“The Grandest Game in the World”

John Dickson Carr‘s essay was first published in Ellery Queen’s Mystery Magazine in March 1963, and reprinted in full in the second edition of The Door to Doom and Other Detections, ed. Douglas G. Greene, International Polygonics, 1991. My thanks to Ronald Smyth for scanning the article.


“Do you solemnly swear never to conceal a vital clue from the reader?”

“I do.”

That is the first article in the oath taken by members of the Detection Club. The candidate, placing his hand upon Eric the Skull, swears this with fervency. He swears it with stern looks fixed on him. He swears it while Eric’s eyes (thanks to John Rhode) glow with red electric lights. He swears it even before he promises to honour the King’s English, use legitimate detective methods in his stories, and refrain from pinching his fellow members’ plots.

And this rule, the sine qua non of the profession, must be emphasized at the beginning to explain my choice of stories in The Ten Best Detective Novels.

For the once humble detective novel has come a long way. It has gone up hill, down dale, over the plain and through the sewer. In fifty years it has undergone so many changes, not to say disguises, that sometimes we quite literally don’t know what we are talking about. A new novel is praised because it is well written, because the characters are admirably drawn, because it is “tough”, because it is experimental in technique, because it is written sideways or upside down: on any grounds, in short, except that it is a good detective story.

If the term means anything at all, it means this:

The detective story is a conflict between criminal and detective in which the criminal, by means of some ingenious device – alibi, novel murder method, or what you like – remains unconvicted or even unsuspected until the detective reveals his identity by means of evidence which has also been conveyed to the reader.

That is the skeleton, the framework, the Christmas tree on which all the ornaments are hung. If the skeleton has been badly strung, or the tree clumsily set on its base, no amount of glittering ornament will save it. it falls over with a flop. Its fall may create a momentary sensation, especially among children; but adults are only depressed when they see the same sort of thing happen in fiction.

The author of the book hasn’t bothered. He has decided that good construction is of no consequence, or that nobody cares anyway. Far from planning in advance every move, every speech, every detail, he has roared ahead on inspiration and trusted to luck. And his attitude is understandable if he is writing a straight thriller, where rapid-fire action swallows up everything. But it becomes merely bad craftsmanship if he thinks he is writing a detective story.

We might postulate, to begin with, that the detective novel at its best will contain three qualities seldom found in the thriller. It will contain the quality of fair play in presenting the clues. It will contain the quality of sound plot construction. And it will contain the quality of ingenuity.

Ingenuity? Do we start an argument here?

It seems remarkable that this need for ingenuity in the outstanding detective novel has been so strangely overlooked. Perhaps the reason is that you cannot turn it into a “must”; you cannot lay it down as a rule of the game. You cannot say to an author, “Look here, sit down and be ingenious.” Maybe he can’t be. Maybe he doesn’t want to be. His interests may lie along other lines, such as the hero slugging the police or (more pleasant to read about) the police slugging the hero.

But though this quality of ingenuity is not necessary to the detective story as such, you will never find the great masterpiece without it. Ingenuity lifts the thing up; it is triumphant; it blazes, like a diabolical lightning flash, from beginning to end.

It is not of intrinsic interest to read that X has been stabbed to death in a hotel room, and that the police – after rewinding the clock, or studying the bloodstains, or any of the stock tricks in vogue since the time of Gaboriau – have proved the guilt of Y the waiter. This is all very well; it may be competent work; it will serve to be read if we have nothing better at hand. But in pitting our wits against the masters of the trade, we require something very different.

We require, for instance, the superb explanation of the clock alibi, in A.E.W. Mason’s The House of the Arrow. Or the means used to conceal the identity of the criminal, in Agatha Christie’s Murder in Mesopotamia. Or the reason why the corpse wore its clothes the wrong way round, in Ellery Queen’s The Chinese Orange Mystery. Or the ironic brilliance of Anthony Berkeley’s Trial and Error, in which a man who has confessed to the murder tries to prove himself guilty and can’t do it.

These writers (with others like them) are the aristocrats of the game, the old serpents, the gambit-devisers and trap-baiters whose strokes of ingenuity make the game worth playing at all.

For what, after all, is the game itself?

It is a hoodwinking contest, a duel between author and reader. “I dare you,” says the reader, “to produce a solution which I can’t anticipate.” “Right!” says the author, chuckling over the consciousness of some new and legitimate dirty trick concealed up his sleeve. And then they are at it – pull-devil, pull-murderer – with the reader alert for every dropped clue, every betraying speech, every contradiction that may mean guilt.

Nothing, in fact, shows more clearly the difference between the expert craftsman and the novice than his manner of presenting this evidence. The novice, even when he is anxious to include a clue, develops a case of acute self-consciousness about it. He feels naked before the reader’s eye. He is much too afraid of being caught with the goods. So he hurls the clue into the story and then runs like a maniac, as though he had thrown a bomb.

The result is that the clue, one or two words at most, will flash past and become lost among sixty or seventy thousand other words. This is painfully evident during the detective’s summing up in the final chapter.

“The whole question of Dagmar Doubledick’s guilt,” declares the detective, “turns on the kind of necktie he was wearing when we met him that day at Wemmerly Park. Of course you remember it was a green tie?”

To which the honest reader is compelled to answer: “No, I’m damned if I do!”

And then, if he is conscientious, he will turn back through the book to discover whether Dagmar Doubledick’s tie really was green. Perhaps he finds this clue, a violet by a mossy stone, half hidden somewhere in the dusky recesses of Chapter Six; perhaps he misses the page and does not find it at all. In either case he is left with a vague feeling of dissatisfaction: as though he has been, if not swindled, at least outtalked.

Now, it may be argued, and reasonably, that the author here was playing perfectly fair. He was not compelled to repeat it, or even stress it. Thus when the whole solution of Earl Derr Biggers’ Charlie Chan Carries On is based on the single word “stuffy”, or when Carolyn Wells in The Luminous Face argues guilt from the thesis that no gentleman would wear a wrist watch with evening clothes, these novels are at least technically within the rules.

But the masterpiece of detection is not constructed from “a” clue, or “a” circumstance, or one single inconsistency of any kind. Such methods, dubious enough in a short story, become grotesque when they are applied to a full-length novel. It is too reminiscent of those minute mysteries, vignettes accompanied by paralytic-looking photographs, with which we are so familiar in magazines.

“You stated, Leonard Andreas,” thunders the Inspector, “that you drank a Scotch-and-soda in the bar parlour of The Flaming Bishop at nine o’clock, whereas we know the pub ran out of spirits at half-past eight. It proves, Leonard Andreas, that you committed the murder.”

Now, this is a bit rough on poor old Leonard Andreas, because it doesn’t prove anything of the kind. It proves only that the witness told a lie, or that the landlord (as usual) was keeping his whisky under the counter for favoured customers. We are dealing, here, with murder; and we can hardly let a man’s life, even that of a character in fiction, depend on such flimsy evidence.

The fine detective story, be it repeated, does not consist of “a” clue. It is a ladder of clues, a pattern of evidence, joined together with such cunning that even the experienced reader may be deceived: until, in the blaze of the surprise ending, he suddenly sees the whole design.

Your craftsman knows, as Dr. R. Austin Freeman long ago pointed out, that it is not at all necessary to mislead the reader. Merely state your evidence, and the reader will mislead himself. Therefore, the craftsman will do more than mention his clues: he will stress them, dangle them like a watch in front of a baby, and turn them over lovingly in his hands. He will give not only the clue physical, but the clue psychological and the clue atmospheric.

No speech in the book is included just because it sounds mysterious, or because it makes a given character look guilty, or because the author doesn’t know what the devil his character does mean and simply throws in the words to fill up space. Not at all. In turning over the pages afterwards, the reader can see for himself – how rare it is! – just what each character was thinking at any moment.

And the result?

That is why the story pulses with vitality all the way through, and springs into living vividness at the end. The veil is twitched away; the masks are removed. Human beings walk here, and no sawdust dolls, because the author has described voice inflections, shades of feeling, as well as Inspector Hogarth’s discovery of the blunted thumbtack under the sofa. He has not forgotten to study his characters merely because he is writing about them in reverse. That turn of the eyes – of course! That momentary hesitation, when Betty puts her hand on the window ledge as though to steady herself – naturally!

Each small detail glitters now with an effectiveness it should have had, and would have had, if the story had been written straightforwardly. It is in the mood, in the tempo, an arrow whang in the gold. And when, in addition to this, we find ourselves flumdiddled by some master stroke of ingenuity which has turned our suspicions legitimately in the wrong direction, we can only salute the author and close the book with a kind of admiring curse.

There, good friends, is a detective story.


But who writes such stories nowadays?

In considering this question, on a terrain where it is to be feared that bricks are apt to fly, we might do worse than examine the wide difference which has developed nowadays between the British and the American type of detective novel.

During the good (or bad) old days twenty-five years ago – let’s speak first of the everyday mediocre practitioners rather than the great ones – these novels were of much the same kind. Both sides were content to write the English language, even when they wrote it badly. Both sides made some mumbling acquiescence in the matter of rules, even when they broke rules all over the place.

Their plots, too, were the same. Alter the locale from Long Island to Surrey, substitute “baronet” for “industrial magnate”, and the stories were almost interchangeable. This change, in fact, was actually made when the thrillers about Frank L. Packard’s Jimmie Dale were published in England, with the redoubtable Jimmie living in Park Lane and battling against an evil, conscienceless gang of robbers called (it is regrettable to state) the Crime Club.

