Biology · Species, Speciation & Macroevolution
Diversification, adaptive radiations and extinction
A clade adds two surviving branches to its history but loses one lineage. Its living count rises by only one. Looking only at today's tips conceals some of what happened: diversification includes both lineage origin and disappearance.
Evolutionary Processes & Adaptation supplies allele frequencies, selection, drift, gene flow and the distinction between observation and explanation. Heredity, Variation & Evolution introduces branching ancestry. Meiosis & Patterns of Inheritance helps with the chromosome-set model; Development, Differentiation & Stem Cells helps with constraints and historical context. The examples require proportions, weighted averages and reading simple trees; each rule is explained where it is used. Species definitions for sexual organisms are introduced with their limits rather than assumed to classify every form of life.
Back to the Biology mapBy the end, you can
Count lineage splitting and extinction separately, and distinguish ecological diversification from an increase in species number alone.
A lineage count has gains and losses
Macroevolution concerns patterns and processes at and above the species level, including lineage diversification, extinction and large-scale trait change. It connects population processes with histories of lineage persistence and splitting. It does not require an organism to aim for a higher rank or a more elaborate body.
In a binary splitting model, one lineage becomes two, increasing the count by one. An extinction removes a lineage, decreasing the count by one. A change in a trait along one lineage does not necessarily add another lineage. Keep branching and change within a branch distinct.
The constructed ten-million-year history starts with A,B,C,D: four lineages. A splits at time 2, giving five. B splits at time 6, giving six. A2 ends at time 7, leaving five at time 10. The tally is 4+2−1=5. Neither the final count nor the net increase of one identifies the two gains and one loss by itself.
Two splits and one loss give a net increase of one
Read the diagram in words
A and B each split once; C and D continue without splitting. A's daughters are A1, surviving to 10, and A2, ending at 7. B's daughters B1 and B2 both survive to 10. Counts are 4 during 0–2, 5 during 2–6, 6 during 6–7 and 5 during 7–10. Exposure is 4×2+5×4+6×1+5×3=49 lineage-million-years. Two splits and one extinction yield final count 5, a net gain of one.
| Lineage | Start | End | Lineage-time contribution |
|---|---|---|---|
| A | 0 | 2 | 2 |
| A1 | 2 | 10 | 8 |
| A2 | 2 | 7 | 5 |
| B | 0 | 6 | 6 |
| B1 | 6 | 10 | 4 |
| B2 | 6 | 10 | 4 |
| C | 0 | 10 | 10 |
| D | 0 | 10 | 10 |
| Time in million years | Lineage count | Lineage-time contribution |
|---|---|---|
| 0–2 | 4 | 8 |
| 2–6 | 5 | 20 |
| 6–7 | 6 | 6 |
| 7–10 | 5 | 15 |
| Time in million years | Event | Lineage | Descendant lineages |
|---|---|---|---|
| 2 | Binary split: one lineage becomes two | A | A1, A2 |
| 6 | Binary split: one lineage becomes two | B | B1, B2 |
| 7 | A2 ends at time 7 | A2 |
| Net count change per million years | 0.1 |
|---|---|
| Splits per lineage-million-year | ≈ 0.04082 |
| Extinctions per lineage-million-year | ≈ 0.02041 |
| Net change per lineage-million-year | ≈ 0.02041 |
4 + 2 − 1 = 5; Observed lineage-time: 49 lineage-million-years; No ecological roles are assigned
A completely known invented ten-million-year history starts with four distinct lineages A, B, C and D. Binary splits replace one lineage with two (net +1); extinction removes one. No immigration, hybrid merger or taxonomic reclassification. A splits at 2, B at 6, and A2 goes extinct at 7. All five surviving descendants are sampled at 10.
An adaptive radiation combines ancestry with ecological diversification
An adaptive radiation involves diversification of a lineage into forms associated with different ecological roles, often with substantial lineage multiplication over a relatively short interval. Species number, relatedness, ecological differences and the timescale all matter. A fast increase in names alone does not establish the adaptive part.
To investigate a radiation, test whether the organisms descend from a relevant common lineage, identify differences in resource or habitat use, and examine how traits relate to those differences. Newly accessible environments, resources or interactions can provide ecological opportunity. Opportunity does not guarantee a particular outcome; available variation, competitors and historical conditions also matter.
Comparative work on Greater Antillean Anolis lizards found repeated origins of similar habitat-associated forms on different islands. Morphological similarity and ancestry comparisons addressed whether those forms originated once and spread or evolved repeatedly. The example shows how related evolutionary questions need different evidence, rather than making every radiation identical or predictable.
The ledger's branches have no ecological characters attached. It therefore illustrates lineage diversification but cannot be labelled an adaptive radiation from its counts alone. Ecological evidence would be an additional layer of the explanation.
