The traditional benchmark of 98.6 °F (37.0 °C) is outdated. A landmark 1992 study of healthy adults found that the true mean oral temperature is closer to 98.2 °F (36.8 °C), and more recent data suggest it may have drifted even lower since then. Beyond the shifting average, the “normal” range turns out to be surprisingly wide and shaped by dozens of variables, from the time of day you check to where on your body you measure.
Where 98.6 Came From and Why It Stuck
The number traces back to Carl Reinhold August Wunderlich, a German physician who published roughly one million armpit temperature readings in the 1860s. He declared 37.0 °C (98.6 °F) the human norm, and the figure became embedded in medical teaching for more than a century. In 1992, researchers at the University of Maryland revisited his conclusions with modern oral thermometers and found that the actual mean in their subjects was 36.8 °C (98.2 °F), not 37.0 °C. They also found that the upper limit of normal was about 99.9 °F (37.7 °C) rather than the 100.4 °F (38.0 °C) that Wunderlich had proposed. Their recommendation was blunt: 98.6 °F “should be abandoned as a concept relevant to clinical thermometry.”1PubMed. A critical appraisal of 98.6 degrees F, the upper limit of the normal body temperature, and other legacies of Carl Reinhold August Wunderlich
Despite that call, the number persists in popular culture, on fever-check instructions shipped with thermometers, and even in some clinical guidelines. Understanding why it is wrong helps you interpret your own readings more accurately.
The Modern Normal Range
A 2019 systematic review pooled data from dozens of studies to pin down normal ranges by measurement site. For oral temperature, the range spanning most healthy adults was roughly 35.7 °C to 37.4 °C (96.3 °F to 99.3 °F). Rectal readings ran higher, from about 36.3 °C to 37.8 °C. Armpit (axillary) readings ran lower, from about 35.0 °C to 36.9 °C. Tympanic (ear) and urine-based measurements fell in between.2PubMed Central. Normal Body Temperature: A Systematic Review The key takeaway is that a single threshold applied to every person and every measurement site is too blunt a tool. Your “normal” depends heavily on where and when the reading is taken.
Why Human Temperatures Are Still Dropping
Even setting aside Wunderlich’s 19th-century data, there is evidence that average body temperature has been falling for more than a century. A study analyzing records from Union Army veterans through to modern adults found a steady decline of roughly 0.03 °C per decade in both men and women born between the early 1800s and the late 1990s. Over the full span, men’s temperatures dropped by about 0.59 °C and women’s by about 0.32 °C.3PubMed Central. Decreasing human body temperature in the United States since the Industrial Revolution A separate study among the Tsimane, an indigenous population in Bolivia, found a similar decline over a much shorter time frame, suggesting this is not a quirk of industrialized societies alone.4PubMed Central. Rapidly declining body temperature in a tropical human population
The leading explanation is that reduced chronic infection and inflammation have lowered the baseline metabolic “heat load” humans carry. Better sanitation, antibiotics, and vaccines mean fewer people walk around with low-grade infections that silently elevate temperature. Other proposed contributors include more stable indoor environments and changes in physical activity levels. Whatever the mix of causes, it means that if you consistently read 97.5 °F or so, you are not abnormally cold. You may simply be tracking the modern average.
Your Temperature Fluctuates Throughout the Day
Body temperature follows a roughly 24-hour rhythm. It bottoms out in the early morning, around 6 AM, and peaks in the late afternoon or early evening, around 4 to 6 PM. The swing between those extremes averages about 0.5 °C (roughly 0.9 °F), though it can be larger in some people.1PubMed. A critical appraisal of 98.6 degrees F, the upper limit of the normal body temperature, and other legacies of Carl Reinhold August Wunderlich This means a reading of 99.3 °F at 5 PM might be perfectly normal, while the same reading at 6 AM would be worth paying attention to.
Sleep amplifies this pattern. Core temperature drops with every transition into deep sleep, falling by about 0.2 °C each time. It rises slightly during dream sleep, then drops again in the next deep-sleep cycle.5PubMed. Body temperature and sleep That overnight dip is why morning readings tend to be lower and why clinicians often prefer an afternoon measurement if they want to catch an early fever.
How Age Affects the Range
Older adults tend to run cooler. A systematic review confirmed that healthy elderly people have lower average temperatures than younger adults across almost every measurement site.2PubMed Central. Normal Body Temperature: A Systematic Review The decrease is thought to reflect a slowing metabolic rate and a diminished ability to adjust to environmental temperature swings. One prospective study found that older people’s oral temperatures tracked more closely with the ambient room temperature than younger people’s did, especially those aged 85 and above.6PubMed. Normal body temperature and the effects of age, sex, ambient temperature and body mass index on normal oral temperature: a prospective, comparative study
This has a real clinical consequence. If your baseline runs at 97.2 °F and you spike to 99.5 °F, that represents a bigger jump than it would for someone whose baseline is 98.4 °F. Yet 99.5 °F falls below the traditional fever threshold of 100.4 °F, so your fever could be missed. Clinicians working with elderly patients increasingly treat a smaller rise from that individual’s baseline as clinically meaningful rather than waiting for a universal cutoff to be crossed.