But the patten of the average detective story ran thus: The victim, on the eve of making a new will, was found murdered in his library. He had been stabbed with an Oriental dagger, customarily used as a paper knife on his desk. The whole room was strewn with cuff links, bus tickets, lace handkerchiefs and cigarette ends, in the fine artistry of a paper chase.

Inspector Brace, summoned hastily to the scene of the crime, found only the beginning of his troubles. The baronet or industrial magnate – in addition to his ne’er-do-well son, his rebellious daughter and his invalid wife – was afflicted with such a household as nobody, even in the days of the servant shortage, would tolerate for five minutes. The butler was a blackmailer, the chauffeur an ex-convict, the housekeeper a religious maniac. If this were not enough, investigation discloses that no less than eight other suspects, at the time of the murder, were skulking in one long procession past the library windows.

“This situation,” says Inspector Brace, “is hopeless!”

And it is difficult not to agree with him, since the various cuff links and cigarette ends are proved to have been dropped innocently by one or other of the suspects, popping at intervals in and out of the windows like Box and Cox. Inspector Brace, desperate, is about to arrest the ne’er-do-well son when the latter’s fiancée calls in that gifted gentleman, the private detective Reginald Du Kink.

Then we get real business. It is Du Kink who discovers that the established time of the murder is all wrong, due to an effect of ventriloquism or a phonograph record of a voice, and at a dramatic gathering of suspects he fastens the guilt on the dead man’s secretary. The secretary, haggard and foaming, waits only to scream out a confession before he drinks off the contents of a small vial and instantly falls dead.

And that was that.

Now the above, so help me, is not written in ridicule. It is not meant as burlesque. You and I, who have been improving our minds with sensational fiction for so many years, are much too fond of detective stories. We are aware that all the above plot tricks were used long before 1920, have been used since, and are still in use today – often by the very best practitioners in the business.

Seldom are they lumped together in one story, as was formerly the case, nor is the clue so naïve as a broken cuff link. And the ghost of Dr. Freud haunts everything today. But the old elements remain. The millionaire’s home, the threatened disinheritance, the rebellious family, the enigmatic servant, the multiplicity of suspects, the wrongly accused, the wrong time of death – how many novels can you name in which not one of these elements is to be found?

Why, then, do we protest at the adventures of Inspector Brace and Reginald Du Kink? Why do their frenzied activities hover always on the edge of comedy, not to say broad farce?

We don’t find them funny because they are what our age likes to call “period pieces”. Far from it. One glance at a list of the detectives who were practising long before them, a list which includes short stories as well as novels, will convince us of that.

There is nothing in the least funny about the great stories of Sir Arthur Conan Doyle. Nobody smiles today at G.K. Chesterton’s Father Brown, though the stumpy little priest first appeared in 1911. The same applies to Inspector Hanaud, whom A.E.W. Mason introduced in At the Villa Rose a year earlier; and Dr. Freeman set the experienced John Thorndyke to solve his greatest problem, The Eye of Osiris, in the same year. E.C. Bentley, in 1913, was a comparative latecomer with his brilliant tour de force of Trent’s Last Case. On the other side of the Atlantic, an underrated genius named Jacques Futrelle had created Professor Augustus S.F.X. Van Dusen as early as 1907, whereas Melville Davisson Post was already an old craftsman when he gave us the classic book of short stories about the far-from-comic Uncle Abner in 1918.

And here we begin to see the explanation of why, as early as the 1920s, the intelligent reader was getting fed up with the adventures of Inspector Brace and Reginald Du Kink.

“Oh!” said the reader. “I’m tired of just guessing who the criminal is. instead of these sleight-of-hand half clues, so that it’s never properly explained at the end how the detective knew, let’s have some real evidence.

“Furthermore,” continued this reader, “it’s all very well to have your eight suspects parading in their endless ring-around-the-rosebush outside the library. That’s fine. But give some sensible reason why they were there. If you must shower the room with bus tickets, provide a reason for that too. In other words, construct your story. Your present problem is not to explain the villainy of the guilty; it’s to explain the stupidity of the innocent.

“Finally, your ‘amazing revelation’ at the end was so soggy, so lacking in essential cleverness, that I couldn’t care less. Haven’t you a new idea tucked away somewhere? Can’t you wield even a minor thunderbolt? It was far different, believe me, from that joyous shock when Father Brown unmasked the Invisible Man, or Uncle Abner showed the meaning of the Straw Man, or Sherlock Holmes, in an unforgettable moment, swept the disguise from the Man with the Twisted Lip.”

Holà! Wow!

Please pardon these exclamations. It is only that I, who write this introduction, feel warm with pleasure merely to recall, and taste in memory, those great moments of fictional crime. Once more, in memory, we see the gaunt figure of Holmes with the bath sponge in his hand, and shock-haired Hugh Boone writhing on the bunk. Or Father Brown, under a lurid sky in the waste of snow, with the giant hall porter between whose very feet runs the straggle of tracks where no man has passed; and out across the snow rings that despairing cry:

“God! The invisible man!”

Such moments, then, aid us in summing up the reasons why an imaginative reader required somebody more enterprising than Inspector Brace or Reginald Du Kink. He required a skilful story told in reverse by a skilful storyteller. He required (need it be repeated?) the quality of fair play, the quality of sound construction, and the quality of ingenuity. And already, at the beginning of the 1920s, this decade saw new writers who possessed just such qualities.

It saw the debut of Agatha Christie in The Mysterious Affair at Styles, based on the (then) startling novelty that the person first suspected turns out to be the murderer after all; he has wanted to get himself tried and acquitted so that he can’t be tried again – a device later used by so many other writers. It saw Freeman Wills Crofts, with The Cask and its grisly contents, inaugurating the new fashion of the Unbreakable Alibi.

It saw John Rhode, in The Paddington Mystery, present a victim dead from no apparent cause – while telling us for the first time (and almost the last time) that Dr. Priestley’s Christian name is Lancelot. It saw Anthony Berkeley’s initial effort take the form of a “locked room” in The Layton Court Mystery. It saw Dorothy L. Sayers – with Whose Body? – setting an unfledged Lord Peter Wimsey to solve the puzzle of a strange corpse, clad only in a pair of pince-nez, found stabbed in a dry bath.

These 1920s, whatever may be said against them, thronged with sheer brains. What would be one of the best possible settings for violent death? J.J. Connington found the answer, with Murder in the Maze. Has anybody ever used the camera obscura, that eerie periscope device, for witnessing sinister events at a distance? Mr. Connington again, with The Eye in the Museum. In the 1920s, too, Philip MacDonald made his notable advent with The Rasp. R.A.J. Walling, in Murder at the Keyhole, demonstrated how you can force a reader to look literally in the wrong direction. And always with us during those days, cherubic, dependable and moaning like an animated cream bun, was H.C. Bailey’s Mr. Fortune.

Now look towards the other side of the Atlantic. It must be acknowledged that America, during the same period, produced only two detective-story writers of the first class.

Regarding those who were not first class or anywhere near it, there is no need to mention names. Most of them were women, one or two of whom are still writing today. These ladies waltzed gracefully, waltzed well; but they waltzed always in the arms of Inspector Brace or Reginald Du Kink. We have pleasant memories of them all; theirs is the scent of arsenic and old lace. They call to mind coloured frontispieces from their own books: the yellow gowns sweeping the floor, the padded rooms cozy with crime.

But there is one name which must be mentioned, because it belongs to a man who came dangerously near being first rate and who had more influence on his medium than anyone seems to have realized. That is the name of Arthur B. Reeve.

Arthur B. Reeve, who began in an earlier era – as, indeed, did most of the lady waltzers – entered the 20s with his once immense popularity fading away. Nevertheless his tales of Craig Kennedy had been read by hundreds of thousands, praised by Theodore Roosevelt, and turned into early film serials which held us petrified.

Craig Kennedy was Professor Kennedy of, presumably, Columbia University. Like Dr. Thorndyke, he was the scientific detective. His laboratory flashed with stranger sparks, and bubbled with more weird beakers and test tubes, than the laboratory of the late Dr. Frankenstein. For each occasion he had some new gadget, guaranteed sensational, to clap on somebody’s wrist or wire underneath the chair. Square-jawed Kennedy in his high collar, whom we remember so well from the illustrations in the Harper editions, has marched into limbo with all his gadgets loaded on him. Much of his scientific knowledge, I believe, has been discredited. Nobody reads about him now. And yet…

He was first in the field of fiction with the lie detector, with murder by electrolysis, with radium poisoning, with death from liquid air. He taught writers the use of the Maxim silencer, and neither tears nor prayers nor curses can induce them to give it up. As a final achievement among many, in a story called “The Dream Detective” and later in a novel called The Soul Scar, it was he who introduced the profession to psychoanalysis.

This, in its way, is a solemn thought. For the humble annals of the detective story, it is like Watt studying the boiling kettle or Franklin flying the kite in the thunderstorm. In these days when every other mystery novel depends on a neurosis or a phobia or a fixation or whatnot, we can see now what wild vegetation has grown from that small seed. Psychoanalysis has been the most widely used contribution to the detective story since the days of Poe and Conan Doyle; and we might do worse than remember who planted the jungle in which our contemporaries lose themselves.

Well, never mind. We were discussing the American situation in the 20s.

