The Dove Shells, and Why Almost Nobody Identifies Them

The most abundant shells nobody picks up

Tip a double handful of shell hash onto a towel. Past the broken bits and the coquina valves there will be small, glossy, spire-shaped snail shells, three to fifteen millimeters long, some banded, some checkered, some plain white. There will be a lot of them. Almost nobody keeps one.

Those are dove shells, family Columbellidae: one of the most diverse and locally abundant families of sea snails there is, and one of the least identified. Abundance is not the same thing as visibility, and an entire dominant family passing under the collector's notice is the cleanest demonstration of that on any beach.

The specialists say so themselves, and bluntly. In their 2022 molecular study of the family, Marta deMaintenon and Ellen Strong note that columbellid biology is poorly known and that relationships among columbellid taxa are essentially unresolved. They give two reasons: the animals are too small to capture the attention of most collectors, most species being under 15 millimeters as adults, and they lack obvious shell features that could be used to characterize the group.

Too small to be collected, too featureless to be classified. That is this page in one line, written by the two people best placed to know.

What abundance looks like when somebody samples the bottom

The abundance is checkable. A museum series gathered on a single day in March 2019 on this stretch of the Gulf coast holds lots of 64 and 29 fat dovesnails, and separate lots of 49, 14 and 8 of the snail now called Astyris lunata. Those are museum lot counts from targeted collecting, not densities. Columbellids also come up routinely in benthic grab samples, including National Benthic Inventory samples taken in the Ten Thousand Islands in the mid 1990s.

The neglect is checkable too. Of the six dove shells listed further down this page, only two have any standardized English common name. The other four, including two of the most frequently recorded columbellids in Southwest Florida, have none. Nobody has needed one badly enough to settle on it.

Where the dove shells sit, and what to call them

Columbellidae Swainson, 1840 is an accepted family name, AphiaID 23009 in the World Register of Marine Species, whose mollusk component is MolluscaBase.

The classification runs Gastropoda, Caenogastropoda, Neogastropoda, then the superfamily Buccinoidea, AphiaID 382214, then Columbellidae. Dove shells are buccinoideans, in the whelk superfamily, next door to the whelks and the nassa mud snails rather than off on a branch of their own, and the 2022 work supports a close relationship between columbellids and buccinids specifically.

The register lists exactly two English names for them: dove shells and dove snails. German has Taubenschnecken and the diminutive Taubchenschnecken, so the dove image is not an English-only invention.

Sizes are small throughout. Published ranges put the family at roughly 3 to 20 millimeters in shell height, most species under 15 as adults, with extremes from a little over two centimeters down to species maturing below four and a half millimeters.

The teeth define the family, not the shell

Membership in Columbellidae is traditionally confirmed by the radula, the ribbon of teeth a snail feeds with, not by the shell. deMaintenon and Strong say so directly. The shell does not make a dove shell a dove shell. The teeth do.

The radula is rachiglossan, meaning three teeth per transverse row. The middle tooth is a flat rectangular plate with no cusps at all, so it does no cutting. The two flanking lateral teeth are tall and S-curved on a narrow base, and the largest cusp on each sits on the inner side and carries a row of two to four smaller cusps set at right angles to the main cusp axis. Picture a comb mounted crosswise on a hook. Nothing else among the neogastropods has that combination.

Radular form in this family is more conservative than shell form, and in some genera it genuinely separates species. The character nobody can see is the reliable one. The one everybody can see is the noise.

A thick lip is a full stop

The classical description of the family is a list of aperture features: the shells are small, the aperture edge thickened or striated on the inside, turned inward and toothed, the inner lip doubly toothed, and the operculum very small.

Why that lip exists is the useful part, and one clean experiment inside the family shows it. Bruno Pernet worked on Amphissa columbiana Dall, 1916, AphiaID 511361, a columbellid, and published in 2007. Thin-lipped snails were growing fast and not sexually mature. Thick-lipped snails had slowed to a stop, and were mature.

That is determinate growth. The mantle adds shell at the aperture, so a snail shell lengthens from its opening outward, and how a shell is actually built covers that machinery. In a dove shell it stops. The animal grows with a thin, raw lip, reaches maturity, lays down the thick inflexed toothed lip, and never builds shell again. So the check is free: a thick-lipped dove shell is a finished adult, a thin-lipped one a juvenile that died partway through.

Two cautions. Size does not substitute for the lip check: among mature animals the largest weighed four and a half times what the smallest did. And effluent from a predatory crab produced no change in shell form or weight, so the thickening is a maturation event, not something switched on when the snail smells trouble.

