A Smell

On the third day, the titan arum’s spathe began to open outward, and the smell in the greenhouse became obvious. A four-day observation then intensified: before flowering, two air samples were taken each day; after the spathe opened, the sampling interval was shortened to about two hours. Again and again, the researchers removed air from near the spadix, while the inflorescence continued to emit a stench that could be perceived without first consulting a chart. Away from the names of the instruments, people would probably describe what was there simply as a smell. That description is not wrong, but it can easily lure the next sentence too far: if what is smelled is one thing, should the things that compose it also grow stronger or fainter together?

Some sulfur-containing compounds and aldehydes dropped sharply in the samples, then rose again later. Following the numbers downward, it might seem possible to say that the stench first weakened. But what this study measured was several volatiles in the air; it did not station people beside the flower to continuously rate, with each sampling, the intensity of the whole smell. The numerical decline therefore can only show that these components became less abundant; it cannot supply, on behalf of the nose at that moment, a missing stretch of sensation. In adjacent phases, the intensity of most alcohols and hydrocarbons was increasing. Nor do they give the article a convenient way out: an increase does not mean the whole stench grew stronger, still less that they took over the place left by sulfur-containing compounds and aldehydes. In research on odor mixtures, components can sometimes still be recognized, sometimes mask one another, and may also synthesize a new odor quality that is absent when any single component is smelled alone. A concentration table does not hand over these perceptual relations along with the numbers. Nor is olfaction a ruler laid beside a test tube: doubling the concentration does not guarantee that the sensation will double as well; one substance may already be vivid at a very low concentration, while another, even if its value rises, may remain hidden in the mixture. Different kinds of substances have different thresholds, and after they meet one another they can also change how salient each other becomes. Even if one row of numbers falls and another rises, there is no arithmetic that can be completed without smelling.

This does not reduce “a smell” to a crude illusion. The nearby stench is a real experience, and the multiple changes in the chemical analysis are real measurements; the trouble begins only when one side is taken without warrant to continue writing the other. If two sets of values are forcibly translated into “first fainter, then stronger,” what results is not the scene itself, but a voice-over prepared for the scene. We can restate which components declined, rebounded, or increased, but we cannot use these values to reenact exactly how a given moment smelled. During sampling, greenhouse personnel had opened a small window at the base of the spathe for manual pollination. The opening of the window and the abrupt changes in volatiles occurred close together; the study did not confirm that the former caused the latter. The spathe opened around the raised spadix, and a heavy stench spread around the entire inflorescence.