Shortly past the middle of the decade, S.S. Van Dine published The Benson Murder Case, in which Alvin Benson was shot to death under circumstances which suggested the fate of Joseph Elwell the bridge expert. It was not a reconstruction of the Elwell case, as we can see for ourselves if we read the real-life account of the police officer in charge of that affair. But it brought forward a new writer who juggled suspects with such dexterity, like whirling Indian clubs, that we could only stare in admiration; and a new detective, Philo Vance, who said his method was psychology and scorned the cigarette ends found near Benson’s body. Three years later, when a crooked lawyer was poisoned with some villainous new stuff called tetra ethyl lead in The Roman Hat Mystery, we saluted Ellery Queen.

Though these were the only practitioners of the front rank, both were so good that they held the scales almost level against their British confreres. It looked, in those far-off days, as though the golden age of the detective story had come. It now played strictly fair. It was adult. It had lost its clumsiness and grown to maturity.

Then came the 1930s. Then came the cleavage. The hard-boiled detective story, which for some years had been lurking in the magazines without anybody suspecting its inherent genius, suddenly blossomed out until it shadowed the whole field. Few writers, even experienced ones who had been dealing with a different type of story, were completely untouched by its influence. Novices rushed to get aboard the bandwagon. And there began, between the school of Sherlock Holmes and the school of Sam Spade, a difference which has been widening for more than thirty years.


Let us consider the hard-boiled type of story.

Whether you prefer this kind of writing is a matter of personal taste. Whether you acclaim it as good, on the other hand, depends on how it is done. If anybody wants to see how “economical, astringent, muscular prose” should really be handled, let him reread the best stories of Melville Davisson Post. Post was a great master of prose style, whereas most of the moderns are fairly answerable to some other description.

But we are not here concerned with literary quality. We are concerned with the detective story, and what goes into it. Dashiell Hammett has been praised as “a creator of the first rank”, belonging among “the small handful of others who brought something really new to their chosen field of effort”, and as one whose “lean, dynamic, unsentimental narratives created a definitely American style, quite separate and distinct from the accepted English pattern”.

These are the words of Howard Haycraft, a sound critic, an admirable critic whose opinions we are bound to respect, and whom we can accuse of eccentric or unbalanced literary judgement only when he praises an undeserving hound named Carter Dickson.

But this originality, this glory of breaking fresh ground, again depends on what you do. You could get a finely original effect, for instance, by sending a whole procession of kangaroos across the stage during a performance of Lohengrin at Covent Garden or the Metropolitan Opera. You would be, definitely, a creator. You would have brought something really new to your chosen field of work. Or, to be more restrained about it, you could decide that the trouble with musical shows was the use of music, and the thing to do was have the musical show without any music at all; just as you can decide to have the detective story without any clues to follow or any rules to observe.

As we earlier discussed the saga of Brace and Du Kink, let’s take a typical American detective novel of the late 30s. Its plot runs something like this:

The hero, Chip Hardstone, is a wisecracking private detective with an attractive blond stenographer. To Chip’s office, in violent agitation, comes the lean, elderly, aristocratic J.T. Witherspoon, a millionaire with a country house in Sundown Hills.

Mr. Witherspoon’s daughter, it appears, has got herself involved with a notorious character called Smooth Ed Spumoni. A priceless crystal flask, with goldwork by Benvenuto Cellini, has been stolen from the millionaire’s collection. Matters at home are tense, since – in addition to his ne’er-do-well son, his rebellious daughter and his neurotic young wife – Mr. Witherspoon has further grounds for suspicion in that the butler is a blackmailer, the chauffeur an ex-convict, and the housekeeper a hophead. What he wants, he says, is to recover the Cellini crystal and free his daughter from the clutches of Smooth Ed Spumoni.

“But no scandal, Mr. Hardstone!” pleads the millionaire. “Above all things, no scandal!”

Already, before going to the country house, Chip has accumulated a lot of information. Practically every character in the story calls on him and tries to retain him. These he first bluffs and then insults – all except the representative of an insurance company, whom he merely insults.

Arrived at the house in Sundown Hills, Chip finds the “mad family” of earlier fiction now so completely nuts as to require a psychiatrist rather than a detective. The daughter removes her clothes; the wife intimates that she is willing to do so; the son tries to knock Chip’s head off on sight. Other friends swing punches at the son, at Chip, or at each other; and Chip, who replies by insulting everybody he has previously missed, is interrupted with the discovery that one of the guests has been found dead – his throat mangled – in the swimming pool.

(Observe the departure of originality here. The millionaire himself is seldom murdered. He must be kept alive to pay Chip’s fee.)

But one of the guests is murdered. No less than eight persons, it appears, know some vital secret about the murder. All of them have disappeared. It being Chip’s job to find them, in a roulette-ball spin round the city, he concentrates first on a mysterious red-haired girl who has been traced to an apartment house at the corner of Pineapple and Banana.

Racing to the apartment house, Chip finds the girl gone but a corpse on the floor. He flies to a second apartment house, only to find the girl gone again and another corpse on the floor. By the time he has reached the third apartment house and the fourth corpse, he is in a spot. The police are after him, the reporters are after him, Smooth Ed Spumoni is after him, even the millionaire is after him to call him off. Chip won’t be called off. He intimates, with something very like blackmail, that the old s.o.b. can’t get out of it after bringing him in.

“All the same,” says Chip, “this setup is hopeless!”

And again we agree, since the vital secrets turn out to be innocent side games in which everybody is chiselling everybody else, and have nothing to do with the murders. Chip, on the point of being arrested by Captain Hooligan of the Homicide Bureau, suddenly gets an inspiration – it is never very clear how – that the murderer is J.T. Witherspoon’s wife. He confronts her; there is a gun fight all over the house; and the wife, waiting only long enough to scream out a confession, falls dead at his feet.

This is the end of the story, leaving the reader in some doubt as to just what did happen after all.

Now why, at the outset, are the adventures of Chip Hardstone so vaguely familiar? What strikes a reminiscent note? Despite the original kind of hero, despite the spit-in-your-eye style of writing, despite the chases and sluggings and kidnappings, we seem to have met this motiveless and clueless method somewhere before.

Don’t we see that it’s Inspector Brace and Reginald Du Kink all over again?

Instead of cuff links, bus tickets and lace handkerchiefs which bear no relation to the problem, we have “secrets” which bear no relation to the problem, either. Instead of the suspects doing their ring-around-the-rosebush outside the library, they now rush away from capture in cars and aircraft; but they still act either for no reasons at all, or for no reasons that are ever explained.

As for the fairness of the evidence, or the quality of the solution, the same test can be applied.

The American wheel, in these hard-boiled stories of the 30s, had turned full circle. We were back again among the whiskers and mothballs of an earlier era. Those very detective-story features of which the reader complained most bitterly in 1920, the features which were its essential faults, the features which craftsmen had worked so hard to eliminate ever since, were triumphantly hailed as a daring new departure from convention.

This period in America, it is true, produced its own first-raters. In 1934, with a story called Fer-de-Lance, Rex Stout by sheer power of characterization and plot construction at once joined the company of Ellery Queen and S.S. Van Dine. There was Anthony Abbot, whose grim first novel, About the Murder of Geraldine Foster – based on a legend of Courvoisier and Lizzie Borden – never seems to have achieved the full critical acclaim it deserves. In the front rank, or very close to it, were Clayton Rawson and C. Daly King.

But these were all practitioners in the great tradition, the clue-serpents and trap-baiters. Their narratives moved as fast as you could wish for; yet they ranged beside their British confreres of the same period, Margery Allingham and Ngaio Marsh and Nicholas Blake, in the vital business of presenting new ideas. In Nicholas Blake’s first novel, A Question of Proof, you will find one instance of what is meant by the great tradition. The murder knife unaccountably vanishes; and the investigators can’t find it because it has been hidden, in front of their eyes, by being used as a tent peg.

Yes; but what about the weaknesses in the English type of novel?

The fault here is just the same, though expressed in a different way. The “literary” type, like the hard-boiled, is too often apt to mistake style for substance. It imagines that with good writing, which sometimes becomes merely pretentious writing, you can disguise the lack of an original plot.

“Come, now!” the author seems to be saying. “I’m really a straight novelist, you know, indulging in this funny little medium of the detective story because nowadays it’s become respectable. It’s true I haven’t got much of a mystery, or any very clear idea of how to handle it; but if I give you strong characterizations and much talk-in-a-mist, you won’t mind that.”

To which the answer is: Sir or madam, we do mind. Either you neglect the plot, which is bad; or else you fall off those stilts with a crash, which is worse.


The present anthology has been called The Ten Best Detective Novels – a title for whose arbitrariness you must blame the publishers. I myself should have preferred to call it ten “of the” best, assigning to this introduction a task less responsible than that of the recording angel.

Nevertheless – if we postulate our necessary qualities as to fair play, sound construction, and ingenuity – there will be many to ask why this anthology does not begin with Wilkie Collins’s The Moonstone.

The Moonstone, published in 1868, has been put at the top of the list by both G.K. Chesterton and Dorothy L. Sayers. And with reason. The problem of who stole the moonstone, that baleful yellow diamond, is only one feature of a story so skilfully woven that it remains a lesson in technique even today. It is also a greenhouse of Victorian charm, with its flower-painted door and its rose-growing detective and its amiable whiskered hero, who gives up smoking to please the heroine and by this small gesture partly brings about the whole catastrophe.

