Speciation, in sexual organisms anyway, occurs when two genetic groups no longer intersperse by interbreeding. This can happen in the following ways.

1. Instantaneous speciation. This tends to occur mainly in plants through hybridisation or a duplication of the number of chromosomes at conception. Hybrids will have asymmetric chromosomes to begin with but a process of doubling and lack of secondary reduction through meiosis (or the formation of haploid, or half complement of chromosome-containing, cells). So, when you get a hybrid, the chromosomes of the two parent forms don’t match up, but if you then double the chromosomes and don’t halve them again, you can pair them up. In the case of duplication of one species’ chromosomes, you end up with a 2n (or 3n or 4n…) set of already paired chromosomes.

Sometimes these polyploids, as they are called (many-numbered), are able to reproduce by selfing, or the polyploid is a common enough event that there are two or more individuals that can cross, and away you go. Secondary selection eliminates genes that are less fit.

2. Allopatric speciation. A bit of terminology to begin with – “patris” means “country” or “homeland”. It comes in flavours ranging from “sym” (together) through “peri” (next to) to “allo” (separate). So allopatric speciation is “speciation that happens when the populations are isolated geographically”. We’ll encounter some other patrises later.

In allopatric speciation, a species’ range is divided, say, by a river or mountain range or desert or currents in the ocean, etc. Once this happens, they are adapting to novel conditions and also a process of more or less random (stochastic) processes of sampling the genetic variants leads to a population that is rather different from the parental one. The stochastic process is called “random genetic drift”. Although reproductive infertility or isolation is not something that selection “aims for”, it is often a byproduct of changes made to the developmental cycle of the isolate population. So, when the two get back in sympatry, they either can’t interbreed (are isolated) or they can but the hybrids are not as fit as either of the parental variants (lowered hybrid fitness), and so they are then subjected to “reinforcing selection” to maintain isolation. [Of course they may just end up going extinct as well, due to lowered fitness.]

3. Peripatric speciation. In this case the isolated population is not entirely genetically isolated, but because it is on the periphery of the main population, a local population (caled a “deme”) may have unusual genetic variants (called alleles”) that get established for a mix of selective and drift reasons to the point where hybrids between them and the main population are less fit, causing reinforcing selection. This occurs because the rate of interdeme crossing is lower than the rate of intrademe crossing, and so it is able under certain conditions to form novel genotypes and developmental sequences. [Technical note: when the rate of migration between populations is less than 50%, it is parapatry. When the geographical isolation is less than 100% it is peripatry. So para- and peri-patry can be the state of the same population. This is a nusiance piece of confusing terminology.]

4. Sympatric speciation. This is controversial (and was Darwin’s preferred view). In this view, a variant form reaches a new “adaptive peak”, and reinforcing selection selects against hybrids with the prior form. This is thought to happen in a couple of ways. One is through the evolution of novel mating systems, such as calls (the case of the electric signalling fishes, for example; the classical case is the mating calls of Rana pipiens). Another way is through the adaptation to a new host, such as when Rhagoletis fruitflies started breeding on apple trees in California, which flower at a different time of year, causing selection to isolate the older hawthorn-breeding developmental cycle from the newer apple-breeding developmental cycle. This is sometimes called “host-race” speciation.

5. Statsipatric speciation (in-place). This is like the instantaneous case above, although the chromosomal variants are able to interbreed with the original chromosome count individuals. Selection takes the form of inviable developmental cycles.

6. Introgression. In this case a population of, say, flowering plants, is able to cross with some other species, but the progeny aren’t thereby members of a new species, but backbreed into the population, changing its genetic constitution and adaptive niche so that when it is in sympatry with the original species, the population is now isolated.

Speciation studies focus a lot on “reproductive isolating mechanisms” (RIMs) which are the particular mechanisms that keep populations from sharing their genes. These are either a byproduct of evolution (say, in allopatry), or are secondarily subjected to selection (in peripatry), or are the direct result of selection (in sympatry).

There are also cases of speciation being caused by parasites. A parasitic cell called Wollbachia can infect the sex cells of arthropods (insects etc.) so that uninfected individuals, though themselves genetically identical, cannot interbreed with infected ones. In this case the infection acts like a kind of allopatry, even if they are in the same home region, allowing the genes to evolve incidentally in their own way through drift and selection.