The textbook figure of 98.6 °F (37.0 °C) is not the true average human body temperature and hasn’t been for a long time. A landmark 1992 study found the mean oral temperature of healthy adults was closer to 98.2 °F, and a large systematic review placed the overall average even lower, around 97.9 °F when measured orally. But even that number is misleading if you treat it as a fixed point, because “normal” shifts throughout the day, differs by age and sex, changes depending on where you stick the thermometer, and appears to be slowly declining across generations.
Where 98.6 Came From and Why It Persisted
The 98.6 °F standard traces back to a German physician named Carl Wunderlich, who published a massive study of axillary (armpit) temperatures in the 1860s and declared 37.0 °C the human norm. That number was converted to 98.6 °F and became medical gospel for over a century. The problem is that Wunderlich’s thermometers were less precise than modern instruments, his measurement site (the armpit) typically reads lower than oral readings, and his patient population lived in a very different era of infectious disease burden.
In 1992, researchers at the University of Maryland tested this figure against modern data and found that Wunderlich was off. Their mean oral temperature came in at 98.2 °F, and the upper limit of normal was 99.9 °F overall, not the traditional 100.4 °F. They recommended that 98.6 °F “should be abandoned as a concept relevant to clinical thermometry.”1JAMA. A Critical Appraisal of 98.6°F, the Upper Limit of the Normal Body Temperature, and Other Legacies of Carl Reinhold August Wunderlich That paper is now over 30 years old, and more recent data suggests the average has drifted even lower since then.
The Numbers From Modern Research
A systematic review pooling data from 36 studies and over 7,600 subjects calculated average temperatures by measurement site. Oral readings averaged about 97.9 °F (36.57 °C), rectal readings about 98.7 °F (37.04 °C), tympanic (ear) readings about 97.9 °F (36.64 °C), and axillary readings about 96.7 °F (35.97 °C). The spread around each of those averages was substantial. For oral measurement, the calculated normal range ran from about 96.3 °F to 99.3 °F (35.73–37.41 °C), meaning a perfectly healthy person could register anywhere in that band without anything being wrong.2PubMed Central. Normal Body Temperature: A Systematic Review – Section: RESULTS
An earlier systematic literature review found even broader ranges when accounting for less common readings. Oral temperatures in studies with strong evidence spanned from 91.8 °F to 100.8 °F (33.2–38.2 °C), though the extremes were rare. That review also noted that women had a slightly wider oral range than men.3PubMed. Normal oral, rectal, tympanic and axillary body temperature in adult men and women: a systematic literature review
How Your Temperature Changes Through the Day
Your body temperature is not a fixed number at any given moment. It follows a reliable daily rhythm driven by your internal clock. A large-scale analysis of body temperature cycles found that mean temperature reaches its lowest point between about 6:00 and 8:00 in the morning and peaks between 6:00 and 8:00 in the evening.4PubMed. The daily, weekly, and seasonal cycles of body temperature analyzed at large scale The difference between your daily low and high can be a full degree Fahrenheit or more, which means a temperature reading that would look slightly elevated at 7 a.m. could be completely normal at 5 p.m.
The 1992 Maryland study recognized this pattern and proposed different upper limits depending on when you take the reading: 99.0 °F in the early morning and 99.9 °F at any time of day.1JAMA. A Critical Appraisal of 98.6°F, the Upper Limit of the Normal Body Temperature, and Other Legacies of Carl Reinhold August Wunderlich Whether you are a natural early riser or a night owl can also shift the timing of your temperature trough: research tracking body temperature continuously for six days found that morning types hit their lowest point almost two hours earlier than evening types.5PubMed. Individual differences in the phase and amplitude of the human circadian temperature rhythm: with an emphasis on morningness-eveningness
The brain region responsible for managing all of this is the preoptic area of the hypothalamus, which works like a thermostat. It reads temperature signals from both the brain itself and sensors throughout the body, then coordinates responses such as sweating, shivering, or adjusting blood flow to the skin to keep things in range.6PubMed Central. A hypothalamic circuit that controls body temperature
Age, Sex, and the Menstrual Cycle
Older adults run cooler than younger adults. The same systematic review that established modern averages found that people under 60 averaged about 98.0 °F (36.69 °C), while those 60 and older averaged about 97.7 °F (36.5 °C).2PubMed Central. Normal Body Temperature: A Systematic Review – Section: RESULTS A study focused on people in their 80s found results that were more dramatic: the average morning temperature in nursing home residents was just 97.3 °F, and over 90% of readings fell below the old 98.6 °F benchmark.7PubMed. Older is colder: temperature range and variation in older people The daily temperature swing also flattens with age. The oldest adults in that study showed almost no change in temperature between morning and evening, losing the pronounced rhythm that younger people have.
