Normal human body temperature averages somewhere around 97.5°F to 98.2°F (36.4°C to 36.8°C) in most modern studies, not the 98.6°F (37.0°C) that has been repeated in textbooks for more than 150 years. That famous number traces back to a single researcher in the 1860s, and both his methods and our bodies have changed since then. What we now know is that “normal” is a range, shaped by time of day, age, sex, hormonal status, and even which thermometer you use and where you point it.
Where 98.6°F Actually Came From
The 98.6°F standard originates with Carl Reinhold August Wunderlich, a German physician who published a landmark paper in 1868. Wunderlich collected temperature readings from a large number of patients and concluded that the average axillary (armpit) temperature was 37.0°C, with the upper limit of normal at 38.0°C.1Oxford University Press. Normal Body Temperature: A Systematic Review His work was a genuine breakthrough at the time, linking measured temperature to clinical diagnosis for the first time. But there were problems lurking in his data that took over a century to fully appreciate.
For one, Wunderlich measured axillary temperature, which tends to read lower than oral temperature. The thermometers available in the 1860s were also slow and potentially poorly calibrated by modern standards. When a team led by Philip Mackowiak revisited the question in 1992 with modern instruments and oral measurements, they found the mean temperature was 36.8°C (98.2°F), not 37.0°C, and the upper limit of normal was 37.7°C (99.9°F) rather than 38.0°C (100.4°F).2PubMed. A critical appraisal of 98.6 degrees F, the upper limit of the normal body temperature, and other legacies of Carl Reinhold August Wunderlich That half-degree discrepancy between Wunderlich’s number and modern measurements was the first major crack in the 98.6°F dogma. A large-scale analysis of over 35 million temperature readings confirmed the point, finding that the mean was below 37.0°C during all times of day, all days of the week, and all seasons of the year.3PubMed. The daily, weekly, and seasonal cycles of body temperature analyzed at large scale
Human Bodies Have Been Getting Cooler
Measurement error and outdated instruments explain part of the gap between Wunderlich’s era and ours, but there is growing evidence that human body temperatures have genuinely dropped over the past two centuries. A study examining three large cohorts of Americans born between the early 1800s and the late 1990s found a steady decline of roughly 0.03°C per birth decade in both men and women.4eLife. Decreasing human body temperature in the United States since the Industrial Revolution Over the full 197-year span, men’s temperatures dropped by about 0.59°C total, and women’s by about 0.32°C. This wasn’t an artifact of changing measurement methods; the decline persisted even within individual cohorts measured the same way over time.
The trend isn’t limited to industrialized countries. A study of the Tsimane, an indigenous population in Bolivia with limited access to modern medicine, found the same downward trend over a much shorter period.5PubMed Central. Rapidly declining body temperature in a tropical human population The leading explanation is that reduced chronic infection and inflammation, better nutrition, and improved living conditions have lowered the metabolic fires that once kept body temperatures higher. In the 1800s, most people were fighting off untreated infections like tuberculosis and gum disease at any given time. Chronic low-grade inflammation pushes body temperature up. As those burdens lightened, temperatures drifted down.
The Daily Temperature Swing
Even within a single day, your temperature is not a fixed number. It follows a circadian rhythm with an oscillation of about 0.8°C to 1°C between the nighttime low and the daytime high.6PubMed Central. Sleep and 24 hour body temperatures: a comparison in young men, naturally cycling women and women taking hormonal contraceptives Temperatures typically bottom out in the early morning hours, around 4:00 to 5:00 a.m., and peak in the late afternoon or early evening. If you take your temperature right after waking, you might see something like 97.3°F. By dinnertime the same day, you could read 99.0°F, and neither reading would be abnormal.
