How many different scents can your nose pick up? A trillion, according to a highly publicized study published last year in Science magazine.
Before that, conventional wisdom was that people could only discern around 10,000 smells. The study made headlines everywhere, and was nominated for "Breakthrough of the Year" by the prestigious journal.
But that breakthrough is probably fiction, according to two studies released Tuesday in the online journal eLife. One, led by Richard Gerkin of Arizona State University, suggests that people might only discern some 5,000 smells. The other, by neuroscientist Markus Meister of the California Institute of Technology, says the math model behind the one trillion estimate rests on unlikely assumptions about how we smell.
"It's not so much they made a math mistake as they used the wrong mathematics," Meister told BuzzFeed News.
The original study created a wide variety of smells by mixing 128 different odor molecules — up to 30 per batch. The researchers then gave 26 volunteers blind smell tests to see how well they could distinguish one smell from the next.
The volunteers could discriminate between two smells so long as their molecular make-up overlapped less than 51%. From that, the team calculated the total number of smells available to human perception as at least a trillion.
But Meister argues that this approach is meaningless. That's because nobody knows how the brain puts together signals from the nose's 400 or so smell "receptor" cells to create our perception of smell. So the brain might perceive two different molecular mixtures as identical scents, even if it can tell they are different from others.
If the authors of the Science study had used the same methods for vision, it would suggest we could see 1,000 trillion trillion — that's 10^27 — different colors, Meister added, which is "dramatically at odds with what we know."
What's more, Meister suggested that everyone's nose is likely different, given the way our brains tend to fine-tune themselves in response to stimuli throughout life. A wine taster and a football coach, for example, likely have very different senses of smell. "We need to figure out how smell works first before we make this estimate," Meister said.
The scientists behind the original study, headed by Leslie Vosshall of Rockefeller University in New York, wrote a response to these critiques this week.
They agreed that their model hinges on the assumption that molecular signals from nasal cells combine to create many smells. "We should have stated this explicitly in our paper," they wrote.
But that is as much as they are willing to concede. "We stand by our estimate," Marcelo Magnasco of Rockefeller University told BuzzFeed News by email. Most smell researchers believe the assumption that different molecular combinations produce distinct smell perceptions, Magnasco added.
"Methodological errors" probably do dog the original Science study, Noam Sobel of the Weizmann Institute of Science in Israel, told BuzzFeed News by email. Sobel reviewed the original paper and is cited in the critiques. But he still thinks it's likely the nose can distinguish among a trillion smells, even if it hasn't been proven.
"My scientific intuition (and this is my field) tells me that the bottom line of the Science paper is generally correct, yet there may have been some errors in how it was reached."
Sobel also pointed out that even if the study is flawed, it's likely not the result of any kind of misconduct. The study authors had provided their data for other researchers to check, Sobel said, "and may have simply made a mistake along the way. We all make them all the time."
Meister agreed that their shouldn't be any "stigma" in research over errors or disagreements. But he did express frustration that high-profile journals tend to favor studies that break with conventional wisdom, risking blow-ups when they don't pan out.
"Stuff happens and sometimes things slip through the cracks," he said. "This one trillion smells paper has been cited all the time and appeared headed into the textbooks. That seems worth correcting."