Hard though it may be to imagine, the Observer has not always been a scrupulously factual newspaper intent on avoiding sensationalism. Back in 1824, it  or should I say "we"?  advertised an execution special at double the usual length to mark the hanging of John Thurtell, a former lieutenant in the marines, sometime fight promoter and notorious killer.

Thurtell had been found guilty of the murder of a gambler named William Weare, and the case had fired the public imagination. Melodramatic plays based on the story performed to sell-out audiences, and newspapers and specially produced "broadsides" displayed a macabre fixation with the grim details of the murder.

Little time was spent worrying about journalistic accuracy or laws on sub judice. Even in newspapers such as the Times, facts were ignored or made up. As Judith Flanders shows in The Invention of Murder, the current fascination with certain murders has a long historical precedence, and it goes back further than the Victorian era. As readers will have noted, 1824 was 13 years before Victoria came to the throne. The dates are significant for Flanders's argument: "Princess Victoria became Queen Victoria," she writes, "and public opinion began to change."

But how? Did it really amount to a new understanding of murder? This is a question with which this absorbing and well-researched book never quite gets to grips. Instead, Flanders piles one murder case on top of another to create a vivid picture of 19th-century Britain, with its extremes of wealth and poverty, its fledgling police forces and forensic science, and its popular public hangings.

The picture is enhanced by a wealth of cultural and literary references. For example, Flanders quotes from Herman Melville's diary, after the Moby-Dick author had witnessed the joint hanging of husband-and-wife murderers Maria and Frederick Manning in 1849: "Paid half a crown each for a stand on the roof of a house adjoining... Men & women fainting... The mob was brutish. All in all, a most  wonderful, horrible, and unspeakable scene."

Yet a coherent argument fails to emerge. Statistical analysis of murder in 19th-century Britain is fraught with difficulties, owing to missing data and vague definitions, and the problems increase in the pre-Victorian era. Flanders does cite occasional statistics, but they are isolated, spread out and potentially misleading.

At the start of the book, she quotes Thomas de Quincey, writing in 1826, on how enjoyable it is to read about murder  "for crime, especially murder, is very pleasant to think about in the abstract: it is like hearing blustery rain on the windowpane when sitting indoors". Flanders explains that at the beginning of the 19th century, it was easy to think of murder as an abstract concept unlikely to impinge on real life, because in England and Wales in 1810, "just 15 people were convicted of murder out of a population of nearly 10 million: 0.15 per 100,000 people". As Flanders notes, that compares favourably with current homicide rates, or it would do if conviction rates and homicide rates were the same thing. They're not. It might be that detection rates were very low in 1810  which would not be surprising, given that there were no real police forces or detectives.

It is another four chapters before we learn that in 1849 there were more than 20,000 deaths in England and Wales that were unexplained or suspicious. And yet earlier in the book, Flanders suggests that in seeking to institute his plan for a new police force in 1829, the home secretary Sir Robert Peel convinced the nation that there had been a rise in crime. "There probably was no such rise," writes Flanders. "There was a rise in prosecutions, the consequence of a change in social expectations, and a growing intolerance of disorder."

So, given these confusing figures, were the Victorians radically different in their approach to murder? Emotionally, there doesn't appear to have been any great change after 1837  some murders still captured the public's fancy and some did not  but there were distinctive political, legal and literary developments.

Chief among these, perhaps, was the introduction of Peel's police force, followed by the idea that the police should be concerned with detecting as well as preventing crime. The late Victorian period saw an industrial expansion in crime fiction, accompanied by a belief in the power of rational thought  perhaps the apotheosis of both trends was Sherlock Holmes.

For those interested in the real-life inspiration behind Victorian writers such as Wilkie Collins, this is an invaluable work of anecdotal and cultural history. But anyone drawn by the promise of the title should be prepared for a variation on Alan Bennett's celebrated line about the study of history. In this case, it's just one bloody murder after another.

"Scratch John Bull and you find the ancient Briton who revels in blood, who loves to dip deep into a murder, and devours the details of a hanging." So said the Pall Mall Gazette in 1887. Its immediate justification was the success of Robert Louis Stevenson's Strange Case of Dr Jekyll and Mr Hyde, which had been published the previous year and had already sold 40,000 copies. But it would be just as easy to prove the same point at any time during the last couple of centuries. And in our own time as well, as every bestseller list and TV schedule reminds us. Murder is as much a British preoccupation as football or the weather.

