Deep time does not draw dotted lines between species

An umbilical-view micrograph of Globorotalia tumida from IODP Site U1559 in the South Atlantic. A shell can preserve exquisite morphology; one shell cannot reveal whether a named boundary records gradual change or a lineage split. Micrograph by Chris Lowery, University of Texas at Austin.[7]

Under an electron microscope, Globorotalia tumida looks like a species with nowhere left to hide. Its calcite shell swells through a sequence of chambers. Pores stipple the surface. A sharp outer margin gives the whole test a recognizable profile. This is the kind of anatomy from which a name can be made.

Put that shell back into a sediment core, however, and certainty acquires a vertical dimension. Older samples contain related shells whose outlines are a little flatter or whose chambers are arranged a little differently. Sample densely enough and each population may overlap the next. The shell at the bottom and the shell at the top can be diagnosably different even though no neighboring pair supplies an obvious place to draw the border.

That is the chronospecies problem. In its narrow sense, a chronospecies is a named segment of an evolving lineage, distinguished from earlier or later segments by morphology. The name can be repeatable and useful. The precise cut may still be conventional. It does not, by itself, identify the instant when one breeding population became another, prove that a lineage divided, or show that the change happened suddenly.[1][5][6]

The cover specimen comes from IODP Site U1559 in the South Atlantic and was photographed for a different ocean-history study; it is not one of the shells measured in the classic G. plesiotumida–G. tumida transition research discussed below. Its role here is exact but limited: it shows the real, intricate morphology that taxonomists must convert into characters, samples, and names.[7]

A name can be a slice without being a branch

Living species can be investigated through genes, mating, geography, ecology, and behavior as well as form. Most fossils withhold nearly all of that evidence. Paleontologists usually begin with preserved characters and ask whether a specimen belongs with the variation already known or carries a distinctive combination that deserves another name.[6]

This works cleanly when separate forms occur together and remain distinct. It becomes harder when one population follows another through time and the whole series changes in one direction. Imagine a lineage whose average shell becomes progressively more tightly coiled. One taxonomist could retain a single species and describe the trend. Another could name an early loose form and a later tight form. A third could recognize an intermediate segment too. All three schemes may sort the specimens consistently, yet only the fossils can decide whether there was one changing lineage or several branches—not the number of Latin binomials printed beside them.

Terminology does not remove the difficulty. Some researchers call successive diagnosable forms chronospecies; others call them morphospecies and reserve “species” in a deeper biological sense for independently evolving lineages. Forey and colleagues went further, recommending that paleontologists abandon chronospecies and stratospecies as special concepts and recognize fossil taxa through unique combinations of characters.[6] The disagreement matters, but it contains common ground: an operational taxonomic boundary and an evolutionary event are different claims.

“Conventional” therefore does not mean careless or imaginary. A map grid is conventional, yet it locates a quarry precisely. A chronospecies boundary can likewise provide an address inside a continuous history. Trouble begins when the address is mistaken for a fork in the road.

One rib pattern can arrive before its species

Jerzy Dzik made the problem tangible in 1985 with latest-Jurassic ammonites from Brzostówka in central Poland. Across a sequence of fossil-rich samples, shell involution and the branching pattern of ribs changed gradually. The populations were also broad: unusual forms occurred at the edges of the variation within a bed, not only at the center.[1]

Dzik contrasted two ways of naming the transition. A “vertical” or typological method picks a diagnostic feature and assigns every shell carrying it to the younger chronospecies. But a new rib pattern does not have to appear in every animal at once. It may begin in a rare extreme, become common in later populations, and eventually become typical. Under the typological rule, the younger name starts early while the older morphology is still common. Two named ranges then overlap even if only one evolving biological lineage existed.[1]

The alternative is a population method. Instead of asking whether one shell crosses a character threshold, compare whole samples: their means, spread, growth stages, and combinations of traits. Dzik proposed separating chronospecies when the distributions of a diagnostic feature no longer overlapped by a defined statistical amount. That makes the decision testable, but it also leaves a transition zone in which a sample can reasonably be assigned to either side.[1]

His worked example makes the limit visible. A nearly complete ammonite from Świętoszewo could be matched most closely to one level in the Brzostówka sequence, yet near-identical individuals occurred as uncommon variants above and below it. Correlation was probabilistic, not an exact timestamp stamped into the shell. Preservation complicated the comparison too: some ammonites were three-dimensional, others partly flattened, and still others compressed in shale. The population method improved the biological question; it did not abolish deformation, small samples, or judgment.[1]

The important shift is from portrait to distribution. A type specimen anchors a name, but it cannot contain the range of a living population. In a changing lineage, the rare shell that resembles a later type may be an early member of a trend, a separate coexisting lineage, or simply one end of ordinary variation. Its name depends on evidence outside its own outline.