Seven other families' shapes, inside one small family

Shell characters in this family are, in the specialists' phrase, highly homoplastic. Homoplasy means the same appearance keeps evolving independently in unrelated lineages, so a shared look is not evidence of shared ancestry. In most groups that is a nuisance. Here it is the central problem.

George Radwin, writing in 1977, listed 14 columbellid genera whose shells converge on the shapes of six other caenogastropod families: the augers, the whelks, the olives, the true conchs, the turrids and the cones. deMaintenon and Strong add a seventh, the miters. Seven families' worth of look-alikes, produced inside one small family of mostly sub-centimeter snails. If you have read the page on the olives, you have met one of those shapes already.

An earlier anatomical study found homoplasy running through anatomical, radular and conchological characters alike. That does not cancel what the radula is good for: it still identifies a dove shell as a dove shell. It is simply not enough on its own to sort the family into parts.

Color, in this family specifically

Dove shells are patterned, sometimes strikingly, and the patterns vary within a single species. The general lesson about color as an identifier belongs to the Florida identification guide. What matters here is how far the problem runs in this family.

Mitrella is the largest genus in Columbellidae, with well over two hundred accepted species. deMaintenon and Strong describe it as lacking any shared derived characters that would diagnose it, then list what its species do differ in: size, shape, how inflated the whorls are, proportions, and often color or pattern. Read that list again. Those are the only characters a person on a beach has. The specialists give them not as identification tools but as the reason the genus cannot be diagnosed.

One honest limit. We went looking for a published, quantified study of color variation within any Southwest Florida dove shell species and did not find one. A 1996 item by deMaintenon on phenotypic plasticity in Caribbean Columbella appears to be a meeting abstract, and we have not read it. So the variability described here comes from the homoplasy work, not from a color study we could verify, and we are not going to invent color forms or percentages to fill the gap.

The dove shells recorded on this coast

Every name below was checked against the World Register of Marine Species and carries its AphiaID. The sourcing is occurrence data for Southwest Florida and the eastern Gulf, not a survey of any one beach, and this page carries no where-to-find layer.

Columbella rusticoides needs a warning. Its name means resembling rustica, and that is the trouble. Columbella rustica is a Mediterranean and northeast Atlantic animal, and every locality in its accepted distribution sits on that side of the ocean. A few western Atlantic museum lots carry the name and are best read as mislabeled or misidentified. If a source tells you Columbella rustica lives in the Gulf of Mexico, the shell it means is almost certainly C. rusticoides.

Your field guide probably still says Anachis

If your shell guide is more than about fifteen years old, most of the dove shells in it are filed under Anachis. Most of them are not Anachis anymore. Here is the current state, plainly; why names move at all is covered on the identification guide.

• avara, semiplicata, floridana and hotessieriana are now in Costoanachis.
• obesa is now in Parvanachis.
• iontha is now in Suturoglypta.
• lafresnayi is now in Cotonopsis.
• Nassarina minor is now Steironepion minus, AphiaID 420010.
• Zafrona idalina is now Falsuszafrona idalina, AphiaID 1397761, in a genus erected only in 2020.

Two things get stated wrong in the other direction. Anachis H. Adams and A. Adams, 1853, AphiaID 137802, is still a perfectly accepted genus; these species left it, the name did not die. And Mitrella Risso, 1826, AphiaID 137804, is not obsolete either. It is the largest genus in the family, and what moved to Astyris was Mitrella lunata specifically. Be careful with that word in a search: there is a Mitrella in the miter family that is an unaccepted junior homonym, and a third that resolves to an unrelated genus, Calyptraea.

How big is the family, and does it have any parts

Two straightforward questions with unsatisfying answers. First, how many species. Three figures are in circulation and none is wrong. Araya and colleagues, writing in 2016 and citing deMaintenon's own work, give about seven hundred extant species in seventy genera. A count taken from MolluscaBase in 2021 and quoted in the 2022 phylogeny is higher, in the low nine hundreds across roughly seventy-five genera. Our own live query of the register in August 2026 returned a higher number again, though that total cannot be split into living and fossil species, because the database's extinct flag is unfilled for most records. The three were counted at different moments and are counting slightly different things. This page says there are getting on for a thousand named species, and prints no single hard number, because a hard number would claim a precision nobody has.

Second, does the family have subfamilies. Formally, yes: MolluscaBase recognizes Columbellinae Swainson, 1840, AphiaID 411755, and Atiliinae Cossmann, 1901, AphiaID 411756. In practice almost nothing lives in them. When we enumerated the family's direct children in August 2026, 103 records came back and 101 were genera sitting outside both subfamilies.

Two older schemes are best treated as history: Cossmann's 1901 split on the length of the siphonal canal, and Radwin's 1977 split into Columbellinae and Pyreninae. The 2022 study tested that structure and could not recover it, stating that its results do not support any major clades that could be called subfamilies. So the nomenclature says two subfamilies, and both practice and the phylogeny say none.

