An extinction occurs when a population decreases to zero size. It is an inevitable event in infinite time given that zero is in effect an absorbing boundary. Thus, all populations must eventually go extinct; it is a natural process, as Jet-Poop has stated above. For natural populations of living organisms, extinction then becomes a question of when.

Current estimates of the rate of extinction indicate that in the next decade or so, the human population will have outdone a planetary asteroid strike as far as the proportion of species wiped out. Way to win one for the Gipper, folks!

human-engineered extinction

Tonight I saw a rather disturbing documentary, State of the planet made by Sir David Attenborough of the BBC.

It dealt with a (perhaps) imminent sixth wave of mass-extinction on our planet (in fact it's a 3 part series, so I'll probably even more worried when I see the other parts...).

According to scientists mankind causes this mass-extinction (mentioned in other writeups in this node) in 5 ways:

1. killing species at a greater speed than these species can reproduce themselves.
   eg. thanks to new technology (eg. sonar, better fishing gear...), man fishes too efficient, making it harder every year for mother nature to recover.
2. destroying the unique habitat of species.
   eg. by destroying the forest, some species depending on it will disappear.
   Even worse: other species who rely on this animal (for food, symbiosis, ...) will go extinct too. It's a kind of Domino -effect.
3. importing species to places where they didn't occur before.
   Because lack of predators, they become too succesful and replace the original inhabitants, or they are too efficient in hunting, eliminating the totally unprepared original species. Animals are conditioned to deal with their own predators, and are not familiar with techniques used by the imported ones, or simply don't recognize the predator as a threat.
   eg. cats, rabbits in Australia
4. creation of islands.
   Man tries to preserve pieces of nature in national parks. While this is of course a good thing, it doesn't stop the extinction of some species. Populations get separated from eachother if the distance is too big (this distance depending on species of course) or some other hazard has been created by man. They end up being on islands. When a population on an island dies for some reason (eg. disease or weather conditions), it will not be replaced by animals from another island. To illustrate the fragility of the balance: some species consider even a road as a border between islands.
5. pollution.
   The list wouldn't be complete without this one, it's also the best known I guess. eg. Global warming (use of fossil fuels) causes climatical changes (in spite of what mr. Bush says...). Species react on this by moving to other areas. Only this time, the climate changes at a faster rate...

Types of extinction

Pseudoextinction, or phyletic extinction - One species evolves into another and there is no loss of species. An example of this would be the missing link between humans and apes.
True extinction, or terminal extinction - A species lineage becomes extinct, resulting in a loss of species.

Modes of extinction

Field of Bullets
Field of Bullets is one possible extinction scenario, where the removal of an organism is caused by an accident that has no connection to a species' adaptability or adjustment to its environment, or the fitness of individuals. It is a random and purposeless process, like standing in a field of bullets.

All species are effected by the same probability of extinction, regardless of classification, habitat, size of population, etc. This is the type of extinction expected after a large meteorite strike, which is one theory for the mass extinction of the dinosaurs 65 million years ago.

A Field of Bullets extinction opens up scope for new life to flourish, as demonstrated by the rise of mammals after the death of the dinosaurs.

Fair game extinction
This is Darwins's survival of the fittest. Extinction occurs as a response to the normal conditions a species may face, and that species' ability to adapt to those conditions.

Wanton extinction
Wanton extinction selects species for extinction through some method other than natural selection, which is based on the fitness of the species. For instance, geological events may wipe out all of the living members of a species because their habitat was located at the point of that event.

Just when you thought there wasn't anything worse to fear than the hydrogen bomb and the Ebola virus, The Economist brings you an article ("Exterminate, exterminate", March 20th, 2003) on something new and chilling. Apparently, a new technique exists with the potential to "lead to the extinction of almost any sexually reproducing species."

The technique exploits site-specific selfish genes which have a tendency to be particularly selfish, in keeping with Richard Dawkins' theory of evolutionary selection. Basically, SSSG's use a cell's own machinery to introduce themselves into new chromosomes: a means of dispersal not unlike that of a virus. If an SSSG falls in the middle of a gene, that gene will no longer function - denying the organism the ability to produce that protein, unless, of course, the gene on the other chromosome is fine. In fact, that backup system is why these genes don't cause death or horrible cancer in everyone all the time.

If, however, the particular SSSG in question is a homing endonuclease gene (HEG), this safety system doesn't work. That is because when a HEG gets into a chromosome, it produces an enzyme called endonuclease. This enzyme cuts the DNA strand at any place where a particular string of nucleotides is found.

