Macroevolution, Extinction, and the Incomplete Record

What changes through deep time can be inferred robustly despite uneven sampling, changing taxonomy and geological loss?

Synthetic field investigation. Synthetic stage-level occurrences across three basins display an apparent diversity crash; students distinguish extinction from rock-record and database effects. These observations are invented for teaching, not reports from a real locality.

The unit builds a connected argument rather than a specimen catalogue. Its key themes are origination and extinction, sampling standardization, mass extinction, ecological turnover, preservation and rock availability. Every conclusion begins with a documented fossil, trace or associated rock unit and proceeds through a set of explicit assumptions. A fossil is not a direct recording of a living ecosystem: death, transport, burial, chemistry, uplift, erosion, collecting and taxonomic recognition all filter the evidence before it reaches a classroom. The scientific challenge is to separate the signal of past life from the processes that determined which signals survived. When several explanations fit the observations, the appropriate result is a constrained set of hypotheses rather than unwarranted certainty.

Imagine investigating the following setting: Synthetic stage-level occurrences across three basins display an apparent diversity crash; students distinguish extinction from rock-record and database effects. The first responsibility is to document material provenance, field associations and the sedimentary architecture. Then define what constitutes an independent measurement: a specimen, a shell valve, a fossil bed, a cataloged collection or a stratigraphic sample. An apparent increase in individuals can be created by longer time averaging or preservation of more skeletal fragments, and an apparent loss can arise when the rock containing a former habitat is absent. A robust field notebook therefore records negative searches, changed collecting effort, restricted access and preparation decisions alongside spectacular observations.

The chronology of a fossil occurrence is tied to the geological unit, and may depend on radiometric constraints, magnetostratigraphic information or biostratigraphic comparisons; no one method should be assumed to work in every sedimentary setting. The 2026/06 International Chronostratigraphic Chart is a dated reference, not a promise that numerical boundary ages will never change. Treat the classification of a species, the assignment of an age and the identification of a depositional setting as separately revisable judgments. An assessment that clearly states what is known, what is uncertain and what would decide among alternatives is more scientifically useful than a seemingly exhaustive list of names.

Lesson 7.1: Origination, extinction, and taxonomic ranges

Core idea and biological mechanism

A taxon’s first and last observed occurrences bracket a record that is normally shorter than its true lifespan; rates need time bins, taxonomy and effort controls. This topic belongs to the broader problem of origination and extinction: an observed fossil pattern is a product of former biology and later geological filtering. Identical apparent shapes or abundance patterns can arise from more than one cause. Rather than treating the textbook category as proof, identify the observations that must be true if the working mechanism is relevant at this particular site.

Mechanism lens

A working explanation has to link the material cause, the trace that survived, and the inference drawn from it. The physical evidence comes first; a familiar evolutionary narrative should not be used as an unstated premise. In this lesson, origination, extinction, and taxonomic ranges is not an isolated vocabulary item. Separate observed range endpoints from biological origins and extinctions, tracking occurrence density by stratigraphic interval. Place that procedure beside the principle: A taxon’s first and last observed occurrences bracket a record that is normally shorter than its true lifespan; rates need time bins, taxonomy and effort controls. A defensible explanation names the preserved feature, what is missing, the transformation linking original organism to specimen, and the circumstances under which that interpretation would cease to be plausible. This chain of reasoning is particularly important whenever the focal specimen is rare or visibly altered.

Counterfactual lens

Good explanations predict cases in which the preferred interpretation should fail. Researchers can look for those cases in an independently collected outcrop, a different depositional environment or a carefully chosen replicate. One potential failure mode is specific: A missing occurrence between two fossil-rich levels may simply be a sampling gap, not a local extinction and recolonization. An alternative model could therefore preserve the measured numbers while changing the biological explanation. Make a prediction that differs between the two explanations, specify the material that would distinguish them, and record in advance what observation would count against your favored inference. This prevents a positive identification or an apparent temporal pattern from becoming unfalsifiable.