This rhythm is driven by a cluster of nerve cells in the hypothalamus called the preoptic area, which acts as a thermostat. These neurons are sensitive to changes in core temperature and also receive information from temperature sensors in the skin and spinal cord, combining internal and external signals to coordinate the body’s heating and cooling responses.7PubMed. Role of the preoptic-anterior hypothalamus in thermoregulation and fever The same brain region is also tightly linked to sleep regulation, which is one reason your temperature drops as you fall asleep.8PubMed Central. Role of the Preoptic Area in Sleep and Thermoregulation When it senses rising temperature, it triggers a negative feedback loop involving specific signaling molecules that brings the temperature back down.9Cell. Thermoregulation via Temperature-Dependent PGD2 Production in Mouse Preoptic Area
Beyond the internal clock, ambient conditions also play a role. In controlled experiments, core temperature climbed higher on days when the surrounding air temperature was gradually increased throughout the day, compared with days when room temperature was held steady or decreased.10PubMed. Effect of change in ambient temperature on core temperature during the daytime The effect was modest, on the order of a few tenths of a degree, but it means your reading on a sweltering summer day may genuinely differ from one taken in a cool, air-conditioned room. That said, the circadian rhythm typically dominates: in women studied under varying ambient conditions, the diurnal pattern held as long as the surrounding temperature wasn’t continuously rising.11PubMed. Effect of change in ambient temperature on core temperature of female subjects during the daytime and its sex differences
How Age, Sex, and Hormones Shift the Baseline
If you menstruate, your body temperature follows a second rhythm layered on top of the daily one. Core temperature is about 0.3°C to 0.7°C higher during the luteal phase (the roughly two weeks after ovulation) compared with the follicular phase before it.12PubMed Central. Temperature regulation in women: Effects of the menstrual cycle This shift, driven by progesterone, is most pronounced first thing in the morning before any activity, which is why tracking basal body temperature has long been used as a retrospective marker of ovulation. The relationship between progesterone levels and temperature is roughly linear at lower hormone concentrations but plateaus at higher levels, meaning temperature doesn’t just keep climbing as progesterone rises.13PubMed. Descriptive analysis of the relationship between progesterone and basal body temperature across the menstrual cycle
Estrogen complicates the picture further. Research comparing different phases of the menstrual cycle with women taking oral contraceptives found that estrogen modifies how progesterone affects temperature. Women in the natural luteal phase (high progesterone, moderate estrogen) had higher core temperatures than women in the follicular phase, but women taking a combined estrogen-plus-progesterone contraceptive did not show the same elevation, suggesting estrogen can blunt progesterone’s warming effect.14PubMed. Estrogen modifies the temperature effects of progesterone
At the other end of the age spectrum, older adults tend to run cooler than younger people and are less able to mount an effective fever when infected.15PubMed Central. Altered Febrile Responses in Older Adults: A Systematic Review This is a clinically important problem. In elderly patients, fever may be absent or blunted roughly 20% to 30% of the time during an infection.16PubMed. Fever in the elderly Without that warning sign, serious infections can go undetected until they are harder to treat. Anyone caring for an older relative should know that waiting for a reading of 100.4°F to sound the alarm may mean waiting too long.
Where You Put the Thermometer Changes the Number
Your body doesn’t have a single temperature. Your core, deep inside the chest and abdomen, is warmer than your mouth, which is warmer than your armpit, which is warmer than your skin surface. The thermometer site matters more than most people realize. A systematic review of studies with strong evidence found the following ranges of normal temperature across measurement sites: oral readings spanned roughly 33.2°C to 38.2°C, rectal readings 34.4°C to 37.8°C, tympanic (ear) readings 35.4°C to 37.8°C, and axillary readings 35.5°C to 37.0°C.17PubMed. Normal oral, rectal, tympanic and axillary body temperature in adult men and women: a systematic literature review The ranges also differed between men and women, with women showing slightly wider oral ranges.
In a study of children with Kawasaki disease, mean oral temperature was about 0.25°C higher than axillary temperature, and mean rectal temperature was about 0.43°C higher than axillary.18PubMed Central. Axillary, Oral, and Rectal Routes of Temperature Measurement During Treatment of Acute Kawasaki Disease These differences are small in isolation but can matter when you are trying to decide if a temperature of, say, 99.5°F should worry you. Under the arm, that might be slightly elevated. In the mouth, it is likely normal. For clinical and research purposes, rectal temperature is generally considered the closest non-invasive proxy for true core temperature. Ingestible telemetric sensors, small capsules that transmit readings from inside the gastrointestinal tract, provide an even more accurate ambulatory measure and have been validated in athletes and military personnel, though they are obviously not something you would use at home.19PubMed Central. The ingestible telemetric body core temperature sensor: a review of validity and exercise applications
Why Forehead Thermometers Often Miss Fevers
Non-contact infrared thermometers (the gun-style devices pointed at the forehead) became ubiquitous during COVID-19 screening. Their appeal is obvious: fast, no physical contact, no discomfort. The problem is that they are not very good at detecting fever. A study in an African outpatient clinic found that a contactless thermometer had a sensitivity of just 13% for fever, meaning it missed the vast majority of truly febrile patients while giving reassuringly normal readings.20PubMed Central. Are all thermometers equal? A study of three infrared thermometers to detect fever in an African outpatient clinic A study in a tropical inpatient setting found sensitivity ranging from about 37% to 41%, still failing to catch more than half of fevers.21Heliyon. Accuracy and acceptance of thermometers in triage and inpatients in a low-resource tropical setting – The MaTe study That study concluded that non-contact infrared thermometers are “an inappropriate choice of device for fever recognition” in settings without air conditioning.
