The Cockles, and What This Family Did With One Body Plan

The Largest Bivalve on Earth Is a Cockle

Start with the result that reorders everything else. The giant clams of the Indo-Pacific, the largest bivalves alive, are not a family of their own. They are a subfamily of the cockles.

The World Register of Marine Species carries Tridacnidae Lamarck, 1819 under AphiaID 196336, and its status field reads superseded rank. The valid name it points to is Tridacninae Lamarck, 1819, AphiaID 382150, rank subfamily, family Cardiidae.

The honest version is narrower than the headline. Nobody was proved wrong. Same author, same year, same animals: Lamarck's name dropped one rank into a family that already existed. A rank change, not a refutation.

Independent lines of work kept landing there. Schneider placed the giant clams as a cardiid subfamily on morphology in 1998. Maruyama and colleagues published a molecular phylogeny of the algae-hosting bivalves the same year, and Herrera and colleagues later called their result startling: the giant clams came out closer to the ordinary cockles than to the small photosymbiotic ones. Schneider and O Foighil sequenced mitochondrial 16S for all eight giant clam species in 1999, and later studies agreed.

The detail worth keeping from 1999 is the outgroup. To anchor the phylogeny of the largest bivalve on the planet, they used Cerastoderma, the edible cockle sold by the pint in British seaside towns.

A Heart, in Five Languages

The family name comes from its type genus, Cardium, New Latin on the Ancient Greek καρδία, kardia, heart. Less noticed is how far the metaphor travels: Cardiidae sits inside the superfamily Cardioidea, inside the order Cardiida. Very few bivalve families get to name the order above them.

Then ask a database what ordinary people call these animals. The vernacular list WoRMS holds for AphiaID 229 runs to ten names, and five are the same idea reached separately: German Herzmuscheln, Swedish hjärtmusslor, Norwegian Bokmål hjerteskjell, Nynorsk hjarteskjel, Dutch hartschelpen. Every one means heart shell or heart mussel.

Northern Europe did not borrow that from Lamarck. Fishing communities around the North Sea and the Baltic got there in their own words while a French naturalist got there in Latin. The Japanese ザルガイ科 and the French coques went elsewhere entirely, which is what makes the convergence worth pointing at.

Two features produce the shape. Cardiid valves are strongly inflated, each bulging rather than lying flat, and the outline is close to equilateral, the beak near the middle rather than pushed toward one end. The closed shell is then a rounded body with a cleft down the hinge line.

What Actually Unites Them

A family is a claim about shared ancestry, and the shell evidence here is unusually tidy. The Digital Atlas of Ancient Life describes cardiid valves as equivalve and inflated, rounded to trigonal or quadrangular in outline, most often taller than long, and built of aragonite.

Sculpture is the character people actually use. Cardiids are either smooth or carry strong radial ribs, beak to margin, sometimes with spines or scale-like ornament on top. Radial is the operative word: the ribs run the length of the shell, not around it.

The inside repays a second look. Cardiid interiors carry crenulations that correspond to the exterior sculpture, so every rib reaching the shell edge leaves a matching notch on the inner rim and the valves close rib into notch the whole way around. It is tempting to finish that with a function, something about resisting twist when a predator works at the margin. We are not going to: the correspondence is documented, the mechanical reading of it is not.

The hinge is heterodont or cyclodont, typically two cardinal teeth with anterior and posterior laterals. Most cardiids are heteromyarian, two adductor muscle scars with one slightly larger. The giant clams are the exception: adult Tridacninae are monomyarian. On the road to becoming the largest bivalve on earth, that lineage dropped an adductor muscle entirely.

None of it holds without strain. Papyridea soleniformis (Bruguière, 1789), AphiaID 225385, the spiny paper cockle, is thin, elongate and gapes at both ends. It is a cardiid that declines to be heart-shaped.

The Short-Siphon Bargain

Everything else this family does follows from one commitment made early: stay near the surface.

Cardiids are shallowly infaunal to epifaunal filter feeders in soft sand and mud. Herrera and colleagues give their range as depths to 500 meters, most species above 150. Read that carefully: it is water depth, how deep the sea is where they live, not how deep they bury.

Burial depth is a separate and much smaller number, set by siphon length. The Marine Life Information Network describes the European edible cockle as having short siphons, needing to stay in contact with the sediment surface, and burrowing no deeper than about five centimeters. That figure belongs to that species, and we are not transplanting it onto a Gulf of Mexico animal nobody has published the equivalent measurement for.

Now set the family beside its neighbors. A tellin buys safety with depth, below the reach of most of what hunts it, and pays with long siphons running back to the surface and a deep pocket inside the shell to house them retracted. That is a one-way commitment, because an animal down there cannot outrun anything.

