A few selected transitional fossils

Apes - humans | Fish - tetrapods | Dinosaurs - birds

Synapsids - mammals | Land mammals - whales | Protohorses - horses

Miscellaneous | Other | Credits | References

Introduction: All species undergo gradual change over time, but in the fossil record we find evidence of some changes that are particularly striking. This website is dedicated to some of these so-called transitional fossils.

Warning 1: The images are only artist's conceptions and might contain errors; so I keep a page with links to photos or diagrams of the fossils themselves.

Warning 2: When a fossil is called "transitional" between two types of animal, that means it shows some of the traits of both, but it does not mean it links those animals by direct descent. Evolution is a branching process - by which we mean that species often split in two. Therefore:

"Because evolution is a branching process that produces a complex bush pattern of related species rather than a linear process that produces a ladder-like progression, and because of the incompleteness of the fossil record, it is unlikely that any particular form represented in the fossil record is a direct ancestor of any other."

— Rusty Cashman / Wikipedia

In short, transitional fossils are best thought of as being close relatives of the species which actually link two groups. They may have lived at the same time as those actual links, or they may not have (this confuses many people). As long as these problems are borne in mind, transitional fossils give a rough indication of what evolutionary changes were occurring. But don't be misled into thinking that fossils are the only evidence for evolution. They're not even the strongest evidence for evolution.

Apes - humans

Most ape-like at the top (though in a technical sense, humans are still considered apes). Images and diagrams of the fossils here.

Sometimes called "the only transition which matters", this mustn't be thought of as a transition from chimps to humans, but rather, as a transition from the-common-ancestor-of-chimps-and-humans to humans. Chimps themselves have had time to evolve and change since we parted ways, and so "the ancestor we last shared probably differed substantially from any extant African ape" (White et al, 2009).

Ardipithecus ramidus ~4.4 million years ago
Ardipithecus ramidus had a brain the size of a chimp's, but probably walked upright on the ground, while still able to go on all fours in the trees, where it would find its opposable big toe useful (Gibbons, 2009).
Australopithecus afarensis ~3.6 mya
Australopithecus afarensis was a more advanced walker, with nongrasping feet (White et al, 2009), but it still had the brain size of a chimpanzee (Dawkins, 2009). Probably not a direct ancestor of modern humans (Rak et al, 2007).
Australopithecus africanus ~3 mya
Homo habilis ~2 mya?
Homo habilis had a brain about 50% bigger than a chimp's. The fossils are found with a variety of stone tools; this is the earliest human which we're sure used tools (Coyne, 2009).
Homo erectus ~1 mya
A tool-maker, Homo erectus had a brain size of about 1,000 cc, still smaller than our own (Dawkins, 2009).
Homo heidelbergensis ~0.5 mya
Homo heidelbergensis had a brain size approaching our own, and shows a mix of Homo erectus and modern human features (Coyne, 2009).

Fish - tetrapods

Most fish-like at the top. Images and diagrams of the fossils here.

Update: This group of fossils were thought to be roughly contemporary with the transition onto land. However, recently tracks of a four-footed animal were discovered in marine sediments firmly dated at 397 million years old (Niedzwiedzki et al, 2010). If that animal was a genuine tetrapod, then creatures like Tiktaalik may have been "late-surviving relics" exhibiting transitional features that actually evolved somewhat earlier.

In short, these are not the actual ancestors of modern land animals; but they are related to the actual ancestors, and so they do show us the sort of creature that evolved during the great move onto land.