A meta-analysis pooling data from multiple studies estimated a somewhat higher sensitivity, around 81%, for forehead-based infrared devices.22Journal of Travel Medicine. Diagnostic accuracy of non-contact infrared thermometers and thermal scanners: a systematic review and meta-analysis But that pooled figure masks enormous variation by setting. Forehead readings are heavily influenced by ambient temperature, wind, sweat, and how recently the person was outdoors. In a climate-controlled hospital lobby, the devices perform reasonably well. At an outdoor checkpoint in a humid tropical city, they are close to useless for catching all but the most extreme fevers. If your concern is whether you or a child genuinely has a fever, an oral or ear thermometer will give you a much more reliable answer than a quick forehead scan.
Rethinking What Counts as a Fever
The conventional fever threshold of 38.0°C (100.4°F) was inherited from Wunderlich, and it assumes everyone starts from the same baseline. That assumption is wrong. A pilot study of cancer patients found that the average baseline body temperature was 36.7°C, meaning the standard 38.0°C cutoff was more than four standard deviations above average.23PubMed Central. Personalized Cutoffs for the Diagnosis of Neutropenic Fever Based on Patients’ Baseline Body Temperature: A Retrospective Pilot Study In practical terms, this means that a patient who normally runs cool at, say, 97.0°F might already be mounting a significant immune response at 99.0°F, well below the official cutoff. That patient’s fever could be invisible to a nurse looking only at whether the thermometer crosses the 100.4°F line.
This is more than a theoretical concern. Research has argued that using 37.0°C (98.6°F) as the assumed normal for everyone can cause healthcare workers to miss serious fevers in people with low baselines, and produces false alarms in people who naturally run a little warm.24PubMed Central. One size does not fit all: Assuming the same normal body temperature for everyone is not justified The problem is especially acute for immunocompromised patients, whose suppressed immune systems may already dampen the normal inflammatory signals that produce fever.25PubMed. Sepsis in the severely immunocompromised patient For these populations, some clinicians advocate personalized fever thresholds based on a patient’s own measured baseline, rather than a one-size-fits-all number.
Even if you are healthy, it is worth knowing your own pattern. Take your temperature a few times when you feel well, at different times of day, and note the range. If your healthy afternoon reading is consistently 97.8°F, you will know that 99.5°F is meaningful for you personally, even though it would not technically qualify as a fever by the traditional definition.
Why Mammals Evolved to Run This Hot
Given that maintaining a high body temperature is metabolically expensive, burning through food at a rate that cold-blooded animals can avoid, why did mammals evolve to keep their core temperatures near 37°C in the first place? One compelling hypothesis centers on fungal defense. Research has shown that the number of fungal species capable of infecting an organism drops by about 6% for every 1°C increase in body temperature.26PubMed. Fungal virulence, vertebrate endothermy, and dinosaur extinction: is there a connection? Cold-blooded animals like reptiles and amphibians are susceptible to tens of thousands of fungal species. Mammals, running hot, are vulnerable to only a few hundred.
This thermal barrier against fungi may have been especially important in the aftermath of mass extinctions, when fungal blooms thrived on decaying organic matter. Small, warm-bodied mammals that could resist those fungi had a survival advantage. Modeling suggests that 36.7°C represents a roughly optimal trade-off between the metabolic cost of running the furnace and the immune benefit of thermal protection against fungal pathogens. Going hotter would burn more calories for diminishing returns; going cooler would open the door to an exponentially growing number of potential fungal invaders. It is an elegant example of evolution settling on a compromise, one that we inherited and now measure with a thermometer under the tongue without thinking about why the number is what it is.
Body Weight and Temperature
A question that comes up often is whether body size affects temperature. The relationship is surprisingly muddled. A large Swiss study found that in men and postmenopausal women, body mass index was positively associated with body temperature, along with waist circumference, resting heart rate, and markers of glucose metabolism.27International Journal of Obesity. Association of body temperature with obesity. The CoLaus study In other words, heavier people in that study tended to run slightly warmer. But a separate study using continuous core temperature monitoring found no meaningful difference between the daily temperature profiles of obese and non-obese subjects.28PubMed Central. Core body temperature in obesity The discrepancy may come down to what was being measured: a single spot reading versus a full 24-hour core temperature profile. Larger bodies may produce slightly higher peripheral readings due to greater insulation, without their actual core temperature differing. For most people, body weight is not going to meaningfully shift your reading on a home thermometer in a way that changes any clinical decision.