Cardiids took the opposite trade. Short siphons force shallow burial, and shallow burial means a sea star or a ray can actually reach them. So this family kept what deep burrowers gave up: a large, powerful, mobile foot. A cockle does not hide deeper. It leaves.

The anatomy of the empty valve itself is a page of its own.

The Foot That Digs, and the Leap We Only Half Know

The cardiid foot is large, muscular and sharply bent, described in the secondary literature as L-shaped or hook-shaped. It comes out through the pedal gape at the lower margin of the closed shell, and it runs two opposite jobs.

Going down, the active stage is contraction: the foot probes into sediment, swells at the tip to anchor, and the pedal retractor muscles haul the shell down onto it. Coming up, the cycle reverses, and the active stage becomes protraction and straightening, the foot extending and levering the body upward. Ansell's 1967 work on three European cardiids describes both.

Run that upstroke past its limit and you get the leap. Plant the bent foot, straighten it hard, and the animal leaves the sediment altogether. Not a jet, not a slam of the valves. A pole vault.

The Digital Atlas of Ancient Life says "most species" are mobile, with very strong feet capable of leaping to escape predators. Three things about that sentence deserve to survive into ours.

• Most, not all. We found no census behind the phrase, and the most heavily studied cockle in the world, the European edible cockle, has a detailed species account that never mentions leaping.
• The classic work is European. All three species in Ansell 1967 are European and no Gulf of Mexico cockle appears in it. The full text sits behind a paywall we could not get through, so the mechanism above reached us through secondary summaries.
• The trigger is thinner still. Sea stars are the predator usually named, but that reached us as a search snippet from a regional museum whose pages refused every attempt to load. Regionally reported, not a verified finding, and nothing establishes it for a Gulf species.

Two Ways to Farm Sunlight

Herrera and colleagues make the family-level claim directly: as far as anyone knows, no other marine bivalve family has evolved an obligate symbiosis with photosynthetic dinoflagellate algae.

That is not the same as saying no other bivalve keeps partners. The lucines run one of their own, an entirely different arrangement: bacteria in the gills, living on seafloor chemistry, in the dark. That page covers it.

What the cockles built runs on light, and inside this family it was solved twice. McCoy and colleagues open their 2024 paper in Nature Communications by setting the two side by side: the giant clams open their valves so the sun reaches their symbionts; the heart cockles do not have to, because sunlight gets through their shells.

The window is far more engineered than a window. McCoy and colleagues measured what passes through the shell of Corculum cardissa (Linnaeus, 1758), AphiaID 216387, and its relatives: 11 to 62 percent of photosynthetically active radiation, averaging 31 percent, against only 5 to 28 percent of harmful ultraviolet, averaging 14 percent. Not a hole in the shell. A selective filter.

Inside each window the aragonite forms narrow fibrous prisms perpendicular to the surface, roughly a micron across and oriented along the crystal's optical axis. Bundled, they behave like fiber optic cable, which the authors demonstrate by projecting images through the shell at better than 100 lines per millimeter. Microlenses beneath condense the light deeper into the symbiont-rich tissue.

The paper claims two firsts, carefully worded. To the authors' knowledge this is the first instance of fiber optic cable bundles in an organism, and the condensing lenses are a second, independent evolution of a solution otherwise known from flowering plants.

Twice, Independently, in One Family

The two solutions are not two versions of one inheritance. They are two separate inventions.

Herrera and colleagues found photosymbiosis evolving independently twice inside Cardiidae: once in the giant clams, at roughly 21.1 million years ago, and once in a clade holding Fragum, Lunulicardia and Corculum, at roughly 41.8 million years ago. Li and colleagues, working from whole transcriptomes, recovered the same two origins.

The dates are the counterintuitive part. On Herrera's estimates the small windowed heart cockles got there roughly twenty million years before the giant clams, the reverse of what most people would guess. Handle it gently, though. The giant clam figure is tightly bracketed, 20.2 to 23.3 million years; the heart cockle figure carries an interval running from 28.7 out to 57.1. The gap is real on these numbers. Its size is not precise.

One further number gets misquoted. Li and colleagues date the split between the giant clam and Fraginae lineages to the Cretaceous, a different event from either origin of the symbiosis: when the branches parted, not when either acquired algae. The two sets of dates describe different nodes and must not be blended.

Kirkendale's 2009 work adds a correction inside Fraginae. Examining live material from more than half the species in the group, she found fewer than half of the derived fragine genera and subgenera host photosymbionts. Even within the branch that invented it, most members do not.