Uncertainty lens

Distinguish measurement precision, model choice, taxonomic uncertainty, age control and sampling incompleteness. These sources can affect a conclusion differently and should not be hidden in a single unexplained error bar. A missing occurrence between two fossil-rich levels may simply be a sampling gap, not a local extinction and recolonization. Where dates or boundaries enter the argument, give the chronostratigraphic source and its version. Where taxonomy enters, identify the revision followed. Where anatomical measurements enter, describe landmarks and observer repeatability. Always distinguish an interval bounded by first and last observed occurrences from the unobservable true lifespan of the taxon; this distinction protects conclusions against an overly complete picture of deep time.

Field notebook and worked problem

Worked quantitative case 7.1: Apparent percentage disappearance. 95 taxa last reported in older bin, 61 observed in younger bin; percent not observed after?

Calculation. Use the explicit arithmetic expression (95-61)/95*100. With the given hypothetical measurements the result is 35.79 %. The calculation is reproducible from the stated values; precision of the displayed result is not a measurement uncertainty. Interpretation boundary: False absences and migration inflate raw disappearance.

Independent task. Compute per-capita apparent disappearance rate in a hypothetical interval. Produce a figure with named axes or a specimen-character table, a statement of the sampling unit, and a two-sentence comparison between your preferred interpretation and a defensible alternative. Include the observation most likely to revise your conclusion.

Check for understanding. What would change if a portion of the fossil record were never preserved? How might an independent collector find a different result? What information is required before the number calculated above can become a biological claim? How will the specimen and locality be recoverable for a later researcher?

Lesson 7.2: Sampling standardization and rarefaction

Core idea and biological mechanism

Observed richness rises with sample size, so assemblages from different formations often require equal-effort comparison or model-based correction. This topic belongs to the broader problem of sampling standardization: an observed fossil pattern is a product of former biology and later geological filtering. Identical apparent shapes or abundance patterns can arise from more than one cause. Rather than treating the textbook category as proof, identify the observations that must be true if the working mechanism is relevant at this particular site.

Mechanism lens

A working explanation has to link the material cause, the trace that survived, and the inference drawn from it. The physical evidence comes first; a familiar evolutionary narrative should not be used as an unstated premise. In this lesson, sampling standardization and rarefaction is not an isolated vocabulary item. Subsample occurrence lists at matched numbers of specimens or collections and plot accumulation curves with uncertainty. Place that procedure beside the principle: Observed richness rises with sample size, so assemblages from different formations often require equal-effort comparison or model-based correction. A defensible explanation names the preserved feature, what is missing, the transformation linking original organism to specimen, and the circumstances under which that interpretation would cease to be plausible. This chain of reasoning is particularly important whenever the focal specimen is rare or visibly altered.

Counterfactual lens

Good explanations predict cases in which the preferred interpretation should fail. Researchers can look for those cases in an independently collected outcrop, a different depositional environment or a carefully chosen replicate. One potential failure mode is specific: A higher raw species count from a thoroughly excavated site is weak evidence of higher biological richness. An alternative model could therefore preserve the measured numbers while changing the biological explanation. Make a prediction that differs between the two explanations, specify the material that would distinguish them, and record in advance what observation would count against your favored inference. This prevents a positive identification or an apparent temporal pattern from becoming unfalsifiable.

Uncertainty lens

Distinguish measurement precision, model choice, taxonomic uncertainty, age control and sampling incompleteness. These sources can affect a conclusion differently and should not be hidden in a single unexplained error bar. A higher raw species count from a thoroughly excavated site is weak evidence of higher biological richness. Where dates or boundaries enter the argument, give the chronostratigraphic source and its version. Where taxonomy enters, identify the revision followed. Where anatomical measurements enter, describe landmarks and observer repeatability. Always distinguish an interval bounded by first and last observed occurrences from the unobservable true lifespan of the taxon; this distinction protects conclusions against an overly complete picture of deep time.

Field notebook and worked problem

Worked quantitative case 7.2: Richness ratio. Richness from two study intervals is 88 and 55 after a fixed nominal collecting protocol.

Calculation. Use the explicit arithmetic expression 88/55. With the given hypothetical measurements the result is 1.6 ratio. The calculation is reproducible from the stated values; precision of the displayed result is not a measurement uncertainty. Interpretation boundary: Equal nominal effort may still mask unequal preservation.