The next part is sufficiently tricky that I defer to the superior writing ability of The Economist's writers:

"Cells have two copies of most chromosomes (one deriving from the mother, and one from the father). If only one of these carries an HEG, the other will be cut by the enzyme which that gene produces. The site of the cut corresponds to the site of the HEG on the "infected" chromosome. The infected chromosome itself is not cut, because the HEG is in the middle of the enzyme-recognised sequence, and thus disguises it. But, since cells repair chromosomal damage by replacing the corrupted DNA with a copy from the same place on the partner chromosome, the HEG is copied over as part of the repair process. Now, instead of one copy of the broken gene, the cell has two."

Once this process has taken place, every gamete (sperm or egg) produced by the organism will contain the HEG disrupted gene, even though the parent only has one copy. This process leads to the gene spreading very rapidly through the population. Finally, if you choose a gene where the loss of one copy does little harm, but the loss of both is fatal, you end up with a situation where 80-95% of embryos produced by parents who each have one HEG gene will die before they come to term. The number of offspring falls below the natural replacement rate and the species goes extinct.

Now, there are some potentially useful applications for this technology. The Economist suggests the worthy idea of wiping out malaria carrying mosquitoes. At the same time, the potential danger of this technique seems to be considerable, especially if it proves relatively easy to do.

Extinctions are a relatively common occurrence in the geologic past, and are the result of biologic and environmental pressures and processes. Mass extinction (along with mass radiation) forms the boundary of geologic time units. Indeed, abrupt extinctions often allow for surviving organisms to enter, exploit and adapt to new environments. Major mass extinctions have included:

Cambrian-Ordovician boundary (510 mya)
Ordovician-Silurian boundary (440 mya)
Devonian-Carboniferous] boundary (360 mya)
Permian-Triassic boundary (245 mya) - over 90% of life was lost in this, the largest extinction ever.
Triassic-Jurassic boundary (208 mya)
Cretaceous-Tertiary boundary (65 mya) - and there go the dinosaurs!

Mass extinction is often caused by competition, reduction of available food, atmospheric changes, climate changes, sea level changes and plate tectonics - all of which can be wondrously intertwined at times.

A bit more on the behavioral psychology definition of extinction:

If a given action is ignored or receives a negative consequence, it will be less likely to be produced next time the same situation arises. This process of decreasing the probability of a behavior occurring is called the extinction of the behavior.

Stan cries when he wants an apple, but instead of getting the apple, we all ignore him. Next time he wants an apple, he will probably try something other than crying.

In this example, the unwanted behavior is crying, the reinforcer for this behavior is the apple (or would be, if we were to give him the apple), and the negative consequence is not getting the apple. Of course, it gets a lot more complicated than that in practice. Stan is probably crying because he has cried before, and it worked. He'll probably try it at least a couple more times before he gives it up, and he'll probably be louder next time. And ignoring people isn't always the best way to get them to do what you want.

For example, let's say that I want to extinguish my daughter's obsessive playing of computer games. If I ignore her while she's playing, there will be no change in her behavior. This is because the reinforcer of her behavior does not come from me, but directly from the computer. I need to pair the playing of computer games with a negative consequence (or, more realistically in this example, take control of the reinforcer, i.e., get control over the computer).

On the other hand, if one of my students likes to swear in order to piss me off and get attention, I may not get far if I punish him for swearing. That only lets him know that his pissing-off-teacher program is right on track, and rewards him with the attention he desires. The stronger I make the punishment, the better he thinks he's doing! I'm better off ignoring his behavior.

Here's some factors that influence the effectiveness of extinction:

• The control of reinforcers: Make sure you know what the reinforcers are, and that they can be controlled. If you want to stop Stan from crying when he wants apples, you need to make sure that no one gives him apples when he cries. Of course, you may be able to extinguish his crying behavior when it comes to dealing with you, but he'll still cry for everyone else.
• Control the setting: Ignoring behavior can be very difficult in public places (restaurants, stores, movie theaters), especially when the behavior involves crying and screaming. Be prepared! Be ready to leave or wait it out, but don't give in. If Stan gets an apple only when he cries in public, then he's 1. more likely to try crying again, and 2. especially likely to try when he's in a public place. Also keep in mind that he might not want to go to the grocery store (for example), and if you decide to leave rather than make a scene, you've just taught him that, while he may not get an apple when he cries, he does get to leave the store. Now he has a new reason to cry!
• Reward appropriate behavior: Extinction will work much better if it is paired with reinforcement for a positive behavior that meets the subject's needs -- that is, if Stan can get what he wants by asking politely, it will be much easier for him to give up crying. Extinction of bad behaviors will work much faster if good behavior is consistently and quickly rewarded (and praised). Don't make the extinction a battle with a winner and a loser -- there should be two winners, you and the subject. Make sure Stan can get his apple, and feel good about doing so.
• Full disclosure: Things will probably move along a lot faster if the subject of extinction knows exactly what the rules are. "If you cry, you don't get an apple. To get an apple, you should ask 'can I please have an apple?'" If it looks like the behavior is about to start, feel free to remind the subject of the rules. ("You look hungry. Do you remember how to ask for an apple?")
• Consistency!: Extinction will go faster for a behavior that has been continuously reinforced than for one that has been only intermediate reinforced. If you have intermediately reinforced a behavior, the subject has become more used to persisting, and will have learned that failure this time does not indicate that the behavior wont get him what he wants next time... or the time after that... or the time after that... This is another reason to Control the Reinforcers! Sometimes we accidentally move from continuous reinforcement (get an apple every time you cry) to intermediate reinforcement (get an apple occasionally when you cry) when trying to extinguish a behavior. This is bad, as it will make it harder to get rid of the behavior in the long run.
  
