An interesting twitter conversation got me thinking about how to "photograph" smells. So much of our experience and our exchange of information is visual, but smell resists visualization. Unlike the cartoons, we don't have smell-o-vision and we don't have stink lines and clouds of green smoke around smelly things. How can we turn smells into a form that's easier to store as data, not in a glass vial but in a computer?

One of the biggest challenges is that not only are smells hard to visualize, they are hard to describe--try describing the smell of strawberries without saying "smells like strawberries." Are there primary smells the way there are primary colors, reference smells that can fully describe a new odor, base compounds that can be mixed to form any other smell? Since Aristotle, scientists and perfumers have been trying to split up smells into these fundamental pieces, not just "smells good" or "smells bad" or "smells like strawberries" but categories like Aristotle's list of sweet, sharp, pungent, bitter, and oily. Perfumers today split up odors into a larger group of core odors, from citrus to woody:

These descriptions rely on the noses of professionally trained perfumers, picking apart mixtures of many chemicals and using them to describe and create new fragrances. Human "wetware" noses can identify many different chemicals, but have a hard distinguishing mixtures with more than a handful of individual compounds. How can perfumers use technological hardware and software to identify all of the individual odor molecules that can make up a smell? The tools of chemistry are used to capture and pull apart smells into their individual components, identifying each molecule as a peak on a gas chromatographer. This is technology my Synthetic Aesthetics partner Sissel Tolaas uses to "X-ray the air" for her cityscape projects. The video below from Design Indaba follows Sissel as she travels through Cape Town, South Africa to identify smells that she will recreate in a city guide. She shows her "camera for smells" starting at about 2:50.

When the samples are sent to the lab and analyzed, the smell becomes a series of peaks on a graph, each peak giving the identify and relative concentration of a molecule. These peaks can be retranslated back into chemicals, recreating the original smell.

A much more abstract kind of smell photography was presented at the 83rd Annual International Exhibition of the Royal Photographic Society in 1938. The first photographs of a smell were taken by H. Devaux, showing the action of volatile molecules on a surface of liquid mercury:

"The emission of an odor involves volatillisation of material. If an odiferous material is enclosed in a cell a few millimeters above a clean mercury surface, it is possible to collect on the surface of the mercury a monomolecular layer of the volatillising or odoriferous substance. If the mercury surface initially is covered with talc powder, the gradual formation of the monomolecules layer may be observed as the talc is gradually pushed away from the point immediately below the specimen of material."

The images below are of the smell of lily and camphor, and although they don't convey a lot of information about the odor, they are a beautiful way of representing the physical nature of smells. Every time we breath in we take in a pieces of the objects we're smelling, the volatile molecules in the air interacting with our olfactory machinery to give us a picture of the chemical world around us.

(smell photographs via John Ptak; thanks to Natalie B. for the link!)