What a five-gene phylogeny found, and why the old order broke

In 2022 deMaintenon and Strong published a broad molecular phylogeny of the family in PeerJ, sequencing five mitochondrial and nuclear loci across 70 columbellid species in 31 genera. It is open access, and it is the best evidence there is on how this family is put together.

The good news came first. The family is real: monophyly of Columbellidae is strongly supported. Two long-standing members turned out not to belong. Pseudanachis basedowi moved to the Nassariidae, and Parviterebra brazieri turned out not to be in the whelk superfamily at all, but a member of the Mangeliidae.

Then the inside of the family came apart. Fourteen species of Mitrella, the largest genus, scattered across eight or more separate clades, and the genus's own type species did not group with the rest in any supported way. Seven species of Anachis fell into four different clades, and across the backbone of the tree the interior nodes were almost uniformly without robust support. The authors declined to fix it, stating that relationships within the family are not resolved well enough to propose either a subfamilial or a generic classification. The names in the table above are the accepted ones, and the people who work on them know they are provisional.

Two reasons for that collapse. The first this page has already made: the shell lies, and the historical classification was built on shells. The second is timing. The genes used here are standard workhorses for neogastropod relationships and they work elsewhere, which is why the authors call the lack of resolution somewhat surprising. Their explanation is that Columbellidae appears to be entirely Cenozoic, with its great diversification from the Miocene onward. If several hundred species appeared in a geological hurry, the order in which they appeared may simply not be recoverable, not from shells and not yet from these genes.

Which leads somewhere worth sitting with. A visitor who tips out a handful of tiny glossy spires, gives up and calls them all dove shells has arrived, by a much shorter route, at roughly where the science is. The difference is that the specialists can show their work.

What they eat, and the ones that went back to plants

There is no single family diet. deMaintenon and Strong describe most species as probably opportunistic carnivores, with some herbivorous, and the word probably is theirs. Araya and colleagues describe them as active epibenthic carnivores or scavengers. Documented herbivory covers only four named taxa in the whole family: Columbella mercatoria, its Mediterranean relative C. rustica, Rhombinella laevigata and Euplica bidentata. Four, out of getting on for a thousand. For the four species most often recorded in Southwest Florida we could find no primary diet study at all.

The teeth track the diet where anyone has looked. The herbivorous Columbella carries large, strap-like lateral teeth the authors suggest are adapted to scraping algae, against the smaller, more delicate teeth of the carnivorous taxa: a lateral tooth around 189 micrometers long in one Columbella species, against 75 and 91 micrometers in two carnivores. The authors write hypothetically and ostensibly when they draw that link, and we are keeping their hedges.

The reversal is genuinely odd. The Neogastropoda are the great predatory radiation of the sea, the cones and whelks and murexes and olives and augers, and dove shells sit inside it. A few went back to eating plants. deMaintenon devoted a 1999 paper in Invertebrate Biology to exactly that, titled for the evolution of herbivory from carnivory in this family. We have not been able to read it, because it sits behind a paywall, so we can give you the title, the direction of the finding, and the tooth evidence above, but not specific anatomical claims from a paper we have not opened.

Sorting a handful, and the animals still in it

Columbellids are mostly epibenthic on hard bottoms, with some burrowing in soft sediment. Epibenthic is the load-bearing word: they live on surfaces, on rock, rubble, algae, seagrass blades and shell hash, rather than buried inside them. Each one is a few millimeters of glass-smooth calcium carbonate, and when they die that is what goes into the wrack line, where at that size it reads as grit.

So the way to meet this family is to sit down with a handful of shell hash and good light. Look at the outer lip: thick, turned inward and toothed means a finished adult, thin and sharp means a juvenile. Then look at shape and aperture rather than color. A photo identification tool is worth trying, and iNaturalist is the credible one, but if a five-gene phylogeny cannot sort Mitrella into genera, a phone photograph of a six millimeter shell will not hand you a confident species. Family level is a real answer in this group, not a failed one.

This comes up with shell hash more than with any other kind of shelling. Our rule on the boat is our own and it is short: empty shells only. A scoop of hash off a live bottom can easily hold living animals at this size, small enough to miss. Look before you keep anything. The page on what the law actually says covers the legal side, which is separate from our house rule. Starting a collection covers why a shell with a locality and a date is data while one without is decoration. That matters more here than almost anywhere: a labeled lot of unidentified dove shells is a real record, and an unlabeled one is a jar of grit.

The Shell Monographs hub collects the rest of the families, and you can reserve a seat to see the kind of bottom this material comes off.