The Arks, and the Clam Family That Invented Red Blood

A Bivalve With Red Blood in It

Cut open almost any mollusk and the fluid inside is not red. Most run on hemocyanin, a copper protein that is colorless once it has given up its oxygen and faintly blue when it has not. Most bivalves carry no respiratory pigment at all. They are slow, they sit still, and diffusion across the gill is enough.

The arks are the exception, in the most surprising available way. Arcoid bivalves carry hemoglobin, iron based and red, and not dissolved in the fluid. They pack it into nucleated red blood cells circulating in the hemolymph, which is the vertebrate arrangement. In the species whose genome has been sequenced, those red cells are close to 90 percent of all circulating blood cells, and more than 90 percent of the protein inside them is hemoglobin.

Two boundaries first, because both get crossed constantly. The pigment is an arcoid trait, shared with near relatives outside this family. And it has never been documented in the one ark everybody can name.

What Makes an Ark an Ark

Arcidae Lamarck, 1809 is the accepted family name, carried under AphiaID 208 by the World Register of Marine Species and by MolluscaBase. Lamarck's name then propagated upward through two ranks: the superfamily is Arcoidea and the order is Arcida, both built on it.

The placement that matters most sits a step further out. Arks are Pteriomorphia, the byssate, largely epifaunal wing of the bivalves, so they are cousins of mussels, scallops, oysters and pen shells rather than of the cockles, venus clams and tellins beside them in the drift line. A heavy ribbed box that looks like a clam is filed with the mussels, and most of what is odd about it follows.

WoRMS lists 30 accepted living genera. Below the family the ground softens: five subfamilies are carried, but all five return with the status alternative representation rather than accepted, so the honest phrasing is that Arcidae is usually divided into about five subfamilies, not that it has five accepted ones.

Two ages, and they must not be blended. The order Arcida reaches the lower Ordovician, around 450 million years ago. The family Arcidae is far younger: Jurassic to Recent, questionably Triassic. Anyone saying ark shells are 450 million years old is quoting the order.

Bivalve basics are covered on how to read an empty bivalve. This page is about what an ark does differently.

The Hinge That Works Like a Zipper

Turn an ark valve so the hinge faces you and you are looking at the family's best character: a long straight row of many small teeth, all much the same shape and size, interlocking with an identical row in the other valve. That is taxodont dentition.

The design philosophy is what makes it worth a section. A cockle or a venus clam is heterodont: a few differentiated teeth, cardinals and laterals, each with its own shape and job, a bespoke joint, a lock and a key. A taxodont hinge is the opposite idea, dozens of near-identical chevrons repeated along a straight line, sharing the closing load instead of concentrating it on two or three large teeth. A zipper rather than a key. The straight dorsal margin and squared-off outline of an ark follow from that row, which needs a straight run to seat in.

Two limits on the claim. Do not hunt for a tooth count: taxodont means numerous and similar, not a fixed number, and counts vary by species and change as an animal grows. And taxodont hinges are not unique to arks. Bittersweets and nut clams have them too, so the defensible statement is that taxodont dentition is diagnostic of the order Arcida, alongside the ligament and the shell microstructure.

It is also the character that survives. Color goes, the outer coat goes, sculpture wears flat, and the tooth row is the last thing on a bleached valve to go unreadable. It is why the group is recognizable in the fossil record at all.

The Same Idea Again, in the Ligament

Between and behind the beaks there is a wide flat deck, usually diamond shaped, scored with fine chevrons. That is the cardinal area, and the chevrons are the print left by a duplivincular ligament: instead of one elastic band, the lamellar part is repeated as a series of bands, each with its two edges seated in narrow grooves cut into that deck on both valves.

The mechanics are the point. A single-band ligament is a spring acting at one place. A duplivincular ligament is a stack of springs spread across a broad plate. The arks solved the elastic half of the hinge with the move they used on the rigid half: take one structure, repeat it many times, spread the load along a straight line. Hinge and ligament are the same idea told twice.

