That's not hyperbole. I really mean it. How else could I react when I open up the latest issue of Bioessays, and see this: Cephalopod origin and evolution: A congruent picture emerging from fossils, development and molecules. Just from the title alone, I'm immediately launched into my happy place: sitting on a rocky beach on the Pacific Northwest coast, enjoying the sea breeze while the my wife serves me a big platter of bacon, and the cannula in my hypothalamus slowly drips a potent cocktail of cocain and ecstasy direct into my pleasure centers…and there's pie for dessert. It's like the authors know me and sat down to concoct a title where every word would push my buttons.
The content is pretty good, too. It's not perfect; the development part is a little thin, consisting mainly of basic comparative embryology of body plans, with nothing at all really about deployment of and interactions between significant developmental genes. But that's OK. It's in the nature of the Greatest Science Papers Ever Written that stuff will have to be revised and some will be shown wrong next month, and next year there will be more Greatest Science Papers Ever Written — it's part of the dynamic. But I'll let it be known, now that apparently the scientific community is aware of my obsessions and is pandering to them, that the next instantiation needs more developmental epistasis and some in situs.
This paper, though, is a nice summary of the emerging picture of cephalopod evolution, as determined by the disciplines of paleontology, comparative embryology, and molecular phylogenetics, and that summary is internally consistent and is generating a good rough outline of the story. And here is that story, as determined by a combination of fossils, molecular evidence, and comparative anatomy and embryology.
Cephalopods evolved from monoplacophoran-like ancestors in the Cambrian, about 530 million years ago. Monoplacophorans are simple, limpet-like molluscs; they crawl about on the bottom of the ocean under a cap-like shell, foraging snail-like on a muscular foot. The early cephalopods modified this body plan to rise up off the bottom and become more active: the flattened shell elongated to become a cone-like structure, housing chambers for bouyancy. Movement was no longer by creeping, but used muscular contractions through a siphon to propel the animal horizontally. Freed from its locomotor function, the foot expanded into manipulating tentacles.
These early cephalopods, which have shells common in the fossil record, would have spent their lives bobbing vertically in the water column, bouyed by their shells, and with their tentacles dangling downward to capture prey. They wouldn't have been particularly mobile — that form of a cone hanging vertically in the water isn't particularly well-streamlined for horizontal motion — so the next big innovation was a rotation of the body axis, swiveling the body axis 90° to turn a cone into a torpedo. There is evidence that many species did this independently.
The tilting of the body axes of extant cephalopods. This was a result of a polyphyletic and repeated trend towards enhanced manoeuverability. The morphological body axes (anterior-posterior, dorso-ventral) are tilted perpendicularly against functional axes in the transition towards extant cephalopods.
We can still see vestiges of this rotation in cephalopod embryology. If you look at early embryos of cephalopods (at the bottom of the diagram below), you see the same pattern: they are roughly disc-shaped, with a shell gland on top and a ring of tentacle buds on the bottom. They subsequently extend and elongage along the embryonic dorsal-ventral axis, which becomes the anterior-posterior axis in the adult.
In extant cephalopods the body axes of the adult stages are tilted perpendicularly versus embryonic stages. As a con- sequence, the morphological anterior-posterior body axis between mouth and anus and the dorso-ventral axis, which is marked by a dorsal shell field, is tilted 908 in the vertical direction in the adult cephalopod. Median section of A: Nautilus, B: Sepia showing the relative position of major organs (Drawings by Brian Roach). C: shared embryonic features in embryos of Nautilus (Nautiloidea) and Idiosepius (Coleoidea) (simplified from Shigeno et al. 2008  Fig. 8). Orientation of the morphological body axes is marked with a compass icon (a, anterior; d, dorsal; p, posterior; v, ventral; dgl, digestive gland; gon, gonad; ngl, nidamental gland).
The next division of the cephalopods occurred in the Silurian/Devonian, about 416 million years ago, and it involved those shells. Shells are great armor, and in the cephalopods were also an organ of bouyancy, but they also greatly limit mobility. At that early Devonian boundary, we see the split into the two groups of extant cephalopods. Some retained the armored shells; those are the nautiloids. Others reduced the shell, internalizing it or even getting rid of it altogether; those are the coleoids, the most successful modern group, which includes the squids, cuttlefish, and octopuses. Presumably, one of the driving forces behind the evolution of the coleoids was competition from that other group of big metazoans, the fish.
The nautiloids…well, the nautiloids weren't so successful, evolutionarily speaking. Only one genus, Nautilus has survived to the modern day, and all the others followed the stem-group cephalopods into extinction.
The coleoids, on the other hand, have done relatively well. The number of species have fluctuated over time, but currently there are about 800 known species, which is respectable. The fish have clearly done better, with about 30,000 extant species, but that could change — there are signs that cephalopods have been thriving a little better recently in an era of global warming and acute overfishing, so we humans may have been giving mobile molluscs a bit of a tentacle up in the long evolutionary competition.
There was another major event in coleoid history. During the Permian, about 276 million years ago, there was a major radiation event, with many new species flourishing. In particular, there was another split: between the Decabrachia, the ten-armed familiar squid, and the Vampyropoda, a group that includes the eight-armed octopus, the cirroctopodes, and Vampyroteuthis infernalis. The Vampyropoda have had another locomotor shift, away from rapid jet-propelled movement to emphasizing their fins for movement, or in the case of the benthic octopus, increasing their flexibility to allow movement through complex environments like the rocky bottom.
Time for the big picture. Here's the tree of cephalopod evolution, using dates derived from a combination of the available fossil evidence and primarily molecular clocks. The drawings illustrate the shell shape, or in the case of the coleoids, the shape of the internal shell, or gladius, if they have one.
A molecularly calibrated time-tree of cephalopod evolution. Nodes marked in blue are molecular divergence estimates (see methods in Supplemental Material). The divergence of Spirula from other decabrachiates are from Warnke et al. , the remaining divergences are from analyses presented in this paper. Bold lineages indicate the fossil record of extant lineages, stippled lines are tentative relationships between modern coleoids, partly based on previous studies [41, 76, 82] and fossil relationships are based on current consensus and hypoth- eses presented herein. Shells of stem group cephalopods and Spirula in lateral view with functional anterior left. Shells of coleoids in ventral view with anterior down. The Mesozoic divergence of coleoids is relatively poorly resolved compared to the rapid evolution of Cambro- Ordovician stem group cephalopods. Many stem group cephalopod orders not discussed in the text are excluded from the diagram.
The story and the multiple lines of evidence hang together beautifully to make a robust picture of cephalopod evolution. The authors do mention one exception: Nectocaris. Nectocaris is a Cambrian organism that looks a bit like a two-tentacled, finned squid, which doesn't fit at all into this view of coleoids evolving relatively late. The authors looked at it carefully, and invest a substantial part of the review discussing this problematic species, and decided on the basis of the morphology of its gut and of the putative siphon that there is simply no way the little beast could be ancestral to any cephalopods: it's a distantly related lophotrochozoan with some morphological convergence. It's internal bits simply aren't oriented in the same way as would fit the cephalopod body plan.
So that's the state of cephalopod evolution today. I shall be looking forward to the Next Great Paper, and in particular, I want to see more about the molecular biology of tentacles — that's where the insights about the transition from monoplacophoran to cephalopod will come from, I suspect.
Kröger B, Vinther J, Fuchs D (2011) Cephalopod origin and evolution: A congruent picture emerging from fossils, development and molecules: Extant cephalopods are younger than previously realised and were under major selection to become agile, shell-less predators. Bioessays doi: 10.1002/bies.201100001.