(image) Tetrapods


© Nobu Tamura

Eusthenopteron ~385 million years ago
A pelagic fish, Eusthenopteron is probably representative of the group from which tetrapods evolved. It had a tetrapod-like skull and spine (Prothero, 2007).
Panderichthys ~385 mya
Panderichthys had a tetrapod-like braincase and tetrapod-like teeth, and had also lost its dorsal and anal fins (Prothero, 2007).
Tiktaalik ~375 mya
Though still a water-dweller, Tiktaalik had fins that were halfway towards being feet, and ears capable of hearing in air or water (Prothero, 2007). It was capable of crawling around in very shallow water, and it had a neck, unlike fish but like tetrapods (Coyne, 2009).
Ventastega ~365 mya
The bones of Ventastega are intermediate between Tiktaalik and Acanthostega (Ahlberg et al, 2008). Sadly, the fossil is incomplete and we can't see its fins/feet.
Acanthostega ~365 mya
Possessing four definite legs, Acanthostega was presumably capable of movement over land (Coyne, 2009), though the legs were still better suited for crawling along the bottom of the water (Prothero, 2007). Its tail was still adapted for propulsion through water, and it still had gills (Ridley, 2004).
Ichthyostega ~365 mya
Slightly more like a land animal, Ichthyostega had powerful shoulders implying it did indeed use its legs to move over land, at least sometimes (Clack, 2005). Even now, the skull still closely resembled that of Eusthenopteron (Futuyma, 2005).
Pederpes ~350 mya
The foot of Pederpes "has characteristics that distinguish it from the paddle-like feet of the Devonian forms [i.e. the above animals] and resembles the feet of later, more terrestrially adapted Carboniferous forms" (Clack, 2002).

These creatures were related to the lungfish of their time, and almost certainly all had lungs themselves.

It would be a mistake to think that the first tetrapods moving on land needed limbs capable of bearing their full weight; legs sprawled to the side would be enough to move about with. One thing the above fossils seem to show is that legs first evolved for crawling over the bottom of the water; only later did their use on land become paramount.

Dinosaurs - birds

Most dinosaur-like at the top. Images and diagrams of the fossils here.

Anchiornis ~155 million years ago
Although many feathered dinosaurs are known, Anchiornis is the first to be found that probably predates Archaeopteryx. The feathers were "not obviously flight-adapted" (Hu et al, 2009).
Archaeopteryx ~145 mya
The famous Archaeopteryx had feathers and was probably capable of at least gliding, but it also had dinosaur-like teeth, claws, and a long bony tail. Its skeleton was "almost identical to that of some theropod dinosaurs" (Coyne, 2009). Precisely how closely related it is to the main line of bird evolution remains the subject of controversy (Xu et al, 2011).
Confuciusornis ~125 mya
Confuciusornis had a bird-like tail and a pygostyle, which is a feature of modern birds. It retained dinosaur-like claws (Prothero, 2007). It had strong shoulder bones, but was probably not capable of true flapping flight (Senter, 2006). It may have glided. It is the earliest known bird with a toothless beak, but other lineages continued to have teeth for a long time.
Sinornis ~110 mya?
Sinornis "still had teeth, an unfused tarsometatarsus, and an unfused pelvis" (Prothero, 2007) but resembled modern birds in other ways, with reduced vertebrae, a flexible wishbone, a shoulder joint adapted for flying, and hand bones fused into a carpometacarpus (Prothero, 2007).
Vorona ~80 mya?
The legs of Vorona are all that we have (Benton, 2005), but they show a combination of bird characteristics and maniraptoran (dinosaur) characteristics (Forster et al, 1996).
Ichthyornis ~80 mya
A strong flyer, Ichthyornis was very nearly a modern bird (Prothero, 2007), and yet it still had teeth.

As birds evolved from dinosaurs, and required feathers to fly, the existance of non-flying, feathered dinosaurs is a prediction of evolution. Happily, we have now discovered a significant number of such dinosaurs, one of which, Mei long, was even found curled-up in a remarkably bird-like sleeping position (picture here).

The first feathered dinosaurs found were more recent than Archaeopteryx — feathered dinosaurs didn't die out as soon as birds evolved — but we now have Anchiornis, which has shown that feathered dinosaurs did indeed exist before Archaeopteryx.

One will occasionally encounter vague suggestions that birds evolved from (non-dinosaur) reptiles. John Ruben and colleagues are the main supporters of this idea, but it remains very much a minority view. Michael Benton (who literally wrote the book on vertebrate paleontology) considers the controversy artificially extended by the popular media's obsession with giving both sides "equal time", even though the dinosaur-bird view is supported by the vast majority of working palaeontologists (Benton, 2005). The University of California Museum of Paleontology has a reasonable overview of the dinosaur-bird relationship.

Synapsids - mammals

Most synapsid-like at the top. Images and diagrams of the fossils here.

Strictly speaking, the group that gave rise to mammals were not true reptiles (though they were closely related). Therefore, there is no transition from reptiles to mammals, but rather from synapsids to mammals. However, the terms mammal-like reptiles and reptile-like mammals are still sometimes used for these transitional fossils.

Archaeothyris ~305 million years ago
Mostly lizard-like. However Archaeothyris is one of the earliest known synapsids; a group defined by possession of a single temporal fenestra (Ridley, 2004).
Dimetrodon ~280 mya
Dimetrodon had specialised canine teeth (Prothero, 2008) akin to those of modern mammals.
(image) Lycaenops

© Dmitry Bogdanov

Lycaenops ~260 mya
More mammal-like, especially in how it held its limbs: closer to its body like modern mammals, rather than sprawled to the side like Dimetrodon (Prothero, 2007). It still had a great many "primitive" features, such as ribs in the lumbar area (Prothero, 2007).
Thrinaxodon ~245 mya
Had the beginning of a secondary palate in its skull (Prothero, 2007); in modern mammals, this allows eating and breathing at the same time, and is a sign of a more active lifestyle (Ridley, 2004). Its more advanced skull also allowed it to chew its food; and indeed it had premolars and molars with which to do so (Prothero, 2007). The skeleton was not yet fully mammal-like, but it had lost those lumbar ribs.
Probainognathus ~225 mya?
Probainognathus still possessed a reptile-like jaw articulation (Macdonald et al 2009) but also had "the initiation of the articulation which was later to become the more highly developed glenoid-condyle articulation of the mammal" (Romer, 1969). It had a well developed zygomatic arch (Macdonald et al 2009). However, its braincase was very unlike that of modern mammals (Romer, 1969).
Diarthrognathus ~210 mya?
The fascinating Diarthrognathus had a jaw that contained both the old reptile-like joint as well as the new mammalian joint (Prothero, 2007).

This is merely a small selection of fossils which could be named as transitional synapsids. Wikipedia has an impressive list of therapsids, a subset of the synapsids.

Land mammals - whales and dolphins

Most land-mammal-like at the top. Images and diagrams of the fossils here.

Indohyus ~48 million years ago
Although only a cousin species of the ancestor of whales, Indohyus had bones denser than normal mammals, indicating it was partially aquatic: heavy bones are good ballast (Thewissen et al, 2009). Its ears shared a feature with modern whales: a thickened wall of bone which assists in underwater hearing; non-cetaceans don't have this (Thewissen et al, 2009).
(image) Pakicetus and Ambulocetus

Pakicetus and Ambulocetus
© Sharon Mooney, based on images from National Geographic (see details)

Pakicetus ~52 mya
Perhaps the actual ancestor, Pakicetus was probably semi-aquatic; like Indohyus, it had dense bones for ballast (Thewissen et al, 2009). Its body was "wolf-like" but the skull had eye sockets adapted for looking upwards, presumably at objects floating above it (Thewissen et al, 2009). Although initially known from just a skull, many more bones were found later (Thewissen et al, 2001).
Ambulocetus ~50 mya
With a streamlined, elongated skull and reduced limbs, Ambulocetus probably spent most of its time in shallow water. Its reduced limbs meant it could only waddle on land (Coyne, 2009). It resembled a crocodile in some ways.
Rodhocetus ~45 mya
The nostrils of Rodhocetus have started to move backwards (towards the blowhole position) and the skeleton indicates a much stronger swimmer (Coyne, 2009). On land it would struggle, moving "somewhat like a modern eared seal or sea lion" (Gingerich et al, 2001). Its teeth were simpler than its predecessors (Futuyma, 2005), a trend that continued to the present.
Maiacetus ~47 mya
Seems similar to Rodhocetus. One fossil was found with what appeared to be a foetus, in a position indicating head-first birth (Gingerich et al, 2009) unlike modern whales. However this is disputed; the "foetus" might just be a partially digested meal (Thewissen and McLellan, 2009).
Basilosaurus ~40 mya
The whale-like, fully aquatic Basilosaurus had almost lost its (tiny) hindlimbs, but they had not yet vanished entirely (Prothero, 2007).
Dorudon ~40 mya
Also fully aquatic, Dorudon also had tiny hind limbs, which "barely projected from the body" (Futuyma, 2005).
Aetiocetus ~25 mya
The blowhole in Aetiocetus is about halfway to its position in modern whales on top of the head. Aetiocetus also represents the transition from toothed whales to the filter-feeding baleen whales, being similar to baleen whales in most respects, but possessing teeth (Van Valen, 1968).
(image) Basilosaurus

Basilosaurus © Sharon Mooney, based on an image from National Geographic (see details)

Whales evolved relatively quickly. As Coyne (2009) explains, "adapting to life at sea did not require the evolution of any brand-new features - only modifications of old ones". Thewissen et al (2009) give a good overview of whale evolution, which is freely available.

