EEB 122. Principles of Evolution, Ecology and Behavior?

Lecture?03.?Adaptive Evolution: Natural?Selection

https://oyc.yale.edu/ecology-and-evolutionary-biology/eeb-122/lecture-3?

Evolution can be either adaptive, in which case it has been driven by and shaped by natural selection. It can be neutral, in which case it's been dominated by drift. Or it can be maladaptive. So evolution does not only produce things that work well. Evolution produces things in which stuff can go wrong, and sometimes evolution just wanders around.?

Now adaptive evolution is not about the survival of the fittest. That is a phrase invented by Herbert Spencer, in the nineteenth century, that has had a long shelf life, and it's wrong. Adaptive evolution is about a design for reproductive success. It's all about how many children and grandchildren you have, and whether you do it better than somebody else that's in the population. It's always relative. It always depends on what the picture is of reproductive success in that population at the time that it's happening.?

Both of those things really do happen: nothing for a long period of time, or incredibly fast. So here's incredibly fast: antibiotic resistance. In fact, this is the poster child for rapid evolution.?

Most resistant bacteria live in hospitals, because that's where most antibiotics are used. However, if we develop a new drug and release it in the UK in 2009, resistant strains of bacteria will have evolved and will be in hospitals in the UK within six months, and those resistant strains of bacteria will be observed in hospitals in Hong Kong within two years. The bacteria will have moved around the earth as people move around the earth. The drug industry is in a co-evolutionary arms race to try to keep up with the bacteria that evolve resistance, and we have gradually been losing the arms race.?

So if you read in the newspapers about multiply resistant staphylococcus aureus, which is MRSA, it is starting to crop up now out in the community. It's not just confined to hospital emergency wards and intensive care units anymore, it's starting to spread. And if staph aureus picks up resistance to vancomycin, which is one of our last lines of resistance against it, it's going to be very difficult for surgeons to do operations in the confidence that they can keep their patients from dying after they have surgery.?

This is a problem which is caused by strong directional natural selection, eliciting a rapid evolutionary response. So in the next few slides I'm going to be talking about what I mean by directional and what I mean by rapid.?

By rapid, in this case, if you have a normal sized bacterial population in your body and I give you an antibiotic, the probability is that if you don't do your antibiotic treatment correctly, within about a week or two you will have resistant bacteria in your body.?Here's a good example of rapid evolution in nature in a fish in ecological time.?

• This was done by doing an experiment in which guppies interacted with predators. This is a cichlid fish Crenicichla. This is the pike killifish Rivulus, and it was done in Trinidad, by David Reznick who is at Riverside.

• Now the setup in Trinidad basically is that there is a mountain range on the north end of the island, and there are lots of little streams that are going down the mountain range into a river and they go over a waterfall.

• And the fact that the stream goes over a waterfall has prevented the large predators from getting up above the waterfall, and above some of the waterfalls there were no fish at all.

• So what Reznick did was he took fish that had evolved for a long time, with predators below the waterfalls, and he put them above the waterfalls, and he did replicates. It was a nice system. There were lots of streams. You could do it four or five times, to make sure it was a consistent pattern.

• And these are the results. The life history traits--that means how big they are when they're born, how old, and how large they are when they mature, how many babies they have and how long they live--all evolved rapidly. So they responded quickly.

• The fastest rates of evolution were measured in things that occur early in life. So the number of babies in the first brood, how big the babies were in the first brood, how fast the babies grew, that all changed quickly.

• And basically the pattern was this. If the guppies are under a high predation regime, they mature earlier and they have more smaller offspring, they have a shorter life--this all has something to do with the evolution of aging and why we grow old and die--and they had more smaller offspring.?The males were less colorful and they displayed more discretely.

• Guppy courtship is normally a fairly elaborate thing. The male, who you can see is really brightly colored, also has an elaborate display behavior, and he will dance up in front of the female and he will wave his fins back and forth and then he will dart in and try to mate. And the females prefers males who have bright orange spots. The bright orange spots probably were originally a direct indication that the male was good at catching crustaceans, because crustaceans have carotenoids in them. So they catch amphipods and shrimp and things like that, and then reprocess the chemicals and they can make orange with it. That was an indication that a male was a good forager, and so the female might select that male because then her babies also would be good at catching food.

• However, the male is dancing in front of the female, and that makes him a sitting duck for Crenicichla (Pike Cichlids).?And as we'll see a little bit later in the lecture, Sexual selection involves a direct tradeoff between mating success and survival, and these guys were displaying frantically to get mating success, at the risk of being snapped up by a predator, and the ones that survived were the ones that simplified their display behavior.?

