When we think of evolution, guided by natural selection, we tend to think it leads to “better” organisms. And in some sense, it often does. Forget the idea of a ladder where you always get “higher” organisms as you go up. It’s way more complicated than that. But even putting that aside, even if we just consider organisms adapting to new environments rather than becoming “better”, it seems natural selection guides things in a purposeful direction.
That does happen, but it’s easy to get the wrong impression. If you look at some feature of an organism that’s well adapted to its environment, like a monkey’s hands and tail used for climbing or a flower’s colours that attract insects, you’ll get an explanation like this: In these circumstances, it was useful for the species to have that trait, so natural selection favoured that trait and the species developed in that direction. Those who had the trait could outcompete those who didn’t, so they had more descendants, and so natural selection gave the species that trait.
It seems like this is the very basis of evolution by natural selection. Well, kind of. But it’s oversimplified. The problem is that it often leads to the reasoning “If trait A is good for survival and its alternative trait B is not, then natural selection must end up choosing trait B.” But if you ask real evolutionary biologists, there are a number of ways in which it does the opposite, and those make perfect sense once you look at the details too…
1. Lions have evolved to kill their cubs: Competition harms the species
If natural selection makes the species more fit, it should avoid anything that harms the species, right? The species will do better if there are no behaviours that threaten it.
But this isn’t really what natural selection does. Sure, often improving individuals for some given environment simply helps the species as well. But the genes that are competing for survival don’t know anything about species lines. If a gene (or allele, but let’s not go into that) succeeds in spreading better than its rival genes, it will do so, regardless of whether those rivals are in the same species or not. And if a gene can spread itself better while harming the species — there’s nothing to stop it from doing so.
An example of this is that male lions taking over a herd often kill the young, unweaned cubs that are already there. It’s bad for the lion species, because there are fewer lions to go around — but so what? It’s good for the genes of these males because they get to breed with the females faster with their old cubs out of the way. So a gene promoting such behaviour will gain an advantage, and will be selected to the detriment of the species.
Apparently it’s an ongoing controversy in evolutionary biology what the “unit of selection” is. Is there such a thing as group selection, for example — where a group that’s more fit than another gains the advantage and thus “selfish” behaviour like described above doesn’t pay off after all, because it harms the group? Or is it “selfish genes” all the way, so anything that a gene can do to get an advantage gets selected, even if it harms the group — and even if harms the gene in the longer run? Or not genes but individuals? I’m not qualified to judge, but the examples in this article certainly seem to speak for genes or at least individuals as the unit where selection happens.
2. Trees really shouldn’t be so tall: Competition harms the individuals
Okay, so natural selection can harm the species, but that’s because it makes the individuals able to breed better. Surely there’s no way natural selection can harm the individuals themselves?
It’s time to ask a question you probably never thought to ask: Why do trees grow so tall?
The thing here is that the basic environment for the tree is the forest. Sure, you also get them growing alone — in which case they grow differently and spread out more, though I’m guessing they still grow big because of how they’re evolved to grow in forests. But trees are evolved to grow among a lot of other trees. And they are plants that need sunlight. Other trees get in the way of sunlight.
You can’t have some trees growing taller than others, because the others wouldn’t get any light, so they won’t evolve to be shorter than others. So all are going to be about the same height. But how tall should they be? Well, growing tall takes resources; it’s harder than being short. If we follow the “If A leads to better fitness than B, then A will be selected,” then if A is trees being short and B is trees being tall, A should be selected.
But it’s not. Suppose all the trees were short. And mutations leading to evolution are still going on. That would mean that trees that grew a little longer and got more sunlight at the expense of the others would prosper better. But then any mutation leading to that would be favoured. So being taller would become the norm. But then being a little taller still would be even better. So that kind of genes would be favoured. And so, trees would keep getting taller.
There’s a limit: at some point, growing any taller would waste so much energy it’s not worth it even to get sunlight better. So that means that the kind of height trees evolve to grow to is the maximum height beyond which competing wouldn’t pay off any more. This harms all of them because they waste resources doing so. If they could only be shorter, they would all benefit.
But the trees can’t make a deal to all stay short. What would make the deal binding? As soon as you’d get a taller mutant, it would be favoured by selection again. So trees have evolved to be tall in a completely futile arms race against each other, and they have no way out of this situation.
The same problem can be seen in the “Prisoner’s Dilemma” kind of situations and the evolution of human morality — which has found a clever way around it, as you can see behind the link.
3. Arms race to the death: Sabre-toother cats and titanotheres
A question that might occur to you is why natural selection doesn’t just put a stop to this kind of silliness. It has a pretty powerful means of doing that; those who aren’t fit die.
