Evolution: the Facts.

[Dinwar]

Yeah....the definition is the hard part. There are biological definitions, but to use "populations" as a biological term would require an overhaul of the definition, from the ground up...

I understand that there are some animals (I recall some grasshoppers) that can potentially mate but do not.

Well, a simple example is my mother and myself. She's female, I'm male; however, we will never mate. Another example is myself and any random person in Africa who never visits the USA. Or, considering I'm in a monogomous relationship, myself and anyone who's not my wife. I have the potential to mate with any woman on the face of the Earth. I simply won't do it.

A better biological example is fish. Let's say you have fish in a lake. Some fish are dropped by a bird into another lake--alive, or with eggs that hatch. This creates a new population of fish (geographically confined and cut off almost or entirely from the original population). After a generation or two the new fish can be said to be a stable population. However, since there's a physical barrier between the two populations, they cannot interbreed.

So it's pretty easy to get animals that CAN interbreed, but that DON'T interbreed. And that's assuming sexualk reproduction. Asexual reproduction makes things even weirder. It may be justifiable to, in some cases, consider each organism of such species a new species--after all, any one of them may branch off into an entirely new direction (sexual selection increases the odds of genetic variation, but variance can also be washed out relatively easily).

 

[Lowpro]

Well a more concrete example are rivers that have a single fish population becoming two populations after a physical separation, and then time passes, both populations being limited to the genes they have and pass against the selection pressures of their respective populations and then general mutation rates as well. When do they become new species?

I'd say they are new species once the separation exists in the first place because "species" makes more sense when you recognize that the gene flow is what determines it. Although this is contentious of course, and impossible to determine for ancient animals to a great degree, but it makes sense.

 

[Dinwar]

I'd say they are new species once the separation exists in the first place

You're talking alopatric speciation. Sympatric speciation works differently. In fact, there are numerous types of speciation--just within sexually reproducing species like us. As an aside, if you read down that page you'll see another problem with the whole issue of using mating to define species--sometimes what foods you eat dictates mating preferences (it's more complex than that, but food was the independant variable in the well-replicated experiments).

Speciation can be all KINDS of fun once you add time to the equation....

Also, what do you do if there's never any separation or substantially constricted gene flow? For some ring species, there's never any clear break between populations--no more so than between populations of non-ring species, anyway (ALL populations experience some restricted gene flow between other populations). If you cut the middle populations out, the living members would clearly be two separate speceis; however, until that happens, what do you call them?

And if you're going to say that since all populations experience constricted gene flow, well, you're where I am now, and with the same problems. At what poinit does the constriction become sufficient? Would genes flowing at 95% of the rate seen within a population be sufficient? 90%? 50%? And why choose that number in the first place? Is it because there's a natural break somewhere along that line, or is it because we've decided we have to put it somewhere? Is it the same for each taxa, or different for different taxa? And so on. I don't mean to say that this idea is wrong--I ascribe to it most days!--just that it has a LOT of work that needs done before we can call it "true" in any real sense.

The issue is that a lot of the concepts involved with species make sense--except when they don't. And when we talk about "speciation" we're really talking about one of a suite of events, each with its own forcing mechanisms, timelines, characteristics, etc.

 

[Acleron]

I've often wondered if we shouldn't discuss populations, rather than species. The population is, after all, the unit of evolution (an individual doesn't evolve, and a species as such doesn't evolve--the population changes over time). It would clear up a lot (but certainly not all) of the ambiguity with the species concept. For example, two organisms are considered (by the biological species concept) to be the same species if they can potentially interbreed. They're the same population if they're both part of the group that actually interbreeds (this sentence is very sloppy, but I think it gets my point across). There's have to be some limit, and some other rules--a single member of a population breeding with a single member of another population does not mean that those two are the same population, and ring species play merry havoc with pretty much any definition of anything--but it would remove some of the more silly consequences of the term "species".