Now how do we measure it? Currently the preferred unit of measurement is a haldane, and a haldane is a change in the mean value of the population by one standard deviation per generation.?

• A standard deviation?is an empirical observation, supported by an elementary theorem of mathematical statistics, called the Central Limit Theorem, that most population distributions look like a bell-shaped curve. It's called the normal distribution. It was formalized by Gauss; sometimes it's called a Gaussian distribution. And the shape of the curve and its spread are basically measured by the standard deviation.

• The mean value is at the center here and the distribution is theoretically symmetrical--in practice it's quasi-symmetrical--and the degree of spread is measured by the standard deviation units. So within 1 standard deviation you will find, on each side, 34.1, or with both 68.2% of all of the individuals observed in the population.

• 1 haldane basically would take a population that say had a mean value--suppose it was for body size, maybe a body size of 10 grams--and if it had a standard deviation of 2 grams, and it was evolving at a rate of 1 haldane, it would move that mean from here to here; 1 standard deviation unit up, and the population mean in the next generation, instead of being 10 grams, would be 12 grams. So that's the meaning of the haldane.?

For those guppies in Trinidad, that were evolving pretty fast, the number of spots in the?area of orange spots--when you took away the predator and suddenly being brightly colored wasn't risky anymore and females liked it; so it was good to be brightly colored--those spot numbers increased quickly. They were increasing at about .7 haldanes.?

In the Galapagos finches that Peter and Rosemary Grant studied, they go through El Nino, and during El Nino--it's a strong selective event, so about every ten years there's a strong selective event on the Galapagos finches--and during El Nino they were evolving at about .7 haldanes in body size; they're getting bigger. And then in the other years they were getting smaller. So they fluctuate, they go up and down, depending upon the El Nino conditions in the Galapagos.?

There have also been lots of measurements of slower rates; for example, since the extinction of their competitors in the late nineteenth century the surviving Hawaiian honeycreeper, the I'iwi has been evolving a shorter bill, and that's been a very slow rate of evolution. The migratory timing of Columbia River salmon has been changing as a result of the human fishery on them.?

All of the fished populations of the world are evolving under the pressure of human fishing. Most of the fish in the world are getting smaller. Many of the stocks are collapsing. It's producing a change in the time of year that the Columbia River salmon run up the Columbia. This is also due to the building of dams on the Columbia. So this is a human induced selection process. These are fairly slow rates.?

So what does this mean, if we just try to think about these rates and evolutionary times? A Galapagos finch is about 25 grams, about the size of a house sparrow. They evolved during El Nino at about half a gram a year. What if the El Nino conditions persisted forever? What if it wasn't the southern oscillation that was driving the rainfall pattern in the Galapagos? What if it just stayed warm and wet for a long time in the Galapagos? Well that would produce directional selection, and if you did it for a hundred years, it would turn a 25 gram finch into a 75 gram finch. Basically it would take a finch and turn it into a small robin.?If you did it for 10,000 years, it would turn it into a turkey.?

Now finches as big as turkeys don't do very well in a finch habitat. They are living in a place where they hop around in bushes. They are living in an environment in which food is sometimes very hard to come by. I've been observing the turkeys that live near my garden in Hamden, trying to get up into the trees next to Lake Whitney to pick the berries off as winter has come on and it's gotten very cold. They're pretty clumsy.?

So what will happen if you keep a strong directional selection going on finches? What would happen to humans if there were strong directional selection on humans to increase in body size? What would happen if we got turned from say 50 to 80 kilo primates into three-ton primates? How long could that go on??

One of the fastest rates of evolution ever measured in the fossil record was when elephants went onto islands in the Mediterranean and turned from twelve-ton elephants down into things about the size of a Saint Bernard. Okay? They did it in less than 100,000 years. They did it because they were food limited and they'd been released from predation pressure.?

So how far can that process go? If you look over the whole spread of evolutionary time, over hundreds of millions of years, things stay within a fairly narrow envelope of body sizes. Why does that happen??

They vary from very fast to very slow. The fastest are in the finches and in the Trinidad guppies. There have been lots of rates measured in Hawaiian mosquitofish and Hawaiian honeycreepers. So there are lots of estimates available. And interestingly, the shorter the period over which the rate is measured, the greater the maximum rate. So if you measure a rate by making comparisons between two populations that have been separated for hundreds of years or hundreds of generations, it's usually fairly slow, and if you focus in and you just look at a brief period, it can be very fast.?

Evolution can be very fast when populations are large and selection is strong. And the reason for that is that big populations have lots of genetic variation. So there's a potential for a big response to selection. Small populations don't have so much genetic variation. So even though selection might be strong, they can't respond so well.?