Well, obviously with the lions and trees, the handicap created by natural selection is not big enough to kill them off, and the intra-species competition is just a bigger deal than the negative effects. But there’s no reason in principle why natural selection can’t be so stupid that it kills off a species. Again, this is seemingly perverse, but entirely in accordance with the logic of natural selection.
An example of this may have happened with creodonts — sabre-toothed cats — and titanotheres — big ancient rhino things with weird horns. It’s hard to find a lot of sources of this, but it’s been cited as fact by a legitimate, even notable biologist.
What apparently happened was this: The creodonts would prey on the titanotheres, so they had a good old-fashioned predator–prey arms race — which sounds more familiar than these weird species-hindering intra-species ones. Both got better at eating and not being eaten. This involved the titanotheres evolving thicker hides and bigger horns, and the creodonts growing bigger teeth to better pierce those hides.
But the teeth, the horns and the thick skins were cumbersome and made it more difficult for the animals to survive otherwise. So both species’ fitness was going down — since they competed with each other and both were getting better at that, each gained no advantage over the other, but they did get the disadvantages. Yet, apparently the pressure from the immediate competitors was so great that neither species could afford to tone it down. If you stopped overdeveloping the teeth or the hide, you would not get to eat/would be eaten. But since both species were becoming so awkward otherwise, they eventually died out. I would guess it was also a case of short-term gain being selected but then leading to long-term loss.
This didn’t even happen just once, but several times with different species among the creodont and titanothere groups.
So, yes: natural selection can weed out the creatures it takes to a dead end. But that doesn’t mean it would always find a way to make the organisms better to get them out of there alive.
4. The panda principle, or stuck with what you’ve got: Human childbirth could be safe and painless (and giraffes are ridiculously designed)
Consider again the (false) principle “If trait A leads to better survival than its alternative trait B, then natural selection will lead to A, not B.”
Now consider human childbirth. Here, it seems we can give a story for why it has to be so difficult, even with natural selection choosing the best way. Human babies have big heads, and the humans giving birth to them can’t have a broader pelvis than they do because they need to be able to walk on two legs. Since the baby has to come through the pelvis, there’s just no alternative: even though it’s presumably been optimised by evolution, human childbirth is difficult.
Okay. So… why does the baby have to come out through the pelvis? Why can’t you have, say, an exit hole for babies below the belly but above the pelvis? Then you wouldn’t have to squeeze it out through a narrow opening.
The reason for this lies in our ancestors. We’re descended from animals that went on four legs and whose babies didn’t have big heads. So when the system for giving birth through the pelvis evolved, it wasn’t an issue that it happened through the pelvis.
But then we got these big-headed, bipedal humans, which evolved that same system further — and evolution couldn’t make it work so well any more. The reason it can’t is that you’d have to make a big change that wouldn’t give any advantage before it was finished. A mutation moving the birth canal or whatever a little in a different direction? It just wouldn’t work, and it wouldn’t be selected.
One thing a lot of people don’t realise about evolution is that it has no goals set in advance. (A lot of them wouldn’t say they think it does, but they still make the assumption in practice — it relates to the principle about A being better than B.) Big changes to some new, fit form only happen as a result of a lot of little changes that each made the organism a little more fit. (This is exactly the opposite of what many “creationist” “arguments” assume.) You can get whales out of land mammals because the land mammals are becoming more suited to living in the water bit by small helpful bit, but you can’t get any change that would require a single big step to become useful, or a lot of small steps that first make you unable to give birth.
So sometimes you’re stuck with what you’ve got because “you can’t get there from here” for some better option. Stephen Jay Gould called this the panda principle after the panda’s false thumb. It can also be seen in human technology.
Another example from evolution is the recurrent laryngeal nerve. Basically, it’s a nerve that leads directly from the brain to the larynx and controls it. Only, when I say “directly”, I mean “through a pointless detour inside the chest and around one of the arteries of the heart.” It could take a straighter path, but it doesn’t. The reason is basically that our ancestors were fish who were shaped a lot more simply and our body plans were made by twisting theirs around, without anyone stopping to optimise them along the way. Now what we have is a rather pointlessly complicated arrangement with possibly maladaptive and certainly weird effects — like the possibility of losing your voice from a blow to the chest.
There’s more. If you think that’s weird with us, it gets ridiculous with the giraffe. You know, the animal with the really damn long neck. Yep, for the giraffe as well, the nerve goes all the way down along the neck and then up again, for absolutely no reason other than the panda principle. (Don’t talk to me about “intelligent” design.) It may even be that here, the cost of the silly design would outweigh the cost of awkwardly evolving a new design — but as the neck evolved to be longer bit by small bit, the marginal cost of each step was never that big, so selection couldn’t get a foot in.
