Now is the winter of our fish content

once again the dreaded lurgy has struck our family this winter with Anwen needing a stay in hospital. So this post and the next few to follow were supposed to be posted a month ago.

Teleost fish outnumber all other modern vertebrates two to one. Despite this staggering diversity it is accurate to call fish palaeontology the poor cousin of amniote palaeontology, particularly when it comes to grabbing publicity. Nevertheless with such a huge diversity it is not surprising that the clade has thrown up more than a few subgroups that do grab public attention. What is surprising is that the austral winter’s edition of Journal of Vertebrate Paleontology (volume 29, number 2) carries articles on no less than three of these attention grabbing teleost groups.
One is the giant ocean-going sunfish (Molidae). These giant jellyfish-suckers include the largest living bony fish but due to their poorly developed skeletons and pelagic habits have left a scrappy fossil record. Thus the finding of three articulated skeletons that exceed modern sunfish in size (one skeleton reaches 4 metres from fin tip to fin tip) is a big deal.
Another curious fossil fish reported in this issue is a deep-sea anglerfish (Ceratoidea). These bizarre fish are well known for their disproportionately large mouths lined with needle-like fangs and the ability attract prey with a luminescent lure. Finding a fossil of one of these is also quite unusual for the ceratoid clade is not all that old by geological standards. There simply aren’t that many places where such recent sediments have been laid down at great depth but have since been brought to a position above sea-level where someone might find any fossils that they might contain.
Both of these discoveries are blogworthy finds, however it is the fossil piranha that I want to highlight in this post (Cione et al. 2009). Piranhas are, of course, the fabled freshwater fish of South America that are said to be able to skeletonise a cow in a matter of minutes when a school is whipped up into a feeding frenzy.
The fossil piranha was found late Miocene (about 5 to 11 million years old) river sediments in northeastern Argentina. It has a concave dorsal margin of the premaxilla and tall, sharp triangular teeth that indicate that its affinities lie with the piranhas among the serrasalmids. Aptly named Megapiranha the fish is immediately striking for its great size. Known from a single jaw bone (the premaxilla) and some isolated teeth it is about two and a half times larger than the premaxilla of an average modern piranha. Assuming similar proportions to a modern piranha may have approached a meter in length.

A modern piranha against a silhouette scaled up to fit the size of the Megapiranha premaxilla.

Now the thought of a school of those getting into a feeding frenzy is worthy of any Hollywood B-grade creature feature. However it is far from certain that Megapiranha would have indulged in such hypercarnivorous behavior. It should be noted that piranhas form a clade of closely related species amongst a broader family of fish known as Serrasalmidae. Most serrasalmids (for example pacus) and even some piranhas are vegetarian, indicating that the herbivory is the ancestral diet of serrasalmids. Given the primitive position of Megapiranha (the sister group of all other known piranhas) it is quite likely that Megapiranha was at least partly vegetarian.
What makes Megapiranha interesting, other than its size, is that it gives us some idea how the piranhas evolved their famous dentition. Primitive herbivorous serrasalmids have seven rounded, flat-topped teeth arranged in two rows. In contrast piranhas have a single row of six double-cusped, blade-like teeth. One would expect that the single rowed condition evolved from the double rowed condition by simple suppression of one of the rows, most likely the inner row which contains just two teeth. An alternative, proposed by Gosline (1951) is that the two rows integrated to become one. Megapiranha provides evidence that supports Gosline’s hypothesis, for it shows just a single tooth row but with the teeth placed in a staggered arrangement as if two rows were merging.

The premaxillae of a pacu (top), Megapiranha (middle), and a modern piranha (bottom)in lateral (left) and ventral (right) views. Scale bars equal 1 cm. Images from Cione et al. 2009.

The teeth themselves are intermediate as well for although they bear tall, sharp-edged triangular cusps like modern piranhas, there is no secondary cusp and the bases are broad, perhaps supporting the idea that Megapiranha was not a hypercarnivore. It is a pity really, I find the idea of a giant ground sloth or an astrapothere being stripped to its bones by a school of meter-long piranhas somehow appealing.

References

Cione, A.L., Dahdul, W.M., Lundberg, J.G. & Machado-Allison, A.(2009) Megapiranha paranensis, a new genus and species of Serrasalmidae (Characiformes, Teleostei) from the upper Miocene of Argentina. Journal of Vertebrate Paleontology 29: 350-358.

Gosline, W. (1951) Notes on the Characid fishes of the subfamily Serrasalminae.
Proceedings of the California Academy of Sciences 27: 17–64.