With taxodont dentition and a microstructure of outer crossed-lamellar and inner complex crossed-lamellar layers, the duplivincular ligament is a defining character of the order Arcida. It survives on beached valves, so the chevron deck is checkable.

The Box That Cannot Quite Close

An ark is built to clamp shut: heavy, thick, coarsely ribbed, squared off, with a hinge designed to spread closing load along its whole length. And then many of them will not close.

Along the ventral margin, opposite the hinge, the attached arks carry a byssal gape, a notch where the two valves do not meet. Through it passes the byssus, a bundle of protein threads the animal secretes to tie itself to rock, coral rubble or another shell. This is the Pteriomorphia inheritance showing. A mussel does the same thing, and a young scallop does it before growing out of it.

A permanently open slot in an armored box is a real design admission, and it is visible on empty valves. What the animal buys is a life spent gripping something rather than burrowing away from trouble. What it gives up is the ability to seal.

On the scallops the byssus is a life stage most species discard. Here it is the whole life.

Two Ways to Be an Ark

That gape is where the family splits, and the split runs through everything else about these animals.

One half stayed on top. The epifaunal arks, the Arcinae type, including Arca, Barbatia, Acar and Lamarcka, live byssally attached, nestled into crevices and under rubble. They keep the gape, and they are often lopsided, because a shell that grows inside a crevice grows to fit the crevice.

The other half went down. The infaunal arks, the Anadarinae type, including Anadara, Lunarca and Tegillarca, burrow into sediment. They are heavier, more symmetrical, more evenly ribbed, and the byssal gape is reduced or gone.

Audino, Serb and Marian put a mechanism under that split in the Biological Journal of the Linnean Society in 2019. Scoring mantle-margin morphology for 64 species across all six living arcidan families against a molecular phylogeny of 54, they found that enlargement of the posterior inner mantle fold, acting as a functional siphon, permitted the repeated shifts into burrowing during the Mesozoic.

That is the good part. Arks never evolved true fused siphons the way heterodont burrowers did. They went underground with a workaround, a flap of mantle enlarged until it could do a siphon's job, arrived at independently more than once.

The Eyes They Kept and the Eyes They Lost

Arks have eyes. Not one or two: rows of tiny ones along the mantle edge, and in some genera they are compound.

Dan-Eric Nilsson described them in 1994 in Philosophical Transactions of the Royal Society B, in a paper on fan worms and ark clams that framed them as optical alarm systems rather than eyes in the ordinary sense. He worked on three arcacean bivalves, among them Arca zebra and Anadara notabilis. Each ommatidium is one or two ciliary receptor cells wrapped in layers of pigment cells, and there are no lenses anywhere in the system: all the directionality comes from the pigment tube shadowing what each receptor can see. Visual motion sets them off. Nilsson's own word for them was burglar alarms.

One caveat: that paper was paywalled during this research pass, so the widely repeated figure of up to about a hundred ommatidia per eye comes here from the abstract and from work citing it. Treat it as approximate.

The evolutionary half is better documented. Audino, Serb and Marian surveyed mantle photoreceptors across 197 species and 22 families in Evolution in 2020. Arcidae carry two organ types: pigmented-cup eyespots toward the front, and compound eyes further back on the outer mantle fold. Both have a single origin, the cups first. Both were then lost, repeatedly and independently, in lineages that gave up living on the surface, including Bathyarca, Trisidos and Anadara. Across Pteriomorphia, loss rates far exceeded gain rates.

So the split above is written twice over. The arks that stayed in the open kept their alarm system. The arks that buried themselves let it go.

Where the Red Blood Came From

The best paper on this family is a genome paper, and its subject is a clam that lives nowhere near Florida. Bao Yongbo and colleagues published a chromosome-level genome of the Indo-Pacific blood clam, Tegillarca granosa, in Molecular Biology and Evolution in 2021. That animal is research material here and nothing else. It is not a Gulf species and it is not on any beach in this state.