A microscopic tooth records two kinds of growth

A 2023 study of latest-Devonian conodonts shows that the same problem can be attacked at a much smaller scale. Conodonts were eel-like early vertebrates whose tooth-like feeding elements accumulated in marine sediments. At Kowala Quarry in Poland’s Holy Cross Mountains, Przemysław Świś traced the proposed sequence Dasbergina marburgensis to D. trigonica through successive rock samples.[2]

Of 127 assigned elements, 85 were complete enough for morphometric measurement. Their platforms shifted from a more oval outline toward a polygonal one with new branches; the basal cavity changed with them. Thirty-four especially well-preserved elements also retained growth increments. In the oldest samples, one branch began developing around the eighth to tenth increment. Near the top of the sequence it began around the fourth or fifth. The lineage was not merely changing its adult shape. The same structure was appearing earlier in individual development, a form of evolutionary change in developmental timing known as heterochrony.[2]

The endpoint forms remain diagnosable, so two names do real work. Yet the intermediate populations form a continuum. An isolated element near the middle cannot always be identified from its outline alone; the other elements found with it and the population occupying that layer help place it. The study interprets the sequence as anagenesis—transformation within an unbranching lineage—and its growth data make that interpretation richer than a simple before-and-after comparison.[2]

That conclusion is still an inference from one sampled section, not recovered DNA or observed reproduction. Gaps, migration, ecological variation, convergence, and uneven preservation can all imitate or interrupt a morphological sequence. The value of the Kowala case is not that it finds a magical natural seam. It shows how multiple layers of evidence—population shape, position in rock, and growth within each element—can make a continuous transition biologically specific.

A smooth curve can conceal a fork

Dense sampling does not always rescue anagenesis. Sometimes it overturns it.

In 1983, Björn Malmgren, W. A. Berggren, and G. P. Lohmann described a ten-million-year Indian Ocean record of the warm-water foraminifer lineage leading from Globorotalia plesiotumida to G. tumida. They reconstructed a relatively rapid but still gradual transformation lasting roughly 600,000 years, bracketed by long intervals of morphological stability. Because they found no split, they called the pattern “punctuated gradualism.”[3]

Pincelli Hull and Richard Norris revisited this famous sequence in 2009 with different measurements and a different core. They analyzed 1,140 shells from 38 horizons at Ocean Drilling Program Site 806B in the western equatorial Pacific, recording outline shape while separating right- and left-coiling forms. A previously unrecognized flattened, right-coiling population emerged from what had looked like one smooth transition. It dominated for 414,000 years before G. tumida appeared, while evidence from other sites indicated that G. plesiotumida persisted afterward. The first fully formed G. tumida population appeared within less than 45,000 years in their age model.[4]

The revised interpretation favored branching, or cladogenesis. A changing average had hidden changing proportions of overlapping forms. Once the researchers asked whether the distribution contained mixtures—not merely whether its mean drifted—the supposed ramp resolved into a more complicated population history.[4]

This was not a genetic test, and the authors did not pretend it was. Their mixture analysis had limited power in some intervals, and no single transition sample supplied every coexistence relationship one might want. Coiling direction also had to be treated as biologically informative rather than as a trivial variant. The result was therefore a stronger case for cladogenesis, not a film of speciation happening. Its force lies in showing that an apparently continuous fossil curve can be partly a property of which traits were measured and how individuals were pooled.[4]

Species counts inherit the cutting rule

The stakes extend beyond the label on one drawer. A 2011 synthesis by Tracy Aze and colleagues built two parallel histories for Cenozoic planktonic foraminifera: one tree of named morphospecies and another of inferred evolutionary lineages. Where successive forms intergraded, the lineage tree joined multiple taxonomic names rather than treating every morphological handoff as a speciation followed by an extinction. The authors called those false biological events pseudospeciation and pseudoextinction.[5]

That bookkeeping changes the story a dataset can tell. Count every named temporal segment as a separate branch and a gradually changing lineage can manufacture biodiversity, originations, and extinctions. Collapse every similar sequence into one lineage and real branching can disappear—the exact failure exposed in the G. tumida reanalysis. Neither maximal splitting nor maximal lumping is safe.[4][5]

The answer is not to strip names from fossils. Names let researchers compare collections, communicate diagnoses, and return to the same portion of a long transition. The answer is to carry two records at once: the taxonomy used to sort forms and the evolutionary hypothesis used to connect populations. When the latter changes, the specimens do not become less real. The historical claim built from them becomes better specified.

Draw the line in pencil—and record the pencil

An honest chronospecies boundary begins with populations rather than a spectacular individual. It follows several characters instead of selecting the one that produces the tidiest break. It compares like growth stages, because a juvenile and an adult can differ more than neighboring species. It tests whether crushing, dissolution, or preparation has shifted the measurements. It looks sideways into other localities as well as upward through one section, searching for genuine coexistence that could reveal a branch.

Most importantly, it publishes the cutting rule. If a boundary follows a change in population distributions, readers should be able to see the sample sizes, overlap, and threshold. If two named forms are treated as one lineage, the intergrading specimens should be documented. If branching is inferred, the claim should survive alternative explanations such as geographic replacement, environmental shape change, or a mixture of variants.[1][2][4][5]

Forey and colleagues’ preference for distinctive character combinations remains a valuable counterweight: no time label should substitute for diagnosis.[6] Dzik’s population approach supplies the complementary warning that a diagnosis applied specimen by specimen can create false coexistence inside a continuous lineage.[1] Together they make the boundary more demanding, not less. The cut must be morphologically explicit and historically modest.

Deep time does not come with dotted lines. It gives paleontologists shells, tooth-like elements, populations, deformed samples, gaps, and—on rare occasions—sequences dense enough to watch average form move. Taxonomy turns that evidence into stable handles. Evolutionary analysis asks what those handles correspond to.

A chronospecies is most useful when it keeps those jobs separate. Its boundary can be conventional without being arbitrary in the everyday sense; evidence can constrain the placement even when no unique natural instant exists. Draw the line, explain the rule, and leave it in pencil. A later sample may move it. A better measurement may reveal that it was a branch all along.

cronfeed.work