Differences Between Women and Men
Women tend to have slightly higher resting temperatures than men. The 1992 Maryland study confirmed this, and numerous studies since have agreed.1PubMed. A critical appraisal of 98.6 degrees F, the upper limit of the normal body temperature, and other legacies of Carl Reinhold August Wunderlich Part of that gap is driven by the menstrual cycle. After ovulation, rising progesterone pushes core body temperature up by roughly 0.3 °C to 0.7 °C compared with the first half of the cycle.7PubMed Central. Temperature regulation in women: Effects of the menstrual cycle During sleep, one study measured this luteal-phase increase at about 0.27 °C, accompanied by a rise in energy expenditure of roughly 7 percent.8PubMed Central. Changes in sleeping energy metabolism and thermoregulation during menstrual cycle
The mechanism involves reduced blood flow to the skin. During the luteal phase, skin thermal conductance drops, meaning less heat is shed at the surface and more is retained in the core.9PubMed. Decreased thermal conductance during the luteal phase of the menstrual cycle in women This is why many women report feeling warmer in the second half of their cycle. Fertility awareness methods exploit this predictable shift, treating the sustained temperature rise as a marker that ovulation has occurred.
Where You Measure Matters More Than You Think
Rectal temperature is generally treated as the clinical reference for core body temperature. Every other site is an approximation, and the approximation can be surprisingly loose. In one emergency-department study, the average difference between rectal and oral readings was about 1.1 °F, but individual pairs could disagree by nearly 3 °F in either direction. Ear (tympanic) readings showed even wider scatter, with a positive predictive value for detecting fever of only 55 percent compared with 97 percent for oral thermometers.10PubMed Central. Oral and Tympanic Membrane Temperatures Are Inaccurate to Identify Fever in Emergency Department Adults
Axillary readings run lower still. In a study of children with Kawasaki disease, armpit temperatures were on average 0.43 °C below rectal and 0.25 °C below oral readings.11PubMed Central. Axillary, Oral, and Rectal Routes of Temperature Measurement During Treatment of Acute Kawasaki Disease Forehead and temporal artery thermometers, popular for their speed and convenience, carry similar caveats. A clinical commentary in the Annals of Internal Medicine stated plainly that peripheral thermometers do not have clinically acceptable accuracy for measuring core body temperature.12PubMed. Peripheral thermometers do not have clinically acceptable accuracy for measuring core body temperature This does not mean your forehead scanner is useless for home screening, but it does mean you should not treat a “normal” reading from it as proof that everything is fine if you feel genuinely unwell.
Exercise, Food, and Other Everyday Shifts
Physical activity raises core temperature. This is simply the heat generated by working muscles. During prolonged exercise in hot conditions, core temperature can climb well above resting levels, and the main influence on that rise is exercise intensity rather than air temperature.13PubMed. Exercise under heat stress: thermoregulation, hydration, performance implications, and mitigation strategies If you take your temperature within an hour of a run or a gym session, expect it to be elevated.
Eating has a subtler effect. Digesting food increases metabolic heat production, a phenomenon sometimes called the thermic effect of food. A mixed meal at normal calorie balance raises energy expenditure by roughly 5 to 15 percent.14PubMed Central. Diet induced thermogenesis Animal research has traced part of this postprandial warming to a chain involving the hormone leptin, adrenaline from the adrenal glands, and activation of brown fat.15PubMed Central. Leptin mediates postprandial increases in body temperature through hypothalamus-adrenal medulla-adipose tissue crosstalk In humans, women appear to show a larger skin-temperature increase after a meal than men do.16PubMed. Skin temperature response to a liquid meal intake is different in men than in women
Even what you drink right before a measurement can throw it off. Drinking a hot beverage caused an immediate average oral temperature spike of 2.6 °F in one study, enough to make seven out of ten subjects look feverish. Cold drinks pushed readings down, and some subjects had not returned to baseline even 30 minutes later.17PubMed. The effect of hot beverages, cold beverages, and chewing gum on oral temperature The standard advice to wait at least 15 to 30 minutes after eating or drinking before taking an oral temperature exists for good reason.
Stress Can Raise Your Temperature Without Infection
Psychological stress activates the sympathetic nervous system and can push core temperature up, sometimes markedly. In healthy people, acute stress produces a small, transient increase that stays within the normal range.18PubMed. Stress-induced hyperthermia and infection-induced fever: two of a kind? In certain clinical populations, the effect is more dramatic. Some individuals develop temperatures as high as 41 °C (about 106 °F) during acute emotional events, while others show a persistent low-grade elevation between 37 °C and 38 °C during or after chronic stress.19PubMed. Stress-induced hyperthermia and hypothermia
This stress-induced hyperthermia works through different pathways than an infection-driven fever. Infectious fever depends on inflammatory signaling molecules, while stress-driven temperature spikes rely on the sympathetic nervous system and heat production from brown fat tissue.20PubMed Central. Psychogenic fever: how psychological stress affects body temperature in the clinical population Standard antipyretics like ibuprofen do not bring down a psychogenic fever the way they would an infectious one, which can be a diagnostic clue. If you find yourself running a low-grade temperature during a stressful stretch with no other signs of illness, this phenomenon could be the explanation.
Your Personal Baseline May Not Match the Population Average
One of the most underappreciated aspects of body temperature is how much it varies from person to person, even after accounting for time of day, measurement site, and all the factors above. A large-data study of patient records found that individual baseline temperatures had a 95 percent range spanning about 0.6 °C. That may sound small, but it means two healthy people measured under identical conditions can routinely differ by more than a full degree Fahrenheit.21BMJ. Individual differences in normal body temperature: longitudinal big data analysis of patient records
Another study using wearable sensors found that mean temperatures among participants ranged from 35.2 °C (95.4 °F) all the way up to 37.4 °C (99.3 °F). The overall average was 36.1 °C (97.0 °F), well below 37 °C. Fully 77 percent of participants had mean temperatures at least 0.55 °C (1 °F) lower than the old textbook benchmark.22PubMed Central. One size does not fit all: Assuming the same normal body temperature for everyone is not justified The practical implication is that knowing your own baseline is more useful than memorizing a universal number. A few readings taken at the same time of day over a week, when you feel healthy, gives you a personal reference point that makes future readings much easier to interpret.
When Does “Warm” Become “Fever”?
The body’s thermoregulatory system normally keeps core temperature within narrow limits despite large swings in the environment and in physical activity.23PubMed. Thermoregulatory disorders and illness related to heat and cold stress Fever happens when the brain’s thermostat is actively reset upward, typically by signaling molecules released during an immune response.24PubMed Central. The pathophysiological basis and consequences of fever Most clinical guidelines still define fever as a core temperature at or above 38.0 °C (100.4 °F), but the 1992 Maryland study suggested that 37.7 °C (99.9 °F) is a more defensible upper boundary for healthy adults measured orally, and that the threshold should be even lower for early-morning readings (37.2 °C, or about 98.9 °F).1PubMed. A critical appraisal of 98.6 degrees F, the upper limit of the normal body temperature, and other legacies of Carl Reinhold August Wunderlich
On the cold side, clinical hypothermia is defined as a core temperature below 35 °C (95 °F).25PubMed. Physiological Impact of Hypothermia: The Good, the Bad, and the Ugly Between that floor and the fever ceiling lies the full band of “normal,” and as the evidence above shows, where any individual sits within that band depends on a combination of age, sex, time of day, recent activity, and personal biology.
Medications That Shift the Thermostat
Several classes of prescription drugs can nudge core temperature up or down, which is worth knowing if you take daily medication and are trying to interpret a thermometer reading. Under heat-stress conditions, drugs with strong anticholinergic properties (such as atropine) raised core temperature by an average of about 0.42 °C, partly by reducing sweating. Non-selective beta-blockers produced a smaller but measurable increase of about 0.11 °C, and anti-Parkinson’s agents raised temperature by about 0.13 °C. Antidepressants and diuretics, by contrast, did not significantly alter core temperature in the studies reviewed.26PubMed Central. The effect of prescription and over-the-counter medications on core temperature in adults during heat stress: a systematic review and meta-analysis These effects are modest in mild weather but can become clinically relevant during heat waves, particularly for older adults already at thermoregulatory disadvantage.
Wearable Sensors and the Future of Personal Temperature Tracking
Consumer wearable devices now claim to estimate core temperature continuously from the wrist or chest, which could in theory help people establish their personal baseline without the fuss of an oral thermometer. The technology is promising but imperfect. A validation study of one popular chest-worn sensor found that it measured core temperature reasonably well during passive heat exposure without a fan, but significantly underestimated temperature when an electric fan was blowing. The limits of agreement between the wearable and a rectal reference were on the order of plus or minus half a degree Celsius, meaning individual readings could miss the mark by a meaningful margin.27Building and Environment. Validity of the wearable CORE temperature sensor during 8-hour indoor heat exposure with and without electric fan use For spotting trends over days and weeks, wearables can be genuinely useful. For determining whether a specific reading qualifies as a fever, they are not yet reliable enough to replace a conventional thermometer at the measurement site you are comparing against clinical guidelines.