Finishing Fish Fortnight

Ok, this is way late but you know how the real world has a habit of intruding on blogging time. Anyway this drawing is one of mine. It depicts what is perhaps the wierdest coelacanth known, Allenypterus montanus, from the Carboniferous of Montana. The bright red coloration was inspired by the resemblance between Allenypterus and the modern Pataecus fronto, a wierd ray-finned fish that is also bright red. There is a lot more I could say about this guy and its relatives - but not today.

Worst case of mistaken identity since Aachenosaurus!

OK on with Fish fortnight,
Aachenosaurus is an infamous a case of misidentification where some pieces of petrified wood were mistaken for dinosaur bones, and a name was coined for them in the literature. The small pieces of bone in the photos above were also strikingly misidentified in the BPI catalogue as belonging to a dinosaur (though fortunately never published as such). They actually belong to a fish, a ray-finned fish (actinopterygian) to be a little more precise. As fish they are very interesting because the come from the upper Elliot Formation, and as far as I can tell are the first recognized ray-finned fish from this unit (lungfish are known from the odd small toothplate here and there). The upper Elliot Formation was deposited in arid conditions with most of the streams being small and ephemeral. Nonetheless ray-fins can’t cocoon themselves when their pond dries up the way some lungfish can and their presence indicates that some permanent water bodies, however small, existed on the upper Elliot floodplain. That’s all I can say about this fish right now - the fossil will be subjected to further prep and study.

Misleading Mitochondria and Ancient Neopterygian Fossils

Last post we looked at the basics of ray-finned fish classification and some of the problems associated with them. Foremost among these are two rather dramatically different topologies. Morphology supports a clade called the Neopterygii which includes ginglymods, halecomorphs and teleosts whereas mitochondrial genetics support an ‘Ancient Fish Clade (AFC)’ that groups chondrosteans, ginglymods and halecomorphs to the exclusion of teleosts. Divergence dates based on the molecular clock are also dramatically older than minimal dates based on the fossil record. Hurley et al. (2007) tackle both problems with a two-pronged approach. Firstly they relook at the early ray-finned fossil record, scrutinizing it for the first appearance of derived characters diagnostic of these major groups and incorporating the data into a new cladistic analysis. Secondly they assembled a new, comprehensive molecular data set of four nuclear genes 29 species covering all the relevant clades (except the cladistians), thus is the first analysis capable of addressing the timing of the whole genome duplication event.
The tree based on the morphological data found strong support for the Neopterygii and virtually no support for the ‘Ancient Fish Clade’ at all. Indeed when the AFC topology was enforced upon an analysis that included only the living taxa the tree length grew by 80 steps (125% of the number of steps in the shortest possible tree) and found just one character that could be interpreted as a synapomorphy of this clade. Clearly the morphology doesn’t just fail to support the molecular ‘AFC’ it is strongly contradicting it. Analysis of the nuclear gene data also strongly supports the neopterygian clade over the AFC. Thus the signal for the AFC is coming from the mitochondrial genes alone. Given that this data set is so at odds with morphology, nuclear genes and the fossil record it seems likely that the source of error is the mitochondrial data. Perhaps more interesting is the morphological analysis that includes the fossils. Neopterygii continues to be strongly supported but the divergence date estimates have changed. Two fossils in particular were found to be significant: Brachydegma and Discoserra. Brachydegma from the Early Permian (285 million years) of Texas was previously regarded as a basal actinopterygian that diverged before the chondrostean-neopterygian split. However Hurley et al. found that it had a number of characteristics of Halecomorpha (that is the bowfin and its fossil relatives), such as an enlarged gular plate, a medial shelf at the front end of the maxilla, and possibly a posteriorly indented maxilla. The latter character is less secure because it depends on the interpretation of a small elliptical patch of differing ornament on the rear edge of the maxilla. If this patch is interpreted as a fused-on scale, then the maxilla does have the classic halecomorph indented maxilla (see figure below).



Brachydegma (from Hurley et al. 2007) on the left and the modern bowfin (Amia) on the right (not to scale; from Grande and Bemis 1998). Two diagnostic features of the Halecomorphi are colourised – the posteriorly indented maxilla (red) and the very large gular plate (green).

Sure enough Brachydegma comes out as the basal most member of the Halecomorphi in their analysis. As an halecomorph, Brachydegma is part of the neopterygian crown-group and pushes the origin of this clade back the Palaeozoic Era, before the big extinction event at the end of the Permian Period. The previous oldest known crown-group neopterygians were the parasemionotids (also halecomorphs) from the Early Triassic of Madagascar and Greenland. Discoserra from the Early Carboniferous (320 million years) of Montana (the famous Bear Gulch Limestone fish deposits) is a far older fish, and is not apparently a member of the neopterygian crown group but it has many of the synapomorphies of the crown-group indicating that by this early stage the neopterygian bodyplan was mostly in place. Previously Discoserra was thought to be an early cladistian.



