Electric Fish Genomes Reveal How Evolution Repeats Itself | Quanta Magazine

The Problem of the Tape

Stephen Jay Gould famously asked what would happen if you rewound the tape of life and let it play again. Would evolution converge on the same solutions, or would the cascade of historical contingency produce something wholly unrecognizable? It is a question that sounds philosophical but is, in fact, deeply empirical — and electric fish may be among the most elegant experimental systems nature has accidentally designed to answer it. The research coming out of Harold Zakon’s group, combining genomic sequencing across South American and African lineages of weakly electric fish, takes this old puzzle and gives it molecular teeth.

Darwin himself was unsettled by electric fish. In On the Origin of Species he wrote that “It is impossible to conceive by what steps these wondrous organs have been produced” — not just once, but repeatedly, across taxa separated by ocean and tens of millions of years. That confession of bewilderment from someone not easily bewildered is worth pausing on. Electric organs presented a double strangeness: they seemed almost too complex to arise by gradual selection, and yet they had apparently done so multiple times independently. The convergence amplified the mystery rather than resolving it.

Sodium, Muscle, and the Simplest Signal

To appreciate what convergence means here, you have to understand the mechanism. Electric fish derive their voltage from modified muscle cells called electrocytes, arranged along extended organs that run through much of the body. These cells establish sodium ion gradients, and when the sodium-gate proteins in their membranes open in coordinated fashion, a burst of current discharges — a signal that Zakon describes as about the simplest you could imagine. This is not some baroque biochemical novelty. It is ion physics, the same electrochemical logic underlying every action potential in every nervous system. What evolution discovered was not a new principle but a new deployment of an ancient one: take muscle, suppress its contractile function, amplify its electrical output, and you have an organ capable of navigation, communication, hunting, or defense.

The fish in South America and Africa independently hit on exactly this strategy. Both groups modified muscle tissue. Both groups recruited sodium pump machinery. The high-level architectural solution is strikingly convergent. And yet the molecular details diverge — the precise regulatory mechanisms governing where and when the sodium pumps operate differ between the two radiations. This combination, convergence at the level of organ logic and divergence at the level of molecular implementation, is the scientific payload of the study.

The Genome Duplication That Unlocked Everything

Underpinning this whole story is a much older event: a whole-genome duplication that occurred somewhere between 320 and 400 million years ago in the ancestor of all teleosts. Whole-genome duplications are catastrophic more often than not in vertebrates, but when they survive, they hand evolution an extraordinary resource. As Gavin Conant puts it, “Suddenly, you have the capacity to make a whole new pathway, instead of just one new gene.” Redundant copies free one duplicate from its original constraint. It can drift, accumulate mutations, take on new expression patterns, and eventually be co-opted for functions the original gene never performed. Without that ancestral accident, the electrocyte may never have been possible — or at least not in the form we observe. The innovation sits on a platform of contingency, a lucky catastrophe that became prolific.

This matters for the convergence question because it reframes what “independent” means. The African and South American lineages were working from the same expanded molecular toolkit, the same duplicated gene families, the same menu of regulatory possibilities opened up by that ancient duplication. Their convergence is not two systems stumbling onto the same answer from completely different starting materials. It is more like two composers independently writing in the same key because they trained on the same harmonic tradition. The constraints are shared even when the histories diverge.

Convergence as Experimental Design

What connects this work to adjacent fields is the broader question of biological possibility space. In developmental biology, the concept of canalization describes how developmental pathways get channeled into robust, repeatable outcomes despite genetic variation. In protein biochemistry, studies of enzyme evolution show that similar active-site geometries recur across unrelated protein folds. In evolutionary theory, the debate between selectionism and contingency maps almost perfectly onto the question Zakon is asking: how much of what we observe was inevitable given the physics and chemistry available, and how much is accident frozen into lineage?

Electric fish offer something rare — a replicated natural experiment. Rather than looking at a single innovation and wondering whether another path existed, you can compare multiple independent solutions to the same ecological problem. The divergence in regulatory mechanisms between the African and South American fish tells you that there was not only one way to wire the pump. The convergence in overall strategy tells you that some ways are strongly preferred. That combination, as Zakon observes, gives a much richer picture of how predictable and how quirky evolution can simultaneously be.

Why It Matters

The deeper reason to care about this work is what it does to our intuitions about life’s repeatability. If evolution mostly converges, the universe of possible organisms is constrained, almost lawlike. If it mostly diverges, every lineage is a historical accident, and the tree of life is less a map of possibility space than a record of frozen chance. Electric fish suggest neither extreme. The tape, replayed, would produce electric fish again — but they would regulate their sodium pumps differently. The headline solution recurs; the fine print does not. That seems right, and it is far more interesting than either determinism or pure contingency alone.