This is a remarkable book, but not a very readable one. Judith Flanders seems to have read every novel, newspaper story, play (the Lord Chamberlain's office kept a record), broadside and treatise that pertains, however glancingly, to her chosen subject - and that includes work she describes as "almost unreadable today". The problem is that she wants to share all this with the reader. At times, The Invention of Murder resembles an extremely detailed list of every violent crime that took place during Victoria's reign (and not only then, because the book reaches back into the 18th century).

Flanders starts by considering the appeal of murder - "very pleasant to think about in the abstract" - but this is all the analysis we get. By the start of the second paragraph, we are being told about the murder of Timothy Marr, his family and his apprentice in the East End in 1811. The author recounts this and dozens of other murders, as well as the attention they received in the press and, when applicable, the arts, the relevance of class, the means of detection, the court case and the testimonies of great men of the day (such as Dickens and Thackeray).

The book's subtitle, How the Victorians Revelled in Death and Detection and Created Modern Crime, suggests a narrative or even thesis rather than a survey; but those 12 words are made to do a lot of work. It's as if the subtitle doesn't announce the shape of the book, but, in fact, has been assigned the task of shaping.

In moments of desperation, the reader may use it as a mental weapon with which to tame the book's amorphousness.

Chapters in a book dealing with one broad subject may be used to separate chunks of material artificially, but with the exception of "Panic" (about the mid-century fear of poisoners) and "Middle-Class Poisoners", the chapters here have a porousness that limits their ability to aid the reader's memory and comprehension. Flanders is able to note that the plot of Wilkie Collins's novel The Woman in White seems more complex than it is because of the "overlapping narratives" without identifying the greater damage such overlap might do to a compendious work of history.

Murder attracted so much attention in the 19th century because journalists and novelists were alert to its juicy dramatic possibilities, but whenever Flanders has an opportunity to give a chapter some propulsion she squanders it. During her account of the murder at Road Hill House (the subject of Kate Summerscale's The Suspicions of Mr Whicher), she observes that "Many novelists turned to the shattered Kent family" before giving us a paragraph each on works by Mrs Oliphant, Charlotte M Yonge and Mrs Braddon. And this is ever Flanders's way. The closing sentence of her opening paragraph - an ideal place to get the reader in your grip - is a statistic-heavy parenthesis. Instead of delivering a knockout conclusion to the strange fact that murder was virtually non-existent at the beginning of the 19th century, she recites the murder figures for 2007-2008 for Canada, the EU, Moscow and Cape Town.

At the other end of the book, in the opening paragraph of her "Modernity" chapter, Flanders follows the enticing thought that "Jack the Ripper brought with him a new kind of crime, and a new kind of fear", with the story of Israel Lipski, the murderer "popularly described as being the precursor to Jack the Ripper". Lipski was a Polish-Jewish immigrant; he arrived in London in 1885; and so on, for eight pages.

By the time we get to the murder of Martha Tabram, the Ripper's first victim, that introductory frisson has been crushed by the weight of background detail.

Flanders's ability to connect things is astonishing, and she has made all kinds of marvellous discoveries (some of them reported casually, in footnotes), but again and again across this vast book, she proves a more dutiful servant of her research than her readership.