How Science Works | How New Species Form as Barriers to Gene Exchange Evolve

For many sexually reproducing organisms, the biological species concept emphasises populations that interbreed in nature and are reproductively isolated from other such populations. It is useful because it connects a boundary with a process: which organisms contribute inherited variation to a continuing population system?

The definition cannot be applied in exactly the same way to asexual organisms, fossils or populations whose potential interbreeding cannot be observed. Other species concepts emphasise separately evolving lineages, diagnostic features, ecology or phylogenetic relationships. The Natural History Museum’s explanation of species describes why no single operational test works equally well for all organisms.

Even in sexual organisms, one successful laboratory cross does not necessarily erase a species boundary. Conditions that allow mating in captivity may rarely occur in nature. Conversely, failure of a particular pair to reproduce does not prove that their populations are different species; age, condition, compatibility and chance can affect that pair.

The unit of interest is therefore a population or lineage over time, not an isolated animal or one unusually coloured specimen. Researchers ask whether differences are stable, inherited and connected with independent evolutionary trajectories.

This does not make species arbitrary. It means the evidence and definition must fit the organism and question. A careful article can report strong reproductive isolation while acknowledging occasional hybrids. It can also describe an incipient split without declaring that every early difference already deserves a separate species name.

Contents · Next chapter

Gene flow occurs when genetic variants move between populations and contribute to subsequent generations. An individual crossing a valley is not enough by itself. If it never reproduces, its movement may be dispersal without effective gene flow.

Regular gene flow can reduce differences in allele frequencies, although its effect depends on selection, drift, population size and the genetic architecture of the traits. An allele is a version of a genetic sequence. Different alleles can be common in different populations even while some exchange continues.

Original one-generation mixing model. Suppose an allele has frequency 0.80 in population A and 0.20 in population B. Assume that each population receives a 10% contribution from the other’s original gene pool, with no selection, drift or other change. The updated frequency in A is 0.90 × 0.80 + 0.10 × 0.20 = 0.74. In B it is 0.90 × 0.20 + 0.10 × 0.80 = 0.26.

The gap falls from 0.60 to 0.48, a 20% reduction in this single symmetric mixing step. This is an accounting example, not a measurement or a universal migration threshold for speciation. Different population sizes, asymmetric contributions or selection could produce different outcomes.

The calculation explains why reduced exchange can matter: it weakens one force that tends to connect the gene pools. It does not show that any isolated population must become a species. Populations may remain similar, reconnect, disappear or diverge without developing strong reproductive barriers. Isolation creates possibilities; the subsequent evolutionary history determines what follows.

Contents · Previous chapter · Next chapter

Geographic separation can reduce gene exchange. A new barrier, a colonisation event or a change in habitat distribution may divide a formerly connected population. Divergence in separated populations is commonly discussed as allopatric speciation.

The barrier’s effect depends on the organism. A river may block one terrestrial animal while providing a route for another. A mountain range may strongly separate lowland populations but have little effect on a species capable of moving across it. The relevant quantity is realised exchange, not the dramatic appearance of the landscape.

After separation, populations can experience different selection pressures and different random genetic changes. Differences may accumulate in habitat use, physiology, reproduction or many other traits. Some of those differences can later reduce interbreeding if the populations meet again.

A useful test asks what happens at secondary contact. Do the populations merge readily? Do they remain largely separate? Is there a hybrid zone? Are hybrids viable and fertile, and how do they perform under natural conditions? Historical separation and present reproductive independence are connected but distinct claims.

For an original scenario, imagine two island bird populations with different plumage after a long separation. Their appearance alone cannot tell us whether they would maintain separate lineages on the same island. Behavioural observations, genomic evidence and ecological comparisons would bear on that question. The scientist should avoid converting the phrase “they have been apart a long time” into an automatic species verdict.

Contents · Previous chapter · Next chapter

Prezygotic barriers reduce the chance that members of different populations produce a fertilised egg. Populations may occupy different habitats, reproduce in different seasons, respond to different mating signals or have reduced compatibility between their reproductive cells.

Several modest barriers can combine. A difference in flowering time reduces encounters; different pollinators further reduce transfer between plant groups; pollen interactions can alter which fertilisations succeed. The process need not depend on one absolute wall.

In research on the monkeyflowers Mimulus lewisii and M. cardinalis, investigators measured multiple reproductive barriers, including geographic and ecological distribution, pollinator behaviour, pollen competition and hybrid performance. This approach asked where gene exchange was reduced rather than merely noting that the flowers looked different.