5. Sexual Fetish selection
Sexual selection is where one sex in a species evolves certain traits because of selection pressures coming from the other sex. The mechanism behind this is that in many species, one sex invests more in the raising of the offspring, so it pays for genes in an individual of that sex to make it pickier of whom it mates with, whereas the other sex can afford to mate around more since it has to invest no more than that. Usually it’s the females that end up being the choosers and the males that end up looking fancy to attract them — just think of the peacock or the widowbird with their huge tails. In some species, the males are the ones who look after the young, and the roles are reversed in sexual selection too. In other species, both parents put in equal effort, and they tend to look more alike rather than either one being fancy.
The peacock’s tail and other traits like it can hinder survival considerably — imagine having to lug something like that behind you. So we’ll want to know why the “chooser sex” favours a particular trait — does the process have to make any sense? Basically, there are different possible reasons, which may also combine in the same case, and we may not be know which one applies where. With the peacock tails, they may be signals of fitness, indicating that the male has such good genes as to survive with such a hindrance (why the hindrance itself doesn’t count as “bad genes” here I don’t know, but this is what I’ve heard), or signals that the male doesn’t have parasites or some such. This would still be an example of wasteful competition: it would save everyone effort if the males could just signal their quality in some less costly way, but then they’d just evolve to give that signal regardless, so the only stable strategy is to make the signal so hard to give you can’t fake it.
But there may also be no reason at all other than natural selection accidentally enforcing arbitrary mating preferences. This process is called Fisherian selection or the Fisherian runaway. The basic idea is this: a female (let’s talk for simplicity as if it’s always females doing the choosing, just don’t start applying this stuff to humans as a crappy excuse for gender stereotypes) might have a genetic disposition to like a particular trait like a long tail. Then, she is likely to mate with males with that trait. Then the offspring are likely to have both the gene for liking that trait if female and for having it if male. If some such preference happens to spread in the population enough, there’s suddenly a selection advantage in the population as a whole to having those traits. A longer tail becomes advantageous for the males because there are so many females around liking it, and the process only reinforces itself in a positive feedback loop. You get longer and longer tails, and preference for longer and longer tails, until the tails get so likely to get you killed even the breeding advantage is not worth growing them even longer.
If and when traits get selected like this, and as we have seen such selection can have quite powerful effects, the logic of natural selection is again followed perfectly. Yet what happens is very far from purposeful and is instead a result of the weird dynamics of interacting systems.
6. Extra-selfish genes kill whole populations of mice
We have seen that natural selection can produce some really stupid results. It can harm the species or even all the individuals in pursuit of short-sighted advantage. It can’t reach better results because of pointless competition and because it can’t cross some gaps. It can even kill off entire species due to its “short-sightedness” and create its own objectively harmful rules that organisms have to obey.
I’ve saved the best for last: natural selection may even select genes that offer no advantage to the organism whatsoever and are in fact entirely antithetical to the survival of its offspring.
How could that possibly make sense? It’s because those genes have an unusual advantage at the point of reproduction. In sexual reproduction (at least in humans and the cases relevant here), the offspring gets a mix of genes form both parents and has a 50% chance of getting any given gene one its parents had. If a gene manages to cheat at this point, it gets a totally unfair advantage — it improves its chances of being “selected” at this point, never mind effects on survival.
An example of this kind of gene is the “t gene” in mice. When a male mouse possesses just one copy of the t gene (genes come in pairs and you get one from each parent), almost all of its sperm are going to contain the t gene, so it’s sure to be passed on. A single t gene has no direct harmful effect, so this trick helps it spread, and when the mutation arises in a population of mice, it really does spread.
And then the whole population may well die, because when a mice gets two of the t gene, from both its parents, it dies young or is sterile.
Talk about selfish genes.
After all this, it’s worth it to bear in mind that natural selection can and does lead to what we would call progress and good solutions. Perhaps it does that most of the time. Though A won’t always be chosen over B just because A would be better in terms of fitness, it does happen all the time that natural selection selects something because it really does help fitness. It’s just that it’s a gradual process that works by its own rules and limitations. And while conscious design could often see how something could be improved from what evolution did, natural selection also has a huge advantage over conscious design: natural selection will take everything into account (in its own short-sighted way) because everything that affects survival affects selection. In contrast, conscious design is limited by the designer’s ideas of what the system needs to be like, which may fail to take into account small vital details or complex interactions. (Obviously I’m talking about finite designers, not the idea of a god designing nature.) Human designers have even learnt to use evolutionary algorithms to come up with designs they couldn’t think of themselves.
Natural selection is a wonderful and powerful process capable of the most amazing things. But to understand it properly, one needs to understand that it’s not a goal-oriented process in the naïve sense we tend to assume it is. Its “limitations” are written all over its results. The above examples should help to understand what this means.
All the images are from Pixabay.