Independent task. Estimate expected taxa in a simple random subsample from specimen counts. Produce a figure with named axes or a specimen-character table, a statement of the sampling unit, and a two-sentence comparison between your preferred interpretation and a defensible alternative. Include the observation most likely to revise your conclusion.

Check for understanding. What would change if a portion of the fossil record were never preserved? How might an independent collector find a different result? What information is required before the number calculated above can become a biological claim? How will the specimen and locality be recoverable for a later researcher?

Lesson 7.3: Mass extinction and background turnover

Core idea and biological mechanism

Major extinction intervals combine severe biological change with ecological and environmental restructuring, but their magnitudes depend on taxonomic and time-bin definitions. This topic belongs to the broader problem of mass extinction: an observed fossil pattern is a product of former biology and later geological filtering. Identical apparent shapes or abundance patterns can arise from more than one cause. Rather than treating the textbook category as proof, identify the observations that must be true if the working mechanism is relevant at this particular site.

Mechanism lens

A working explanation has to link the material cause, the trace that survived, and the inference drawn from it. The physical evidence comes first; a familiar evolutionary narrative should not be used as an unstated premise. In this lesson, mass extinction and background turnover is not an isolated vocabulary item. Compare extinction intensity across successive bins and connect hypotheses to stratigraphic, geochemical and sedimentological evidence. Place that procedure beside the principle: Major extinction intervals combine severe biological change with ecological and environmental restructuring, but their magnitudes depend on taxonomic and time-bin definitions. A defensible explanation names the preserved feature, what is missing, the transformation linking original organism to specimen, and the circumstances under which that interpretation would cease to be plausible. This chain of reasoning is particularly important whenever the focal specimen is rare or visibly altered.

Counterfactual lens

Good explanations predict cases in which the preferred interpretation should fail. Researchers can look for those cases in an independently collected outcrop, a different depositional environment or a carefully chosen replicate. One potential failure mode is specific: A correlation with a global environmental perturbation does not establish a unique causal mechanism without temporal and mechanistic tests. An alternative model could therefore preserve the measured numbers while changing the biological explanation. Make a prediction that differs between the two explanations, specify the material that would distinguish them, and record in advance what observation would count against your favored inference. This prevents a positive identification or an apparent temporal pattern from becoming unfalsifiable.

Uncertainty lens

Distinguish measurement precision, model choice, taxonomic uncertainty, age control and sampling incompleteness. These sources can affect a conclusion differently and should not be hidden in a single unexplained error bar. A correlation with a global environmental perturbation does not establish a unique causal mechanism without temporal and mechanistic tests. Where dates or boundaries enter the argument, give the chronostratigraphic source and its version. Where taxonomy enters, identify the revision followed. Where anatomical measurements enter, describe landmarks and observer repeatability. Always distinguish an interval bounded by first and last observed occurrences from the unobservable true lifespan of the taxon; this distinction protects conclusions against an overly complete picture of deep time.

Field notebook and worked problem

Worked quantitative case 7.3: Morphospace spread ratio. A standardized trait variance summary increases from 3.0 to 4.5.

Calculation. Use the explicit arithmetic expression 4.5/3.0. With the given hypothetical measurements the result is 1.5 ratio. The calculation is reproducible from the stated values; precision of the displayed result is not a measurement uncertainty. Interpretation boundary: Different dimensionalities invalidate naive ratio comparisons.

Independent task. Compute percent loss between two synthetic taxonomic inventories. Produce a figure with named axes or a specimen-character table, a statement of the sampling unit, and a two-sentence comparison between your preferred interpretation and a defensible alternative. Include the observation most likely to revise your conclusion.

Check for understanding. What would change if a portion of the fossil record were never preserved? How might an independent collector find a different result? What information is required before the number calculated above can become a biological claim? How will the specimen and locality be recoverable for a later researcher?

Lesson 7.4: Adaptive radiation and ecological restructuring

Core idea and biological mechanism

Lineage diversification, ecological disparity and geographic spread need not peak simultaneously; post-crisis reorganizations can occupy many timescales. This topic belongs to the broader problem of ecological turnover: an observed fossil pattern is a product of former biology and later geological filtering. Identical apparent shapes or abundance patterns can arise from more than one cause. Rather than treating the textbook category as proof, identify the observations that must be true if the working mechanism is relevant at this particular site.

