Catfish and the Lazarus Taxon Problem

Some time ago I mentioned the phenomenon of Lazarus taxa. For those who don’t remember, Lazarus taxa are taxa that are known from fossils that appear to disappear from the fossil record and then re-emerge (“from the dead”) significantly later. The concept gets tossed around a lot when discussing the possibility that groups thought to be extinct might still persist. One of the central questions in evaluating the odds of Lazarus taxa occurring is the completeness of the fossil record. Now, nothing I’m going to say is going to be surprising to a paleontologist, but the fossil record is, as far as completeness goes, pretty terrible. I’ve sometimes heard the fossil record compared to a TV show that is missing chunks. It’s probably more like five non-consecutive frames from an hour long show.

One of the clades that demonstrates this well, and that I happen to know about, is catfish. By catfish I mean the Siluriformes, a quite large group of mostly freshwater fish, some of whom don’t look very much like the “classic” catfish. Catfishes often have a heavy-duty first fin ray in their pectoral fins which happens to preserve well and is identifiable to major group. (This fin spine is meant to stab potential predators, as more than a few recreational anglers have found out.) Because of this catfish fossils are sometimes identifiable even in a severely fragmented state. This is good because fish bones are pretty lightweight compared to tetrapods and fish skeletons are easily scattered and the bones broken.

Diogo (2004) discusses a fossil from the catfish genus Corydoras dated to the Paleocene. Corydoras is a common modern genus, found not only as many species in South America but also as the “cory catfish” in pet stores across the world. The Paleocene period is just past the massive extinction event that brought the Mesozoic1 to a cataclysmic end. To give an idea how unlike our modern world the Paleocene was consider that mice, bears, and cats had not yet evolved, and that the ancestors of whales were walking on land at this time. And yet here’s a modern genus, Corydoras. Moreover, Corydoras is endowed with bony, armored plates that wrap around its body, making it potentially much more fossilizable than other fish (although it is a small fish). Given the success of the modern genus we might expect to find that the world is littered with Corydoras fossils. Instead, Diogo (and I have seen nothing more recent to suggest that new discoveries have changed the picture) notes that there are no other Corydoras fossils until, perhaps, recent sub-fossils. There’s a 56 million year fossil lacuna for this genus.

However, the story gets more interesting. Diogo was arguing for what now seems to be a seriously minority position about the age of the Siluriformes but what makes this an argument is that everyone is sure that we do not have the earliest Siluriformes in the fossil record. A short digression into geologic timelines (really just what I needed to teach myself to follow the arguments I was reading) is needed here. We live in a large era called the Cenozoic. The prior large era is the Mesozoic, known for dinosaurs. The Mesozoic is coarsely divided into three pieces: the Triassic (oldest), Jurassic (middle), and Cretaceous (most recent2). The Cretaceous is divided more finely into a number of periods which I can’t remember because only two have catfish fossils, the last two, the Campanian and the Maastrichtian. The Maastrichtian is the very last sliver of time before the disaster that closes out the whole Mesozoic.

The very earliest catfish fossil comes from the Campanian in Argentina but skip forward just briefly in time to the Maastrichtian and catfish fossils are found in Bolivia (de Muizon et al., 1983; Gayet et al., 2001), India (Cione & Prasad, 2002), Niger (possibly, I cannot find the original description, just a mention in Cione & Prasad [2002]), the western United States (possibly, the species Vorhisia is not unanimously agreed to be a siluriform, Frizzel & Koenig, 1973), and Spain (Pena & Soler-Gijon, 1996, possibly not actually Maastrichtian but just beyond, the authors place the fossil on the border between the Maastrichtian and the following era). Also importantly, even within just the undisputed (as far as I know) Bolivian finds at least two major catfish clades are present. de Muizon et al. identify these two clades as the families Ictaluridae and Ariidae. Even if these exact assignments were to be disputed it seems unlikely that the total diversity of catfish at this site will be reduced. What this means is that the snapshot we have of early catfish is a diverse, widespread clade. Clades don’t start this way. Clades start as small, localized groups with low diversity. They then spread and diversify. It’s a bit like looking through a photo album and finding that the earliest photo you have of someone who you are researching is a photo of them holding their first child. You know they are not a child themselves but you also don’t know quite how old this makes them. All you know is that the beginning is further back.