This point, the shorter the time interval over which you measure the rate, the higher the maximum rate. And here's one reason why you can't take Galapagos finches and turn them into turkeys and then turn the turkeys into ostriches and then turn the ostriches into moas and then have the moas turn into tyrannosaurus rex. Okay? As you push things very far, in any direction, there's an internal process that converts the directional selection into stabilizing selection. And those are the tradeoffs, the linkages among traits.?

If you try to make a finch very large, then although it may be gaining something in terms of say food capturing ability, it is giving up maneuverability. If you try to take elephants and make them very small, then at some point they are not going to be able to compete with other elephants for food supply, even though there may not be any predators there. There are all kinds of biomechanical linkages within bodies where tradeoffs are involved.?

So if you look within the organism, you see that it's a bundle of linkages and compromises, and every time you try to change one trait you have a byproduct, you have an implicit selection going on, on other traits. So although you may be realizing a benefit in one, or a place, you are paying a cost in the others.?

The most striking example we've seen of it in the lecture so far is the guppy, the male guppy. If he evolves to be bright and a wonderful dancer, so that females just love to mate with him, he will get killed by a predator. That is about as straightforward and brutal a tradeoff as you can imagine.?But these go on all over the place and some of them are very subtle.?

Now why is it that sometimes traits evolve very fast and sometimes very slow? This is a picture of clubmoss, lycopodium. If I were to take you out into the woods of Connecticut in the springtime, you would see them all over the place, and if I were to put you a time machine and take you back 400 million years, they wouldn't look any different.?

This is latimeria, this is a Coelacanth. If I were to put you into a research submarine off the Comoro Islands in the strait between South Africa and Madagascar, between Malawi and Madagascar, and we went down at night to a depth of 300 to 600 feet off the volcanic slope of the island, we would find these guys cruising around in mid-water. They have spent the day in caves and they come out at night, into the mid-waters of the earth's oceans, and apparently they have been doing this now for going on 150 million years.?

Look at what's happened to their relatives. The clubmosses had relatives at the time that looked about like them, that since then have turned into redwood trees, orchids, wheat fields--you name it, these guys still look the same. Latimeria had relatives that since then have turned into marlin and reptiles and birds, mammals; it hasn't changed.?

So we have these two things to understand. We have to understand how evolution can go really fast--antibiotic resistance, guppies, finches--and why sometimes it is so slow.?

1. ?An externalist explanation: it relies on aspects of the habitat and the way natural selection is operating on the organisms.

• The reason these guys haven't changed is that they're really good at always finding the same kind of environment, so that they are never exposed to change. So if their environment moves around the globe, they track it.

• this?argument is actually probably most convincing for marine invertebrates, that make larvae that can go out and spread through the ocean for thousands of kilometers.

• And, in fact, we know from the behavior of marine invertebrate larvae--so now I'm talking about worms, barnacles, clams, stuff like that--that they like to settle on places where there are successfully growing adults of their own species. They smell that out very carefully, and that's where they settle. So basically the larvae are selecting the habitat in which the adults will be selected by natural selection. That means that they manage themselves to generate stabilizing selection over hundreds of millions of years.

? ?2. ??An internalist explanation:?developmental constraints; that development has constrained the organisms so that they couldn't evolve in certain ways.?

• That's plausible for certain major features of the body plan, that are determined very early in development, and involve developmental tissue relationships and things like that, that are obviously hard to change.

• It's not so plausible for some of the smaller details of these creatures.?

• The actual explanation is probably a mixture of these things. There probably is some phylogenetic or developmental constraint. Things that happened a long time ago, in the way organisms were built, are hard to change, and they've been constraining the things that can change more rapidly.?

We can talk about directional, stabilizing and disruptive selection; natural and Sexual selection; frequency dependent selection; and then selection acting on individuals, on kin, on groups and on species. So each of these is cutting the selection cake in a different direction; but it is all of these different things.?

• Directional selection, that's?making the Galapagos finch into a turkey, is that the fitness gradient is linear. That means that if the fitness of something over here is low and up here is high, that means that natural selection is selecting for say bigger things--this body size on the X-axis is going to the right--and it will take a distribution that looks like this and it will move it to the right. So if this is 1 standard deviation, then this amount of movement is 1 haldane, right here.?trong directional selection will produce very rapid evolution. We saw that with antibiotic resistance and the guppies. It can't continue. It usually gets converted into stabilizing selection.

• Stabilizing selection is actually what we were just invoking to argue that the coelacanth didn't change. It was living in a habitat where it was always good to be like a coelacanth, and natural selection was selecting out things that didn't look like coelacanths; whether they were larger or smaller, or their fins were different shapes, or things like that. So they tended to stay the same. That means that we were selecting for the mean of the population and we were discarding the extreme values.