Mechanism lens

A working explanation has to link the material cause, the trace that survived, and the inference drawn from it. The physical evidence comes first; a familiar evolutionary narrative should not be used as an unstated premise. In this lesson, adaptive radiation and ecological restructuring is not an isolated vocabulary item. Plot richness, trait variance and geographic distributions separately through successive dated horizons. Place that procedure beside the principle: Lineage diversification, ecological disparity and geographic spread need not peak simultaneously; post-crisis reorganizations can occupy many timescales. A defensible explanation names the preserved feature, what is missing, the transformation linking original organism to specimen, and the circumstances under which that interpretation would cease to be plausible. This chain of reasoning is particularly important whenever the focal specimen is rare or visibly altered.

Counterfactual lens

Good explanations predict cases in which the preferred interpretation should fail. Researchers can look for those cases in an independently collected outcrop, a different depositional environment or a carefully chosen replicate. One potential failure mode is specific: Rapid increase in named taxa can result from improved collecting or a shift in taxonomic practice. An alternative model could therefore preserve the measured numbers while changing the biological explanation. Make a prediction that differs between the two explanations, specify the material that would distinguish them, and record in advance what observation would count against your favored inference. This prevents a positive identification or an apparent temporal pattern from becoming unfalsifiable.

Uncertainty lens

Distinguish measurement precision, model choice, taxonomic uncertainty, age control and sampling incompleteness. These sources can affect a conclusion differently and should not be hidden in a single unexplained error bar. Rapid increase in named taxa can result from improved collecting or a shift in taxonomic practice. Where dates or boundaries enter the argument, give the chronostratigraphic source and its version. Where taxonomy enters, identify the revision followed. Where anatomical measurements enter, describe landmarks and observer repeatability. Always distinguish an interval bounded by first and last observed occurrences from the unobservable true lifespan of the taxon; this distinction protects conclusions against an overly complete picture of deep time.

Field notebook and worked problem

Worked quantitative case 7.4: Recovery fraction. Toy diversity recovery is 1-exp(-t/6), with t=8 Myr.

Calculation. Use the explicit arithmetic expression 1-math.exp(-8/6). With the given hypothetical measurements the result is 0.7364 fraction. The calculation is reproducible from the stated values; precision of the displayed result is not a measurement uncertainty. Interpretation boundary: No fitted timescale or causal recovery mechanism inferred.

Independent task. Compare richness growth factors and morphospace variance before and after a modeled event. Produce a figure with named axes or a specimen-character table, a statement of the sampling unit, and a two-sentence comparison between your preferred interpretation and a defensible alternative. Include the observation most likely to revise your conclusion.

Check for understanding. What would change if a portion of the fossil record were never preserved? How might an independent collector find a different result? What information is required before the number calculated above can become a biological claim? How will the specimen and locality be recoverable for a later researcher?

Lesson 7.5: The quality of the fossil record

Core idea and biological mechanism

Sedimentary rock area, facies distribution, exposed outcrops, taxonomic practices and collecting history all influence diversity curves. This topic belongs to the broader problem of preservation and rock availability: an observed fossil pattern is a product of former biology and later geological filtering. Identical apparent shapes or abundance patterns can arise from more than one cause. Rather than treating the textbook category as proof, identify the observations that must be true if the working mechanism is relevant at this particular site.

Mechanism lens

A working explanation has to link the material cause, the trace that survived, and the inference drawn from it. The physical evidence comes first; a familiar evolutionary narrative should not be used as an unstated premise. In this lesson, the quality of the fossil record is not an isolated vocabulary item. Construct a causal diagram linking preservation, publication, field access and true abundance to observed counts. Place that procedure beside the principle: Sedimentary rock area, facies distribution, exposed outcrops, taxonomic practices and collecting history all influence diversity curves. A defensible explanation names the preserved feature, what is missing, the transformation linking original organism to specimen, and the circumstances under which that interpretation would cease to be plausible. This chain of reasoning is particularly important whenever the focal specimen is rare or visibly altered.