  Confused by intermediate reinforcement? A common example is a slot machine. You may lose $20 in an afternoon, slowly feeding it quarters. No one would walk up to a row of payphones and start popping quarters randomly into them, even though the end effect will likely be the same, and much faster. The difference is that the slot machines are intermediate reinforced. You don't really expect to get rich, but you do want to try just one more pull... because sometimes it pays off! Now compare the slot machine (intermediately reinforced) to an ATM (continuously reinforced). When the slot machine stops paying off, it could be quite a while before you notice -- and even when you notice, you may think that it only means that the big payoff is coming up any second now. But if an ATM doesn't give you cash, you get worried really quickly. If it happens even twice, you will likely start avoiding that machine. The intermediately reinforced slot machine will be given more chances than the (formally) continuously reinforced ATM. Likewise, intermediately reinforced bad behavior will reoccur more often than continuously reinforced behaviors once the reinforcers have been removed.
  
  
• Extinction burst: a behavior will often become more common for a while before it starts to decrease (and hopefully disappear). When something stops working, people tend to try extra hard for a little bit, then give up. This may mean an increase of crying over days or weeks as Stan tries to get an apple, or a big 30 second flurry of scribbling as I try to get one more burst of ink out of my ball point pen. It would be a big mistake to give up on extinguishing a behavior in the middle of an extinction burst. That would mean that you have reinforced the more extreme level of behavior (it didn't work when Stan was whimpering pitifully for 2 minutes, but crying loudly for 10 minutes worked! Stan ain't no fool, and he now knows to go strait to the screaming, and stick with it!).

So, what with continuous vs. intermediate reinforcement and extinction bursts, there is one big rule to follow when you're trying to extinguish a behavior. Don't Give Up. If you start to withhold the reinforcer (i.e. the apple), but then relent and provided it after a later incidence of the behavior, you will have both taught the subject that if they try long enough they can break you down, and since the behavior was probably more intense and/or more frequent than usual during the extinction burst, you will have reinforced a worse, more obnoxious level of behavior than you originally had been dealing with.

Extinction of behaviors often produces frustration and anger, which produces aggression. Be ready for this, and be ready to wait it out. Other behaviors may increase along with the behaviors targeted for extinction (i.e. fits, crying, hitting, non-compliance, etc., etc.)

Behavior that seems to have been completely extinguished may reappear at a latter date. (tech speak: spontaneous recovery). Usually this spontaneously recovered behavior is more mild than the original behavior. Extinguish it again! It'll be easier this time.

We tend to unwittingly extinguish good behaviors all the time. A child tries to ask for something from his mother; she ignores him and keeps talking to her friend. He keeps saying Mom! Mom! Mom!. No response. Grabs hands. Starts crying. Gets attention! Talking in a normal voice and waiting patiently are both extinguished. Crying is reinforced, so crying will happen all the sooner next time. Eventually, it may be the first, and only, thing the child bothers to try. (I've seen this a hundred times. It's amazing that some children bother to speak at all).

I've talked about extinction as applied to kids, but it also applies to adults and animals. I've also focused more on ignoring behavior rather than punishing it, but punishment is also a fine path to the extinction of behaviors. My cat stopped eating the houseplants because every time he tried, I squirted him.

Ex*tinc"tion (?), n. [L. extinctio, exstinction: cf. F. extinction.]

1.

The act of extinguishing or making extinct; a putting an end to; the act of putting out or destroying light, fire, life, activity, influence, etc.

2.

State of being extinguished or of ceasing to be; destruction; suppression; as, the extinction of life, of a family, of a quarrel, of claim.

© Webster 1913.