A too-enthusiastic reader, however, must be warned that the first chapters are tolerably heavy going. Old Gabriel Betteridge, the first of several narrators, requires no less than fifty-seven closely printed pages to set his stage, and encourages you with, “Cheer up!” at the end of the fifty-seventh (in the edition from which I quote: Chatto & Windus, 1905). The reasons against including it are mainly those of space: The Moonstone runs to more than a quarter of a million words, and would overweight even a formidable volume like this.

There are very different reasons for excluding another story which has been handed down from generation to generation as a classic: Anna Katharine Green’s The Leavenworth Case.

This American landmark, which appeared in 1878, must be treated with reverence as being the first novel in history where our unfortunate millionaire (in this case the retired tea-merchant Mr. Leavenworth) is found shot through the back of the head in his library. Listen, now, to a description of the beautiful Mary Leavenworth, one of the dead man’s nieces:

Seated in an easy-chair of embroidered satin, but rousing from her half-recumbent position, like one who was in the act of launching a powerful invective, I beheld a glorious woman. Fair, pale, proud, delicate; looking like a lily in the thick creamy-tinted wrapper that alternately clung to and swayed from her richly-moulded figure; with her Grecian front–

Don’t laugh; Miss Leavenworth’s Grecian front isn’t what you think it is. And, in any case, young Raymond was really impressed.

–her Grecian front, crowned with the palest of pale tresses, lifted and flashing with power… her whole appearance was so splendid, so startling, so extraordinary, that I held my breath in surprise, actually for a moment doubting if it were a living woman I beheld, or some famous pythoness conjured up from ancient story.

It would be unjust to make evil-minded comments on this picture, merely because the author wrote in the popular style of her day. In the best detective stories we shall find antiquated writing, which doesn’t matter when there is a dexterous story to tell. What does matter is that we plod for hundreds of pages beside a detective who never unearths a single clue against the right person, and refrains from arresting the wrong person only because of his notion that no woman would think of cleaning a revolver after firing it.

Anna Katharine Green at her best could devise brilliant plot-tricks, like the ice-bullet of Initials Only, or the unearthly penance of Dark Hollow, or the “portrait” in The Filigree Ball, which is not a portrait but a face drawn in microscopic lines of handwriting. All excellent; yet something more is required than this. It is a real pleasure when we go on to Sir Arthur Conan Doyle and Sherlock Holmes.

The position occupied by Sherlock Holmes among present-day detectives is a trifle curious. He still towers over the rest of them, huge and ageless, with those piercing eyes fixed on a would-be critic.

“Yes?” he seems to be saying. And then again: “Yes?”

It is no good for his enemies to attack him as a character; he exists. He is as real as the pavement of Baker Street and as unmistakeable a personality as Mr. Churchill is. So his detractors (as the fashion is) will acknowledge he is a great man but deny he is a good detective. Even his admirers, while praising Holmes and Watson as living people, sometimes intimate that the stories themselves are rather thin stuff. “Holmes,” we read on more than one occasion, “is always picking up some small object, which he conceals from Watson and from the reader.”

The operative word is “always”; it has become customary to say this, and the only objection is that it simply isn’t true.

Our difficulty is that we have been, most of us, so familiar with these stories from early years as to forget their essential cleverness. They are a part of life, a part of youth, as entwined with boyhood as our first long trousers or our first girl friend. A happy light mists everything. We are too close to the stories, too hypnotized by Holmes; we can’t see the plot for the personality.

It is true that with A Study in Scarlet (1887) Conan Doyle had not yet found his method. But then he had not found Sherlock Holmes either. In the brilliant study, published in 1945, by Adrian Conan Doyle, son of Sir Arthur, you will read of a first draft of A Study in Scarlet in which Holmes did not appear at all.[1] The great man was an addendum, an afterthought; in A Study in Scarlet Conan Doyle was concerned far less with detection than with the Mormons and Jefferson Hope’s vengeance. Even in The Sign of Four (1890) he was still searching. Conan Doyle found his true method only when he had leisure to sit down and write the short tales of The Adventures (1892) and in his best stories he never lost it again.

Taste may cause debate here. But it can be suggested that, in any list of the dozen best short stories, six of most people’s choices would be “The Red-Headed League”, “The Man with the Twisted Lip”, “The Speckled Band”, “Silver Blaze”, “The Naval Treaty”, and “The Reigate Squires”.[2] And in each of those stories you will find every vital clue fairly and honourably set forth.

You doubt it?

You say that in “The Red-Headed League”, for example, we are not told what Holmes sees when he looks at the knees of Vincent Spaulding’s trousers?

But we have already heard that Holmes recognizes “Spaulding”; we know Spaulding’s trick of diving into the cellar; we see Holmes hammering on the pavement with his stick; we are carefully informed that one of the buildings in line with Wilson’s pawnshop is the Capital and Counties Bank. In this story of a tunnel dug for a bank robbery, we must realize that the stains on the trouser-knees are not a clue at all; they are proof, clinching proof, which Holmes legitimately keeps back until the solution.

Again, in “The Speckled Band”, it is putting the cart before the horse to maintain we should have been told about the snake’s fang-marks on the victim. The whole problem was to find those marks, to discover what could have killed Julia Stoner. We are given the ventilator – note how deftly to begin with, by the scent of cigar-smoke – the clamped bedstead, the dummy bell-rope, the locked safe, the saucer of milk near the looped dog-leash, the carefully repeated insistence on Dr. Roylott’s fondness for Indian pets. If few readers of The Strand Magazine in 1892 thought of a snake, it was not for want of evidence. It was because young Dr. Conan Doyle had been the first to use one.

Now take a scene which contains perhaps the most famous words in the saga. The scene occurs in “Silver Blaze”, a puzzle honourably displaying every card from the curried mutton to the dressmaker’s bill; and its famous words, at risk of over-repetition, must be quoted again here.

“Is there any other point to which you would wish to call my attention?”

“To the curious incident of the dog in the night-time.”

“The dog did nothing in the night-time.”

“That was the curious incident.”

This is generally quoted as an instance of Sherlockismus, or some other fancy name. But it isn’t a character-twist; it is a clue, and one of the cleverest clues in fiction at that. It is the trick by which the detective, while making you wonder what in the name of sanity he is talking about, nevertheless gives you fair opportunity to think for yourself: the sort of clue which Conan Doyle invented, and which nobody has ever managed so well since.

Of the two remaining novels, choice for this anthology lay between The Hound of the Baskervilles and The Valley of Fear. The latter has been chosen for two reasons: first, because its opening chapters contain perhaps the best Holmes-Watson dialogue in the whole series; second and more important, because its main problem turns on a clue which ought to be as celebrated as the dog in the night-time, or (if we include the Hound) both dogs in the night-time.

“The dumb-bell!” exclaims Sherlock Holmes, this time without reference to any official policeman. “All my lines of thought,” he declares, “lead me back invariably to one basic question – why should an athletic man develop his frame on so unnatural an instrument as a single dumb-bell?”

It is a legitimate question, a puzzle to set the police biting their nails. Do you remember our discussion a while ago, about clues repeatedly dangled in front of the reader? And the story in which, afterwards, you can tell what each character was thinking at a given time?

In the investigation of the murder at Birlstone Manor, observe the happy significance which attaches to Dr. Watson’s umbrella. (No, I am giving nothing away.) Re-read afterwards the interviews with Barker and Mrs. Douglas; consider the testimony of the half-deaf housekeeper. In the second part, which so shocks moderns and pleases your obedient servant, consider each double-innuendo. And let detractors claim, if they can, that Conan Doyle does not play fair.


The Yellow Room! A Supernatural Crime!

With some such explanatory terms were the good feuilleton-readers of Paris set buzzing and chattering when there appeared, in the supplement to L’Illustration for September 7th, 1907, the first instalment of Gaston Leroux’s new serial, The Mystery of the Yellow Room.

This noble enigma was first published in England in the same year through the medium, it is a little disconcerting to find, of the Daily Mail Sixpenny Novels; perhaps it was inevitable that the Yellow Room should get into the Yellow Press. But it is still the best novel in the fashion it made so deservedly popular – the apparently supernatural crime revolving round a “hermetically sealed room”.

True, there had been locked rooms before Leroux’s time. Edgar Allan Poe, prodigal of invention as always, had included one in “The Murders in the Rue Morgue” as he included nearly everything else. But his sealed room was not really sealed; we discover that one of its windows was fastened only with dummy nails; and no present-day artificer (I hope) would cheat his readers with any such unspeakable swindle.

This is not said in deprecation of Poe; the father of the detective story must not be called off-side in a game he invented himself. The interesting thing, here, is why Poe introduced this situation at all. For the problem of the locked room – which some may suppose to be a fantasy of fiction writers – actually has happened more than once in real life. A curious reader may be referred to the strangling of the Prince de Conde behind a bolted door, and the stabbing of Rose Delacourt in a locked attic: both causes célèbres, both occurring in France, and both in the first three decades of the nineteenth century. Students might speculate as to whether Poe, who published “The Murders in the Rue Morgue” in 1841, had ever seen any account of them.[3]

Again, in the field of fiction, there was Israel Zangwill with a novelette called The Big Bow Mystery. It contains an adroit dodge, first use of the victim who is presumed to be dead before he (or, in this case, she) really is. But each step of the story grows progressively more cumbersome and confusing, becoming somehow entangled with politics and allowing the villain at last to commit suicide on the steps of Number Ten Downing Street.

It remained for Gaston Leroux, a Parisian journalist, to produce a masterpiece with The Mystery of the Yellow Room.

