Speciation

Speciation is the process by which one population splits into two or more populations that can no longer interbreed. It is produced by the same mechanisms — mutation, selection, drift, and gene flow — that drive smaller changes within a population; see What is Evolution? for those mechanisms.

The claim is testable directly: has a population ever been observed splitting into two that no longer interbreed? It has, in the field and in the laboratory, in plants, insects, and vertebrates, and the cases are documented well enough that the original papers can be read directly. What follows is how the process works, the observed examples, and what those examples do and do not establish.

What counts as a species#

Biologists use several species concepts rather than one, because life does not sort itself into tidy boxes. The most familiar is Ernst Mayr's biological species concept: populations that actually or potentially interbreed in nature and are reproductively isolated from other such groups. It works well for sexually reproducing animals and badly for bacteria, which swap genes sideways, and for organisms known only from fossils.

Fuzzy boundaries are informative rather than an embarrassment. If species were separately created kinds with fixed boundaries, defining them would be straightforward. Boundaries are fuzzy because they are still forming, and a theory of gradual divergence predicts exactly the messy, in-between, hard-to-classify cases that exist.

How populations split#

Geographic isolation#

The most common route. A population is divided by a river, a mountain range, a lava flow, or a rising sea. Mutation and drift then push the two halves in independent directions, and if they meet again after enough time, they no longer interbreed. This is what happened to the Isthmus of Panama's fish populations, split into Atlantic and Pacific pairs roughly three million years ago.

Polyploidy#

In plants especially, a whole-genome duplication can produce offspring that are reproductively isolated from their parents in a single generation. This is not a slow process and it is not a hypothesis — it is how a great many crop plants arose, including bread wheat, and it has been produced deliberately in greenhouses.

Ecological divergence#

Populations sharing a range can still split if they specialize on different resources and mate where they feed. The apple maggot fly is the standard case: a fly that historically laid eggs in hawthorn began using introduced apples in the mid-1800s, and the two host races now show measurable genetic differences and differing emergence times.

Observed cases#

The London Underground mosquito#

Culex pipiens molestus lives in the tunnels of the London Underground, which were dug starting in 1863. Byrne and Nichols found in 1999 that the underground populations differ genetically from surface populations, breed year-round rather than seasonally, feed on mammals rather than birds, and produce almost no viable offspring when crossed with the surface form. The isolating barrier arose inside a documented human timeframe.

Some later work argues the molestus form entered the tunnels already partly differentiated rather than originating there. The reproductive isolation itself is not in dispute; the timeline is.

Drosophila in the laboratory#

Diane Dodd's 1989 experiment split a fly population into groups raised on starch and maltose media. After several generations, when the groups were recombined, flies preferentially mated with flies from their own medium. This is incipient reproductive isolation produced by nothing but a change in food and time.

Rice and Salt's 1988 experiment went further, using habitat choice as the selective pressure and producing strong assortative mating in around thirty generations.

Faeroe Island house mice#

House mice introduced to the Faeroe Islands, likely by Norse settlers, diverged sufficiently within roughly 250 to 800 years that some populations were described as distinct species on morphological grounds. Whether they warrant full species status is debated; the rate of divergence is not.

Hybrid plant species#

Tragopogon miscellus, a goatsbeard, arose in eastern Washington and Idaho in the 1940s from hybridization between two introduced European species followed by chromosome doubling. It is reproductively isolated from both parents, it is fertile, and its origin is dated to within a couple of decades because botanists were watching the region at the time. A new, self-sustaining species formed in living memory.

Ring species#

Around the Tibetan plateau, greenish warblers form a chain of populations that interbreed with their neighbors all the way around — except where the two ends of the chain meet in Siberia, where they behave as separate species. This is speciation caught mid-act and laid out geographically rather than temporally. Later work has shown the ring is broken in places, which complicates the textbook version; the gradient in reproductive compatibility remains.

Common objections#

"That's still microevolution — variation within a kind"

The standard response, made by Answers in Genesis and the Institute for Creation Research: speciation is accepted, but only within an originally created kind, and no mechanism crosses between kinds.

Part of the reply is correct. No biologist claims a single speciation event produces a new class of animal, and evolution does not predict a dog turning into a cat — cats are not descended from dogs but are cousins. Such an event would count against the theory rather than for it.

The claim being tested is whether a barrier exists. The observed cases show populations splitting into reproductively isolated lineages with no mechanism arresting further divergence. Once that happens, the difference between "within a kind" and "beyond a kind" is elapsed time.

The position has also moved. Young-earth creationism now requires very rapid speciation, since a few thousand kinds on the ark must produce present-day diversity within a few thousand years — a faster rate than any biologist proposes.

"What is a 'kind' biologically?"

Baraminology is the creationist attempt to define the category, developed by Frank Marsh and continued by Kurt Wise and Todd Wood.

It has not produced a consistent method. Different practitioners place the boundary at different taxonomic levels — sometimes family, sometimes genus, sometimes order — and the placement tends to shift when new hybrids are discovered.

Todd Wood, working within young-earth creationism, has written candidly that the evidence for common descent is substantial and that creationists should not pretend otherwise. That is a notable admission from inside the position.

"If speciation happens, why don't we see it constantly?"

A fair question. Typical speciation in vertebrates takes tens of thousands to millions of years, which is why direct observation is uncommon.

The documented cases are the fast ones: plants that double their chromosome number and become reproductively isolated in a single generation, insects with short generations, and populations under unusually strong isolation.

Tragopogon miscellus and Tragopogon mirus both originated in the Pacific Northwest during the twentieth century and are documented in the botanical literature from the decade they appeared.

"Reproductive isolation isn't a real species boundary"

There is genuine disagreement among biologists about species definitions, and the biological species concept does not apply to asexual organisms or cleanly to ring species.

The disagreement is about where to draw a line on a continuum, which is what would be expected if populations diverge gradually. A world of separately created kinds would produce sharp boundaries requiring no debate.

The difficulty of defining species is therefore evidence for gradual divergence rather than against it.

What the evidence shows#

New species have been observed forming, in the field and in the laboratory, in plants, insects, and vertebrates. The mechanisms — geographic isolation, polyploidy, and ecological divergence — are directly documented.

What cannot be observed directly is divergence over millions of years, which is why the case for deep common ancestry rests on fossils and genomes rather than on watching.

What the observed cases establish is that no barrier has been found that variation cannot cross.