Two men killed Rasputin, then sat down and wrote the scene as if it were a novel — and the novel they wrote is the version the whole world has repeated for a century. Felix Yusupov published his account in Paris in edition after edition; the Duma deputy Vladimir Purishkevich scribbled his all night on a hospital train pulling out of Petrograd, promising "photographic accuracy." 1 What they produced reads less like testimony than like the pulp thrillers popular in their day: a monstrous demon fed poisoned cakes that don't kill him, who survives a bullet, who springs back to life foaming at the mouth, who is chased across the snow and finally brought down by heroic gunfire. 2 Even Trotsky, no friend of the regime, sniffed the trashiness and called the story "tasteless." 3 Radzinsky's point in The Rasputin File is blunt: the more carefully you read the murderers' account, the more it falls apart — not because the men were liars about everything, but because they built a story to do specific work. They needed to hide the women who were present, glorify themselves as slayers of a supernatural evil, and above all keep Grand Duke Dmitry Pavlovich's hands clean of blood. [POPULAR SOURCE] The genre wasn't accidental. The thriller form was the cover-up.
The fabricated narrative draws its raw material from three pressures that converged on the killers the moment Rasputin stopped breathing. First, the legal threat: a murder investigation was already underway, and a beat constable named Vlasyuk had heard the shots and would tell his superiors. 4 Second, the dynastic stakes: Dmitry Pavlovich was a Romanov, an Olympic athlete, a favourite of the Guards, and — critically — a plausible pretender to the throne if a coup followed. A grand duke who organized a patriotic deliverance could aspire to the crown; a grand duke who personally shot a peasant could not. 5 Third, the women. Felix's wife Irina was meant to be the lure, but she refused to come, so the conspirators staged her presence — a gramophone, abandoned pastries, voices upstairs — and at least one and probably two real women were in the house that night: the ballerina Vera Karalli, Dmitry's lover, and Marianna Derfelden, Grand Duke Pavel's stepdaughter, mockingly nicknamed "Malanya." 6 [POPULAR SOURCE, per Vyrubova and police-file evidence]
Feed all three pressures into a single document and you get the thriller. Each fabrication solves a problem. The "Countess N." pseudonym shields Irina. 7 The all-male guest list erases Karalli and Derfelden to protect their honour from the police. 8 And the relentless insistence that Purishkevich, not Dmitry, fired the killing shots — repeated so many times it grows suspicious — protects the royal youth's claim to the throne. 9 What the narrative ingests is danger; what it emits is literature.
The engine of the fable is a single device: make the victim superhuman, and any inconsistency becomes proof of his demonic power rather than a hole in the story. The cakes are the clearest case. Lazavert chipped enough potassium cyanide into the pink petits fours "to kill an ox," and Felix says Rasputin ate them and felt nothing. 10 But Rasputin didn't eat sweets. He kept a strict diet he tied to his healing powers — fish and no sugar — and called candy "scum," a fact confirmed by his daughter Matryona, by Simanovich, by Beletsky and Khvostov in their File testimony, and by Chikhachev's account of the boxes of candy Rasputin shared but never touched on the train. 11 [POPULAR SOURCE, multiply corroborated] Even Felix's own text records Rasputin refusing the pastries — "I don't want any; they're too sweet" — before the narrative implausibly has him change his mind. 12 He never ate them. The poisoned-cake superman is pure invention, added afterward to dress the killing as the destruction of a man-devil.
The wine is subtler. The cyanide solution in the glasses may have been real but too weak; Grand Duke Nikolai Mikhailovich, who poisoned insects as a hobby, noted dryly that a weak solution explains everything without recourse to demons. 13 So the device runs in two registers: where the poison was fake (the cakes), invent the eating; where the poison was real but feeble (the wine), invent the supernatural resistance. Either way the victim becomes "an extraordinary person," and the heroes who finally fell him become correspondingly larger. The resurrection-and-chase — Rasputin springing up, seizing Felix's throat, foaming, crawling up the stairs, fleeing across the yard while Purishkevich fires four times — is the genre's climax: "The devil revived." 14 Radzinsky reads the real event underneath as far more boring: Felix, no marksman, only gravely wounded Rasputin, who lay unconscious until the cold yard and the passage of time revived him, exactly as he had survived Guseva's knife in 1914. 15
This page hands the vault a clean worked example of how a primary historical record can be authored fiction rather than evidence — a document that looks like first-person testimony but functions as a literary artifact built to specification. It supplies the forensic contrast the cyanide-failure page needs: the poison "failure" is not a chemistry mystery but a narrative requirement. It feeds the autopsy and drowned-alive claim the reason the killers needed Rasputin indestructible. It gives the prophecy-letter debunking a sibling case of textual analysis exposing a fabrication. And it offers the cross-domain pages on historiography as ideology and suppressed-testimony epistemology a concrete instance: the canonical source is the least reliable, and the police constable's bored deposition outranks the published memoir.