The order of barriers matters when estimating their combined effect. A later barrier acts only on the opportunities that passed earlier stages. Adding several percentages of isolation without accounting for that sequence can produce an impossible or misleading total.

A barrier can also be asymmetric. Pollen from population A may perform differently on B than B’s pollen does on A. Likewise, one direction of an animal cross can differ from the reverse. Reporting only one direction can hide important biology.

For an explanatory diagram, draw a sequence of gates: encounter, mating or pollen transfer, and fertilisation. Label what was actually observed at each gate. Avoid writing “different species” above the first gate as though the name itself caused the failure. The mechanism lies in interactions among organisms and their environment.

Contents · Previous chapter · Next chapter

Postzygotic barriers affect hybrids after a fertilised egg has formed. A hybrid may develop poorly, survive less well in a particular environment or produce fewer successful offspring. Some hybrids survive but have reduced fertility; others perform well in one setting and poorly in another.

Intrinsic incompatibility concerns problems arising from the combination of the parental genetic systems. Ecological disadvantage concerns how a hybrid’s traits perform in its environment. These can coexist, and a greenhouse or laboratory result may not capture the full ecological effect.

An original two-factor model makes the logic visible. Imagine an ancestral combination of factors a and b that works well. In one separated population, a derived A works with b. In the other, a derived B works with a. Suppose A and B interfere when brought together in a hybrid. Each parental population can function well even though the new combination performs poorly.

This is a simplified illustration of genetic incompatibility, not a claim about particular genes or a universal dominance pattern. Its key point is that neither population needs to pass through the poorly functioning hybrid combination while diverging separately. Interactions among changes can create a barrier that no single change created alone.

The monkeyflower study also measured hybrid performance, showing why viability and fertility should not be collapsed into one “can reproduce” label. A hybrid can germinate, grow and flower while still contributing fewer successful gametes. Strong inference follows the life cycle far enough to measure the reproductive outcome named in the claim.

Contents · Previous chapter · Next chapter

Original conditional-probability exercise. Start with 1,000 hypothetical opportunities for between-population reproduction. Suppose 40% pass an encounter stage. Of those that meet, 25% achieve fertilisation. Of the resulting offspring, 50% later contribute to the next generation under the model’s conditions.

The expected counts are 400 encounters, 100 fertilisations and 50 contributing offspring. The combined fraction is 0.40 × 0.25 × 0.50 = 0.05, or 5% of the original opportunities. Every stage allows some success, yet the combined pathway permits only a small fraction through.

The multiplication uses conditional fractions: the 25% applies to the group that passed the encounter stage, and the 50% applies to the offspring produced. It does not require assuming that unrelated overall percentages can simply be multiplied without checking their denominators.

Relative to an explicitly hypothetical baseline in which every opportunity passed every stage, the reduction is 95%. A real biological comparison would need a meaningful within-population baseline and uncertainty estimates. We have not measured reproductive isolation in an actual species pair.

Now remove the encounter difference while holding the later conditional fractions fixed. The model gives 1.00 × 0.25 × 0.50 = 0.125, or 12.5%. The first barrier mattered, but it was not the whole explanation. If environmental change also altered later stages, this simple recalculation would no longer suffice.

The lesson is practical: ask what the starting opportunities are, which stage each measurement describes and how earlier filters change the population reaching later ones. A list of barrier names becomes explanatory when those relationships are explicit. No numerical percentage by itself supplies a universal point at which a population officially becomes a species.

Contents · Previous chapter · Next chapter

Populations can become differentiated without complete geographic separation. Different habitat preferences, mating choices or locally favoured traits can reduce effective exchange even where populations occupy overlapping regions. This makes the relationship between selection and gene flow central.

Different parts of the genome can also behave differently. Variants associated with important reproductive or ecological differences may face stronger barriers than variants elsewhere. A species boundary can therefore be permeable to some genetic material without disappearing altogether.

Genome-wide work on Heliconius butterflies examined divergence alongside gene flow. More recently, a 2024 study of hybrid speciation in Heliconius connected ecological and mating-related traits with the persistence of a distinct lineage despite ongoing exchange. These are studies of particular histories, not permission to infer speciation from any genetic cluster.

Hybridisation has several possible outcomes. It can weaken differences, introduce useful variation into an existing population or contribute to the origin of a new lineage. The outcome depends on inheritance, ecology, mating and population history. “Hybrids exist” is therefore neither an automatic disproof of species nor proof of a new one.

For a reading exercise, imagine a genomic analysis showing some regions shared across two groups and other regions strongly differentiated. Ask which histories could create that pattern: recent exchange, ancestral variation, selection or a combination. An attractive coloured plot needs a demographic and biological interpretation. Genetic difference is evidence to explain, not a self-interpreting species label.

