Lots of filler. Weird grandiose writing style. Poorly argued. Many unnecessary appeals to Aristotle.

Night/day cycle is one of the few environmental features that has been constant throughout evolution. Daytime means massive energy supply, but also massive dose of ionizing radiation. Adaptive pressure.

Individual cells have circadian rhythms. Feedback cycle of gene expression -> protein production -> gene inhibition. Not clear how the proteins are removed to restart the cycle. Cycle use different genes+proteins across animals, plants, bacteria, fungi. Possibly parallel evolution.

In animals, cycles are synchronized between cells. Mechanism is not clearly explained - some kind of chemical signaling. A region of the brain receives light/dark signals directly from eyes and uses this to recalibrate the cycle daily.

Disrupting the light/dark cycle in animals can cause health problems including obesity. Human shift workers have higher rates of many major diseases.

Sleep is characterized by physical dormancy along with reduced response to stimulus. The latter part is what makes it mysterious - being unaware of predators is clearly dangerous, so there must be a strong advantage to it.

Energy conservation alone not enough to this - sleep states are observed in various cold-blooded animals including fish and insects, which don’t spend much energy to maintain their awake state compared to mammals and birds.

In dolphins and orcas, newborns and mothers go without sleep for a month after birth. Sharks and several other fish don’t appear to sleep at all. Prey animals which have nowhere to hide (eg horses, buffalo) sleep much less than similar species which have safe sleeping locations. Some species of blind cave fish hardly sleep at all. Fish, ducks, horses etc sleep in groups where members near the edge of the group display more awareness / shallower sleep than members inside the group. So sleep appears to be a luxury rather than a necessity, in evolutionary terms.

Birds and mammals both display REM-like states, but reptiles don’t. Suggests parallel evolution.

Argues that learning acts to smooth the evolutionary fitness landscape. Not super clear, but I think the idea is this: A complex behaviour is made up of many parts. If those parts individually have no benefit, then it’s unlikely that evolution will produce that behaviour. It’s a spike in the fitness landscape that can’t be climbed. Animals that learn can acquire the behaviour and pass it down to their children. If the learned behaviour exists for long enough, mutations that move components of the behaviour into hardware can be selected for. The spike becomes a hill.

Argues that the main benefit of sleep is the ability to learn over time.

Animals that sleep display cognitive impairment when sleep is disrupted. Even bees dance poorly when sleep deprived. Sleep-deprived humans show less cognitive flexibility and rely more on habit and instinct. Sleep deprivation also impairs memory formation.

Against the homeostasis theory of sleep. If sleep is for processing waking experiences, it makes sense that longer waking times result in longer/deeper sleep. Don’t need to invoke an gradual imbalance that is restored in sleep to explain this.