I'm currently reading Richard Dawkins book "The greatest show on earth", which I recommend to everyone, it's a really well written book. I haven't finished it yet, but I have some questions that the book doesn't seem to answer(at least, from what I read), or probably I missed them. Basically:

1)The whole mechanism of evolution is mutation+natural selection, am I correct? That is, let's take the eye or the lungs. They were first maybe a very simple organ for detecting light, and in the passage of many millenia, random mutations happened in the genes that "coded" that organ, and those mutations that helped the organism survive better remained, while others(the most) were discarded. It must follow that happened thousands or millions of "beneficial" mutations? Is that basically it, or am I missing something, is there a more sophisticated mechanism?

2)What I don't understand at all, since my days of highschool, is how completely new species arise. It is talked about in the book how we see evolution real-time in dog or flower breeding(it gives an example of foxes being bred in such a way that they give birth, after some generations, to foxes that look and behave more like dogs), it is talked about experiments in bacteria that show the same things. What I'm lost is if we know how new species arise. Does that happen due to mutation, that is, at some point happen a series of mutation that turn some population of a species to another species, genetically incompatible with the first one? When a species "splits" in half, with the halves being genetically incompatible so we can talk about different species? If so, what happens, do whole chromosomes get added or subtracted(as we know, different species have different number of chromosomes).

Thanks in advance for any helpful answers.

http://en.wikipedia.org/wiki/Evolution

http://en.wikipedia.org/wiki/Speciation

1) Yes, that's basically it.

2) Species boundaries are not always 100%. Think donkeys and horses and their crossbreeds. So for speciation to happen, you basically need the population of a species to be strictly separated for some reason, usually geographic separation, and then wait a sufficiently high number of generations. The separated populations are likely to drift apart genetically over time, making off-spring that results from interbreeding more and more unlikely to survive, until you reach a point where interbreeding has become basically impossible and you can look at the populations as different species.

It must follow that happened thousands or millions of "beneficial" mutations? Is that basically it, or am I missing something, is there a more sophisticated mechanism?

Not necessarily beneficial ones, plenty of bad ones too. Imagine being born blind, how difficult it would be to survive and reproduce (not so difficult now of course). There's still plenty of genetic diseases around which no longer die out because we can treat them (symptoms at least).

An iteresting thing to look up is "convergant" evolution. When two completely different species with no similar ancestry both evolve things that are the same (when neither species had them to begin with).

1. http://en.wikipedia....tion_of_the_eye
Interestingly octopuses have "better designed" eyes than mammals. Basically, the terms "better", "developed" etc are highly arbitrary and very context dependant.

2. I guess usually they got separated by something. By a geological stuff, or maybe they are just highly territorial animals (not interacting with other kennels, or whatever the word is).

Evolution is not a driving force, it's observation. It does not force or direct anything.

DNA replication is imperfect and each replication, whether with offspring or during normal cell replication results in small errors. Since most of DNA is redundant and since most of such defects are harmless, variation occurs on every cell division and on each conception of new offspring.


Now you take a total population of species (100 million, 100 billion, ....) and watch them over a period of time (1 million years). A small part of this population will have slightly different organ, perhaps 2% different than the average. This in turn will cause them to use 1% less energy, meaning they will need 1% less food. When mating, such specimens will be able to perform 1% better and longer and result in 1% more offspring, meaning their genetic material will be passed on in 51% of the cases.

1% may sound very little. But let's say it is year 0 AD and we put $1 into bank at 1% interest rate. How much will we have today? $490097714.

In nature, this happens with each generation of offspring. So a tiny advantage results in huge difference over long periods of time. And since nature changes and species migrate, such advantages may be considerably more than 1%.

Take a fictional example of salmons. They travel upstream to procreate. On year, water is low and only small ones make it, the bigger ones get stuck in shallow water. Whatever the cause of "smallness" (perhaps less agressive and gets less food or slower metabolism), it now becomes a dominant trait in next generation. But there is another school which didn't suffer from that in which the bigger specimens procreate more and generate bigger offspring. And the species diverge.


"Winning" in terms of natural selection does not mean being The Best. It merely means being better than immediate competition.


Mutation is an extreme example and most of the time it results in defective individuals which do not survive. The driving force behind is small tiny changes which allow some to be just slightly better than peers.

Um, I need a clarification about something. If a mutation that doesn't matter (0% advantage/disadvantage) occurs, then statistically, after a long (infinite) time, half of the population would have that mutation (setting aside environmental changes that would make the mutation matter, and setting aside mutations that "build on" the said mutation).

This is an idea from long ago and I'm a bit tired to rethink it....

Um, I need a clarification about something. If a mutation that doesn't matter (0% advantage/disadvantage) occurs, then statistically, after a long (infinite) time, half of the population would have that mutation (setting aside environmental changes that would make the mutation matter, and setting aside mutations that "build on" the said mutation).

Read this http://en.wikipedia.org/wiki/Fixation_%28population_genetics%29
It says that only one version will survive. The chance that a mutation will eventually spread to the whole population is just the frequency of the mutation.

1. Natural selection (as far as evolution is concerned) isn't always tied to survival, but procreation. Guys with super effective sperm for example will cause that genetic trait to proliferate within the species, but isn't directly beneficial to their survival.

2. Yes, slow evolution can lead to such splits. As we learn more about genetics though, we're finding that some genes impact the function and behavior of many others. I'm no expert, but I suspect that mutations to particular genes such as these, or to genes that dictate fetus development right after procreation have drastic cascading effects on the population.

Um, I need a clarification about something. If a mutation that doesn't matter (0% advantage/disadvantage) occurs, then statistically, after a long (infinite) time, half of the population would have that mutation (setting aside environmental changes that would make the mutation matter, and setting aside mutations that "build on" the said mutation).

There is a very recent example of such occurrence - Flash.


In 2003, there were many ways to play video on web. Then along came youtube. At that point, Flash was just an inferior way to play video on web, there were many superior specialized solutions. But as YouTube took off, it inherited Flash which gave it advantage in other areas as well simply by providing enough incentive for majority to install it.

So while Flash was always inferior choice, it piggybacked on something else to dominance until it became defining trait.


Software ecosystem is very susceptible to such behavior. Most software in use today is inferior (local minima), but the cause is always secondary, where something that wasn't relevant at time of inception becomes its defining feature. This is reflected in software startup culture - instead of engineering in-depth solutions, as many trivial attempts as possible are made, seeing what sticks and building from there.