Of course, the REAL problem is that we're trying to apply a somewhat rigid framework to what is by its nature an extremely fluid concept. Where you draw the line is always going to be a bit arbitrary, and there's always going to be viable arguments to be made that the line should be HERE, not THERE.

Would you need to define a 'population' any further than the average genome for the group you are examining?

The membership of your population would not be a yes/no situation for individuals. Instead, an individual would be defined as a closeness to the average genome. The term 'closeness' is the problem though. Some parts of the genome would be more important than others. For instance, that part which determines the ability to breed with a significant part of the population is going to be more important than fur colour.

 

[Dinwar]

Would you need to define a 'population' any further than the average genome for the group you are examining?

I'm not a fan of defining things by their averages--averages are mere mathematical abstractions, and often entirely fictional. Taxonomy got around this problem by defining a species by a type specimen, which gets around the problem of defining a population by a mathematical abstraction but which presents whole new problems (such as, what do you do if you get an abnormal specimen).

The other problem is, two members of the same species may have very similar genomes despite being from nearly disconnected populations. The whole point of talking about populations rather than species is to focus on actual interbreeding.

Still, gene flow may be approximated by similar genetic code across the population....It would be more complex than a simple average, though--you'd have to take into account deviations, and the nature of those deviations. Geographic barriers would also come into play--if a mountain range formed it's unlikely that two groups of critters, no matter how similar, should be treated as a single population (unless they obviously could cross mountain ranges).

 

[ThunderChunky]

The current state Jack Szostak's work in the origin of life (3 very good lectures):

http://www.youtube.com/watch?v=PqPGOhXoprU

http://www.youtube.com/watch?v=CJ5jh33OiOA

http://www.youtube.com/watch?v=jfq5-i8xoIU

 

[Sideroxylon]

I'm not a fan of defining things by their averages--averages are mere mathematical abstractions, and often entirely fictional. Taxonomy got around this problem by defining a species by a type specimen, which gets around the problem of defining a population by a mathematical abstraction but which presents whole new problems (such as, what do you do if you get an abnormal specimen).

The other problem is, two members of the same species may have very similar genomes despite being from nearly disconnected populations. The whole point of talking about populations rather than species is to focus on actual interbreeding.

Still, gene flow may be approximated by similar genetic code across the population....It would be more complex than a simple average, though--you'd have to take into account deviations, and the nature of those deviations. Geographic barriers would also come into play--if a mountain range formed it's unlikely that two groups of critters, no matter how similar, should be treated as a single population (unless they obviously could cross mountain ranges).

Isn't this true of most/all definitions we might come up with? Is it true nature has no convenient joints at which we can carve her, but rather all categories are human constructs and the best we can hope for are tools that are useful most of the time?

 

[Acleron]

I'm not a fan of defining things by their averages--averages are mere mathematical abstractions, and often entirely fictional. Taxonomy got around this problem by defining a species by a type specimen, which gets around the problem of defining a population by a mathematical abstraction but which presents whole new problems (such as, what do you do if you get an abnormal specimen).

The other problem is, two members of the same species may have very similar genomes despite being from nearly disconnected populations. The whole point of talking about populations rather than species is to focus on actual interbreeding.

Still, gene flow may be approximated by similar genetic code across the population....It would be more complex than a simple average, though--you'd have to take into account deviations, and the nature of those deviations. Geographic barriers would also come into play--if a mountain range formed it's unlikely that two groups of critters, no matter how similar, should be treated as a single population (unless they obviously could cross mountain ranges).

You are quite correct, perhaps a range.

But the point about a mean or some mathematical description of population is that a population is an abstraction.

As for breeding groups being separated, although they have the potential to genetically move away from each other, shouldn't they be considered as part of the same population until they do?

 

[Dinwar]

Isn't this true of most/all definitions we might come up with?

