Tom Williamson, then the curator of palaeontology at New Mexico Museum of Natural History and Science, was part of the dig team. He decided to test the hypothesis that the crest could be used to make sound. He took the specimen to a local hospital so it could be imaged with a CT scanner – a machine shaped like a big donut, in which a human patient is passed through a central hole while being bombarded with X-rays. In this case, the machine scanned the P tubicen skull in 3mm slices, and stacked the visual layers on top of one another to create a three-dimensional representation of the creature’s crest.

With the help of Carl Diegert, a computer scientist, Williamson simulated the frequencies of noise that P tubicen could have produced. The men had to infer the location and structure of several missing parts, such as the soft tissues of the nose and throat. But eventually they settled on a note of around 30Hz, a rich bass at the bottom of the range of human hearing, with some unexpected peaks of intensity caused by the internal intricacy of the creature’s airways. ‘You get these weird booming sounds,’ Williamson told me. ‘It sent chills up and down my spine.’ This noise is now thought to have helped P tubicen form complex social groups, via its enhanced capacity to communicate with and identify other members of its species, including males and females.