This matters clinically because a lower baseline means an older person can be running a fever at a temperature that would look “normal” by traditional standards. Research in nursing homes found that rectal thermometry identified fevers in roughly 15% of elderly patients who appeared afebrile by oral measurement.8PubMed. A comparison of oral, tympanic, and rectal temperature measurement in the elderly Researchers have recommended that clinicians either lower the fever threshold to 99 °F for older adults or track each patient’s individual baseline and watch for a rise of more than about 2.4 °F above it.9PubMed. Fever response in elderly nursing home residents: are the older truly colder?
For women who menstruate, body temperature shifts predictably across the cycle. After ovulation, when progesterone rises, core temperature increases by roughly 0.5 to 1.3 °F (0.3–0.7 °C) compared with the first half of the cycle.10PubMed Central. Temperature regulation in women: Effects of the menstrual cycle This is the basis for basal body temperature charting as a fertility-tracking method. The relationship between progesterone and temperature is reasonably consistent, though one study found the correlation was moderate rather than airtight and that temperature stopped rising once progesterone passed a certain level.11PubMed. Descriptive analysis of the relationship between progesterone and basal body temperature across the menstrual cycle
Why the Thermometer You Use Matters
One of the biggest sources of confusion around “normal” temperature is that different measurement sites give genuinely different numbers, and people often compare readings across sites as if they are interchangeable. They are not. Rectal temperatures run about a degree Fahrenheit higher than oral ones on average, while armpit readings run about a degree lower. Ear (tympanic) readings fall somewhere in between and are notoriously variable from one measurement to the next.
An emergency department study comparing methods side by side found that the mean difference between rectal and oral readings was 1.1 °F, but individual agreement was poor: an oral reading could be nearly 3 °F below the rectal reading or almost a full degree above it. Ear thermometers showed even wider swings, with individual readings deviating from rectal by up to 2 °F in either direction.12PubMed Central. Oral and Tympanic Membrane Temperatures Are Inaccurate to Identify Fever in Emergency Department Adults A systematic review concluded bluntly that oral temperature cannot accurately reflect core body temperature, probably because it is influenced by ambient air temperature, probe placement, and fluid intake.13PubMed Central. Is oral temperature an accurate measurement of deep body temperature? A systematic review
Even within the mouth, location matters. Computational modeling has shown that the back of the mouth holds higher and more stable temperatures than the front, and the cheek side of the mouth reads higher than the tongue side. After drinking something cold, the front teeth area takes noticeably longer to recover its temperature.14Results in Engineering. Computational and experimental study of thermal behavior in the oral cavity for biosensing applications The practical takeaway is to place the probe as far back under the tongue as comfortable and to wait at least 30 minutes after eating or drinking anything hot or cold. Research on beverage effects specifically confirmed that waiting at least half an hour after drinking yields a more accurate oral reading.15PubMed. The effect of respiratory rate and ingestion of hot and cold beverages on the accuracy of oral temperatures measured by electronic thermometers
Temperatures Have Been Falling for Two Centuries
Even accounting for better thermometers and different methods, something genuinely physiological seems to have changed. An analysis of three historical cohorts spanning from the Civil War era to the early 2000s found a steady decline in body temperature over time. Men born in the early 1800s ran about 1.1 °F (0.59 °C) warmer than men born in the late 1990s, a drop of roughly 0.05 °F per decade. Women showed a comparable rate of decline across a slightly shorter window.16PubMed Central. Decreasing human body temperature in the United States since the Industrial Revolution
The leading explanation is that widespread reductions in chronic infection and inflammation over the past two centuries lowered the baseline immune activation that keeps the body slightly warmer. Better sanitation, antibiotics, and vaccines mean that most people today are not fighting the constant low-grade infections that were standard in the 19th century. Supporting this idea, a study of the Tsimane people of Bolivia, an indigenous population undergoing rapid epidemiological transition, found a similar temperature decline happening over just two decades, coinciding with increased access to medical care and reduced infection rates.17PubMed Central. Rapidly declining body temperature in a tropical human population If this trend is real and ongoing, it means that textbook normal values may need periodic updating.
What Actually Counts as a Fever
Given all this variation, defining “fever” is harder than it sounds. The traditional threshold of 100.4 °F (38.0 °C) was based on Wunderlich’s data and does not fit well with modern measurements. The 1992 Maryland study proposed 99.9 °F (37.7 °C) as the upper limit of normal for healthy adults under 40.1JAMA. A Critical Appraisal of 98.6°F, the Upper Limit of the Normal Body Temperature, and Other Legacies of Carl Reinhold August Wunderlich In practice, most clinicians still use 100.4 °F as a working cutoff for clear fever while treating temperatures in the 99–100 °F range as a gray zone that requires clinical judgment.
For older adults, this gray zone is especially important. Elderly patients often fail to mount a robust fever even during serious infections, and a review of the literature noted that peripheral thermometers have low sensitivity for catching the low-grade fevers that older patients do produce.18PubMed Central. Approach to Low Body Temperature or Mild Hypothermia in the Geriatric Population: A Narrative Review – Section: Challenges With Body Temperature Measurement The age-related drop in baseline temperature, the blunted fever response, and common use of fever-reducing medications like acetaminophen can combine to mask infections that would have produced an obvious temperature spike in a younger person.
Fever Versus Hyperthermia
Not every elevated body temperature is a fever. Fever and hyperthermia look similar on a thermometer but work through completely different mechanisms, and this distinction changes what you should do about them.
In fever, the hypothalamic thermostat is deliberately turned up. The body’s regulatory system remains fully functional; it is just defending a higher target temperature. That is why you feel chills at the onset of a fever: your body is trying to generate heat to reach a set point that has been pushed upward by immune signaling. Aspirin-like drugs work against fever because they lower that set point back to normal.19PubMed. Fever versus hyperthermia
In hyperthermia, the set point has not changed. Instead, heat production overwhelms the body’s ability to get rid of heat, or the cooling mechanisms have failed. Heatstroke is the classic example. Aspirin does nothing for hyperthermia because the thermostat is not the problem. External cooling is the only effective treatment.20DeckerMed Medicine. Hyperthermia, Fever, and Fever of Undetermined Origin Exercise can push core temperature well above 100 °F in a healthy person. In one study of recreational runners completing a 10 km treadmill run in warm conditions, average core temperature reached about 103.3 °F (39.6 °C) by the end of the run.21PubMed Central. Predicting the body core temperature of recreational athletes at the end of a 10 km self-paced run under environmental heat stress That is a reading that would signal serious illness in someone lying in bed, but it is a normal consequence of vigorous exercise in the heat.
Medications That Shift Your Baseline
Certain drugs can nudge your temperature up or down in ways that have nothing to do with infection. A systematic review and meta-analysis examined the effects of common medications on core temperature during heat stress. Drugs with strong anticholinergic properties, which reduce sweating, raised core temperature by about 0.8 °F (0.42 °C) at air temperatures above 86 °F. Non-selective beta-blockers, adrenaline, and anti-Parkinson’s medications also produced small but measurable increases.22PubMed Central. The effect of prescription and over-the-counter medications on core temperature in adults during heat stress: a systematic review and meta-analysis – Section: Results For people taking these drugs during a heat wave, even a small shift in baseline heat dissipation can tip the balance toward dangerous overheating.
Over-the-counter painkillers can work in the other direction, at least at night. NSAIDs like ibuprofen blunted the normal overnight temperature dip in one study, flattening the natural nighttime cooling that accompanies sleep.23Physiology & Behavior. Nonsteroidal anti-inflammatory drugs alter body temperature and suppress melatonin in humans The clinical relevance of this for most people is probably small, but it is a reminder that common pills interact with the thermoregulatory system in ways we do not always think about.
Brown Fat and Heat Production at Rest
Your body generates heat even when you are not moving, and part of that comes from an unusual tissue called brown fat. Unlike regular fat, which stores energy, brown fat burns energy specifically to produce heat, a process called non-shivering thermogenesis. Brown fat is most abundant in newborns, but adults retain functional deposits, particularly around the upper back and neck. Calculations based on available data suggest that brown fat thermogenesis accounts for roughly 5% of resting metabolic rate.24PubMed. Implications of nonshivering thermogenesis for energy balance regulation in humans That might sound modest, but it is thought to contribute meaningfully to keeping core temperature stable, especially during cold exposure. There is active research into whether activating brown fat could be useful for treating obesity, since it burns calories to generate heat rather than storing them.25PubMed Central. Non-shivering Thermogenesis Signalling Regulation and Potential Therapeutic Applications of Brown Adipose Tissue
Wearable Sensors and Continuous Temperature Tracking
Consumer wearables that track skin temperature overnight are increasingly common, and they are generating a different kind of temperature data than a one-time thermometer reading. Devices like smart rings and wristbands measure peripheral skin temperature continuously, then use algorithms to estimate core body temperature. A systematic review of wearable sensor technology found that most prediction algorithms achieved high accuracy against gold-standard core measurements, with 17 out of 18 tested algorithms meeting clinical validity standards. However, the review noted that few of these algorithms accounted for individual and environmental factors known to affect core temperature, which limits their reliability in real-world conditions.26PubMed Central. Wearable Sensor Technology to Predict Core Body Temperature: A Systematic Review
What wearables do well is track trends over time for the same person. Because your normal baseline might differ from someone else’s by a full degree, a device that learns your personal pattern and flags deviations from it can be more informative than a single spot check against a population average. Some fertility-tracking wearables already use this approach, detecting the post-ovulatory temperature rise that would be invisible in a single reading but stands out clearly when plotted across weeks. The same principle applies to early illness detection: your temperature might climb from your personal baseline of 97.4 °F to 98.8 °F, a shift of nearly a degree and a half that a traditional threshold would call “not a fever” but that, for you, represents a meaningful change.
The gap between what these devices can do and what they reliably deliver to the average user is still real, though. Skin temperature is not core temperature, ambient conditions affect readings, and the algorithms underlying the estimates vary from product to product. For now, continuous wearable data is best treated as a supplement to, not a replacement for, a thermometer reading when you actually feel sick.