It's more true in some cases than in others. The fact that the average animal is usually fictional is only part of the issue, though. The problem with biological systems is that unless there's physical constraints, they always include variance. It's a fundamental assumption of evolution, and something trivially easy to observe (if you ever talk to a hog farmer you'll find that the good ones can identify every one of their pigs, even if they all look the same to you). Two populations with the same average organism can be wildly different populations--one may be far more skewed to the right than the other in some trait, or the total variance can be much larger, or much flatter, etc., than the other population, and that'd lead to entirely different populations.

There's also the issue of practicality. It'd be extremely difficult to accurately assess the average organism--you're finding a single point. It'd be easier to assess population dynamics, as we've developed statistical methods to determine all that.

As for breeding groups being separated, although they have the potential to genetically move away from each other, shouldn't they be considered as part of the same population until they do?

That leads to absurdities (that probably comes off more harsh than I intended it--in physics, when some equation tends towards infinity it's usually a sign you've screwed up; in biology, if some definition produces absurd conclusions it generally--but not always--means there's some flaw, so I tend to look for such cases when I hear a new definition). For example, it leads to the potential for calling American kudzu and Japanese kudzu the same population, despite having extremely limited to no actual gene transfer, and the fact that they're separated by one of the largest geographic barriers on the planet. It would also mean that by some mechanisms of speciation two essentially freely interbreeding groups may be considered two populations, merely because of morphological/genetic differences.

 

[Lowpro]

You're talking alopatric speciation. Sympatric speciation works differently. In fact, there are numerous types of speciation--just within sexually reproducing species like us. As an aside, if you read down that page you'll see another problem with the whole issue of using mating to define species--sometimes what foods you eat dictates mating preferences (it's more complex than that, but food was the independant variable in the well-replicated experiments).

Speciation can be all KINDS of fun once you add time to the equation....

Also, what do you do if there's never any separation or substantially constricted gene flow? For some ring species, there's never any clear break between populations--no more so than between populations of non-ring species, anyway (ALL populations experience some restricted gene flow between other populations). If you cut the middle populations out, the living members would clearly be two separate speceis; however, until that happens, what do you call them?

And if you're going to say that since all populations experience constricted gene flow, well, you're where I am now, and with the same problems. At what poinit does the constriction become sufficient? Would genes flowing at 95% of the rate seen within a population be sufficient? 90%? 50%? And why choose that number in the first place? Is it because there's a natural break somewhere along that line, or is it because we've decided we have to put it somewhere? Is it the same for each taxa, or different for different taxa? And so on. I don't mean to say that this idea is wrong--I ascribe to it most days!--just that it has a LOT of work that needs done before we can call it "true" in any real sense.

The issue is that a lot of the concepts involved with species make sense--except when they don't. And when we talk about "speciation" we're really talking about one of a suite of events, each with its own forcing mechanisms, timelines, characteristics, etc.

Don't jump the gun, what I am saying is that speciation at its simplest is the idea of gene flow restriction. When a population of fish become physically unable to allow their genes to flow be it through a new physical divide in their environment, or maybe different mating habits that restricts gene flow are all parts of speciation. There's no pinpoint time a species is a species because it's in continuum.

 

[Dinwar]

Don't jump the gun, what I am saying is that speciation at its simplest is the idea of gene flow restriction.

My point is that that's not a great way to define it. Chronospecies, for example--it's the same population through time, so there's no restricted geneflow except where one would expect for any other population. I mean, if you call me the same species as my grandfather how can you justify calling two chronospecies different species?

The other problem is, it only helps us look to the past. What do we do about the future? Not every split is permanent--our own species is proof of that.

There's no pinpoint time a species is a species because it's in continuum.

That's the problem in a nutshell. The whole issue of speciation essentially is that we're putting a somewhat arbitrary line in the sand. We know it has to go SOMEWHERE, but the precise location isn't something we can define objectively. Give me any definition for species and I can poke holes in it. I'm sure Kotatsu can do more. Honestly, there's nothing wrong with defining speciation at the point where gene flow becomes restricted--at least, no more so than any other. The problem may be that the concept of species isn't valid.

 

[Paulhoff]

"How did life arise? What were conditions really like at the dawn of life?"

