The Drosophila Beat protein is related to adhesion proteins that contain immunoglobulin domains

recognizing faces presented from different viewpoints, and I came across a paper that had looked at the effects of pose on facial recognition. Laboratory studies had suggested that a three-quarter view leads to better recognition, but Logie et al. wanted to know if this generalized to the real world. In their first experiment they replicated some of the laboratory findings. So far so good. Their second experiment was to publish in a local newspaper, photographs of six different faces from different viewpoints, instruct readers that the six people would be walking around Cambridge at a certain time the following Saturday, and invite the readers to identify them and contact the experimenters by phone or by filling in a response form from the newspaper. Only one of the 100 000 readers responded. The experimenters then published a request for readers to say whether or not they had participated in the task. No responses were received. This was an important study for me. A well-respected group of researchers had carried out a wellplanned, somewhat extravagant experiment and it had turned to dust in their hands. If it could happen to them, it could happen to anyone. Undeterred, they carried out a third experiment in which volunteers were paid to find the targets after viewing photographs. This time the weather intervened: “Unfortunately, on the night prior to the experiment Cambridge received an unusually heavy fall of snow.” As a result, almost 25% of the paid volunteers didn’t take part, and only two reported accurate sightings. The experiment was failing for a good reason — the sheer difficulty of recognizing multiple targets from photographs — and the authors eventually solved this problem, but it is for the documentation of the failed experiments and the circumstances which led to failure that I remember this paper. It was the first lesson I received in how to be wrong, and whenever the first two stabs at an experiment don’t work out, it’s to this paper that I turn for reassurance.