Counterfactual lens

Good explanations predict cases in which the preferred interpretation should fail. Researchers can look for those cases in an independently collected outcrop, a different depositional environment or a carefully chosen replicate. One potential failure mode is specific: A statistical correction can reduce one bias but inadvertently amplify another if its assumptions are violated. An alternative model could therefore preserve the measured numbers while changing the biological explanation. Make a prediction that differs between the two explanations, specify the material that would distinguish them, and record in advance what observation would count against your favored inference. This prevents a positive identification or an apparent temporal pattern from becoming unfalsifiable.

Uncertainty lens

Distinguish measurement precision, model choice, taxonomic uncertainty, age control and sampling incompleteness. These sources can affect a conclusion differently and should not be hidden in a single unexplained error bar. A statistical correction can reduce one bias but inadvertently amplify another if its assumptions are violated. Where dates or boundaries enter the argument, give the chronostratigraphic source and its version. Where taxonomy enters, identify the revision followed. Where anatomical measurements enter, describe landmarks and observer repeatability. Always distinguish an interval bounded by first and last observed occurrences from the unobservable true lifespan of the taxon; this distinction protects conclusions against an overly complete picture of deep time.

Field notebook and worked problem

Worked quantitative case 7.5: Bin-occupancy detection. Taxon occurs in 36 of 60 synthetic equally sampled bins.

Calculation. Use the explicit arithmetic expression 36/60. With the given hypothetical measurements the result is 0.6 fraction. The calculation is reproducible from the stated values; precision of the displayed result is not a measurement uncertainty. Interpretation boundary: A regional occupancy proxy, not global detection probability.

Independent task. Compute observed detection probability across bins with known artificial occupancy. Produce a figure with named axes or a specimen-character table, a statement of the sampling unit, and a two-sentence comparison between your preferred interpretation and a defensible alternative. Include the observation most likely to revise your conclusion.

Check for understanding. What would change if a portion of the fossil record were never preserved? How might an independent collector find a different result? What information is required before the number calculated above can become a biological claim? How will the specimen and locality be recoverable for a later researcher?

Research brief. Synthetic stage-level occurrences across three basins display an apparent diversity crash; students distinguish extinction from rock-record and database effects. Design a student investigation that tests whether an observed difference reflects ancient biology or the way rocks and specimens were sampled. Begin with a map, stratigraphic table or measured section; create unique IDs, lithology fields, preservation grades, sample dates and collector identifiers. Make a distinction between information recorded directly from fossil material and information inferred by a model, and mark fictional values as synthetic at every stage.

Recommended workflow

1. Define a focused question and mutually distinguishable predictions before any calculation. Identify the comparison units and the temporal resolution.
2. Audit original observations, including unknown and unusable specimens. Document facies changes and the risk of reworking or time averaging.
3. Examine at least two admissible analysis choices and preserve the alternative results alongside the selected result.
4. Use the provided synthetic dataset as a calculation exercise, not as evidence about a real outcrop.
5. Report a figure, an uncertainty statement, an explicit limitation, and an observation that would cause revision.
6. Record collection permissions and archive metadata, including any reason not to publish exact locality coordinates.

Reporting and interpretation

A final report should distinguish the statement “we observed this character on this preserved element” from “this animal used this behavior” or “this group diversified at this time.” Explain the inferential bridge that connects each stage. A statistical trend in specimen counts requires assumptions about collection effort, taxonomic consistency and geological availability before it supports an evolutionary claim. Conversely, a well-preserved fossil with excellent anatomical detail can be weak evidence for local abundance if its origin or geological setting is poorly documented. A transparent limitations paragraph is an essential result, not a concession.

Two self-contained browser models accompany this unit: Apparent taxon loss across a boundary and Illustrative diversity recovery curve. Each changes one explicitly described parameter and computes a toy result using the expression provided in MODELS.json. Record values near the minimum, midpoint and maximum, and explain why the output should not be treated as a field-validated biological estimate. In particular, avoid applying a visual trend line to stratigraphic gaps, population histories or laboratory processes for which the underlying simplifications are false.

• Explain the difference between a direct fossil observation and a model-dependent paleoecological interpretation.
• Propose a preservation or sampling mechanism that could reverse a naive abundance comparison.
• Document an alternate character coding, correlation or stratigraphic age interpretation.
• Show the units and arithmetic in one worked case and identify a meaningful source of uncertainty.
• Describe how you would make your analysis reproducible without exposing protected fossil localities.