Especially 19th-century murder  both because it's the template for modern crimes and punishment, and because its stories exist at a distance that allows them to seem shocking but safe. Or better than safe: romantic. Poeticised by dripping shrubberies, curling mist and the echo of footsteps in gas-lit alleys. Judith Flanders certainly thinks so, anyway. The Invention of Murder is her third popular-scholarly book about the period (the other two being The Victorian House and Consuming Passions) and although there are times when her narrative staggers under the weight of detail, and others when her chapter-structures seem too indeterminate (is it really a good idea to have a heading as catch-all as "Violence" in a book on this subject?), she makes us feel its stories are peculiar but relevant. The book is full of gory details that are hard to forget, has an appropriate sense of overview, and is delivered in a tone that is at once seriously attentive and sympathetically appalled.

It has a handy way with facts, as well. Murder in pre-Victorian Britain was a pretty rare event. In 1810, when the population of England and Wales was almost 10 million, only 15 people were reported dead by this means  that's 0.15 per 100,000. (Compare that with 62 per 100,000, which were the figures for Cape Town in 2007-08.) Forty-odd years later the picture had changed somewhat  there were 20,000 unexplained or suspicious deaths. And so on: the figures rise as the century progresses, and alongside them grows the means of policing and detection. The Met, which was founded in 1829, was 3,000-strong to start with and 12,000-strong by 1886.

These are all significant increases, of course, and they help to explain the swelling fascination with murder during the Victorian era. Some of this fascination depended on simple human curiosity. But as Flanders shows, it also stemmed from genuine anxiety and from the efforts of a large supporting industry which stood to benefit from the original crime in various ways.

The case of John Thurtell helped to set the pattern. He was the son of a prosperous Norwich merchant who in 1823, after a short career in illegal gambling, bludgeoned to death a fellow-gambler called William Weare and dumped his body in a pond at Gill's Hill in Hertfordshire. The story was immediately turned into a sensation. Newspapers used it to crank up their circulations (in most cases placing it as the lead item in the several pages they gave over to crime reporting every day), melodramas were written for two London theatres, ghoulish tourists trekked out to the murder scene and paid for a mini-tour (the novelist Walter Scott was shown round by a "truculent-looking hag" as late as 1828), and balladeers did their worst, without any regard for the facts of the matter: "Then pulling out his murderous knife, / As over him he stood, / He cut his throat, and, tiger-like, / Did drink his reeking blood."

Forty thousand people turned out to see Thurtell executed, and the marketing of his story continued long after his death  as did his influence on what these days we would call the media. When Flanders turns to consider the later and more famous cases of Maria Marten and the red barn (where Marten was murdered by William Corder in 1827), and Burke and Hare, and Eugene Aram, we find a very similar pattern of revulsion, grisly relish, appropriation by other forms, then a sort of panting lull until the next shocker comes along. It gives her book an appropriately 19th-century serial feel, and provides a steady stream of cruelties and stupidities for our delectation, just as the press did for contemporary readers.

It also creates a sense of saturation, even of tedium, for which the banality of crime itself is mainly responsible. Tedium, that is, until we reach the most interesting transformations of "true crime"  those managed by writers who are more interested in human motive and social context than in making a quick buck. Thomas Hood's "The Dream of Eugene Aram", for instance, turns its hero into a tormented and repentant sinner by "ignoring entirely the mercenary element, making the criminal more important than the crime".

Several novelists (including Harrison Ainsworth in Rookwood) made a similar sort of attempt. None of them succeeded better than Dickens. Oliver Twist is quite rightly discussed here as a crime novel (to "twist" was underworld slang for to hang), but Flanders does well to trace Dickens's curiosity about the criminal mind, and about questions of detection and punishment, through other novels such as Our Mutual Friend, Great Expectations and The Mystery of Edwin Drood. These  like his public readings  all demonstrate the ways in which his imagination combined with the facts of contemporary murders, and justify the claim that the 19th century turned murder into a fine art. It is only one strand in Flanders's book, but is sufficiently interesting in itself (and in the way it combines with a similar incorporation of Wilkie Collins and Stevenson) to give vitality to the whole project.

So does the sense that all its stories lead to a late-century climax: Jack the Ripper. By the time we reach the account of his crimes, we have grown used to criminals being named and puzzles being resolved  which means that these particular unsolved murders retain a strangely shocking power. They make for an appropriately grave ending to the book, since they force us to confront the grim actuality of murder, and a sharp sense too of the sinister collusions it excites in the minds of those who imagine killing. In the mind of the journalist, for instance, who (probably) wrote the fake letter to Scotland Yard that pretended to come from Jack. "The next job I do I shall clip the ladys ears off and send to the police officers just for jolly wouldn't you. Keep this letter back till I do a bit more work . . . My knife's so nice and sharp I want to get to work right away if I get the chance. Good luck."

Andrew Motion's The Cinder Path is published by Faber.