The great quality of this book, for all its stifled emotionalism, is that it comes alive. Events march and move with real breathlessness and diabolism. We are snared by its very imaginative conception – the lonely pavilion, the cry of the cat called the Good Lord’s Beast, Professor Stangerson and his daughter with their scientific experiments on “the instantaneous disassociation of matter” – when the instantaneous disassociation of matter, apparently, is just what we get. For the criminal, here, does more than escape from a locked room. In a later chapter, as Sainclair writes:

I was utterly bewildered. Indeed, the phenomena of that still unknown science called hypnotism, for example, are not more inexplicable than the disappearance of the body of the assassin at the very moment when four persons touched him!

Its effects land with a crash of impact, just like that. They take you off balance; they startle you into belief. Yet each has a logical explanation. And Leroux, unlike the bloatedly over-rated Simenon of present days, plays scrupulously fair with the evidence. We note with pleasure that Joseph Rouletabille, the young reporter-detective, has picked up Sherlock Holmes’s trick of the enigmatic clue. “Oh, if only she had had her hair arranged in bandeaux!” cries Rouletabille, in a frenzy of despair. And again:

“How do you know the handkerchief was blue, with red stripes?”

“Because, it if had not been blue with red stripes, it would not have been found at all.”

But Gaston Leroux’s greatest contribution was in widening the field of the locked room to include the general “impossible situation”: the thing that can’t happen but does happen. He developed it with the dissolution-act of the Yellow Room, and the incredible murder of the Green Man in a later chapter. He opened up possibilities which the Carnacki stories of William Hope Hodgson, or the scientific conjuring tricks of L.T. Meade and Robert Eustace, had as yet only touched. It would be difficult to underestimate the immense influence these two elements – the enigmatic clue of Conan Doyle, the impossible murder of Leroux – exercised over writers who followed them.[4]


The enigmatic clue, for instance, looms up again in the work of A.E.W. Mason. But here the whole approach is different, adding still another quality to the growth of the detective story. Here it is character, character, and character all the way.

“Your idioms, I know him!”

That is the triumphant cry, accompanied by a posturing attitude, as Inspector Hanaud of the Sûreté airs his knowledge of the English language. Never has a French writer managed a French detective so entertainingly as English writers have done. We know Hanaud. We know his burly figure, his blue chin and blue Maryland cigarettes, the deep-throated chuckle when he tramples all over his prim friend Mr. Ricardo, or complacently indulges in the English spikking.

“Ten thousand pounds on the table, and he had not one silver!”

“Stiver,” Ricardo corrected.

“As I said, one sliver,” Hanaud agreed equably. “Mrs. Hubbard, she was naked. You understand?”

But, paradoxically, the greater Hanaud grows as a character the less impressive he is in the story. In the later novels, especially, we are always waiting for him to fire off a string of gags like a radio comedian. It is significant that in Hanaud’s first appearance – At the Villa Rose (1910) – he shows no trace whatever of the buffoon. He is sinister, satiric, no more. Even with The House of the Arrow (1924) this tendency is only foreshadowed. Which is exactly as it should be.

For Mr. Mason strikes straight at the emotions. The strength of his writing gives us, with almost intolerable vividness, the heat of bricks basking in the sun at Aix-les-Bains or the clash of bells from a Dijon steeple. Against this background we are caught into a whirlwind of emotion, of human beings in peril. There is no detachment; hardly even time to breathe; and, except at odd times, very little opportunity for laughter.

“I tell you,” says Hanaud, as the police stand in Madame Dauvray’s bedroom after that evil night at the Villa Rose, “this is human! Yes, it is interesting just because it is so human.”

We begin, as a rule, with some beautiful girl menaced by doubt or danger. Popularly speaking, this situation is sure-fire. It never fails. But it must be handled by an expert. The human mind recoils, for instance, from what some lady-novelists of the Had-I-But-Known school would have done with Celia Harland in At the Villa Rose or Betty Harlowe in The House of the Arrow or Joyce Whipple in The Prisoner in the Opal. The formula of the Had-I-But-Known school is not to attempt a detective story but endlessly to re-write Jane Eyre, with Rochesters more roaring and gibberings more pronounced than any dreamed of by Charlotte Brontë.

Mr. Mason’s triumph is in arranging matters so that, instead of an arbitrary plot creating the characters, the characters shall create the plot. He starts with his people. He dwells on visual detail from every expression of a face down to the smallest trifle of hat or shoe-button. Everything, including the unexpected ending, arises from the murderer’s own twisted motives. The criminal – though you may never notice this until afterwards – is always alert, always tense, responding on every page to some stimulus of suggested guilt.

This is what the detective novel had not hitherto fully explored, a purely psychological quality in which a turn of an eye, a gripping of a window-ledge, could become as legitimate a clue as any footprint. It remained for this new technique only to find new ways – always new ways! – in How to Conceal the Identity of the Murderer. And, on How to Conceal the Identity of the Murderer, a whole chapter of tips for the beginner might be compiled from the works of Agatha Christie and Ellery Queen.


Agatha Christie is the female Sixteen-String-Jack, the original Artful Dodger: you must never take your eye off her for a second. It has already been indicated how she upset the apple-cart with her first novel, The Mysterious Affair at Styles, in 1920. She even upset it with her choice of a detective.

It is nonsense to maintain, as some have done, that Poirot was modelled after Hanaud. A comparison of dates and books will show that Hanaud did not begin to display any exotic mannerisms until four years after Poirot’s first appearance. What Mrs. Christie did was to take the comic little foreigner of the music halls – always with us, twirling waxed moustaches and shrugging his shoulders amid a firework splutter of “Zut, alors!” or “Mais oui!” – and transform him into the keen-eyed avenger with the little grey cells. As Hercule Poirot, he is a joy for evermore.

“I’m sick of the little blighter!” his creatrix once complained at a Detection Club dinner where a number of us were airing our grievances. But this, we felt sure, was merely a mood: like the startling mood of Mr. Bentley when he announced that any love-interest in detective stories should be confined to an appendix at the back of the book. For nobody else has ever tired of Hercule Poirot, and nobody else ever will.

Poirot is a symbol of Mrs. Christie’s method. Everybody now knows that in The Murder of Roger Ackroyd ROT-13: vg vf gur aneengbe jub gheaf bhg gb or gur pevzvany. What seems not so generally known is that with The Man in the Brown Suit, an inferior novel but a better hoax, Mrs. Christie anticipated her own device by several years.

The Man in the Brown Suit is told from two aspects: straight narration in the third person, intermingled with chapters from the diary of a bluff and likeable old baronet. ROT-13: Vg vf gur oyhss naq yvxrnoyr byq onebarg (lbh abj thrff) jub cebirf gb or gur zheqrere. Ohg Fve Rhfgnpr – abgr gur qvssrerapr – unf pnershyyl jnearq gur ernqre gung ur jvyy jevgr va uvf qvnel bayl jung ur vf jvyyvat gb unir rirelobql frr: a stroke which Roger Ackroyd, for all its oblique fairness, does not contain.

Nevertheless, even this is a tour-de-force, like the superb juggling in Cards on the Table or the island strewn with stiffs in Ten Little Niggers, which may not be repeated.[5] Where Mrs. Christie scores, over and over again, is the use of a trick which we will call the Implied Alibi.

Now the implied alibi is a very different thing from one of those elaborate police-nightmares – “Well, Joe Maunders is out of it; he’s got an alibi” – which Inspector French so painstakingly breaks down. On the contrary, the whole trick consists in not stressing the fact that Joe Maunders has an alibi; in scarcely even mentioning it; in accepting Joe Maunders’s innocence as so radiantly obvious, under the circumstances, that it has never even occurred to the author (bless her heart) to suspect him herself.

At the time the murder was apparently committed, Joe has been sitting on the beach with the detective. Why even consider the possibility?

Once you have lured the reader into this trap, you have got him fast. It is a form of hypnotism. Whole brass-bands of clues can march past without making him blink an eye to notice them. But once let your own eye stray for a second in the direction of Joe Maunders – once intimate by the slightest twitch that you, the author, are thinking of him – and the spell is broken. The reader is after that suspect like a dog after a cat. You might as well concede defeat from the page where it occurs.

How many times Mrs. Christie has rung the changes on this form of alibi it would be spoiling too many stories to reveal. The alibi itself usually consists of somebody dressing up as somebody else, either to establish the wrong time of death (Evil Under the Sun) or to show the real murderer never went near his victim (Appointment with Death).

Yet this is far from being her only dodge. Another good card, which we might call Misdirection As To The Source of Danger, is played well in Peril at End House and comes down as a veritable ace of trumps with Murder in Mesopotamia. Even to the mania theme in multiple murders – where some ostensible lunatic seems to be polishing off victims for no reason, though we realize quite well that the right corpse is hidden in a pile of wrong ones – even to this theme she brings some fresh twist which is more than sanity disguised as madness.

Through such cases struts Poirot, undisturbed by flying red-herrings. His mannerisms never jar the story, which is light and swift and filled with thumb-nail characterizations. You will find him at the top of his form in Death on the Nile, which I venture to think the best of what Torquemada called “the little grey sells”. Poirot, perhaps, might have told us about the bullet-hole in the table. But if you spot the murderer you deserve a medal.


The same statement, with equal fervour of emphasis, can be made about any of the earlier novels of Ellery Queen.

Though this writer belongs to the end of Mrs. Christie’s decade, his first story appearing in 1929, he is bracketed with her at this point because both are specialists in the dazzle of the thunderbolt ending. He can hocus and flummox the most experienced. Let’s see how he does it.

