Measuring the Spiciness of Chili Peppers
by Michael Lombardi, Metrologist - July 2012

Have you ever dined at a Chinese restaurant and accidentally eaten one of those little dried red chili peppers that are used to season the food? Those innocent looking peppers are native to the Chinese province Tien-Tsin, and thus are known as Tien-Tsin hot red chili peppers. If you mistakenly eat one, which I have on several occasions, you might “feel the burn” inside your mouth or lip for a few hours or more. If someone wonders why your face is flushed and your mouth is full of ice cubes, you can either mutter something unintelligible, or correctly state that you swallowed a pepper with a Scoville scale of about 60,000. Not hot enough to harm you, but certainly hot enough to make you uncomfortable for a little while.

The Scoville scale was created by the American pharmacist Wilbur Scoville (1865-1942). It isn’t a temperature scale, and it doesn’t actually measure heat. Instead it is a metric for the piquance of capsicums, better known as peppers. Piquance, better known as “spiciness,” is the taste sensation associated with spicy foods. The Scoville scale produces spiciness readings based on measurements of the amount of capsaicin, which is a chemical compound found in chili peppers. Capsaicin stimulates chemoreceptor nerve endings in the skin, and while it can make food taste better, in large doses it is an irritant to humans and other mammals. If fact, it is believed that peppers produce capsaicin as a defense mechanism against their mammalian predators. Pure capsaicin represents the top of the scale and produces a staggeringly high 16,000,000 Scoville heat units (SHUs).

Scoville developed his method for measuring capsaicin while working at the Park-Davis pharmaceutical company in Detroit. He published the method in a short paper entitled “Notes on Capsicums” (Journal of the American Pharmaceutical Association, vol. 1, issue 5, pp. 453-454, 1912). He noted that if a solution contained one part per million of capsaicin, it would “make itself known to the tongue.” He pointed out that “physiological tests were tabooed in some quarters,” but if the tongue was able to detect less than a millionth of grain of capsaicin, it would have an advantage over analytical balances, which were much less sensitive. He proposed a technique where one grain of a ground pepper was softened by soaking it overnight in 100 cc of alcohol. This alcoholic solution was then shaken and filtered, and added to sweetened water in definite proportions, until a “weak but distinct pungency is perceptible on the tongue.”


This test became known as the Scoville Organoleptic Test (organoleptic refers to the aspects of food that are experienced by the senses, in this case by taste). A Scoville test typically involves a group of at least five testers, each of whom tastes only one sample per session. The greatest weakness and criticism of the test, which Scoville had anticipated in his original paper, is the fact that it relies on human subjectivity.

Today, the human “taste test” methods have largely been replaced by measurements made with high-performance liquid chromatographs (HPLCs). These instruments are designed to separate a mixture of compounds, to identify the individual compounds in the mixture, and to measure their amounts. When used to measure spiciness, the HPLC extracts capsaicin from the mixture, and one part capsaicin per million equals one American Spice Trade Association (ASTA) pungency unit. Typically, one ASTA pungency unit is considered equal to about 15 SHUs, so you can multiply pungency units by 15 to convert to the Scoville scale. However, the conversion is approximate, and interlaboratory comparisons show that results can vary widely between tests. Uncertainty analysis is difficult, not only because of the different laboratory methods used, but also because the piquancy
of the “peppers under test” can vary by more than a factor of 10, even when the peppers are of the same species. Much of the variation is due to how and where the pepper was grown. Climate matters, as does soil conditions, and the lineage of the seeds. References to the Scoville scale are easy to find. Tables that list the approximate SHUs for peppers, sauces, and various food products, can be found in grocery stores and occasionally even on restaurant menus. It is interesting to compare the number of SHUs that common peppers produce. For example, sweet peppers or bell peppers contain no capsaicin at all, thus their Scoville rating is zero. Pepperoncini, which I sometimes pop like candy from a jar in our refrigerator, typically generate only a few hundred SHUs, and can be thought of as a very mild pepper. Some of us thrill seekers think we are living large when we pile green jalapenos on our nachos at the ballpark, but a jalapeno only generates about 2,500 to 5,000 to SHUs. Tobasco packs more punch, typically rated at about 30,000 to 50,000 SHUs. Consider that the next time you sprinkle a few drops.

If your mouth is made of asbestos and your stomach is made of steel, you might try munching on a habanero pepper. The habanero chile pepper is typically colored orange or red, and is grown in North, Central, and South America, with the largest producer being the Yucatan Peninsula of Mexico, where it is used in many food items. The habanero typically produces about 200,000 SHUs, and was once listed in the Guinness Book of World Records as the world’s hottest chili pepper.

In recent years, the Habanero has taken a backseat to other peppers that have been found to be much hotter. Bhut Jolokia, a hybrid pepper grown in northeastern India and Bangladesh, is now generally recognized as the world’s hottest. Seeds transported from India were used to grow Bhut Jolokia in a temperature controlled greenhouse in New Mexico. In 2007, Paul Bosland and Jit Baral of New Mexico State University published results that measured Bhut Jolokia at

1,001,304 SHUs using HPLC analysis. They concluded by noting that “it is indeed the world’s hottest known chili pepper” (P. W. Bosland and J. B. Baral, HortScience, vol. 42, no. 2, pp. 222-224, 2007), and the Guinness Book agreed with them in its next edition.