Contents · Previous chapter · Next chapter

Long-term field studies can follow the formation of reproductive barriers directly. Research on Darwin’s finches combined ecological observations and genomic evidence to document the establishment of a reproductively isolated hybrid lineage. Its value comes from following ancestry, traits and mating over time, rather than inferring the entire process from one later sample.

The result demonstrates that important steps can occur rapidly in a particular setting. It does not mean all species form at that speed or that a newly isolated small lineage is guaranteed to persist indefinitely. “Incipient lineage” and “long-term evolutionary success” address different timescales.

Host-associated insects provide another comparison. Studies of apple- and hawthorn-associated Rhagoletis flies tested how host-related selection can maintain differences despite exchange. Their history also includes geographic and genetic complexities; they should not be presented as a perfectly simple experiment in which one host switch instantly created two fully isolated species.

Chromosome-number changes provide an additional route, especially in plants. Whole-genome duplication can alter reproductive compatibility, while establishment of a persistent population still requires suitable reproduction and ecology. Research on the recently formed allopolyploid Tragopogon miscellus reveals substantial chromosomal variation after formation. The new lineage is a changing system, not a perfectly finished product at its first generation.

Together these cases widen the mechanism: divergence after separation, divergence with exchange, hybrid ancestry and chromosome changes can all contribute. Their relative importance must be established in the lineage under study rather than chosen from the example that makes the neatest story.

Contents · Previous chapter · Next chapter

A useful investigation combines several kinds of evidence. Genomic comparisons can describe ancestry and exchange. Field observations can show mating, habitat and timing. Experiments can test particular reproductive barriers. Morphology and physiology can reveal stable differences. No one method supplies every answer.

Sampling must cover the relevant populations. Comparing one individual from each of two places can exaggerate a difference or miss overlap. Sampling only distant endpoints may conceal a gradual geographic transition. Including contact regions can reveal whether apparent separation persists where exchange is possible.

Environmentally induced differences also need attention. Two groups can look different because they developed under different conditions, without every observed difference being inherited. Common-environment comparisons can help distinguish such effects, although they too have limits and may not reproduce natural conditions.

The number of genetic differences alone is insufficient. There is no single percentage of DNA divergence that divides all species across animals, plants, fungi and microbes. Different genes and lineages evolve at different rates, and recent divergence can leave much ancestral variation shared.

For an original audit, imagine a claim based on three facts: different colours, different average DNA sequences and no hybrids seen during one short visit. Each observation is relevant, but the conclusion remains underdetermined. Were enough individuals sampled? Did breeding occur during the visit? Were potential hybrids recognisable? Was gene flow tested rather than assumed absent?

A strong report specifies what is established, which species concept is being applied and what uncertainty remains. That is more useful than forcing a complex developing boundary into an absolute yes-or-no answer unsupported by the data.

Contents · Previous chapter · Next chapter

Draw two populations connected by arrows representing effective gene flow. Add a barrier that reduces, rather than necessarily eliminates, those arrows. Then show possible changes in habitat use, mating and hybrid performance. Keep a separate branch for outcomes such as reconnection, continued partial divergence or extinction.

This diagram avoids a common mistake: drawing a geographic split and immediately changing the species names. The split begins an investigation. The later evidence must establish whether the populations developed and maintained sufficiently independent histories.

Does isolation guarantee a new species? No. It can permit divergence, but several outcomes remain possible.

Must all gene flow stop first? No. Some lineages diverge and remain distinct while limited exchange continues.

Does a fertile hybrid mean there is only one species? Not necessarily. Natural mating patterns, ecology and the amount and consequences of exchange matter.

Are prezygotic barriers always stronger than postzygotic ones? No universal ordering applies. Their contributions must be measured for the system.

Can one individual become a new species simply by changing during its lifetime? Ordinary developmental or behavioural change is not a lineage split. Even a major heritable change matters evolutionarily through reproduction and the persistence of descendants.

The central explanation is a relationship between connection and independence: gene flow links populations; barriers and divergence can weaken that connection; inherited differences and reproductive outcomes determine whether distinct lineages persist. Return to Evolutionary Biology or explore the wider How Science Works collection.

Sources and example boundaries

The linked primary studies examine reproductive barriers, gene flow, hybrid-lineage formation, host-associated divergence and polyploid history. The museum source explains species concepts. The allele-mixing calculation, conditional-barrier account and two-factor incompatibility scenario are original teaching models, not measured thresholds, gene identities or predictions for a particular population.

Contents · Previous chapter · Continue to Evolutionary Biology