When any character first steps out on the stage of a detective novel, steps out and speaks his first lines, that character is subjected at once to the reader’s pitiless scrutiny. A fierce spotlight beats on him. “Well, now!” says the reader, unmoved by any distracting gestures on the part of the author, “what’s that fellow doing there? Is he part of the action? Is he needed to work the machinery? Because, if he isn’t, that character’s likely to be the murderer.”

Where Agatha Christie is the Artful Dodger, Ellery Queen is the Confidence Man. He preys on human nature. His trick consists in smoothly getting you to accept some character – in reality the criminal – as a necessary, if humble, part of the mechanism. This character must not be dragged in; he must have a right to be where he is. Suppose he appears in some minor official capacity: a coroner, say, or the house-physician of a hotel where the murder occurs? He can then talk his head off, and have a right to. Nobody will suspect him. Or, extending this principle but still keeping it for the sake of argument to the medical profession, he may be the docile assistant to some doctor who is under suspicion; put there by the author, apparently, only to confirm or deny certain vital facts. Instead of this character butting into the story, it is rather as though the author had tried to keep him out.

Here is the secret behind the shattering dénouement of The Greek Coffin Mystery, or French Powder, or Dutch Shoe. Like the implied alibi, it is a form of hypnosis; once you have slipped your murderer past the reader’s glazed eye, long strings of clues can follow him. Ellery Queen presents his evidence fully and fairly, by telling us what is seen. But he does more than this, with a device peculiarly his own. In addition to describing the clue that is there, he makes great capital out of the clue that isn’t there.

Let us suppose that Xavier Claverton, a New York playboy, collapses of cyanide poisoning on Fifth Avenue and is carted off to the morgue. At the morgue appear Inspector Queen, Sergeant Velie, and Ellery himself, to examine the dead man’s clothes and the contents of his pockets. They are shown a raincoat, a soft black hat, grey coat and vest, blue shirt and dark-blue tie, underwear, black socks and shoes, fifty-four dollars and eighteen cents in money, a billfold, a wrist-watch, a fountain-pen, some letters and a bunch of keys.

Whereupon Ellery gives a long, low whistle, whose meaning is not elucidated until the end of the story.

“Don’t you see, Dad,” he explains then, “that Claverton’s belongings did not include a pair of trousers?”

“By jing, son!” mutters Inspector Queen. “You mean the poor fellow wasn’t wearing any britches?”

“Exactly, Dad. You grasp the point at once. Therefore I asked myself: why should Claverton, a man of impeccable taste in other respects, have been walking down Fifth Avenue without his pants?”

This dialogue (I admit) does not actually occur anywhere in the saga. But you can find the principle operating in novels like The Chinese Orange Mystery and short stories like “The African Traveller”. In addition to its fiendish ingenuity, it is also perfectly fair. If the missing item is not spotted by the reader who himself will be wearing a shirt or tie and not improbably a pair of trousers as well, then he has only himself to blame.

Ellery (the detective) grows in stature through the later novels, as the early Van-Dine influence lessens. He is no longer a bouncing young man; he is genuine; he convinces. The startling solutions remain, with more adroit evidence – as in the arrowhead clue of The Devil to Pay, or the concealment of the murderer in The Door Between – and a compactness of story necessitating fewer characters. That is why any critic complaining of too many characters in the earlier stories, an obscuring of the criminal behind mere numbers, is here offered a later novelette called The Lamp of God. The field is narrow; but you will be misdirected. Artifice has become art.


If there is any writer more passionately addicted to upsetting the apple-cart than Agatha Christie or Ellery Queen, that person is Anthony Berkeley.

Mr. Berkeley himself will take no offence, it is to be hoped, if I say that few people seem to have understood him; certainly many have misunderstood what he was trying to do. His famous preface to The Second Shot, in which he argues that the old crime-puzzle should become a puzzle of character, has often been interpreted to mean that it should grow less and less like a detective story. This pronouncement has been hailed with delight, especially by those who can’t write detective stories for beans.

But look more closely at what the creator of Roger Sheringham really says.

It will become (he writes) a puzzle of character rather than of time, place, motive, and opportunity. The question will be not, “Who killed the old man in the bathroom?” but, “What on earth induced X, of all people, to kill the old man in the bathroom?”

Now this is not very revolutionary. In fact, with those key words of all people, it is not revolutionary at all. Its thesis is that we should stick to the formula of the least-likely person, while demonstrating at the end that the least-likely person should have been recognized as the most-likely person after all. Which is, and always has been, the definition of a good detective story. Assuredly it is a definition of Mr. Berkeley’s own method. That is how he kills the old man in the bathroom. And we seldom get out of that bathroom without falling with a crash on its slippery floor.

No; the question that arises here is a question of evidence.

Mr. Berkeley’s contention has always been that purely material evidence in detective stories – a cuff-link or a bus-ticket or a bottle of ink – is unconvincing because it can point two ways, or three ways, or all ways at once. By selecting what you want, you can make it prove anything you like. This is the theme of The Poisoned Chocolates Case, a masterpiece of wit which rises to Gilbert-and-Sullivan grandeur as each wild-eyed sleuth convicts a different person on the same set of facts.

Most of the problems Mr. Berkeley sets are deceptively ordinary-looking to start with. His people and circumstances are the people and circumstances of everyday life. When he introduced Roger Sheringham with The Layton Court Mystery in 1925, he was careful to point out that his detective is “very far removed from being a Sphinx, and does make a mistake or two occasionally”. Against this background of reality – the popular-price hotel of The Piccadilly Murder, the newly-weds’ suburban home in Murder in the Basement – move characters drawn with effortless skill, as though the author were merely reminding you of somebody you knew.[6]

But this is the trap. This is how he gets his effects. An unwary reader, lured into thinking this is all routine stuff, does not see the crafty Apache lurking in ambush to wallop out with a tomahawk.

For Mr. Berkeley must always twist or double-twist something. If he is not doing it with facts, he is doing it with character. Whenever a hint of the grotesque does peep out of his comfortable atmosphere, as in the sinister invitation of Panic Party, or the murderers’ charade of Jumping Jenny, it is a flash of somebody’s character; and it is the disclosure of somebody’s real character which comes with such a shock at the end. Herein lies the strength of the detective story as a “puzzle of character”, and also its danger. After pouring scorn on the ease with which material clues can be twisted in The Poisoned Chocolates Case, Mr. Berkeley then twists the dead woman’s character – not very convincingly, perhaps – to provide his final surprises. It is magnificent; but it is six of one and half a dozen of the other. For psychology must play fair too.


It is in this realism of detail, this preference for a clue in human nature rather than in the dropped cigarette-end, that Anthony Berkeley is akin to the late S.S. Van Dine.

I do not, of course, suggest any similarity between Roger Sheringham and Philo Vance. Nobody, not even Chief Inspector Moresby, has ever wished Mr. Sheringham a kick in the pance; whereas this has been suggested, often vociferously, for the aristocratic detective who looked like John Barrymore, talked like a Dictionary of Familiar Quotations, and dressed (“Is it warm enough for a silk suit? And a lavender tie, by all means.”) like God knows what.

But let’s be honest about Philo Vance. Despite his mannerisms, despite an “English” style of speech which has caused some hilarity in England, he is a triumph of his creator’s method. Few detective-writers of the front rank can ever have possessed less ingenuity than S.S. Van Dine. He had small skill at misdirection; his phonograph alibi of The Canary Murder Case, even when it was published in 1927, could not have deceived a little golden-haired child. His best dodge in The Greene Murder Case is a variant of a real-life incident used with greater effect by Conan Doyle in “The Problem of Thor Bridge”.

S.S. Van Dine’s real success came from taking pains; taking pains, and still again taking pains. You may not be deeply moved to hear that:

Once when Vance was suffering from sinusitis, he had an X-ray photograph of his head made; and the accompanying chart described him as a “marked dolichocephalic” and a “disharmonious Nordic”.

Yet it goes into the picture, small detail built upon small detail, until the reader is battered into submission. Philo Vance’s reality cannot be doubted. You would recognize him in your sitting-room even if you only threw him out. If S.S. Van Dine had ever murdered the old man in the bathroom, we should have been given every aspect of that bathroom from the name of the maker stamped on the plumbing down to a footnote describing the quality of the soap.

Inside this framework, Vance has three-dimensional existence. Once he steps out of it (as in later novels) the whole thing collapses. Its strength lies in being a framework, as rigid as a cage; the author merely destroyed his effectiveness by trying to break the bars. When Philo Vance turns into a man of action and pursues gangsters, we regretfully decline to believe. And when he is permitted to fall in love, as in The Garden Murder Case, the author himself was so dubious that the whole plot is based on mystifying us as to which girl Vance is in love with.

But in The Bishop Murder Case, and still more strikingly in the Greene, his story has deadly impact. Imagination has been added to the accumulation of detail; it is the commonplace made startling by the abnormal. So painstaking has been the description that we could find our way blindfolded through Professor Dillard’s house in West 76th Street. We know every nook and cranny of the old Greene mansion overhanging the East River. It might be a neighbour’s house; and yet into it has come a prowling mass-murderer.

“That book scared me,” a friend of mine once declared, about the slaughterings in the Greene household. “I can’t tell you why, because it’s not a ghost story. But that book scared me.”

A follower of detective fiction could have told him why.