Discoserra, from Lund 2000.

The new fossil data places the origination of the neopterygian crown group into the Paleozoic Era, long before the big Permo-Triassic mass extinction event of 251 million years, and at least 40 million years earlier than the previous oldest crown-group neopterygian and somewhat closing the gap between the molecular and palaeontological dating of the Neopterygian crown-group origination. Furthermore molecular clock dating using the nuclear gene data, rather than the mitochondrial genes yields a more recent range of dates 271-371 million years that actually ecompasses the age of Brachydegma, thus the discrepancy is more or less resolved. Once again it appears that the mitochondrial genes are giving misleading results, but why this is so is not immediately clear. No particularly ancient crown-group teleosts were recognized in this study so the fossil based minimum age for this clade is unchanged. However the new nuclear genetic data were used to estimate the divergence of the crown group. Like the estimates for the age of the neopterygian crown-group the estimates for the teleost crown-group based on nuclear genes are considerably younger than the estimates based on mitochondrial data. Thus using these new age estimates the discrepancy between molecular and palaeontological dates closes to a minimum of 30 million years.
Lastly Hurley et al. look at the timing of the whole genome duplication and how it relates to the explosive radiation of teleost fishes. The duplication event is indeed shown to be a feature of the teleost stem, as its products appear to be present in all living teleosts but are not found in the other surviving actinopterygian groups. This suggests that the duplication event most likely happened sometime in the Permian or Triassic, with the most recent possible occurrence being the Mid Jurassic, before the first appearance of crown-group teleosts in the fossil record. However the explosion in teleost diversity does not get underway until the Late Cretaceous, demonstrating a considerable lag between the duplication event and the rapid diversification event. This falsifies the hypothesis that the duplication was the direct causative agent of diversification.
In conclusion, this is a case where the morphological signal appears to have given a more reliable estimate of phylogeny than mitochondrial genes, whereas the huge discrepancy in divergence date estimates were a product of both overlooked fossil data and misleading signal based on the same unreliable genetic data.

Grande, L. and Bemis, W.E. (1998) A comprehensive phylogenetic study of amiid fishes (Amiidae) based on comparative skeletal anatomy. An empirical search for interconnected patterns of natural history. Society of Vertebrate Paleontology Memoir 4: 1-690.

Hurley, I.A., Lockridge Mueller, R, Dunn, K.A.,Schmidt, Friedman, M., Ho1, R.K., Prince, V.E., Yang, Z., Thomas, M.G. and Coates, M.I. (2007)A new timescale for ray finned fish evolution. Proc. R. Soc. B 274, 489–498

Lund, R. (2000) The new actinopterygian order Guildayichthyiformes from the Lower Carboniferous of Montana (USA). Geodiversitas 22, 171-206.

Problems in Ray-Finned Evolution.

I’ve touched on the tension between genetic and morphologic data in this post and the comments it attracted. As I stated then both sets of data are the products of a single history and should therefore be more or less in accord. When they are not, well then it is our job as scientists to find out why. Gone are the days of simply giving a hurumph and declaring one’s personally preferred set of data to be correct and the other erroneous. I want to feature a 2007 paper by Hurley et al. as an excellent model for the unification of both sources of data. Although a few years old now, it remains one of my favourite bits of palaeontological research. And it doesn’t even feature dinosaurs. No, it is about the phylogeny of ray-finned fish. But before I can get to the paper in question I will need to bring my readers up to speed with the relevant issues in this post before turning my attention to the paper itself in my next post.
For those not aquainted with fish taxonomy, I’ll first give a quick rundown of the groups involved. All vertebrates that replace their cartilaginous inner skeletons (endoskeleton) with bone are known as Osteichthyes (literally “bony fish”). Note that not all fish with bone are osteichthyans. Several extinct groups like placoderms and osteostracans covered their bodies in large plates of bone. However this bone is derived from skin (ectodermal) tissue and remains superficial to the endoskeleton which remains cartilaginous. Also note that we tetrapods also replace our endoskeletons with bone: we are bony fish (albeit highly terrestrially adapted bony fish). Anyway the bony fish clade divides into two great clades: Sarcopterygii (including tetrapods) and Actinopterygii. It is the Actinopterygii, or ray-finned fish that concern us here. Actinopterygians are a clade and can be diagnosed by a bunch of characteristics, the most obvious one being that there is just one dorsal fin. Of course the path of evolution is rarely simple and we find that many modern actinopterygians divide their single dorsal fin into an anterior spiny portion and a posterior portion with soft rays. Sometimes the dorsal fin is bilobed with a dip in the profile between the spiny and soft sections but more often there is a finless gap, thus two dorsal fins are secondarily acquired.
The evolutionary success of ray finned fish cannot be overstated. In terms of diversity they outnumber all other vertebrate clades put together 2:1 and exhibit a jaw-dropping array of specializations, some of them nightmarishly bizarre to our eyes. Living ray-fins can be divided into five clades, the vast majority of ray-fins belong to just one of these the Teleostei. The other four groups are the cladistians (bichirs), chondrosteans (sturgeons and paddlefish, not to be confused with the chondrichthyans which are the sharks, rays and ratfish), ginglymods (gars*) and halecomorphs (bowfin).



