Continental counterparts: mormyroid and gymnotiform electric fishes

Africa and the Neotropics both harbor endemic monophyletic groups of teleost fishes, each of about 200 described species, that communicate, find their prey, and navigate through freshwater habitats at night by means of a sophisticated “active” electric sense. They produce a weak electric field around their body with an electric organ and detect nearby objects as distortions to this field via arrays of receptors distributed over their skin. These are the Mormyroidea (Africa) and the Gymnotiformes (South and Central America). We’ve known for some time that these two clades of fishes are only distantly related to each other and that therefore they must have evolved their very similar active electric sense, morphologies and ecologies independently (i.e., starting from ancestors that lacked electrogenesis). However, we had little idea how old these two groups of fishes were due to their poor fossil record. The few fossils that exist represent modern taxa and are not very old. The living representatives of mormyroids and gymnotiforms of fishes are not especially robustly boned, making fossilization probably extra unlikely.

As someone who has had the privilege to study both groups of fishes on both continents, and being something of a daydreamer prone to idle speculation, I’ve often wondered whether a mormyrid species transplanted from Africa to South America or a gymnotiform species taken from South America and placed in an African stream would “take." (This is one of those unspeakable experiments, like crossing humans and chimps.) If they were to occupy the same area, would one group outcompete the other? Can you see invasive gymnotiform species spreading like wildfire through the Congo, or mormyrids infiltrating the Amazon? (I imagine a huge tanker leaving the port of Matadi, D.R. Congo filled with Mormyrus, Campylomormyrus, Petrocephalus, etc., dumping its cargo at Belem, Brazil in the Amazon delta, as well as the opposite scenario. Yes, I really am that much of a geek.)

I also used to wonder why this experiment hadn’t already happened, naturally. Before about 110 million years ago in the early Late Cretaceous Period, Africa and South America were a single landmass palaeogeologists call Western Gondwana. (The other Gondwanan fragments—Antarctica, Australia, & India-Madagascar—went their separate ways earlier.) The fossil record doesn’t provide any direct evidence that these fishes were around nearly that long ago, but of course the fossil record is very incomplete. Both a number of armchair theorizers and several earlier studies using DNA sequence evolution to estimate the ages of these groups came to the conclusion that gymnotiforms and mormyroids probably existed before that split. If so, why are both clades now restricted to their respective continents? Or why didn’t either the mormyroids or the gymnotiforms, which occupy the same ecological niche, outcompete the other a long time ago on Western Gondwana? In that case, one of them would now reign supreme on both continents as the king of the freshwater electrosensory world.

A new study by Lavoué et al.—on which I am one of the “al.”—entitled "Comparable ages for the independent origins of electrogensis in African and South American weakly electric fishes" has just been published in PLoS ONE that may help answer at least this second question. (I’m not holding my breath for funding to test the transplant experiment.) It’s the first time that ages of both groups of electric fishes have been estimated in a single study from one dataset. The core of the study is a phylogenetic analysis that includes 70 complete mitochondrial genomes from a good sampling of mormyrids and gymnotiforms as well as their sister lineages and other “basal” teleost fishes. That’s over one million aligned A's, C's, G's and T's! (Maybe not so impressive for the “next-gen” genomics folks, but I’m still living in “last-gen”.) We performed a Bayesian "relaxed molecular clock" analysis using software called BEAST that infers not only a phylogenetic tree, but also the relative time at which nodes of the tree, which represent ancestors, appear. By calibrating the tree with a set of fossils from outside these two fossil-poor groups, we are able to estimate the absolute age of origin for each (the point at which each lineage separates from its nearest living sister lineage, this is called the “stem group”), and also the age of the most recent common ancestor of all the modern species (or “crown group”). This was done twice in two separate analyses that implement different assumptions about how much information fossils give us. One assumption is that some fossils give us information about the maximum age of lineages on a phylogenetic tree as well as minimum ages, while the other is that fossils only provide minimum ages. By doing it both ways, we were able to bracket a range of possible ages for gymnotiforms and mormyroids that most folks who concern themselves with this kind of thing will hopefully find reasonable.

So what did we find?

With the more “conservative” analysis, (i.e., the one that tethers maximum ages of certain nodes to fossils associated with them), crown-group gymnotiforms appeared about 100 million years ago (mya), plus or minus a wide credibility interval. At this time, South America and Africa had had just split apart and a narrow seaway existed between them. Thus gymnotiform fishes, which most likely arose on what became South America, may have missed their chance to colonize Africa. (Even a little saltwater may as well be a Himalayan mountain range to your average freshwater fish.)