In addition to the careful detail, Van Dine never allows you to forget the presence of this evil force. Not by the number of corpses; any writer can scatter dead bodies all over the house and have little to show for it; but in many ways far more subtle and more disquieting.

There’s somebody here, and yet we can’t see him. He touches us on the shoulder; we hear him breathing outside a door; there’s a faint creak at the end of the passage; but we can’t see him. Chester Greene talked to him – look at the expression on Chester’s dead face! – yet there’s nothing but a black bullet-hole, a sly step, the smell of burnt candle-grease where somebody’s been reading in the library. It is the mere hint, the half-suggestion (as in a ghost story) which builds up this sense of suffocating evil. And it might be under the bed now. And the goblins will get you if you don’t look out. That is how a fine craftsman weaves the spell in The Greene Murder Case; and those who nowadays labour hopelessly to get such effects are recommended to study his methods. S.S. Van Dine, who professed to regard the detective story as a matter of mechanics, could succeed with horror where he so very often failed with plot.


Atmosphere? For that we must look also to the works of Philip MacDonald.

Mr. MacDonald, so often justly praised for his humorous touches and his sound construction, seems to have escaped too-frequent recognition for another quality he manages even better: the atmosphere of sheer human evil which has breathed out of those country houses and wet woods and darkling landscapes since he published The Rasp in 1924.

G.K. Chesterton, as preface to a not-very-satisfactory novel by Mr. Masterman[7], once wrote a brief essay in which he enumerates the virtues of a good practitioner by listing the things which the good practitioner does not do. The essay is heart-warming; twenty years have not dimmed its vitality or its truth. As it rises to a crescendo in the list of things the good detective-writer does not do, there is one point at which we can all give a special round of applause.

“He does not,” thunders G.K.C. “he does not say it was all a mistake, and that nobody ever intended to murder anybody at all, to the serious disappointment of all humane and sympathetic readers.”

Now this is a mistake which Philip MacDonald very seldom makes either. His, in general, is a fight-to-the-death contest against a malignant enemy. Atmosphere is heavy in The Noose; it becomes a poisonous fog in The White Crow. There is a reason for that wiry staccato prose, which raps like knuckles on a table, and for the tensity which sweeps along Anthony Gethryn as it sweeps along the narrative. Colonel Gethryn, C.M.G., D.S.O., is not fooling. And neither is the criminal.

On the technical side, Mr. MacDonald is master of every trick from the implied alibi – “It was, of course, never recognized as being an alibi,” says Gethryn at the conclusion of The Rasp – to the Chestertonian reversed-identities of The Nursemaid Who Disappeared. He is not free from faults; on two occasions he makes an apparent murder turn out to be a suicide, which seems (to me, at least) the single unforgiveable sin. But there are clues to spare, clues both material and psychological. And through everything runs this driving intensity: it kindles a blaze of emotion even when, as in The Maze, we read only the bare evidence at a coroner’s inquest.

Murder Gone Mad, though it does not contain Anthony Gethryn, is presented here because there are some who consider it the best detective novel written by Mr. MacDonald or any other author. Like The Greene Murder Case, it deals with multiple killing; but observe the differences. The theme of Murder Gone Mad is Jack the Ripper in a modern garden suburb, where Gethryn’s talents would be useless and only police organization can succeed. You are warned that this is no sham mania; it is the true blood-seeking; and even Superintendent Pike walks warily at the heels of “The Butcher”. Its climax, after a miasma of terror, should raise your hair. The reader must walk warily too.


Walking warily, in the physical sense at least, is a necessity which nature has imposed on the person of Rex Stout’s Nero Wolfe.

We would not have it otherwise. Wolfe looms vast and Buddha-like among the beer-bottle caps, with wiggling forefinger and inscrutable face. He was spry enough to kill a snake in Fer de Lance; he survived even the train-journey of Too Many Cooks. But in general we prefer to find him sitting behind his desk, while people rave at him, or absorbed with orchids on the top floor. In recent years there seems a conspiracy to say that his strength as a character is a little eclipsed by that of his assistant. I admire Nero Wolfe too much as an investigator, and like him too much as a person, tamely to hear him slighted in favour of the insufferable Archie Goodwin. Archie Goodwin has been described admiringly by Mr. Haycraft as “paint-fresh”. Another quality of fresh paint, you may have noticed, is that it is wet.

But this again is a matter of personal taste. Even if a critic’s dislike of Archie’s type should be greater than my dislike (which is considerable), he would be compelled to admit that this latter-day Buster Brown is a character. Mr. Stout’s literary skill could turn anybody into a character, just as he fashions clues from the most unlikely materials. During Nero Wolfe’s absences, he is no less nimble with the yellow cleaning tissue which traps the murderer in Red Threads, or the gloves in the watermelon of Crime on Her Hands. All the same, we prefer the Gargantuan trickster to be on the scene.

“All we can do,” Wolfe announces on one occasion, “is to try our luck on the possibilities until we find a fact that will allow only one interpretation. I detest alternatives.”

The massive and sublime self-confidence of this remark would make Roger Sheringham stare. It would draw a goggle-eyed murmur of “Mais quel homme!” even from Hercule Poirot. As a general rule, the detective who seeks after facts with only one interpretation is chasing marsh-lights with a butterfly-net. Yet from Wolfe we accept it; he is like that. In the powerful story called The League of Frightened Men, with its tortured figure of Paul Chapin, you will find Wolfe the psychologist drawing a moral which might well be heeded by many writers of the too-hard-boiled school.

If any criticism may be made, it is an obvious one. There are some of us, maddened, who have occasionally wished Wolfe would stop arguing about who is going to pay his fee, and get on with the business in hand. True, this is an integral part of his nature. He would not be Nero Wolfe without it. He himself has met the objection, with icy and crushing dignity, by saying he is not in business for fun. We find the Neronian philosophy summed up during the investigation of Clyde Osgood’s murder in Some Buried Caesar.

“If all this is true – you knew it last night, didn’t you? Why the hell didn’t you spill it when the sheriff was there? When the cops were there on the spot?”

“I represented no interest last night, sir.”

But he ought to be doing these things for fun. Every really great detective does. This deplorable commercialism will ruin our sense of sport. “I play the game for the game’s own sake,” said Sherlock Holmes, who himself had a living to earn; yet there is only one instance (that matter of the Duke of Holdernesse’s cheque) when the greatest of them all ever stepped out of character. Nor can Wolfe’s attitude be attributed to modern American business methods. It is not recorded that Ellery Queen ever took a fee; and we can imagine the acid remarks that would have ensued if Markham had attempted to slip Philo Vance a retainer. In England, excluding Scotland Yard professionals like French and Alleyn and Bobby Owen, the total amount paid out to detectives cannot have exceeded twenty-five shillings in the past quarter of a century. It would be dangerous even to mention the subject to Gethryn or Dr. Priestley. As for Lord Peter Wimsey…


Enough! That’s enough. It brings us, finally on this list, to Lord Peter Wimsey and Dorothy L. Sayers.

However much the hero’s love-affair has messed up the plot in other careers, it can scarcely be denied that his meeting with Harriet Vane was the making of Lord Peter Wimsey. Allowing a detective-hero violent emotion is either kill or cure. He emerges as a character, and emerges triumphantly, if he has always been more than a rack of neck-ties and a polished top-hat. Between the vapid-looking young man of Whose Body? in 1923, and the figure on the stage of the Comedy Theatre when the play-version of Busman’s Honeymoon was produced in 1937, there is a world of difference.[8]

In Strong Poison, midway between those two dates, Wimsey met the dark-haired Harriet. With Have His Carcase, Murder Must Advertise, The Nine Tailors and Gaudy Night he became as a character progressively stronger, solider, less in need of mannerisms or eyeglass. It would be possible to maintain that the quality of the plots matured as well.

This implies no deep criticism of the earlier stories. Miss Sayers could not have written an undistinguished novel if she had tried. But the “murder” in Clouds of Witness (forgive this return to the subject) is proved to be suicide – amid elaborate incidents including a trans-Atlantic flight, a full-dress trial before the House of Lords, and a finale in which Wimsey, Parker, and the Duke of Denver are seen reeling tight as ticks through Parliament Square. The idea of committing murder with an empty hypodermic, though neatly handled in Unnatural Death, was not a new idea. There are times, in the short stories of Lord Peter Views the Body, when Wimsey becomes a jumping-jack figure flopping all over the place among secret societies and triple impersonations.

Most of all there was occasionally a wavering: a sense of brilliant talents employed at a job which had begun to bore the author, and half persuaded her to drop Wimsey altogether. Then the author wrote a story in which Philip Boyes, the lover of Harriet Vane, was poisoned with arsenic in what looked like a squalid case. Harriet stood on trial at the Old Bailey; Wimsey saw Harriet; Miss Sayers seemed for the first time really to see Wimsey; and there ensued the series of fine novels we know so well.

If Strong Poison is chosen for this anthology, it is not necessarily as the best achievement. There will be many to maintain the superiority of Have His Carcase, with its last twist on how to establish the wrong time of death. Or of The Nine Tailors, where the chill of the fen-country, the boom of bells amid drifting snow, become so vital a part of the plot. Or of Gaudy Night, in which an overpowering sense of menace takes the place of a murder.