I won’t get into the (long, complex) debate about when the first catfish swam the rivers of Earth (and the first catfish probably did swim in rivers, not lakes) but what we know is that the Siluriformes have a significant ghost lineage, the name for a lineage that is unrepresented in the fossil record. When we discuss things like Lazarus taxa we should be aware just how many creatures we know existed during times in which they did not leave fossil records.

Perhaps fittingly, since the last Lazarus taxa article discussed the coelacanth, there are particular biases against bony fish in the fossil record. Becker et al. (2009) discuss our friend Vorhisia and the other osteichthyan fish found in that time and region. They also discuss why the fossil record for Cretaceous fish is so poor. They list four things that bias the fossil record against bony fish.

  1. Fish skeletons are lightly built and fall apart easily. Many are from small animals. The bones can be destroyed easily and what is collected are frequently only a few more durable parts of the skeleton (like teeth).
  2. Fish bones are often hard to identify because a lot of basic work remains to be done. Perhaps someone has already found a Jurassic catfish, which would be a major find, but can’t identify it as such because the work comparing catfish skeletons to other fish skeletons hasn’t been done and so this person can’t determine what they have.
  3. Bias in collection and research. Basically, nobody cares about fish. People grab the big stuff and prioritize that when they do research. Fish bones may get left in the ground or in a file drawer instead of being described.
  4. In the specific area Becker et al. described the rock type was also not a good one for preserving fish.

Now, there are examples where it would be hard to claim that the fossil record is simply too patchy to show the continued existence of a taxon. For instance, the (unfortunately frequent) claims that pterosaurs or plesiosaurs have made it into the modern era require that entire lineages of large creatures with very distinctive bones have made it millions of years without leaving fossils. However, individual species seem quite capable of dodging fossilization for extensive periods of time.


Becker, M. A., Chamberlain Jr., J. A., Robb, A. J., Terry Jr., D. O., & Garb, M. P. (2009). Osteichthyans from the Fairpoint Member of the Fox Hills Formation (Maastrichtian), Meade County, South Dakota, USA. Cretaceous Research, 30(4), 1031–1040.

Cione, A. L., & Prasad, G. V. R. (2002). The Oldest Known Catfish (Teleostei:Siluriformes) from Asia (India, Late Cretaceous). Journal of Paleontology, 76(1), 190–193.

Diogo, R. (2004). Phylogeny, origin and biogeography of catfishes: support for a Pangean origin of “modern teleosts” and reexamination of some Mesozoic Pangean connections between the Gondwanan and Laurasian supercontinents. Animal Biology, 54(4), 331–351.

Frizzell, D. L., & Koenig, J. W. (1973). Upper Cretaceous Ostariophysine (Vorhisia) Redescribed from Unique Association of Utricular and Lagenar Otoliths (Lapillus and Asteriscus). Copeia, 1973(4), 692–698.

Gayet, Mireille; Marshall, Larry G.; Sempere, Thierry; Meunier, François J.; Cappetta, Henri; Rage, J.-C. (2001). Middle Maastrichtian vertebrates (fishes, amphibians, dinosaurs and other reptiles, mammals) from Pajcha Pata (Bolivia). Biostratigraphic, palaeoecologic and palaeobiogeographic implications. Palaeogeography, Palaeoclimatology, Palaeoecology, 169(1–2), 39–68.

de Muizon, C., Gayet, M., Lavenu, A., Marshall, L. G., Sigé, B., & Villaroel, C. (1983). Late Cretaceous vertebrates, including mammals,from Tiupampa, Southcentral Bolivia. Geobios, 16(6), 747–753.

Pena, A. D. E. L. A., & Soler-Gijon, R. (1996). The first siluriform fish from the Cretaceous-Tertiary interval of Eurasia. Lethaia, 29(1), 85–86.