How one species becomes two

Every species alive is the far end of a line of splits going back to a single ancestor. The splitting itself is the part that gets skipped: it is gradual, it has no moment, and by the time anyone can point at two species there is nothing left to watch. This is how it happens, and why drawing the line is harder than it sounds.

One requirement, several routes

Speciation needs one thing: two parts of a population must stop exchanging inheritance. While variants can flow freely between groups, differences get mixed back in as fast as they arise, and the population stays one thing. Stop the flow and the two halves drift and adapt independently, and the differences accumulate instead of dissolving.

Everything else is detail about how the flow gets interrupted.

Three routes by which one population splits into two Three panels. In allopatric speciation a physical barrier divides the range and the two halves diverge in isolation. In parapatric speciation the range stays continuous but a sharp environmental gradient makes the two ends diverge despite contact at the middle. In sympatric speciation the population occupies the same place throughout and splits by using different resources or breeding at different times. Allopatric a barrier divides the range Parapatric a gradient, contact at the join Sympatric same place, different living barrier wet dry overlapping in space
Allopatric splitting is the easiest to demonstrate and was long assumed to be the main route. Sympatric splitting was doubted for decades and is now accepted in cases such as fish in single lakes and insects that shifted to a new host plant.

There is no moment

The awkward truth is that speciation is a process without an event. Two diverging populations pass through every intermediate stage: freely interbreeding, then interbreeding less, then producing offspring that are less healthy, then producing offspring that are sterile, then not interbreeding at all. Every stage is occupied by real populations somewhere right now.

This is why the question "when did they become separate species" has no clean answer, and why ring species are so awkward. A chain of populations around a barrier, where each neighbouring pair interbreeds happily but the two ends meet and cannot, is not a paradox. It is what a gradual process looks like when you catch it partway and insist on sorting it into two boxes.

What a species even is

The definition most people learn, a group that can interbreed and produce fertile offspring, is the biological species concept, and it is useful and limited. It says nothing about organisms that reproduce asexually, which is most life by count. It cannot be applied to fossils. And it is routinely violated by organisms that hybridise in the wild while remaining distinct.

Working biologists use several definitions and pick by circumstance: shared physical form, shared ecological role, or a distinct branch on a genetic tree. These disagree with each other on real cases, which is not a scandal. Species are a category humans impose on a continuous process to make it discussable, and continuous processes do not sort cleanly.

Drawing the line deliberately, and admitting it

A simulated world runs into exactly this problem, and has to answer it explicitly rather than leaving it to judgement.

Organisms here inherit their genes and their network weights from a single parent with mutation, so lineages diverge continuously and by degree. To report anything about lineages at all, a rule is needed for when a group of descendants has drifted far enough from the rest to be counted separately, and this world has one: a stated threshold of genetic distance, fixed in advance, which produces a lineage identifier when crossed.

The important thing about that rule is what it is not. It is a bookkeeping decision, not a discovery. The threshold could have been set elsewhere and would have produced a different number of lineages from the identical run. Nothing in the world becomes reproductively isolated at that moment, because there is no sexual reproduction here to be isolated from. The rule and its exact values are on genome, inheritance and mutation.

Why this is stated so carefully

A lineage count is the easiest number on a site like this to oversell. "Fourteen species have evolved" would be a much better sentence and a false one. What can be said is that fourteen groups have crossed a distance threshold chosen by a person, which is a real measurement of divergence and is not a claim about speciation.

What does carry across honestly is the mechanism underneath. Populations here are separated by distance and by terrain that costs energy to cross, and two groups that rarely mix will accumulate differences that nothing mixes back. That is allopatric divergence, running for the same reason it runs anywhere: interrupted flow plus time.

Keep going

Watch the universe