Will carpers kindly note, by the way, that this sense of menace – of the goblin that might get you – can be just as effective as any corpse? Have bodies, by all means. But they won’t necessarily startle anyone. Whereas the writer of anonymous letters, the wrecker of rooms, the person who hangs up evil dummies as a symbol, can be magnified to terrifying proportions; just as, in The Greene Murder Case and Murder Gone Mad, it is the malignant mind that counts. This is the secret of Gaudy Night. All the characterizations, the working out of the love-affair between Wimsey and Harriet, must not blind us to the fact that it is plot, plot, and plot all the way.

Strong Poison, however, appears here because of its effectiveness in the more purely traditional style. Who killed Philip Boyes? And, if Harriet is innocent, how was it done? Because Harriet’s fate depends on it, because Wimsey’s peace of mind depends on it, everything is whittled down to the sharp urgency of the main questions. Each event, from the callous detachment of the spectators at the beginning – “Everybody looks like a murderer in a wax-works; have a choc?” – to the eating of a more sinister kind of confection when Wimsey confronts the murderer, rests on the eternal basis of who, how, and why. To answer those three questions of who, how, and why, to answer them in some novel and yet legitimate way, will always be the function of the detective story.


So we come to the end of a survey which has covered many years and no inconsiderable number of books. And now (I confess it) I am seized by a horrible temptation. My Better Nature, seraphic with upturned eyes and halo, pleads and whispers, “No!” But the devil won’t be denied; gleefully he beckons. I have enjoyed writing this introduction so much – in contrast, it is to be feared, to the labours of the weary reader – that I want to end it with a list of rules on What to Do and What Not to Do.

Admittedly, this has been done in full tabular style by Carolyn Wells, by S.S. Van Dine, by H. Douglas Thomson, by Basil Hogarth, by Howard Haycraft, and others. More cautious lines have been taken by Monsignor Knox and by Miss Sayers herself. And I think these two latter writers were wise.

Once the evidence has been fairly presented, there are very few things which are not permissible. The oath of the Detection Club, stern though it may sound, does not forbid the employment of conspiracies, gangs, death rays, ghosts, trap doors, mysterious Chinese or homicidal lunatics. It is not so harsh as that. It merely enjoins the writer to preserve “a seemly moderation” in the use of them. The only thing it rules out, and rightly rules out, is the use of mysterious poisons unknown to science.

Those who nail a manifesto to the wall, saying, “The beginner will do this, and must under no circumstances do that,” are in many cases quoting not rules but prejudices. That is the danger. It is a prejudice, like my own prejudice against having the murder turn out to be a suicide; and should freely be indicated as such. With all due respect and admiration for those who have compiled lists, it would not be difficult to show that they were often giving dubious advice and sometimes talking arrant nonsense.

“Disguise,” declares one writer – to take a single instance – “disguise, of course, went out with the bustle.”

To which the answer is: “My dear sir, that is a prejudice. Furthermore, it’s not true. Have the goodness to read, among other stories with whose titles half a page could be filled, G.K. Chesterton’s ‘The Dagger with Wings’, R. Austin Freeman’s The Mystery of Angelina Frood, Q. Patrick’s S.S. Murder, Ellery Queen’s The Dutch Shoe Mystery, Philip MacDonald’s The Wraith, E.C. Bentley and H.W. Allen’s Trent’s Own Case, Anthony Berkeley’s Top-Storey Murder, or Agatha Christie’s Three-Act Tragedy. Disguise is one of the best weapons in the armoury. The test of a device is not whether it is new or old; there’s nothing new under the sun; the test is what novel twist can be put on it.”

Here, then, is my own list of Dos and Don’ts: compiled partly from those of the writers quoted above and partly from my own heart’s blood.

  1. The criminal shall never turn out to be the detective, or any servant, or any character whose thoughts we have been allowed to share.
  2. The criminal shall never at any time be under serious suspicion until the unmasking. If you haven’t the ingenuity to keep his identity a secret until the end, at least pretend you have. Even if the reader outguesses you, and your thunderbolt ending doesn’t come off, the effect is far more satisfying than if you apologize for your murderer by “clearing” him in an early chapter.
  3. The crime shall be the work of one person. While the murderer in some instances may be allowed to have a confederate, you will ruin your story if two or three or four people are dragged in as accomplices. The essence of a detective story is that the one guilty man shall fool the seven innocent; not that the one innocent shall be fooled by the seven guilty.
  4. The crime shall be clean-cut. If a character disappears and is assumed to be murdered, state frankly what has happened to him. If he hasn’t been murdered it’s a pity; but the reader has a right to clear stating of the problem.

Those are four golden maxims. In each one I believe. And each one you will find shattered – shattered admirably, shattered to bits, shattered by a mighty hammer – in the “best” detective novels, while the reader wishes to do nothing but applaud. Because they are not really rules; they are only prejudices.

The greatest trap into which a critic can fall is to maintain that something is being “done” in the current year, as though there were a style in shrouds as well as in hats, or to maintain that something else has gone out of fashion. When Carolyn Wells’s The Technique of the Mystery Story was first published in 1913, the late Miss Wells was already talking about outworn devices. But nothing ever has gone out of fashion, and nothing ever will, provided only that the old trick can be worked in a new way. Yesterday’s fashion may not be today’s; but it may be none the worse for that. On the contrary, it may be a devil of a sight better.

So let them write their stories, the hopeful young men and women!

Let them not be frightened by that worst bogey of all, the feeling that they have got to be innovators. Let them remember that the real test of their mental skill is in the drive and nimbleness and strategy of their play; it does not consist in putting the goal posts in the middle of the field or dashing through half the game with a ball that isn’t here. And you and I, serene in our armchairs as we read a new detective story, can continue blissfully in the old game, the great game, the grandest game in the world.


SECOND THOUGHTS – AFTER 17 YEARS

I well remember when and how the above outburst was written. It was written in London during the cold and gusty spring of 1946, and in my flat on Haverstock Hill. Despite an acute housing shortage, I had obtained that flat for reasons quite apart from my Scottish luck. When I moved into it in 1943, Adolf Hitler still walked the earth. He had yet to unleash the Little Blitz, the flying bombs, the rockets, and other drolleries from his inexhaustible sense of humour. But we were all expecting something of the sort. And nobody except your obedient servant was stupid enough to want a flat on the top floor.

If I were to write again of favourite detective novels, should I change anything above? Not one sentiment, not one author. The ensuing seventeen years have produced no writers who are better, nor any (tell it not in Gath) one-half so good. It may be that for four of the authors I should choose a different novel. A.E.W. Mason, for instance, might be better represented by The House of the Arrow; Philip MacDonald by The Rasp; Ellery Queen by The Chinese Orange Mystery; and Dorothy L. Sayers by Strong Poison.

But this is a minor matter; it is the author and his detective who count. “Then you still believe all that?” will be the whisper of kindly friends. “Haven’t you learned anything in all these years?” Since I have learned wisdom in no other respect, it is useless to hope for it here. And a man should always be willing to defend his prejudices. As he gets on in years, these prejudices may constitute the most satisfactory sum total of all the things he has – or is.


[1] Adrian Conan Doyle, The True Conan Doyle (John Murray, 1945), p. 17.

[2] A poll taken in the London Observer to determine the most popular Sherlock Holmes tales, and including the novels as well as the short stories, listed four of these titles out of eight. See H. Douglas Thomson, Masters of Mystery (Collins, 1931), p. 140.

[3] Louis André, The Mysterious Baronne de Feuchères (Hutchinson, 1925) ; C.L. McCluer Stevens, From Clue to Dock (Stanley Paul, 1927), pp. 213–216. In the first case, contemporary opinion seems to have decided that the trick was managed by a door bolted from outside by manipulating a piece of ribbon after a fashion used in several detective stories. In the second case, no solution has ever been found; though one suggestion was that of an orang-outang or large monkey – just as in Poe’s story later. Gross’s Criminal Investigation (Sweet & Maxwell, 1934) lists several locked-room problems. It is noteworthy also that many of the details of a real-life case quoted by Gross (p. 392) parallel those of the attack on Mlle. Stangerson in the Yellow Room.

[4] Notably G.K. Chesterton, who unfortunately cannot be included here because he did not write a full-length detective novel. From the references to it scattered throughout his essays and stories, we know how deeply G.K.C. was impressed by the Yellow Room, and his fondness for the enigmatic clue needs no emphasis. Of the fifty Father Brown stories, no less than seventeen deal with a seemingly supernatural crime.

[5] Or should not be repeated. We also find the Roger-Ackroyd formula used in Brian Flynn’s ROT-13: Gur Ovyyvneq Ebbz Zlfgrel, Anthony Berkeley’s Gur Frpbaq Fubg, and Anthony Gilbert’s Gur Pybpx va gur Ung-Obk. Though in these instances it is quite legitimate, each author giving a different reason for employing it, the practice at one time became so common that you could never begin a story told in the first person without wondering uneasily whether it might not be That Man Again.

[6] The death of John Bentley, in The Wychford Poisoning Case, is based on the death of James Maybrick at Liverpool in 1889. In Poisoned Chocolates you will find parallel cases cited from the “book” which figures so prominently in the later evidence; and the book itself is not imaginary: it is Edward H. Smith’s Famous American Poison Mysteries (Hurst & Blackett, 1926).

[7] Walter S. Masterman, The Wrong Letter (Methuen, 1926).

[8] The play, by Miss Sayers and Miss M. St. Clair Byrne, antedated the novel. Those who say Busman’s Honeymoon is no true detective story should have watched its fine construction when emphasized by the limitation of stage technique. The shattering of the lamp is the best sensation-dénouement you will ever see.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s