The Near Miss in Gulf Sand

The Gulf of Mexico does have a fragine. Americardia media (Linnaeus, 1758), AphiaID 420847, the Atlantic strawberry cockle, sits in subfamily Fraginae, checked against the classification WoRMS returns for the species rather than assumed from a resemblance. So the strawberry cockle is in the same subfamily as the windowed heart cockles, a real and rather pleasing piece of kinship, and it is exactly as far as it goes.

The strawberry cockle is not photosymbiotic. No cockle in the Gulf of Mexico is. Li and colleagues draw the boundary in one line: the photosymbiotic species belong exclusively to three genera, Fragum, Corculum and Lunulicardia. Herrera and Kirkendale draw it the same way. All three are Indo-Pacific. Americardia is not among them.

The reason to be blunt is that the softer version writes itself: same subfamily, warm shallow water, and one careless sentence leaves a reader believing there is a solar-powered clam in Florida sand. There is not. The relationship is real. The trick did not come with it.

Settled That They Are Cockles. Unsettled Which Cockles.

Are the giant clams cockles? Yes. Settled, on morphology, on one gene, on three genes, and on whole transcriptomes, with the nomenclature updated to match.

Which cockles are they closest to? Genuinely open, and the disagreement is instructive.

• Schneider's 1998 morphological work placed Tridacninae as sister to the Lymnocardiinae, and Maruyama and colleagues' molecular work was read as corroborating it.
• Herrera and colleagues in 2015 could not resolve it, and said so. Their maximum likelihood runs left the placement unresolved, their Bayesian runs put it in one position, and their divergence dating put it in another. Three methods, three answers, one paper.
• Li and colleagues in 2020, with the largest data type of the three, recovered Tridacninae as sister to Fraginae.

If that last result holds it is a better story than a clean answer would have been: the giant clams would sit beside the group that independently invented the same partnership. A page that picked one of these three positions and stated it flatly would read as more authoritative and tell you less.

Six Cockles, With Their Receipts

Six Gulf of Mexico cardiids below, every name checked against WoRMS and MolluscaBase with its AphiaID printed so you can check it. The table holds up the family anatomy above, not anyone's attempt to name a shell.

Two of those rows are the same lesson pointing opposite directions, and getting them backwards is the easiest error here. One prickly cockle moved out of Trachycardium; the other stayed, and a Commission ruling is why.

Note the smooth ones. Laevicardium and Dinocardium sit in Laevicardiinae with shells smooth or nearly so, making strong radial ribbing a tendency rather than a requirement.

A Filing System, Not a Family Tree

WoRMS currently lists nine accepted subfamilies inside Cardiidae: Cardiinae, Clinocardiinae, Fraginae, Laevicardiinae, Lymnocardiinae, Nemocardiinae, Orthocardiinae, Trachycardiinae and Tridacninae. It is tempting to read that as a diagram of how the family branched. It is not.

Herrera and colleagues tested eight traditional subfamilies against three genetic loci across 110 species and found six para- or polyphyletic: those groups do not correspond to real branches. Only two came out monophyletic, Clinocardiinae and Tridacninae. The giant clams are a better-defined group than most cockle subfamilies.

Meanwhile the list keeps moving: Nemocardiinae was erected in 2023, eight years after that paper. The WoRMS list is the nomenclatural standard and the right thing to cite. Parts of the tree it implies are known not to hold.

Even the head count is unsettled: roughly 265 living species in 50 genera by ter Poorten's 2014 figure, 272 in 48 genera by the Digital Atlas. Roughly 260 to 270 is as precise as anyone should be.

One entry deserves a sentence of its own. The WoRMS record for the family is flagged marine, brackish and freshwater, which looks like a data error until you find Lymnocardiinae, the Ponto-Caspian cockles, in the nearly fresh water of the Caspian. A family most people meet as a saltwater beach shell has members far from any ocean.

The fossil record here is unusually good, which is why any of this can be tested. The oldest known cardiid is Late Triassic, in an extinct subfamily, and Herrera's molecular clock puts the family root at about 210 million years. Every one of the 50 living marine genera is also represented as a fossil.

What to Do With the One in Your Hand

A cockle valve is the record of a particular bargain. Short siphons, shallow burial, a big bent foot, a rim that seats rib into notch. It belonged to an animal that could not hide from a sea star and did not try to, in a family that also produced a clam the size of a bathtub and one that grew windows in its own shell.

To put a name to what you are holding, our Florida shell identification guide is built for that job. This page was never going to be.

To make the shell worth something later, write down where you found it, when, and who. A shell with a place and a date attached is a record. Without them it is an ornament. Starting a collection covers the rest.

One rule stands whatever the shell turns out to be, and it is ours rather than an interpretation of any statute: empty shells only.

The Shell Monographs work through the rest of the families the same way. If reading about an animal is a poor substitute for standing over one, we run guided trips daily. Reserve a seat.