First Authors

There's an evolutionary gap between two kinds of primitive bony fish — it separates the sarcopterygian (lobe-finned fishes) from the actinopterygian (ray-finned fishes). Many early sarcopterygian fishes were covered in a hard tissue known as cosmine, which is unknown in any living vertebrates. This tissue has a structure similar to the mammalian tooth (with its pulp cavity, dentine and hard enamel) and the layers are shot through with a network of small canals. Early actinopterygians lack cosmine.

Now, a group of palaeontologists has found fossils that bridge the gap between the two types of fish and explain how they might have evolved. They also illustrate how the enamel structure was formed and eventually progressed to include pore channels. Min Zhu of the Chinese Academy of Sciences and Xiaobo Yu of Kean University, New Jersey, explain what they found from their specimens.

Where and how did you find the fossils?

We found this fish, dubbed Meemannia eos, during 2001 and 2002 field excursions in east Yunnan, in southwestern China. During the expeditions, we had to find the correct rock layers, and use hammers, picks and chisels to split them. Usually we found a mixture of different fossil fish parts. Sometimes, a very bizarre fish immediately caught our attention; at other times we realized the novelty and significance of a find only when preparing it for the lab.

Your paper mentions controversies about the biology of cosmine. What are they?

They include questions about how cosmine arose evolutionarily, and what occupied the cavities of the pore–canal network.

How does your paper help resolve some of these questions?

Our paper shows how the pore network could have been constructed. Meemannia's cosmine suggests that the depth of a pore cavity grows with the deposition of each successive layer of enamel.

What's next?

We expect to find out more about the morphology of Meemannia using further fossils from the site. Its unusual combination of primitive characters suggests that it is close to the common ancestor of the lobe- and ray-finned fishes. We predict that Meemannia will have several other features that reduce the gap between early sarcopterygians and actinopterygians.

Together with materials from ongoing work by palaeontologists in other parts of the world, we are confident that a more complete understanding of the evolution of these major vertebrate groups will emerge. The different pieces of this enticing puzzle are starting to make sense.