What they found is the reason to care. The clam's circulating hemoglobins are phylogenetically closest not to anybody else's hemoglobin but to myoglobins: the blood clam's own, and those of mollusks that are not red-blooded at all. The arks' blood pigment was derived from the muscle oxygen-storage lineage, repurposed muscle protein promoted to a blood protein.

Then it gets stranger. Like ours, the blood clam's hemoglobins sit in a gene cluster, and gene phylogeny plus synteny indicate that the clam cluster and the vertebrate cluster originated independently from a single ancestral myoglobin-like gene. A full suite of heme-synthesis enzyme genes is present too, expressed most strongly in the hemolymph, which the authors describe as resembling those in vertebrates.

None of that is inherited. A clam and a person have no red-blooded common ancestor to inherit it from. The arks built the same solution separately, out of a different starting protein, and landed on the same architecture: hemoglobin, in a gene cluster, packed into red cells, fed by a vertebrate-like heme pathway. Convergent evolution at the scale of a whole physiological system.

What the Pigment Is For Is Not Settled

There are at least two live answers and nothing that adjudicates between them.

The classic reading is oxygen storage. Freadman and Mangum measured it in 1976 in Comparative Biochemistry and Physiology A, working on the ponderous ark. They reported a P50 for intact cells of about 5.30 mmHg at 10 degrees Celsius and 6.25 mmHg at 23 degrees, over a pH range of 6.6 to 7.4. That is a high oxygen affinity: the pigment holds oxygen tightly and gives it up only when supply gets genuinely poor. Blood drawn from clams held at very low ambient oxygen still contained oxygen. The picture is a reserve tank for an animal that clamps shut for hours or sits in oxygen-poor sediment.

The second answer is immune. Bao and colleagues reported in Fish and Shellfish Immunology in 2016 that both blood clam hemoglobins show peroxidase activity, and that both the intact molecules and peptides released from inside them act against Vibrio bacteria. A 2025 follow-up suggests hemoglobin may be substituting for the heme peroxidases other animals use in antimicrobial defense.

These are not mutually exclusive. A molecule can store oxygen and kill bacteria. But no source located in this research pass weighs the two against each other, so this page asserts no single settled purpose, and you should be wary of one that does. Most of the biochemistry is published under superseded genus names, Scapharca above all, which matters if you go looking for it.

The Most Familiar Ark Here Is Not an Arcid

English is the problem. Arcidae has almost no folk-name record: five vernaculars for the whole family, three of them English variants of ark shell, plus the German Archenmuscheln and a Japanese term. Its sibling family Noetiidae has no vernacular names at all. So English took the one word it had and applied it across both.

The consequences are specific. The ponderous ark, Noetia ponderosa (Say, 1822), AphiaID 156909, is a noetiid; WoRMS lists its family as Noetiidae without qualification. It is one of the most familiar heavy ribbed shells in Southwest Florida, everybody calls it an ark, and it is not in this family. The same goes for Arcopsis adamsi, AphiaID 420724, the cancellate or Adams' miniature ark. Nor are the bittersweets arks, though they share the taxodont hinge and the heft: Glycymerididae is its own accepted family, AphiaID 209.

All of them sit in the superfamily Arcoidea, which holds five accepted families: Arcidae, Noetiidae, Glycymerididae, Cucullaeidae and Parallelodontidae. That is the right frame for the ponderous ark. It is a close relative, and the red-blood story genuinely does cover it, since the classic physiology above was done on it. It just is not a member of the family this page is about.

If that reads like hair-splitting, keep going. It may be worse than hair-splitting, and not in the direction you would guess.

The Names in the Field Guides Are Mostly Superseded

This is a bad family for out-of-date names. Most ark-named shells a Southwest Florida collector meets carry a binomial in print that WoRMS and MolluscaBase no longer accept. Every accepted name below was checked against its AphiaID.

