Homeostasis
Body temperature held near 37°C against cold air, hard work and cold water. The card asked whether 37 is something the body makes for itself or a number handed down that the machinery only defends. The physiology gives a sharper answer than the card expected.
The regulation was never in doubt. Take a person out of the cold or put them to work and their core temperature moves a little and comes back; push hard enough and it leaves and does not return. The open question was the target. If 37°C were stored somewhere as a fixed setting, the loop would be a good servant to a number it did not choose. Current thermoregulation research describes something else: there is no single stored setting at all, the resting temperature is where several independent loops balance, and in a fever the body's own signals move where they balance.
On this site a basin is the set of conditions a system comes back from. Push a marble around the inside of a bowl and it rolls back to the bottom; the bowl is its basin, and the rim is its edge, the push beyond which it does not come back. Finding basins, and their edges, is the point of the Laboratory. Body temperature has one basin, and a fever moves it. With the loops cut it has none.
| basin | what it returns to | what pushes it | its edge |
|---|---|---|---|
| 1. The balance point | a core near 37°C, where the three loops cancel | Exercise, Cold plunge, hot or cold air | a load larger than every loop at its limit can answer |
| 2. The fever's balance point | a core near 39°C, set by the body itself | Start a fever, and the fever ending | the same limits, with the bowl moved up |
1. The balance point. As the panel There is no thermostat explains, no 37 is stored anywhere: the bottom of the bowl is where shivering, skin blood flow and sweating pull equally. In the model, air from −10°C to 45°C moved the core only between 36.4°C and 37.6°C. Exercise adds 500 W of heat and the core held at 37.97°C while it lasted. Out of the water at 36°C, the core came back to 36.9°C in 63 minutes. The edge: heat going out faster than shivering at its 400 W limit can replace, or coming in faster than sweating at its 600 W limit can remove. Cold plunge crosses it: the core fell below 35°C in 6 minutes and below 32°C in 19, and did not turn back until the load was removed. In a real body the two crossings are well known: hypothermia on the cold side and heat stroke on the hot side.
2. The fever's balance point. Start a fever moves the bowl itself: all three thresholds rise 2°C, so a core at 37°C now reads as too cold, shivering runs at its 400 W limit, and the core reached 38.9°C in 92 minutes and held at 38.95°C. When the fever ends the thresholds drop back, sweating runs at its 600 W limit, and the core was below 37.1°C in 38 minutes. The body sets this target through its own signal, prostaglandin E2. The edge: the same loop limits, now measured from the higher point. No crossing on record.
With the loops cut, temperature has no basin. Press Cut the loops and nothing pulls the core anywhere: with Exercise on it passed 40°C in 25 minutes and kept rising, and in 0°C air it reached 32.4°C after three hours and kept falling.
The physiological claims are quoted from the two sources at the bottom, with the date read. The panel is a simple model built on the first of them, and every number from it was measured from the code before the page was written. It is a teaching model, not a physiological one.
There is no thermostat
The textbook picture is a thermostat: one controller, one stored set point, and effectors that push the temperature toward it. Imeri's 2017 teaching review states the newer view plainly: body temperature is "NOT 'regulated by a unified system with a single controller'", but by "independent thermoeffector loops, each having its own afferent and efferent branches."
Each loop switches on at its own threshold. Shivering starts below one temperature, sweating above another, and the width of the skin's blood vessels answers in between. The temperature the body sits at is a "temperature balance point", whose position "is determined by the different forces" of those loops pulling on it in different directions.
core temperature → sensed → each loop compares it with its own threshold → shivering, skin blood flow, sweating → heat made and heat lost → core temperature
That changes the card's question. There is no stored 37 to ask about. There are thresholds, and 37 is where the loops they govern happen to cancel out.
A fever is the body moving its own target
It is natural to picture a fever as a disturbance that the loop drives back down. That is not what happens. Blomqvist and Engblom describe fever as involving "an initial upward shift of the threshold for both cold-defense and thermolytic responses", and they identify the signal: "PGE2, through the activation of EP3 receptors in the central nervous system, is the final mediator of fever." Imeri's review says the same in the new vocabulary: a pathological process "shifts the thresholds of the different thermoeffector loops, and, as a consequence, the balance point to a new, higher value."
