Saturday, November 24, 2012

The Smell Test

People discover new species. They discover new stars. They discover new elements.

They also discover new smells, apparently. That's literally what LiveScience has found itself leading with for a headline. 'New Smell Discovered'. I guess it makes sense- otherwise you couldn't have seemingly random celebrities coming out with their own fragrances- but it is just weird to see it written so starkly.

A new study out of Israel's Weizmann Institute of Science (it's available at that link as a PDF) has spurred that headline after finding a smell they're calling 'olfactory white'. What is it made up of? Wrong question. It doesn't actually matter all that much what goes into the smell. The important thing is that there are a lot of things going into it. After you get up to 30 components or so, everything just starts smelling the same. It's white noise for the nose, hence the name olfactory white.

How they went about finding this was, the people at Weizmann took a group of 86 different molecules from across the olfactory spectrum, calibrated to be of equal intensity. (We're not talking 'coffee' or 'orange' or 'grass clippings'. We're talking 'chlorothymol' and 'diethyl sulfide' and 'hexaonic acid'.) They were thrown into a variety of mixtures containing anywhere from 1-43 components. (43 being half of the 86 used.) They then took these mixtures and provided test subjects with them under the made-up name 'Laurax'. After allowing them a few days to get acquainted with their own personal Laurax, they were then given a new set of four mixtures of varying numbers of components and asked which one was Laurax.

None of them was actually the original mixture. In fact, the lion's share of the new mixtures had not even a single component in common with the first mixtures. (A little bit of overlap was inevitable, as the four options had to be different from each other too.) What ended up happening was that the more components something had, the more likely it was to be called Laurax by the subjects. (There was one caveat: the components within each mixture should themselves be more spread across the olfactory spectrum if they wanted to be called Laurax.)

The original setup had the subjects label the four scents 'Laurax', 'Compound 2', 'Compound 3', and 'Compound 4'. They redid it to include an 'other' option so people weren't just calling things Laurax because the researchers were making them do it. (This would also be known as 'the correct answer'.) Didn't matter. Laurax was still being used on the high-component mixtures. Maybe it had something to do with the molecules used-- namely, maybe there weren't enough of them. So the researchers upped it from a pool of 86 molecules to a pool of 144 molecules. That didn't make a difference either. The testing was altered to being given 21 compounds, one at a time and none with any overlap, and being asked 'is this Laurax, yes or no'. If they had enough components, yes they were.

In the 144-component pool yes-or-no test, subjects were decent at distinguishing one smell from the other when there were 15 components per mixture. At 60 components, they might as well have flipped a coin.

So what does olfactory white smell like? As part of the testing, subjects were asked to describe each odor they were given. The researchers were given 146 different descriptors of mixtures that were later determined to qualify as olfactory white. The 20 most common are shown below, sorted by percentage of applicability.

72.9%- Fragrant
66.1- Chemical
63.6- Perfumery
62.3- Aromatic
62.1- Floral
60.7- Soapy
56.8- Sweet
56.2- Fruity, citrus
56.1- Medicinal
53.7- Fruity, other than citrus
51.3- Disinfectant, carbolic
49.8- Cleaning fluid
48.9- Cool, cooling
48.4- Light
45.9- Lemon
45.5- Musky
44.2- Incense
43.5- Alcoholic
43.2- Rose
41.6- Varnish

And for fun, the least common descriptors:

4.8- Sewer odor
4.7- Animal
4.3- Sooty
3.8- Beany
2.9- Burnt paper
2.9- Fishy
2.8- Popcorn
2.5- Kippery (smoked fish)
2.0- Fried chicken
1.9- Meaty (cooked, good)

You will also be pleased to know that only 6.2% described olfactory white as 'fecal'. That can be covered right up by the 'cat urine' smelled 6.7% of the time.

No comments: