Interesting point on the evolutionary conundrum of sleep. I'm not sure I follow the teleological argument, or if I even like teleological arguments, but here's my two pesos on the evolution of sleep:

You are certainly more vulnerable when you sleep, and death is a particularly strong selective force, so it would seem quite obvious that sleeping would be selected against. On the flipside, the advantages of sleeping are hightened levels of performance during wake time. Imagine if you had a mechanic living in your garage who would fix your car every night. You could drive it like a maniac during the day, knowing that at night, your little mechanic/elf would make your car nice and new the next morning. For an animal, the advantage is that the body can focus energy on being active during periods of activity, and focus energy on repair during periods of sleep.

While the disadvantage of being killed in one's sleep is more of a direct problem for prey than for a predator, both parties benefit from sleeping. The predator is a better hunter after a long rest and the prey is more likely to survive if chased after a refreshing nap. Still, extra energy alone isn't enough of a selective advantage for sleep. If energy for activty was the only necessity, why wouldn't a predator reverse sleeping schedule and become nocturnal, while its diurnal prey is napping?

My hypothesis is that the ability to hunt sleeping prey confers too large of a short-term advantage to be stable for a predator-prey relationship. Predator-prey relationships in nature are to some degree symbiotic. Most nocturnal predators hunt nocturnal prey and diurnal predators hunt diurnal prey. If field mice slept out in the open plains so that barn owls could browse around and gobble them up, mice would quickly disappear from that field. But, once all the mice were gone, the owls, if they had no other diet, would die off soon after from starvation. So, if owls evolved the ability to stay awake 24 hours and catch mice at day while they sleep, sure those owls will be fat and happy, but soon they will quickly outstrip their resources and die.

This argument is teleogical only to the extent that it invokes ideas of Gaia Theory where the real selection is not for survival of an individual, or for a species, but for survival of an ecosystem. An ecosystem that has complementary predator-prey relationships is more likely to be stable in the long run.

One more analogy (evil things, analogies) are rules between humans concerning war. Why is it that there are bans on biological weapons? Why is it that armies would cease fighting at night and sometimes even fraternize? Because, being 24 hour and merciless costs a lot of energy. Because using biological weapons threatens the entire species. Maybe this issue should be investigated more based on game theory rather than evolutionary biology. Since I have studied neither very extensively, I hope this speculation I've presented isn't too asinine.

I wouldn't say asinine but still some loopholes which is normal for evolutionary hypotheses. (And I would like to point out that, YEAH!, the Pragmatists & Evolutionists have taken over the teleological argument. Nyah-nyah.) *cough*

Another idea is that sleep-attacks might just have simply selected for other characteristics: light sleep, improved hearing, better vision, habitat-matching fur/feathers/scales/etc. And, it is reasonable enough to say that a species switched from diurnal to nocturnal or the other way around due to evolutionary forces. All of these can be clearly driven by evolution.

Sleep, on the other hand, not so. The argument that an animal after sleep performs better is not answering the question. The problem here is that evolution "should" have selected for an animal that didn't sleep yet still performed. Since it didn't, there are 3 choices: 1) oops, 2) it's not actually more beneficial to survival, 3) sleep is über-evolutionary.

The correct answer lies presumably in (3) due to cellular biology/histology rather than evolution. The study of the daily rhythm in all organisms and, in fact, in every cell has been progressing. Apparently, there is an enzyme-feedback mechanism which approximates the day. This means a certain chemical is produced and builds up until it reaches a critical point where it is no longer produced. Then an enzyme breaks it down until a certain point at which there is so little that it is produced again. For the cycle to occur once is about 22-24 hours.

Though this isn't proof of sleep, these rhythms make it more more reasonable to assume that there's a natural need for sleep. I'm not saying I know the mechanism which requires sleep, but I would say that it's a much more basic necessity than an evolution-selected characteristic. The need for this energy renewal exists at the cellular level, and so is there to stay.

Although sleep has been extensively studied in humans and animals, it is still a poorly understood phenomenon. Sleep involves extensive adjustments of brain function, lowering of body temperature and the buildup of certain substances in the cerebrospinal fluid (CSF) such as serotonin.

Even less understood is the physical motivation behind the evolution of sleep (see: The teleological argument for sleep). An organism is more vulnerable while sleeping to predators, and it seems that there would be a strong selective pressure against falling asleep. So why did sleep evolve? Obviously for small endotherms (animals who need to produce body heat), the closer the body temperature is to the ambient temperature, the less energy they spend on heat generation. During periods of inactivity, reducing metabolism, especially at night helps save energy. For homeotherms, body temperature is maintained at all times. So what does sleep accomplish then? One theory relates evolution of sleep, particularly for mammals and birds, with evolution of the vertebrate brain.