So during a fever the loops are working normally, and they are defending a higher value. That is why a person coming down with a fever shivers under blankets at 37.5°C: every cold-defence loop now reads that as too cold. And when the fever breaks, the thresholds drop back and the same person sweats, because a body at 39°C now reads as too warm.
The signal that moves the thresholds, prostaglandin E2, is made by the body in response to infection. The target is set by the body and reset by the body.
The loops, run here
Three loops with their own thresholds: shivering below 36.7°C, sweating above 37.2°C, skin blood flow answering around 37.0°C. Nothing in the model stores a set point. Ten simulated minutes pass every second.
How to read the bars. Each bar is one loop. Shivering and sweating are switches with a threshold, printed above each bar: at 37°C in mild air both sit empty, which is correct. Shivering fills when the core falls below its threshold (cold air, cold water, the start of a fever); sweating fills when the core rises above its own (exercise, the end of a fever). Skin flow works both ways from the line across its middle: above it the skin vessels widen and dump heat, below it they narrow and keep heat in. Cut the loops and all three are switched off.
How to read the graph. Core temperature over the last three simulated hours. Blue ticks mark a load switched on or off (exercise or cold water), red ticks a fever starting or ending, white ticks the loops cut or restored. A line that bends back toward 37 is the loops absorbing a load; a line that keeps going is a value leaving the basin.
Measured from the panel
| what was done | what happened |
|---|---|
| air from −10°C to 45°C, left to settle | core held between 36.4°C and 37.6°C across the whole range; shivering up to 82 W in the cold, sweating up to 176 W in the heat |
| fever started | shivering at its 400 W limit; 38.9°C reached in 92 minutes; held at 38.95°C |
| fever ended | sweating at its 600 W limit; back below 37.1°C in 38 minutes |
| exercise, +500 W | held at 37.97°C, sweating 309 W |
| exercise with the loops cut | past 40°C in 25 minutes and still rising |
| air at 0°C, three hours | loops on: 36.5°C. Loops cut: 32.4°C and still falling |
| cold plunge | shivering at its limit; below 35°C in 6 minutes, below 32°C in 19 minutes |
| out of the water at 36°C | back to 36.9°C in 63 minutes |
Two things in that table are worth reading twice. The resting temperature is not fixed: it drifts from 36.4°C in freezing air to 37.6°C in hot air, because the loops are proportional and settle wherever their pulls cancel under that load. That is the balance-point picture in action, and it is a property this model was built to have, so it demonstrates the idea rather than proving it. And the plunge shows the edge of the basin: when the heat going out is more than every loop at its limit can make, the value leaves and does not come back until the load is removed.
What would show this reading wrong
A demonstration that core temperature is held by a single controller comparing it with one stored value, which would restore the thermostat picture and reopen the question of where that value comes from.
A fever in which the loops fight the rise, with sweating and widened skin vessels throughout, which would make fever a displacement after all rather than a moved target.
A resting temperature that stays exactly fixed across every load, which the balance-point picture says should not happen and this panel does not show.
Honest limits
The card also names blood glucose, a loop of the same general shape. This page models temperature only, and makes no claim about glucose.
The model has three loops. Real thermoregulation has more, including behaviour: putting on a coat is a thermoeffector too. The model's thresholds, gains and limits are chosen, not measured.
Not read: Romanovsky's 2007 review in the American Journal of Physiology, which the search results tie to the independent-loop view, returned an error; a StatPearls chapter on fever and a 2015 Nature Reviews Immunology review on fever were not reachable. Nothing on this page rests on them.
Sources
Imeri, L. (2017). Thermoregulation as a non-unified system: a difficult to teach concept. Temperature. Source of the independent thermoeffector loops, the balance point, and the account of fever as a shift of the loops' thresholds. pmc.ncbi.nlm.nih.gov/articles/PMC5356210
Blomqvist, A., & Engblom, D. (2018). Neural Mechanisms of Inflammation-Induced Fever. The Neuroscientist. Source of the upward threshold shift in fever and of PGE2 acting on EP3 receptors as its final mediator. doi.org/10.1177/1073858418760481
Related on this site: Life & Physiology; the Damped Pendulum, return that costs nothing; and the Living Cell, the structure a loop like this belongs to.