PG isn’t exactly certain how, but there has to be a book in this.
From The Economist
Dogs greet other dogs nose-first, as it were—sniffing each other from fore to (especially) aft. People are not quite so open about the process of sniffing each other out. But the size of the perfume industry suggests scent is important in human relations, too. There is also evidence that human beings can infer kinship, deduce emotional states and even detect disease via the sense of smell. Now, Inbal Ravreby, Kobi Snitz and Noam Sobel of the Weizmann Institute of Science, in Israel, have gone a step further. They think they have shown, admittedly in a fairly small sample of individuals, that friends actually smell alike. They have also shown that this is probably the case from the get-go, with people picking friends at least partly on the basis of body odour, rather than the body odours of people who become friends subsequently converging.
As they report in Science Advances, Dr Ravreby, Dr Snitz and Dr Sobel started their research by testing the odours of 20 pairs of established, non-romantic, same-sex friends. They did this using an electronic nose (e-nose) and also two groups of specially recruited human “smellers”.
The e-nose employed a set of metal-oxide gas sensors to assess t-shirts worn by participants. One group of human smellers were given pairs of these shirts and asked to rate how similar they smelt. Those in the other group were asked to rate the odours of individual t-shirts on five subjective dimensions: pleasantness, intensity, sexual attractiveness, competence and warmth. The e-nose results and the opinions of the second group of smellers were then subjected to a bit of multidimensional mathematical jiggery-pokery (think plotting the results on a graph, except that the graph paper has five dimensions), and they, too, emerged as simple, comparable numbers.
All three approaches yielded the same result. The t-shirts of friends smelt more similar to each other than the t-shirts of strangers. Friends, in other words, do indeed smell alike. But why?
To cast light on whether friendship causes similarity of scent, or similarity of scent causes friendship, Dr Ravreby, Dr Snitz and Dr Sobel then investigated whether e-nose measurements could be used to predict positive interactions between strangers—the sort of “clicking” that is often the basis of a new friendship. To this end they gathered another 17 volunteers, gave them t-shirts to wear to collect their body odours, ran those odours past the e-nose, and then asked the participants to play a game.
This game involved silently mirroring another individual’s hand movements. Participants were paired up at random and their reactions recorded. After each interaction, participants demonstrated how close they felt to their fellow gamer by overlapping two circles (one representing themselves, the other their partner) on a screen. The more similar the two electronic smell signatures were, the greater the overlap. Participants also rated the quality of their interaction in the game along 12 subjective dimensions of feelings that define friendship. Similar odours corresponded to positive ratings for nine of these dimensions. Intriguingly, however, two participants smelling alike did not mean they were any more accurate at the mirroring game than others, as measured by a hidden camera.
Why scent might play a role in forming friendships remains obscure. Other qualities correlated with being friends, including age, appearance, education, religion and race, are either immediately obvious or rapidly become so. But while some individuals have strong and noticeable body odour, many—at least since the use of soap has become widespread—do not. It is present. But it is subliminal. Dr Ravreby speculates that there may be “an evolutionary advantage in having friends that are genetically similar to us”.
Link to the rest at The Economist