Polypterus, the bichir, a living cladistian from Central Africa. Image from wildwoods.co.uk



Pallid sturgeon, a chondrostean. Image from wikipedia.



Lepisosteus, gar, a living ginglymod from North America. Image from wikipedia.



Amia calva, the bowfin. The sole surviving halecomorph. From North America. Image from wikipedia.

The staggering diversity of modern ray-fins is actually a relatively recent phenomenon compared to the venerable age of the total actinopterygian group. The first ray-fins appeared no later than the Late Silurian, when we find the first fossils of their sister group, the Sarcopterygii (e.g. Psarolepis in China) whereas the steep rise in diversity and disparity didn’t really start until the Late Cretaceous and is restricted to the teleost clade. One intriguing feature of living teleosts is that their entire genome has been duplicated. Some have suggested that this duplication event may have been the impetus for the teleost diversification. With two copies of every gene it is easy to imagine that duplicates of essential genes that have little freedom to vary without compromising the organism would be free to vary, perhaps allowing a more rapid exploration of morphospace.
It is universally agreed on both morphological and molecular data that the cladistians are basal to all other rayfins, but the relationships of the other clades remain contentious. Morphological analyses support a ladder like arrangement with chondrosteans, ginglymods and halecomorphs forming serially closer outgroups to the teleosts. The clade including ginglymods, halecomorphs and teleosts is known as the Neopterygii and is characterized by, amongst other features, loss of the clavicle bone, a vertical hyomandibula (jaw suspensorium) and a mobile maxilla (but not in modern ginglymods). Molecular data based on sequences of mitochondrial genes however gathers these three groups into an ‘ancient fish clade’ that is the sister group of teleosts. The following diagram from Hurley et al. (2007) illustrates these two different arrangements well.



The shape of the tree is not the only source of molecular vs. morphological tension in ray finned fish evolution. Estimates of the timing of the divergences between these clades is also dramatically different when using the different data sources. For example the oldest crown-group teleost (reminder: the crown-group is the group descended from the most recent common ancestor of all extant species) in the fossil record dates from the Late Jurassic (151 million years), whereas mitochondrial data estimates the divergence date at either the Early Permian (285 million years) or Early Carboniferous (334 million years) depending on the method used. That’s a difference of 134 to 183 million years. Now we can always expect that a clade will not show up immediately after its divergence from its sister group, due to low diversity and low abundance but this difference is extraordinary. Other molecular divergence dates are also strongly discordant with the fossil data, for instance the bowfin-teleost split (which would provide a minimum age of crown-group Neopterygii) is estimated to range from 417-390 million years (Late Silurian to Mid Devonian) using mitochondrial data whereas the palaeontological date of neopterygian crown-group is 245 million years (Early Triassic). However the picture isn't quite as bad as it once seemed and we shall look at how improvements in our understanding of both sets of data is straightening out this problem in the next post.

*Note to Australian fisherpeople, the ‘garfish’ you will be familiar with are part of the vast teleost radiation and are not ginglymods of any sort.

Hurley, I.A., Lockridge Mueller, R, Dunn, K.A.,Schmidt, Friedman, M., Ho1, R.K., Prince, V.E., Yang, Z., Thomas, M.G. and Coates, M.I. (2007)A new timescale for ray finned fish evolution. Proc. R. Soc. B 274, 489–498

Where do spiny sharks go?

Hi and welcome to fish fortnight on Dracovenator. You may notice a slight change in style here, as a new year's resolution I'm going to try to write for a more general audience. Does this mean I'm dumbing Dracovenator down? No, I'm just going to try to stop assuming a lot of specialist knowledge on a part of my readers, and will throw in some more basic anatomy for the Form and Function students I'll be teaching later this year.
Anyway by pure coincidence a bunch of fish-related items have come up all at once so I'm running a kind of fish-festival over the course of the next two weeks.
The first of these is a new paper in Nature this week by Martin Brazeau, which desribes the braincase and jaws of an acanthodian (popularly called spiny sharks, although they are not sharks, at least not in the conventional sense). Why is this a big deal? I'll get to that, but first some background.
Most modern vertebrate animals have jaws (only lampreys and hagfish do not). Those that have jaws belong to the great clade Gnathostomata (literally 'jaw-mouths'). Modern gnathostomes can be further divided into two clades: The cartilaginous fish (Chondrichthyes) including sharks and rays and the bony vertebrates (Osteichthyes)including everything else from goldfish to humans.
Both groups are monophyletic, that is they include all descendants from a common ancestor, thus neither group was ancestral to the other, both having split from a common jawed ancestor somewhere back in the past (probably in the Late Ordovician or Early Silurian Period).
Now there are some fossil jawed vertebrates that don't fit into either of these two clades. These are traditionally placed into two groups: The placoderms (armoured fish) and the acanthodians (so-called spiny sharks).



Dunkleosteus, a giant placoderm. Painting by Charles R. Knight. From universe-review.ca

How do these groups relate to the surviving Chondrichthyes and Osteichthyes, or in other words, what is the shape of the first part of the evolutionary tree of jawed vertebrates?
That has been a big mystery indeed. Just about every possible arrangement has been suggested. Placoderms are either regarded as the sister group of all other gnathostomes (that is they branched off first and the others all share a more recent common ancestor) or are thought to share a closer relationship with Chondrichthyes than with Osteichthyes. Nevertheless the plates of dermal bone which encase the front end of these fish has been thought to be an evolutionary novelty (synapomorphy in technical parlance) which marks out the placoderms as a monophyletic group.
The acanthodians have proven even more difficult to place, partly because the internal skeletal anatomy of most species has not been preserved (because like hagfish, lampreys and sharks it was cartilaginous). The favoured hypothesis of more recent years was that acanthodians shared a close relationship with Osteichthyes, that is they split from the line leading to bony vertebrates after the placoderms and chondrichthyans had already split and started their own evolutionary journeys. This was the position taken by Phillippe Janvier in his book 'Early Vertebrates' (1996, Oxford University Press)- an excellent book by the way. Once again the group had been thought to form a clade largely on the characteristic of bony spines supporting the leading edge of their paired fins and some details of their scales. Recent discoveries have shown that both of these features are probably general features for early gnathostomes that were lost in the modern groups, rather than evolutionary novelties unique to acanthodians. It should be noted that even back in 1996, Janvier noted that acanthodian monophyly wasn't all that strong and presented the suggestion that some acanthodians might turn out to be more closely related to chondricthyans (what we would call stem group chondrichthyans) while others might be stem group osteichthyans.



A diversity of acanthodians by Stanton Fink. Image from wikipedia-commons.

Enter Brazeau's recent paper. It includes a cladistic analysis of early jawed vertebrate phylogeny (the branching pattern of evolution) that uses individual genera, rather than large monophyletic groups. Lo and Behold, Janvier was right, or at least on the right track. Indeed Brazeau's analysis finds former acanthodians falling out all over the tree. Some are stem-group gnathostomes that branched away before modern gnathostomes split into Chondrichthyes and Osteichthyes, others are stem-group chondrichthyans others are stem-group osteichthyans. In particular the former acanthodian that forms the centrepiece of the paper, Ptomacanthus, is found in some of his trees to be the closest sister group to the gnathostome crown-group (the crown-group is the clade including the extant species and all descendants of their most recent common ancestor - in this case the common ancestor of Chondrichthyes and Osteichthyes). Ptomacanthus is special because unlike most other acanthodians, including all the early, more relevant species, does preserve some of its internal skeleton, namely the braincase and the jaws. The braincase is remarkably primitive, actually looking more like that of a placoderm than a crown-group gnathostome. This unusual combination of characters (externally crown-group like but internally more primitive) helps to break up the acanthodian group and gives this new tree. There are other odd things in this tree, for instance placoderms are not found to be monophyletic either but that is a topic for a later post (and further investigation methinks).

UPDATE: Check out the ever-informative Catalogue of Organisms for a post that slipped past me before for a little bit more about acanthodians.

Martin D. Brazeau (2009). The braincase and jaws of a Devonian ‘acanthodian’ and modern gnathostome origins Nature, 457 (7227), 305-308 DOI: 10.1038/nature07436