What does this have to do with Evolution, evolution has to do with changes to life forms, not how life started.

Paul

 

[Lowpro]

heh I agree, but I think the most accurate way to define a species shouldn't be a blurb for a textbook, but a tabulation of gene flow

 

[Paulhoff]

I did. The evidence for the Evolution Theory in its current form is very weak. There is currently no evidence whatsoever to conclude that multi-cellular life evolved from unicellular life. The investigations into that belief has been so disappointing that it has been all but abandoned.

Evolution is one of the strongest Scientific Theories, it seems you don't know how it works.


Paul

 

[Dinwar]

heh I agree, but I think the most accurate way to define a species shouldn't be a blurb for a textbook, but a tabulation of gene flow

This gets into an issue that came up in the '60s and '70s--what's the POINT of classifying things? Taxonomy started simply to put things in boxes--the branching hierarchy was a discovery, not something that Linnaeus knew going in. Classical taxonomy actually makes no assumptions about shared ancestry. I've even heard the argument (from my grad professor, though he may have been playing Devil's Advocate) that it's still that way--a phylum does not need to have a common ancestry, merely a common bauplan. In contrast, cladistics and phylogeny were intentionally devised to determine ancestry. When you say that two critters are in the same clade you're also saying that they came from a common ancestor.

Paleontologists (not sure if biologists joined them or not) have been pushing to switch from taxonomy to cladistics for decades. What's held them back is that the cladistic system is so cumbersome that it becomes unusable.

Part of the issue is that classical taxonomy is more stable than cladistics or phylogeny--a new species can be added without upsetting the whole structure. Cladistics and phylogeny, on the other hand, change with each added species because they generate hypotheses about the ancestry of KNOWN taxa. That said, most paleontology papers dealing with species and inter-relationships among them include cladistic analyses.

It's an interesting debate, in that it really digs into the concepts of classification. And it got downright silly. There's an example of a guy creating imaginary species, using a known evolutionary plan, and challanging paleontologists to put them in the right order. The guy that did it right used random computer punch cards--he put them over the picture of the fictional critter and every time there was black in the hole it got a "1", where there wasn't it got a "0" (you've got to code characters to do the math for cladistics/phylogeny). This bears NO resemblance to biological work, but....it worked. And it worked better than the experts did. We're still dealing with that.

 

[Lowpro]

Well that's just a problem with taxonomy. I think cladistics works better, but what it will come down to is ignoring what a species is as we know it and instead use gene pools from populations. It's more informative this way.

 

[ThunderChunky]

I did. The evidence for the Evolution Theory in its current form is very weak. There is currently no evidence whatsoever to conclude that multi-cellular life evolved from unicellular life. The investigations into that belief has been so disappointing that it has been all but abandoned.

Multicellularity has evolved multiple times in the history of life. There was a recent study published where they took the unicellular s. cerevisiae (yeast) and evolved it into a multicellular form.

Experimental evolution of multicellularity
http://www.pnas.org/content/109/5/1595

Here's a nature news article about the study which includes a really cool video: http://www.nature.com/news/yeast-suggests-speedy-start-for-multicellular-life-1.9810

 

[cyborg]

This gets into an issue that came up in the '60s and '70s--what's the POINT of classifying things?

From my perspective I would think ultimately to be helpful to the user of the data.

I would expect that genetic databases will open up whole new ways of doing this - as a computer scientist it simply makes sense to me that a researcher would not necessarily be so concerned with any one categorisation scheme but would probably want to run a specific query for the relationships of interest.

When such a thing is technically feasible I imagine the tradional notion of coming up with one heirachy will become obsolete - machines will categorise automatically.

 

[Dinwar]

How will you get genetic information from, say, Isotilus maximus?

 

[cyborg]

How will you get genetic information from, say, Isotilus maximus?

You don't - but then isn't that one of the problems that DNA research has brought up as far as classification goes? That seemingly related/unrelated species' DNA tell a story contrary to their morphology?