The common ancestor to mammals and birds are reptiles. Most reptiles are strictly diurnal because they are poikilotherms - animals whose body temperature matches that of the environment. At night, when it is cooler, most reptiles enter a state of torpor. In the morning, sunlight brings warmth and initiates activity. This is often why lizards are described as sunning themselves on rocks. They are increasing their metabolic rate by absorbing solar heat. But torpor is different from mammalian sleep because it is directly linked to metabolism. Mammals can often nap at will and although they do adhere to circadian (night/day) rythms, are not limited to them. At night, reptiles do show behavioral signs of sleep including closed eyes, specific resting sites, specific resting body positions, etc ... During the day, reptiles may also sleep with the same behavior patterns, suggesting that sleep is encoded in reptillian neurophysiology.

Behaviorally, if reptillian sleep and mammalian sleep are similar, does this correspond to neurophysical behavior during periods of rest? Electroencephalograms (EEGs) of reptiles and other poikilotherms are apparently fraught with all kinds of technical problems including temperature effects and electrode placement effects, but it appears that levels of electrical brain activity go down significantly during sleep. Also, high voltage spikes are detected in the limbic region of reptiles, very similar to phenomena noted in mammals during Slow Wave Sleep (SWS). There are also reports of eye movement during reptile sleep (results are still highly disputed) suggesting a reptillian analog of REM (rapid eye movement) sleep.

At night, during the age of reptiles, only the biggest of animals would have enough thermal intertia to maintain activity in periods of lower temperature. As a result, the nights were generally free of vertebrate activity and predation wasn't a problem. With the advent of homeotherms such as birds and mammals, this restriction was no longer true. If mammalian sleep is a result of ancestral sleeping behavior in response to nighttime temperature changes, then the interesting question for homeotherms, where the body temperature remains constant is - how did mammals acquire the current sleep behavior?

The explanations in the literature now get rather sophisticated and speculative. It is believed that the first protomammals probably only had homeothermic regulation during the late night hours, in order to conserve energy. During the daytime, they used ambient temperature, just as poikilotherms did, to maintain body temperature. These creatures could now exploit this window of reptillian sleep in order to develop unhindered. However, in the absence of light, mammals needed to develop other senses such as hearing and smell in order to navigate the nighttime environs. Most reptiles are not neurologically capable of processing only one type of stimulus at a time. The protomammals would need to have heightened awareness in order to survive, which would require 'switching on' of cortical activity during the night time. During the daytime, only the reptillian brainstem activities for thermoregulation would be required. However, since low-grade, reptillian behavior would interfere with cortical activity at night, mammals found a way to downregulate it. (instances now of sleep walking are believed to be brainstem reptillian remnants of thermoregulation through behavior where this switch has malfunctioned.)

Ok, so sleep in mammals is derived from reptillian behavior of thermoregulation. REM sleep is akin to total reptillian torpor, where there is no endothermic process at all. SWS sleep is similar to daytime reptillian sleep where some behavioral remnants of primitive motion are retained. Still - why do we sleep?

Its interesting to note that as we get older, we sleep less and less. During early stages of our life, humans and other mammals are reared, giving some degree of protection. It has been found that reduction in the amount of REM sleep, especially at young ages, can impair brain development. So, it seems straightforward that increased sleep would help us develop the adult brain. The physiological correlation between REM sleep and brain development isn't known yet. It is known that sleep induces higher levels of serotonin and other factors in the CSF.

The nice thing about mammalian sleep is its reversability. If extreme predation is an issue, mammals may survive with less or no sleep for extended periods of time. In summary, sleep may have evolved as a remnant of reptillian responses to low temperatures at night. As a result, mammals developed heightened cortical functions at night to survive the nocturnal environment. Now, our brains grow best at night.

For an interesting review - see: "Why we sleep: the evolutionary pathway to the mammalian sleep" by MC Nicolau and colleagues in Progress in Neurobiology (2000), v62, p 379.

This and other resources within provide a comprehensive picture of the current understanding of sleep behavior and its relation to neurological development.

Go read Steven Pinker or Daniel Dennett or someone with something intelligent to say on this topic.

If I recall correctly, their argument is that you are asking the wrong question: Not why do we sleep, but why are we ever awake? Why do you assume that being awake is a desirable thing from the gene's point of view?

Sleeping conserves energy, reduces the chances of harmful accidents, etc. All things that help the gene's chance of staying incarnate. When there is nothing better to do, i.e. after you have fed, mated and found a spot safe from predators, what is so bizarre about conserving energy by not doing much for a while?

---

sleep reduces the time available to reproduce Since when has lack of time to shag been the limiting factor? It doesn't take that long. Pregnancy takes months for most species, how does sleeping reduce the time available to do this?

it opens us up to much more danger from predators Does it? Which is more likely to get you spotted by a predator: hiding in a cave and not moving, or roaming the countryside.