Dr Hare and his colleagues took 43 men and 51 women, all of them straight, and gave them two tasks.
One was to decide whether an androgynous computer-generated face was, on balance, more likely to be female or male.
The other was to rate members of the opposite sex shown in photographs for both their sexual attractiveness and their likelihood of being unfaithful.
The participants completed both tasks twice, on consecutive days.
On one day they were exposed to the appropriate molecule (AND for the women; EST for the men) and on the other to a placebo that ought to have had no effect.
Crucially, the study was double-blinded, which meant that neither the researchers nor the participants knew which day was which.
This should have made it impossible for unconscious biases on the part of the experimenters or the subjects to have had any effect on the result.
If AND and EST really are aphrodisiac pheromones, the researchers reasoned, then they ought to make participants more likely to assume that androgynous faces belonged to the opposite sex.
They should also boost the sex appeal of the people in the photographs—and, because of that boost, increase the perception that those people might be unfaithful,
since the attractive have more opportunities for infidelity than the plain.
In fact, they did none of these things. The study thus found no evidence that either AND or EST is a pheromone.
Those who buy pheromone perfumes based on them would therefore appear to be wasting their money.
Whether the triumph of hope over experience will cause them to carry on doing so anyway is a different question altogether.