Beware: There's a commonly reproduced image from a popular science book (Zimmer, 1998) that depicts two dolphin-like creatures, Takracetus and Gaviacetus (always misspelled Gaviocetus) as having short hind limbs. However, the first fossils found were basically skulls plus a few vertebrae (Gingerich et al, 1995), and I've been unable to find any discussion of more complete fossils. Zimmer's book is explicit that some of the depictions are based on very incomplete fossils, so I don't regard that image as authoritative. Happily, we have much better fossils for the species I mention in my list.

Protohorses - horses

Least horse-like at the top. Images and diagrams of the fossils here.

(image) Horse mass graph

Body mass of horse species over time; click for larger; after MacFadden (1986)

With horse evolution, it's particularly important to bear the warning at the top in mind: we have many fossils, and the known family tree is very bushy, not a straight line. At one time, 13 different genera of horses existed simultaneously (Raven et al, 2008). Regardless, there are definite trends over time towards larger body size; larger, ridged teeth suitable for grazing; longer limbs; and reduction of side toes (Raven et al, 2008). These trends were not absolute, however.

Hyracotherium ~60 million years ago
A cousin species of the ancestor of horses. The forelimb of Hyracotherium had four toes (Raven et al, 2008).
Protorohippus ~50 mya
Bigger. The forelimb had four toes.
Mesohippus ~35 mya
Bigger. The forelimb had three toes (Raven et al, 2008).
Miohippus ~35 mya
The skull and snout of Miohippus are becoming more horse-like (Prothero, 2007).
Parahippus ~23 mya
The skeleton of Parahippus was more adapted to long-distance running, for escaping predators in an open environment (Evans, 1992). About this time, grasslands were becoming common in North America, where horses evolved (Raven et al, 2008). They would later die out in America (Dawkins, 2009).
Merychippus ~17 mya
With bigger teeth, Merychippus was more adapted to the grazing lifestyle of modern horses. Earlier species were likely browsers that ate leaves, but Merychippus could also eat grass (Raven et al, 2008).
Pliohippus ~12 mya
Pliohippus still had three toes, but only the central toe touched the ground; the others being too small. This was probably not a direct ancestor of modern horses.
Dinohippus ~5 mya
Some specimens of Dinohippus have three toes; but some have one, like modern horses (Florida Museum of Natural History).

There are a large number of other fossil species that could be mentioned. There are reasonable pages on horse evolution at Tufts University and Wikipedia.


A few other transitional fossils of interest. This is, obviously, not a sequence. Images and diagrams of the fossils here.

(image) Eocaecilia

© Nobu Tamura

(image) Mixosaurus

Mixosaurus © Nobu Tamura

There was a section on ichthyosaurs (a group now extinct) on this site, but I wasn't entirely satisfied with it. We don't yet have anything as good as Tiktaalik or Ambulocetus for ichthyosaurs. Still, fossils like Mixosaurus can be considered transitional between early eel-shaped ichthyosaurs, and later dolphin-shaped species.


This website mentions many of the more famous transitional fossils, but these represent only a tiny fraction of what we know. We have good fossil evidence for a great many other transitions, some of which might be added here in the future.

Acknowledgements / contact

Tetrapods and Mixosaurus images copyright 2007 Nobu Tamura; released under the Creative Commons Attribution 2.5 License. Eocaecilia image copyright 2007 Nobu Tamura; released under the Creative Commons Attribution 3.0 License. Lycaenops image by Dmitry Bogdanov; released into the public domain. Whale images copyright 2006 Sharon Mooney; based on images from National Geographic, November 2001: "The Evolution of Whales" by Douglas H. Chadwick, Shawn Gould and Robert Clark; but re-illustrated for public access distribution by Sharon Mooney.

The author is a biology graduate. To contact me, send a message to any email address at this domain (except "info", which gets too much spam).


Articles behind a paywall are accessible for free from some universities and libraries. In any case, you can usually get at least an abstract.

Last site update: 2013-12-02