A Short History of Forensic Science

Song Ci (1186-1249 CE) was a judicial official in the Southern Song dynasty who oversaw criminal investigations for decades. His text, Xi Yuan Ji Lu (Washing Away of Wrongs, roughly), compiled from his own observations and from earlier sources, was completed around 1248. It is the earliest known systematic text on forensic medicine.

Song Ci described how to distinguish a drowning death from a death by strangulation that was then placed in water, how to identify the marks left by a rope or hands, and how to use the activity of insects on a body to estimate time of death. He described a technique for identifying an implement used in a killing by laying candidate tools in the sun and watching which one attracted flies first, exploiting the residual blood and tissue that insects would detect. The technique is rudimentary by modern standards. The principle, that biological evidence associated with a body or weapon can be read systematically, is not.

In the Islamic world, the physician and polymath Ibn Qayyim al-Jawziyya wrote in the 14th century on the legal status of medical evidence in courts, addressing how injuries should be described and when medical opinion could support or contradict a witness account. In Europe, medieval courts relied on sworn testimony and trial by ordeal rather than physical examination, and formal forensic medicine was slow to develop. The divergence between Eastern and Western practice in this period is worth remembering: the history of forensic science is not a European story alone.

In 19th-century Europe, arsenic was nearly undetectable in a corpse, dissolved readily in food and drink, produced symptoms resembling cholera, and was sold openly as rat poison. It was consequently used for murder with some regularity. The problem was not that people suspected poisoning; it was that no one could prove it. The stomach contents of a corpse showed nothing. Illness and poison looked identical.

Mathieu Joseph Bonaventure Orfila (1787-1853) changed this. A Spanish-born chemist working in Paris, he published Traite des Poisons in 1813, the first systematic account of the chemistry of poisons, their action on the body, and their detection. Orfila's contribution was not just the chemistry. He established that poisons could be extracted from human tissue after death and detected by chemical means, which was the foundational move that made toxicological evidence possible in court. His testimony in the trial of Marie Lafarge in 1840, where he demonstrated the presence of arsenic in her husband's stomach contents, set a template for how scientific expert witnesses should operate: method stated, results reproducible, conclusions qualified by what the analysis could and could not show.

The 19th century saw rapid urbanisation and mobile populations, and with them a policing problem: a criminal could serve a sentence, leave the city, return under a new name, and be treated as a first offender. Photographs existed but did not help unless someone had seen the person before. People changed their appearance. Names were invented. The criminal justice systems of Europe and North America had no systematic tool for identifying repeat offenders.

Alphonse Bertillon (1853-1914), a records clerk for the Paris Prefecture of Police, proposed a solution in 1879 and was allowed to test it from 1882. His system, anthropometry, rested on the observation that adult bone dimensions, once fixed, do not change, and that the probability of two unrelated adults sharing the same measurements across eleven standard sites (including head length, head width, middle finger length, foot length, and height sitting) was vanishingly small. Each arrested person was measured, recorded, and filed. On the next arrest, the new measurements were compared to the file, and a match identified the person regardless of the name they gave.

The system was adopted by police forces across Europe and North America through the 1880s and 1890s. It was the first time that science, applied systematically, solved the problem of criminal identity at scale. Its eventual replacement by fingerprinting was not because Bertillon's measurements were inaccurate but because fingerprints were easier to take, easier to file, and easier to recover from scenes without the person present. Bertillon also developed portrait parle, a standardised verbal description of facial features that became the template for police-description protocols.

The observation that the skin ridges on human fingertips form unique patterns goes back at least to 1684, when the English physician Nehemiah Grew published a description of fingerprint skin ridges in the Philosophical Transactions of the Royal Society. In 1823, the Czech physiologist Jan Purkinje described nine categories of fingerprint patterns. But observation and utility are different things: no one had built a filing and retrieval system that made fingerprints useful for police work.

Francis Galton (1822-1911) took the critical step. Building on the work of William Herschel, who had used fingerprints for contract authentication in Bengal from the 1850s, and Henry Faulds, who had written to Nature in 1880 proposing that fingerprints could be used to identify criminals, Galton collected over 8,000 fingerprint specimens. In 1892 he published Finger Prints, which established statistically that no two people share the same print, provided a classification system for filing, and demonstrated the permanence of ridge patterns across a lifetime. Galton's estimate of the probability of a coincidental fingerprint match (1 in 64 billion) was based on a simplification, but it was the first time the claim of uniqueness had been attached to a number.

Edward Henry, Inspector-General of Police in Bengal, used Galton's system as a base to build an operational classification system in the 1890s. The Henry classification system, published in 1900 and adopted by Scotland Yard in 1901, divided prints into five pattern types (plain arch, tented arch, radial loop, ulnar loop, whorl) and assigned numerical values to allow rapid filing and retrieval. It remained the dominant manual system for decades and is the direct ancestor of modern automated fingerprint identification systems.

Edmond Locard (1877-1966) opened the first dedicated police forensic laboratory in Lyon, France, in 1910, with two attic rooms and two assistants. His exchange principle, the idea that every contact between two surfaces transfers material in both directions, gave forensic science a unifying theoretical framework that applied to every trace type. Locard's practical cases included identifying a forger from traces of face powder on a typewriter, and demonstrating that earth on a suspect's shoes matched the soil at a murder site. The laboratory as a necessary fixture of serious criminal investigation is largely Locard's legacy.

Blood-group evidence entered forensic practice after Karl Landsteiner's 1901 discovery of the ABO blood group system. Leone Lattes (1887-1954), an Italian scientist, developed in 1915 a method for determining the ABO group of dried bloodstains, enabling blood on a weapon, garment, or surface to be grouped and compared. This was the first time blood evidence could be used for inclusion and exclusion in criminal cases, though it remained a class characteristic.

Alec Jeffreys, working at the University of Leicester, developed the technique he called DNA fingerprinting in September 1984 while studying genetic variation in families. Within months, he recognised that the same technique could identify individuals in a criminal investigation. The method identified variable-number tandem repeats, regions of DNA where a short sequence repeats a number of times that varies between individuals. The resulting pattern is, for practical purposes, unique to each person.

The first criminal application came in 1986, when Jeffreys was asked to test whether a young man who had confessed to two murders of girls in Leicestershire, England, was the real killer. His DNA did not match the biological evidence from either crime. He was exonerated, one of the first uses of DNA evidence to exclude rather than indict. The actual perpetrator, Colin Pitchfork, was identified when police asked 5,000 local men to provide voluntary DNA samples, making it the first mass DNA screening in criminal history. The Enderby murders case established that DNA could work in both directions, exoneration and identification, and that it was reliable enough to anchor a prosecution.

The polymerase chain reaction (PCR), developed by Kary Mullis in 1983 and applied to forensic samples by the early 1990s, allowed DNA profiles to be generated from tiny amounts of biological material, including single hairs, microscopic bloodstains, and touch DNA left by skin cells. This transformed what physical evidence could be analysed. National DNA databases, starting with the UK National DNA Database in 1995, created the infrastructure to compare crime-scene profiles against populations of known offenders. The distance from Bertillon's filing cabinets of body measurements to a searchable national genomic database represents the full arc of forensic identification in little over a century.