Matt Wolfe writing for New Republic:

“Ever since the old practice of marking convicts by mutilating their ears or branding their skin had been abolished, police had lacked a reliable system to determine if a person had a criminal record. Some larger departments assembled photographic rogues’ galleries, but photos were imprecise: People could look alike, or alter their appearance. And matching a suspect to his rap sheet might require leafing, one by one, through hundreds or thousands of pictures. Police chiefs began offering cash bonuses to any officer who successfully recognized a felon.

A solution was engineered in 1881 by Alphonse Bertillon, a sickly, 26-year-old record keeper for the Parisian police. The black sheep in a family of social science luminaries—his father had co-founded the School of Anthropology in Paris—Bertillon worked in the basement at police headquarters, tasked with transcribing physical descriptions of criminals, most of them incomplete or ambiguous. Frustrated, he sought to apply to his work some of his family’s scientific precision. Using a pair of calipers, he studied the physical attributes of prison inmates and found eleven measures unlikely to alter with age or a change in weight, such as sitting height, arm span, and the length and breadth of the head. He concluded that, while two people might share one measurement, the odds of sharing eleven were infinitesimally small. These figures—human flesh rendered into numbers—were placed on a single index card. This system, known as Bertillonage, was quickly put into wide use. An officer could now locate a suspect’s file in minutes, verifying his identity and hitching him, eternally, to his criminal record.”

Statistics and true crime. What’s not to like?