I have been asked this questions by many biology students and even non-biologist without a pretty straightforward answer to give. We are quite accustomed to phylogenetic trees where a common ancestor becomes two different species and these new species become, each one, two more. But, Why is not possible that a species might divide in three different species? Or four? Or more?

What arguments support this conception of the evolutionary process? Also, if somebody has a good article explaining this, please provide me with the link or citation so I can refer my fellow biology students to it.

  • $\begingroup$ Speculation: consider what tends to cause speciation — either some barrier (fully or partially) divides a population, or a selective pressure favours different extremes of a group. With each of these, speciation will be binary unless two such barriers or selective tensions arise simultaneously. So you’d expect it to be unusual for that reason, but you can also imagine situations which would cause it, e.g. a large flood wipes out most of a population (of a sedentary species), and the remnants form three or more isolated groups. $\endgroup$
    – PLL
    Commented Mar 9, 2015 at 11:35
  • $\begingroup$ possible duplicate of Does a fully-resolved phylogenetic tree have to be dichotomous?. See also Evidence & discussions of hard polytomy $\endgroup$ Commented Mar 9, 2015 at 23:40

2 Answers 2


I offer this perspective in addition to what WYSIWYG has provided.

Phylogenetic trees are tools to model what has actually happened, or what some evidence implies has happened, to a population of entities that exchange genetic material.

These models fail, for instance, when describing the phenomenon of horizontal gene transfer since a tree is acyclic by definition. Consider a viral genome that has integrated into a host genome. The host now contains DNA from its ancestral lineage and the viral lineage. This process cannot be modeled by a tree.

Some founder population that is susceptible to exogenous genetic transformation may give rise to many distinct lineages, some of which your species concept may qualify as "new" species.

  • $\begingroup$ +1 This is an important point. I wonder what can be commented about this case in which a virus integrates in two different locations in two different individuals of a population and this leads to phenotypic differences between them (a sort of speciation). This also qualifies as a case of two species coming from one because of anisotropic horizontal gene transfer. $\endgroup$
    Commented Mar 9, 2015 at 6:46

It is not impossible that a species can give rise to more than two lineages simultaneously, but it is just highly unlikely in certain cases.

If you assume that a point mutation in the DNA gives rise to a lineage then the number of two different mutations happening simultaneously (in two different individuals from the population) and also getting fixed, would be low for a small population. In any case a single DNA won't give rise to two different mutants. However this number would increase with increase in population size.

In such cases (small population) two separate events can be assumed rather than a single event that gives rise to two mutants. However, if the time difference between the occurrence of these two events cannot be determined (or it is too small compared to other timescales) then a single event can be assumed.

Phylogenetic trees can have nodes with multiple branches if the distance between all of them is the same (as in the abovementioned example; the distance between all of them is 1).

So I think it can be said that the likelihood of simultaneous formation of multiple lineage increases with population size.

  • $\begingroup$ I agree that the total ordering of events that give rise to new species (lineages) would give a total ordering to the branches in the tree or graph representing the speciations. The last statement may be true only if the probability of speciation-event-mutations increases with population size, like in the case where it were equiprobable and non-zero for all individuals. $\endgroup$
    – mcg256
    Commented Mar 9, 2015 at 7:23
  • $\begingroup$ @mcg256 The likelihoods actually won't change. But the number of mutants will increase with bigger population. $\endgroup$
    Commented Mar 9, 2015 at 8:09

Not the answer you're looking for? Browse other questions tagged .