• Disruptive selection is selection against the mean and for the extremes, and it will take a bell-shape curve like this, it will knock out the mean value, and then the next generation it will push it apart like that.?

So if we look for examples, strong directional selection will produce very rapid evolution. We saw that with antibiotic resistance and the guppies. It can't continue. It usually gets converted into stabilizing selection. Disruptive selection causes, historically, things like the conversion of similar looking gametes into quite different gametes.?

So disruptive selection was involved in the origin of eggs and sperm, back in the day, about a billion years ago, and it may play a role in sympatric speciation.?Disruptive selection is selection to take a population that has a certain mean value and split it in half and turn it into two different things.?

Now, natural and Sexual selection. We've referred to Sexual selection with the guppy. The classic example of Sexual selection is a peacock's tail. In fact, peacocks are eaten by tigers, or they were eaten by tigers before the tiger just about went extinct in India; they're down to a few hundred in India, and the Siberian tiger is under threat right now in Siberia. But the tigers traditionally ate peacocks. So?the display behavior was dangerous. What the male was doing was he was trading off survival for mating success. He was a victim of female preferences.?

Sexual selection is a component of natural selection. Natural selection is all about variation in reproductive success, and you can achieve reproductive success by mating and by surviving and by doing other things. Okay, so it's a component of natural selection. And the tradeoff involved is survival versus mating.?

It's driven by two things. Either the males are competing with each other for access to females, or the females are conniving against each other for access to males--one of those processes may drive intelligence a little bit more than the other--and it's also driven by members of one Sex choosing mates of the other Sex.

Frequency dependent selection is another kind of selection, and that happens whenever the advantage of doing one thing depends on what the other people in the population are doing.?There are some classical examples of this. One is the classical 50:50 Sex ratio, and another is genetic diversity for immune genes.?

• Let's suppose that you have a gene that is resistant to a particular disease, and therefore your offspring survive better and you have more grandchildren, and this gene then spreads through the population until eventually most of the people in the population are resistant.

• That means that there's selection operating on the disease to come up with a variant that can overcome that resistance, and when that variant comes up, it will spread until it is common, and it creates selection to cause the same thing going on in the host population, and back and forth it goes.

• The more frequent something becomes, the more it's subject to very strong negative selection, and the less frequent it becomes, the more it's protected from being selected, because things that are rare aren't very good resources; things that are common are great resources.

• And so what happens is that you get what is now recognized as a classical oscillation of virulence and resistance between the host and the pathogen.

• One of the most interesting things about a human immune system is that the MHC or HLA genes that mediate this kind of resistance against pathogens have some of the highest genetic diversity of any genes anywhere. It looks like variants, rare variants, have been selected again and again and again. So every time something becomes frequent, it becomes useless and another rare one is selected, and eventually a huge supply of variation builds up in the population. So this principle really has had quite a role to play in the selection of the vertebrate immune system.?

Group selection--here's an example of group selection. A bunch of partridges get together in Scotland in late fall. They look around. They notice that there are just a tremendous number of partridges in Scotland in late fall, and they think--you know, speaking anthropomorphically--they think, "Oh, there are too many partridges. Therefore we will all cut back on our reproduction so that our population does not go extinct."?

That's an example of group selection. It won't work because over in the corner is sitting Joe Partridge, who looks at all of these guys and says, "You're idiots. You're cutting back on your reproduction. I'm going to have 50 babies."?Group selection is vulnerable to selfish mutations; selfish mutants invade.?

So they invade for a variety of reasons, and we'll work through all of that. But group selection is not stable. Selfish mutants invade.?There's a lot of selective events in genes and individuals for each selective event of a whole group.?It is bullshit to?extend group selection up to species,?behavior X or morphology Y existing?for the good of the species.?Things don't exist for the good of the species. Things exist because individuals outperformed other individuals in the competition for reproductive success.?

Now there is some large-scale differential species selection that occurs on the phylogenetic tree, and it shapes patterns at a big macro scale across the tree. One of them is Sex. Virtually all aSexual things are relatively young and they had Sexual ancestors. It appears that Sex reduces the probability of extinction, and that aSex makes you more extinction prone. So that is a kind of species selection. But it's not a selection for a precise adaptation. There's no way that species selection could have ever designed the vertebrate eye, the vertebrate brain, any of the detailed, precise, complicated mechanisms that we know of in biology. All that stuff has gone on because of individual and gene selection.?

Some of the big macro evolutionary patterns have been generated by a kind of species selection. For example, the fact that dinosaurs aren't here anymore and that mammals dominate the earth is a kind of selection. It doesn't tell you about how fast the mammals run, why they are warm-blooded.?