of Uranium-238 into Thorium-230. Potassium-Argon and Uranium-Lead dating allow us to accurately date rocks of between 100,000 years old and 4.5 billion (when Earth was formed). The problem with The last three options is that they cannot be used to date fossils directly. Their use comes from dating layers of igneous rock, formed by volcanic activity, that sandwich fossils in between. Radioactive isotopes such as Potassium-40 and Uranium-238 are trapped inside molten rock as it cools to form igneous rock, and can therefore be measured to allow radiometric dating. By contrast, sedimentary rock is formed from the weathering of igneous rocks. Although radioactive decay could be measured for these rocks, it would not be informative about when the sediment (and therefore the fossil) was laid-down, but rather when the original igneous rock was formed. However, by dating the nearest igneous layer below and above a fossil-containing layer it is possible to determine an age-range for the fossil layer.
The DNA Double Helix
Evidence from radiometric dating is also broadly supported by an even newer technology; genomics. The advent of technology that allowed us to sequence first genes and then genomes gave us the opportunity to study evolution at the molecular scale, and lead to the concept of the molecular clock. The molecular clock is a method based on the idea that, at least for certain genes and regions of the genome, evolution (as measured by changes in the nucleotide sequence of the DNA code) is relatively constant. Thus measuring the amount of difference between the genome sequences (or smaller regions, which were commonly used prior to modern genome sequencing technology) of two related species can be used to estimate how recently their common ancestor lived – ie the point at which their ancestor diverged into two separate species. For the majority of major transitions, the molecular clock and fossil record estimated dates agree well. The agreement is never perfect however, but we wouldn’t expect it to be. The two methods measure slightly different things; molecular clocks measure the divergence point itself, whereas the fossil record indicates the latest possible date for the appearance of a distinct new species. Furthermore, both methods are subject to a certain degree of inaccuracy, caused by various unavoidable methodological issues such as preservation bias in the fossil record and difficulties of separating time and evolutionary rate for molecular techniques. There are a few cases in which the fossil record and molecular clocks disagree substantially on the dating of specific divergence times. In some cases this has an obvious explanation, such as the preservational biases in for the fossilised evidence of the early soft-bodied Scleractinian corals. In other cases the explanation is less clear. Nevertheless, there is a broad agreement between the two independent measures of evolutionary time.
However, all this misses the point slightly – it doesn’t matter so much for the theory of evolution itself whether the dates agree precisely – what counts is that the progression is consistent. ie that all methods of measuring divergence dates agree that certain species evolved after others, regardless of the exact timing of this evolution. This condition is necessary to support the theory of faunal turnover, and evolution itself. JBS Haldane famously said, when asked what evidence would falsify the theory of evolution, “a fossilised precambrian rabbit”. Finding a mammalian fossil prior to the apperance of the vast majority of animal phyla in the Cambrian explosion would certainly be extremely difficult to rectify with the theory of evolution. However, no such fossil has been found. There is no evidence of Precambrian rabbits.
A further criticism sometimes used against the fossil record as evidence for evolution is the apparent lack of transitional fossils. A transitional fossil is one that exhibits traits which were found in the ancestral species, and traits of the descendant species. Technically, any fossil could be considered transitional; the concept of a species is an artificial human construct, born of our need to categorise things in order to understand them.
Nevertheless, there are plenty of examples of transitional fossils. We now know of more than 5 separate species of hominid ancestor, including (but certainly not limited to) Ardipithecus ramidus, Australopithecus afrarensis, Homo habilis, Homo erectus, etc. These species show a gradual transition from aboreal to bipedal life, a gradual increase in brain size and changes in sexual dimorphism. Further back in evolutionary history, there are transitional fossils known for the move from water onto land, and evolution of the pentadactyl limb. Eustrenopteron was the first fish to possess internal nostrils, found only in land-mammals and lobe-finned fish, whist Panderichtyshad large, leg-like front fins and showed the begins of a transition from pectoral to pelvic dominance in controlling locomotion. Even closer to the move to land, Acanthostega appeared around 365 million years ago, and clearly possessed four limbs with webbed digits. These species were closely related to the lungfish of their time, and almost undoubtedly had lungs themselves. Although some transitional fossils most likely do not represent the actual ancestor of their modern counterparts, even the record of a sister species is greatly informative in understanding the ancestor itself.
There is also ample fossil evidence for the transition from dinosaur into bird; famously Archaeopteryx, a winged (but probably not flying) dinosaur with teeth, but also interesting was the lesser known,Sinornis, whose pelvis and shoulders were adapted for flying. The evolution of mammals is illustrated by Dimetrodon and Diarthrognathus. Diarthrognathus was particularly fascinating because it possessed both the ancestral reptilian jaw joint as well as a new mammalian joint. The return to the sea of land mammals that gave rise to whales and dolphins is demonstrated by the semi-aquatic Pakicetus and fully-aquatic Basilosaurus. Rather pleasingly, there are also transitional fossils for flatfish, as they developed from normal fish; a move that require one eye to move to the opposite side of their face. Amphistium paradoxum had eyes which are half-way between a modern flatfish and a normal fish, with one eye at the top of the head. Finally, fossils such as Najash show the gradual loss of legs from the ancestors of snakes.
Evidence for Evolution
The fossil record is not a perfect record of every animal that ever lived. The geological processes that drive fossilisation inherently make it biased towards large animals with a short-generation time and a bony interior (or exterior!). Fossilisation is so rare that many species are never represented in the fossil record at all. Many others are simply never discovered. Despite this, the fossils we are aware of are a powerful tool for understanding the evolution of life on Earth. The sequential nature of the geological record is supported by radiometric dating and genomic techniques, and the occurrence of fossils within the rocks is both internally and externally consistent. Transitional fossils are a reminder that species are not fixed but constantly changing and evolving, and our choice of where to draw the lines between species is relatively arbitrary. Criticisms of the fossil record as evidence for evolution are born out of a misunderstanding of the criticism itself or of the theory of evolution. Usually both.
Articles in this Series:
- Intro: Reasons Why Evolution is True
- Part One: The Panda’s Thumb
- Part Two: Parasitoid Wasps
- Part Three: Ring Species
- Part Four: Galapagos Finches
- Part Five: The Quirky Human Eye
- Part Six: Homology
- Part Seven: Coevolution
- Part Eight: PreCambrian Rabbits
- Part Nine: DIY Evolution
- Part Ten: Convergent Evolution
Want to Know More?
- BBC Nature: Fossils
- Benton (2005) Fossil Record: Quality Encylopedia of Life Sciences, John Wiley & Sons Ltd.
- Smith and Peterson (2002) Dating the Time of Origin of Major Clades: Molecular Clocks and the Fossil Record Annu. Rev. Earth Planet. Sci (30) 65 – 88<
- Jensen (2003) The Proterozoic and Earliest Cambrian Trace Fossil Record; Patterns, Problems and Perspectives Intergr. Comp. Biol (43) 219 – 228
- Crimes (1987) Trace fossils and correlation of late Precambrian and early Cambrian strata Geological Magazine (124) 97 – 119
- Transitional Fossils
- Earth Facts: Fossil Dating
- Tulane University, Earth Materials: Radiometric Dating
- Radiometric Dating and the Geological Timescale
Featured image used under a creative commons licence from Wikimedia commons. Image by Günter Bechly