Crown-group mormyroids are very similar in age, with the mean estimate at 94 mya. The mormyroids’ sister group are the Afro-Asian notopteroids that are also absent from the New World. We presume mormyroids arose on Africa, or somewhere on what would become Africa. So mormyroids, too, may have missed the boat, or the continental fragment, that would have put them into contact with their continental counterparts.

With the analysis that imposes softer time constraints, the estimates get a lot older: 125 mya for crown-group Mormyroidea and 144 mya for crown-group gymnotiforms. If these dates are accurate, it’s harder to see why one or both groups couldn’t have bridged the future divide, but keep in mind that the rifting process began well before final separation and freshwater connectivity between the two halves of Western Gondwana would have been limited long before. If the mormyroids and gymnotiforms originated on opposite halves of Western Gondwana, it’s hard to know how easy it would have been to disperse and eventually come into contact. Also keep in mind that this less conservative analysis may be estimating something more like the oldest possible ages for these groups. In other words, they’re probably in fact somewhat younger. The true ages of these electric fish clades most likely lie somewhere between these two estimates. My hunch is that the more conservative analysis with the younger ages may be the closer of the two, but I could be wrong. These fancy analyses using BEAST are nice, but in the end we need fossils to confirm an hyphothesis for the age of a living group of organisms. While obviously rare, we can hope that more mormyroid and gymnotiform fossils will turn up in the future.

There is another really interesting finding in the study, namely that active electrolocation, (the sort that makes use of an electric field generated by an electric organ), may have required about the same amount of time to evolve in each lineage. This follows from the observation that the internode or branch on the tree between the stem- and crown-groups of each electric fish clade are very similar. Interestingly, the sister-groups of both the Momryroidea and the Gymnotiformes are passively electroreceptive, i.e. they have a certain kind of receptor that detects biologically produced fields of prey organisms, but lack electric organs and specialized high-frequency receptors of true electric fish. This implies that both mormyroids and gymnotiforms evolved from ancestors that were already passively electroreceptive. The fact that the internodes in each case appear to represent about 17-24 million years or so (depending on the calibration method) indicates that active electroreception (with electric organs) may have required a similar amount of time to evolve in each group. It’s not clear what if anything is special about this interval of time, but it makes the parallelism of the gymnotiform and mormyroid radiations all the more remarkable. (See more in the paper on how mormyroids and gymnotiforms are examples of both evolutionary convergence and parallelism and what the distinction is.)

Earlier studies by Harold Zakon, & Matt Arnegard and colleagues showed that parallelism extends even to the genetic architecture of electrogensis in gymnotiforms and mormyroids: both have apparently co-opted a copy of the sodium channel gene for exclusive expression in electric organs. These copies are thought to have originated in a whole genome duplication event at the root of the teleost tree. So this co-option of the gene copy had to have happened in this 20 million year-or-so interval in both groups, long after the hypothesized whole genome duplication event in the ancestral lineage of modern teleosts. A summary of the sodium channel gene discovery can be found here.

Thus, originating pretty much contemporaneously from very different ancestors at very different parts of the teleost tree, mormyroids and gymnotiforms evolved remarkably similar electrosensory systems, shapes and habits and diversified into roughly the same number of species in the tropical freshwaters of the two sister continents. I think that’s pretty cool, but I’m admittedly and obviously biased. Why two groups of weakly electric fishes evolved at roughly the same time in such parallel fashion may always be a difficult question to answer. (One may also wonder why electric fishes didn't evolve in the third great tropical freshwater ecosystem of southeast Asia. Perhaps a suitable passively electrosensitive ancestor didn't exist there at the right time.) Hopefully this paper will get a lot of reads and convince more people that African and Neotropical electric fishes represent an excellent “system for investigating pathways to evolutionary novelty, as well as the influences of key innovations in communication on the process of species radiation,” as the authors say.

Again, here's the link to the full article in PLoS ONE and here's the link that will give you the legend to the figure included above.

Lavoué, S., Miya, M., Arnegard, M., Sullivan, J., Hopkins, C., & Nishida, M. (2012). Comparable Ages for the Independent Origins of Electrogenesis in African and South American Weakly Electric Fishes PLoS ONE, 7 (5) DOI: 10.1371/journal.pone.0036287

Add new comment

You must have Javascript enabled to use this form.

Who's online

There are currently 0 users online.

Scratchpads developed and conceived by (alphabetical): Ed Baker, Katherine Bouton Alice Heaton Dimitris Koureas, Laurence Livermore, Dave Roberts, Simon Rycroft, Ben Scott, Vince Smith