Extinction differs from failure to observe a lineage
Extinction ends a lineage's living descendants. Local disappearance from one site is different from global extinction if descendants persist elsewhere. In fossils, a last observed occurrence is a sampling result; it need not be the exact moment when the final organism died. Uneven preservation and collecting affect the record.
Large losses across many lineages over a geologically restricted interval are described as mass extinction, relative to the background context. Their selective effects and later ecological consequences can differ among groups. A decline in a particular site's fossil count alone does not establish a global event or its cause.
After a loss, surviving lineages inherit a changed ecological setting. They may diversify, persist with limited change or later disappear themselves. An extinction is not a process that supplies survivors with whatever beneficial variants they will need.
Living descendants are an incomplete history
The two small histories end with living descendants P and Q at time 6. Both show their shared split at time 2. The second also includes a side lineage R arising at time 3 and going extinct at 4. Remove R and suppress the unobserved split along P's branch: the sampled living tree matches the simpler history.
The complete histories have different gain and loss counts despite the same sampled living relationship. This drawing illustrates why fossils, sampling information and justified model assumptions can add information beyond present-day tips. It is not a full statistical proof, and it does not make every diversification analysis uninformative.
The same living tips can conceal different histories
Read the diagram in words
History 1 has one split at 2, producing P and Q through time 6. History 2 has that same split, plus a split on P's ancestral branch at 3 that produces R; R goes extinct at 4. The complete counts are one split/no extinction versus two splits/one extinction. Pruning R and suppressing the resulting unobserved branch point leaves P and Q with the same shared split at 2 in both histories.
History 1: no hidden side branch
| Lineage | Start | End |
|---|---|---|
| Shared ancestral lineage | 0 | 2 |
| P | 2 | 6 |
| Q | 2 | 6 |
| Number of splits | 1 |
|---|---|
| Number of extinctions | 0 |
History 2: an extinct side branch
| Lineage | Start | End |
|---|---|---|
| Shared ancestral lineage | 0 | 2 |
| Earlier P branch | 2 | 3 |
| P | 3 | 6 |
| R | 3 | 4 |
| Q | 2 | 6 |
| Number of splits | 2 |
|---|---|
| Number of extinctions | 1 |
| Lineage | Start | End |
|---|---|---|
| Shared ancestral lineage | 0 | 2 |
| P | 2 | 6 |
| Q | 2 | 6 |
Remove extinct unsampled R; Suppress the unobserved split along P; Both give the same sampled P/Q tree; A missing-history illustration, not a statistical proof
Two invented complete histories have the same sampled living descendants P and Q and the same visible split at 2 in a 0–6 time window. History 2 also has an extinct unsampled side lineage R. These cartoons demonstrate missing information; they are not a statistical proof of diversification-model identifiability.
Worked example
Net increase is not the number of origins
Read the ledger chronologically: counts are 4,5,6,5 across the successive intervals. Two splits contribute +2; one extinction contributes −1; final minus initial is +1. Dividing the net increase by ten million years gives 0.1 lineages per million years for this window.
That value is a raw count-change rate. It is not yet a per-lineage speciation rate: that calculation also requires how long the different numbers of lineages were present. A separate denominator can change the question even when the numerator comes from the same history.
Check your understanding
A living count rises from four to five. Does this imply exactly one split and no extinction?
Show reasoning
No. Our model has two splits and one extinction, producing the same net increase. Endpoint counts alone do not separate gains from losses.
Why is the ledger insufficient to establish an adaptive radiation?
Show reasoning
It supplies branch history and times but no ecological roles, trait functions or evidence of adaptive divergence. Lineage multiplication is only part of the claim.
What changes when extinct R is removed from the second P/Q history?
Show reasoning
The visible sampled tree loses one extinct tip and its extra split. It retains the P/Q relationship and time-2 split, so the living tree conceals a different complete gain/loss history.
What to carry forward
Count origins and losses separately, then add ecological, fossil and sampling evidence to explain the larger pattern.
References
Resources supporting this lesson’s explanations and examples.
- Contingency and Determinism in Replicated Adaptive Radiations of Island LizardsJonathan B. Losos, Todd R. Jackman, Allan Larson, Kevin de Queiroz and Lourdes Rodríguez-SchettinoOriginal research — selected article text
- Extant timetrees are consistent with a myriad of diversification histories — author overviewStilianos Louca and Matthew W. PennellAuthor research overview and data context
- Biology 2e — formation of species, reconnection and ratesMary Ann Clark, Matthew Douglas and Jung ChoiOpen textbook — selected sections18.2 Formation of New Species18.3 Reconnection and Speciation Rates
- Understanding Evolution — species and macroevolutionUnderstanding Evolution teamUniversity educational guide — selected sectionsDefining a speciesMacroevolutionWhat is macroevolution?