Protobalistum imperiale from around 50 million years old of Bolca, Italy.

Like most evolutionary tales, this one could have started on the Galapagos Islands. Instead we find ourselves in an ancient sea, near the end of the Devonian, 360 million years ago. A mass extinction has struck life underwater. The armoured placoderms, once an abundant class of fishes, have gone extinct. Other groups of fishes have been decimated and are struggling to survive. But, as a Dutch saying goes, one man's death is another man's breath. For the ray-finned fishes (fishes whose fins are supported by a ray of spines) this time of trouble is a time of opportunity. With their direct competitors out of the way, they are free to evolve into a multitude of shapes and species, from stream-lined hunters to plump grazers. The fish are dead. Long live the fish!

Fast forward to today. With over 23.000 species alive, ray-finned fishes are the largest and most diverse group of vertebrates of this day. Their rapid evolution after the Devonian mass extinction was the turning point that ensured them their evolutionary success. Biologists have come across similar explosive patterns of diversification across the tree of life, and call them 'adaptive radiation'. Adaptive radiations are evolution’s way of hitting the jackpot. The payout is twofold: a single lineage spins of many new species (speciation) that adapt to diverse ways of life (adaptation).

A species may radiate when it finds itself in an environment where plenty of ecological opportunities await exploitation, such as when it has just colonized an island or lake, or after mass a extinctions. Darwin’s Galapagos finches are the iconic examples of such an adaptive radiation. A single ancestral species arrived to the Galapagos archipelago and split into a dozen species, each one adapted to the local circumstances of its island. The finches with the heaviest beaks eat the largest seeds, whereas those with slender, sharp beaks ones that catch insects.

The beaks of Darwin's finches are adapted to different food sources.

These little finches have been studied in great detail ever since Darwin first set foot on the Galapagos, but biologists still know little about how adaptive radiations unfold. Some think that species and their different shapes evolve in a single burst. This explosive diversity is then followed by periods of relative stability. Others disagree, and think that radiations occur in stages. They argue that new species first adapt to their environment or habitat, by changing their body shape and size, before they adapt to a specific diet or way of life, by changing their skulls and jaws. In this model, wings evolve before beaks, fins before mouths and legs before teeth.

Biologists have argued both ways, but neither side has delivered convincing evidence so far. This is where fossils of ray-finned fish come in. The biggest advantage dead fish have over living finches is that we know both their past and future. This makes it possible to track their radiation through time and see whether their different shapes evolved in steps or not. The fossil record of ray-finned fish is rich, and they underwent multiple adaptive radiations. For the first time after the Devonian mass extinction, and a second time at the end of the Cretaceous period (around 65 million years ago), when a massive asteroid struck earth and killed off many species of animals, including the dinosaurs.

Lauren Sallan and Matt Friedman investigated both these radiations by digitalizing the shapes of 69 Devonian fish and 304 fish from the Cretaceous. They first mapped several landmarks onto their skulls and skeletons, such as the positions of their jaw joints and fins. In the next step they determined which axes of these landmark maps explain the major differences between fish. They then analyzed how these differences changed through time.

Ray-finned fish evolved their heads before tails. Skull diversity starts to increase right after the Devonian (orange range), in the Tournaisian (darkest green). The diversity of body shapes lags behind.

Sallan and Friedman found that for fish from both era, heads evolved before tails. The heads of Devonian fish started to diversify right in the aftermath of the mass extinction. Some skulls became longer and flatter, while skeletons lagged behind. Only after a couple of million years did some evolve the shape of flat spades, in addition to the classical torpedo-shape. The Cretaceous fish also went through a head-first phase. Their skulls grew more elongated and streamlined before the main extinction event at the end of the Cretaceous, and long before their bodies followed suit.

This head-first trend in the evolution of ray-finned fish contradicts the biological big bang model of adaptive radiations, and it is a direct reversal of the idea that radiating species first adapt to their habitats. So why would fish evolve their skulls before anything else? The answer seems simple: to bite, crush, rip, nibble and suck. Certain niches might have been left vacant after the Devonian and Cretaceous mass extinctions, and ray-finned fishes evolved the jaws to exploit them. Large predatory fish died out after the Cretaceous extinction for example, making room for creatures such as the sword-fish like Blochius to evolve.

But skulls did not only evolve earlier, they also reached their peak diversity in a shorter span of time than body shapes did. This suggests that it's also easier to evolve most variations on a skull than it is to evolve most body forms.

Blochius was a swordfish that lived 56 to 34 million years ago, in what is now Italy.

Friedman and Sallan are careful to generalize their findings to other adaptive radiations. "The real world is more complicated than any model", is what they write in their final paragraph. A lot more studies need to be done before either theory can be discounted. That said, Sallan thinks the head-first model has the potential to explain a large portion of adaptive radiations. "There's anecdotal evidence for many groups: lungfishes, sharks, birds, mammals, insects and even worm lizards", she says. "And in a way it makes sense. When an animal is faced with limited food relative to the total population size, which is likely in a successful group, it has two choices. It can either find a new resource, or move to a new habitat and hope the same resource is there. Changing diets in probably easier and more likely to be successful."

Do these findings mean that shifts in behaviour are the ultimate drivers of adaptive radiations? A bird first has to change its diet before it can change its beak, after all. Sallan thinks this might be the case. "Animals can be plastic in what they eat", she says. "You hear about deer eating squirrels, squirrels eating birds, etcetera. Marginal dietary behaviors could turn out to be beneficial and some individuals might be better at exploiting a new food source than others, due to variation already present in the population. Directional selection then takes hold. So basically, you never know until you try!"


Fossil fish by Lauren Sallan and Matt Friedman.

Darwin's finches by Gould.

Change in head/body shape from first reference.


Lauren Cole Sallan, & and Matt Friedman (2011). Heads or tails: staged diversification

in vertebrate evolutionary radiations Proceedings of the Royal Society B : 10.1098/rspb.2011.2454

Gavrilets S, & Losos JB (2009). Adaptive radiation: contrasting theory with data. Science (New York, N.Y.), 323 (5915), 732-7 PMID: 19197052