National Geographic News
A member of the jury smells a French 'baguette' bread during its consideration as part of the 'Best baguette of Paris' contest, on March 13, 2014, in Paris.

Our sense of smell is much better than previously thought, a new study says.


Jane J. Lee

National Geographic

Published March 21, 2014

Our noses are better at distinguishing smells than we ever knew. A lot better.

In fact, we may be able to detect as many as a trillion different odors, according to a new study published this week. That’s orders of magnitude more than earlier estimates of nasal intelligence.

The study marks the first time that the human sense of smell has been put through a rigorous scientific test, says Leslie Vosshall, a researcher at Rockefeller University in New York City.

A 1920s study concluded that people could smell roughly 10,000 different odors, but that estimate wasn't backed by data, says Vosshall, co-author of the study published March 20 in the journal Science.

Still, that decades-old figure remained unchanged until now, likely giving rise to a scientific under-appreciation of our nose's capabilities. (See "We Can Distinguish Between At Least A Trillion Smells.")

National Geographic spoke via email to study co-author Andreas Keller, a researcher in Vosshall's laboratory at Rockefeller University, about the new research, whether there’s a universal “good” smell, and if scientists might find evidence for "supersmellers."

When did this idea that humans have a poor sense of smell first pop up?

I blame Plato, who wrote in Timaeus—[his account of the formation of the universe]—that vision is the greatest gift given to us by the gods and the basis of philosophy, whereas smell is a half-formed thing about which not much can be said.

What about the idea that we can distinguish 10,000 odors? Where’d that come from?

This number is based on theoretical work from the early 20th century. [It] was based on the assumptions that (a) there are four elementary odor qualities—fragrant, acidic, caproic [or sweaty], and burnt—and (b) that we can discriminate around ten different intensities of each of these qualities, 10x10x10x10=10,000. Unfortunately, both assumptions were wrong.

Did it surprise you to find that humans can smell a trillion odor mixtures?

Not at all. From my experience working with odors, I knew that it is very unusual to find two mixtures of odors that cannot be discriminated. And because the number of possible mixtures is astronomically high, it always seemed to me that 10,000 was much too small a number. I suspect that most other people working with smells would have agreed with me.

You tested whether people would be able to distinguish one mixture of odors from another. Is there a difference between being able to smell those mixtures as opposed to single odors?

There is a condition called specific anosmia, or odor-blindness, in which people with an otherwise OK sense of smell cannot perceive a specific type of odor, like musk. To them, some single odors that have a smell to others are odorless. For mixtures, one will not find this effect because even if people cannot smell some of the components in a mixture, they will be able to smell [the other odors].

Is there a scent that's universally considered a good smell? What about a universal bad one?

I published a paper in which a few hundred people rated how much they liked around 60 different odors. I think vanillin was the favorite and isovaleric acid [a compound that gives smelly cheese its pungent odor] the least favorite.

However, all the subjects were from the New York area, so that doesn't count as universal. But from an evolutionary standpoint, fruit odor would be a good candidate for a universally pleasant odor and the odor of decaying flesh a good candidate for an unpleasant one.

Are there "supersmellers," just like there are supertasters—people with an extremely good ability to distinguish flavors?

There probably are. There is enormous variability in an individual's olfactory abilities. Much of it is probably due to genetic variability in the [odor] receptors.

However, there are different ways to measure how good a person's sense of smell is—sensitivity, the ability to discriminate similar smells, the ability to identify components in mixtures—and it is not clear which one would have to be unusual to qualify as a "supersmeller."

Where will our noses lead your research next?

Like many others in our field, I want to better understand what makes odors smell similar or different. So this is the question my research is now focusing on. I'm especially interested in pairs of mixtures that share none of their components, yet smell similar.

This Q&A has been edited and condensed. Follow Jane J. Lee on Twitter.

rafael molina navas
rafael molina navas

Is there any "home lab" to practice and improved our sense of smell?

Not for billions of mixtures, of course. Just for a certain amount of most frequent odors.

Jay Wells
Jay Wells

It seems to me that every time a so called "scientist" these days thinks about writing an article they tell themselves, "The first thing to do is dream up a really impressively huge number so every reader will marvel at how smart I am."

That stuff makes me marvel at how arrogant they are and what a condescending opinion they have of us mere mortals.

Swiftright Right
Swiftright Right

Well from reading the snippet of the study that lays on this side of the pay wall It looks to me that the 1 trillion estimate (and it is an estimate) is only slight more rigorous then the 10,000 estimate. 

I also have a hard time swallowing such a large number. A little perspective. 

If you started smelling 1 odor per second it would only take you ah, 6 times longer the the ENTIRE history of the human race to work through tho odor combinations. Or stated another way it would take just under 31,689 years to smell through the combinations. From reading on other sites this estimate is based on taking 128 of the most distinct smells and testing at what point volunteers could not discriminate a difference and then assuming that number also compared to every other stimuli.  So basically they take smells we have evolved a sensitivity to and apply those numbers to say, crude oil derivatives. 

Ishani C.
Ishani C.

Wow! Thats really fascinating! 


Popular Stories

  • Lost City Found in Honduras

    Lost City Found in Honduras

    A joint Honduran-American expedition has confirmed the presence of extensive pre-Columbian ruins in Mosquitia in eastern Honduras, a region rumored to contain ruins of a lost "White City" or "City of the Monkey God."

  • Astronomers Find a Galaxy That Shouldn't Exist

    Astronomers Find a Galaxy That Shouldn't Exist

    Small, young galaxies should be free of interstellar dust, but an object called A1689-zD1 is breaking all the rules.

  • Cool Polar Bear Pictures

    Cool Polar Bear Pictures

    Take a peek at polar bears playing, swimming, and sleeping in their changing habitat.

The Future of Food

  • Why Food Matters

    Why Food Matters

    How do we feed nine billion people by 2050, and how do we do so sustainably?

  • Download: Free iPad App

    Download: Free iPad App

    We've made our magazine's best stories about the future of food available in a free iPad app.

See more food news, photos, and videos »