A database quirk goes with this. When a species changes genus its English common names do not always travel with it. Blood ark, mossy ark and white-beard ark all sit on the superseded records rather than the accepted ones, so searching the current name may not return the familiar folk name.

The Hairy Coat, and the Box the Family Is Named After

Two of the names in that table, mossy ark and white-beard ark, describe something you will never see on a beach shell.

The periostracum is the outer organic layer, made of conchiolin, laid over the mineral shell. In arcids it is thick, and in many species drawn out into bristles, hairs or a shaggy thatch, so a living or freshly dead ark can look furred. A beached one never does. The bold brown and white banding on a clean turkey wing is the shell showing through after the hairy coat has gone.

A function often given for that coat is camouflage. This research pass found it only in secondary sources and could not trace it to a primary experimental study, so it is offered here as a suggestion rather than a result.

The family name is simpler than the story usually told about it. Arcidae comes from the type genus Arca, and arca is just the Latin word for a box, chest or coffer, the same root behind English ark and arcane. The shell reads as a box: long straight hinge, squared ends, a flat deck between the beaks. The family is not named after Noah's ark. One Mediterranean species, Arca noae, was named after it much later, and it does not occur in Florida waters.

One last thing the name conceals. WoRMS flags Arcidae as marine, brackish and freshwater, and the freshwater flag is real: the accepted genus Scaphula lives in rivers in South and Southeast Asia.

Whether Arcidae Is a Real Group at All

Here is the open question, and it is a good one.

Molecular work keeps failing to recover Arcidae as a natural group. Kong and colleagues, working from mitochondrial genomes, reported in Molecular Phylogenetics and Evolution in 2020 that Noetiidae, Cucullaeidae and Glycymerididae are nested within a polyphyletic Arcidae. If that holds, the Arcidae and Noetiidae boundary this page has been careful about is a naming convention rather than a division in nature, and the ponderous ark is inside the arks after all.

Other studies are less drastic without being reassuring. Feng, Li and Kong in 2015 recovered the order and Noetiidae as monophyletic while finding several genera and subfamilies were not. Combosch and Giribet in 2016 found support for most arcidan families under some analytical conditions, and named Arcidae, and Arcinae in particular, as the major source of inconsistency, in dire need of taxonomic revision.

The three disagree about how bad the problem is and agree completely about its cause: arcoid classification rests on shell characters that turn out to be widely homoplastic, evolved more than once in separate lineages. Which is exactly what the mantle folds and the eyes were doing above. The morphology that built the classification is the morphology that keeps breaking it.

The standard this site follows is the current one. WoRMS and MolluscaBase treat Arcidae and Noetiidae as separate accepted families, so this page does too. The honest version is that the tree here is live science and the family may not survive it intact.

What to Look at in the One You Are Holding

Most of this page is checkable on an empty valve, which is not true of many families.

Look at the hinge line first. A long straight row of many small similar teeth, rather than a handful of large differentiated ones, puts you in the order Arcida, and that is a real taxonomic read rather than a guess. Then find the flat deck between and behind the beaks and look for the fine chevrons scored across it: the print of the duplivincular ligament, the same repeating-unit trick applied to the elastic half of the joint.

If you have both valves of one animal, run a finger along the bottom edge with the shell closed. A notch where the valves do not meet is a byssal gape, and it says the animal lived tied to something rather than buried in sediment. What you cannot see is the rest: the hairy coat, the mantle edge with its pigment-tube eyes, and the red cells that earned this family a page.

If you want a shell to be worth something later, write down where and when you picked it up. A specimen without a label is a decoration; with one it is data. Starting a collection covers the rest, and the Florida identification guide covers what else washes up here. Our own rule aboard the boat is short: empty shells only.

The rest of the Shell Monographs take the same approach to the other families in the sand. If you would rather be shown than read, we run three trips a day, and you can reserve a seat here.