Walk through the single most load-bearing fabrication. Purishkevich describes the preparation in loving, almost stage-direction detail: Lazavert puts on gloves, chips the cyanide pellets onto a plate, "thickly sprinkles the poison" inside the pink-cream petits fours chosen to match the wall colour, then tosses the gloves into the fireplace, which smokes and forces them to air the room. 16 The chocolate ones are left clean for Felix. [POPULAR SOURCE] It is vivid, specific, and theatrical — and it is the part the daughter, the bodyguard, and the secret-police interrogators all contradict.
The contradiction is not subtle. Matryona Rasputina states flatly that her father "never ate sweets, meat, or pastries." 17 Simanovich confirms it. Beletsky and Khvostov, interrogated by the 1917 Extraordinary Commission, report Rasputin kept a strict diet. 18 Filippov explained that Rasputin tied the diet to his healing gift — fish, no sweets — and didn't break it even drunk; devotees gave him candy boxes he never touched. 19 Konstantin Chikhachev told the File about the train compartment piled with candy Rasputin shared but refused, calling it "scum." 20 Five independent voices, several under interrogation, establish a lifelong habit. Against them stands one self-interested memoirist who has Rasputin reverse that habit at the precise moment required by the plot.
Felix even sabotages his own lie. He records Rasputin declining the pastries with the exact words a real Rasputin would use — "I don't want any; they're too sweet" — and only then writes that the peasant "took first one and then another." 21 The seam shows. The truthful detail (the refusal) survives because Felix half-remembered the real evening; the fabricated detail (the eating) is grafted on because the thriller needs a poisoning to fail. Radzinsky's verdict: Rasputin drank some weak wine at most, and "the story about the pastries" was "made up later as part of his fable about a devil whom ordinary people had heroically destroyed." 22
Strip the cakes out and the whole supernatural architecture collapses. No ox-killing dose survived, so no superman is required to survive it. What's left is a man given thin poisoned wine, shot by an amateur, and left for dead — who woke up. The fable was retrofitted onto a botched killing.
It is past two in the morning and you have two documents open on the desk. On the left, Purishkevich's manuscript, written in a swaying railway carriage, the prose breathless: "I shall try to recount with photographic accuracy." 23 On the right, the flat carbon of Constable Vlasyuk's deposition. Read the memoir first and you are inside a thriller — the gramophone playing "Yankee Doodle," the icon lamp, the demon's clenched fingers sinking like red-hot iron into Felix's shoulder. 24 You want to believe it. That is the danger.
Now do the discipline. Take one concrete, checkable claim — the cakes — and ask: who else was in a position to know, and what did they say? You find the daughter, the secretary, two ministers, a judicial official, and Rasputin's own friend, all saying he never touched sweets. 25 You hold that against the memoirist's lone assertion and you notice he himself recorded the refusal. The seam opens. Take a second claim — that Dmitry was away burning the fur coat during the murder. 26 You check the constables: Vlasyuk and Efimov watched the palace after the shots and saw no automobile arrive, only one leave with the body, a detail Globachyov's coded telegram confirms. 27 No car came back; therefore Dmitry never left. 28
The method is mechanical and unglamorous. Find the witness with nothing to protect — the constable who, asked to keep silent, answers that under oath he'll tell "the whole truth, as it would be a sin to lie." 29 Anchor on his boredom. Then read the thrilling document as a suspect, not a source. Every place the bored deposition and the breathless memoir diverge, the memoir is doing work — hiding a woman, clearing a prince, glorifying a killer. Mark each divergence and ask what it protects. That list of protections is the real story.
You can smell this kind of document before you can prove it false. The first sign is genre creep: the prose reaches for a literary register no honest report needs — "the devil revived," eyes "out of their sockets," a body "raised up by dark powers to avenge his death." 30 When a witness narrates like a novelist, ask what the novel is for. The second sign is the repeated disclaimer. Purishkevich insists again and again that "the hands of the royal youth" were never "stained with blood" — and the very insistence, Radzinsky notes, "looked suspicious." 31 A truth said once is reportage; a truth said five times is a defence brief.
The third sign is the convenient supernatural, the load-bearing miracle that exists only to paper over a physical impossibility. A man who eats an ox-killing dose and feels nothing is not a fact; he is a patch over the fact that he never ate it. The fourth sign is the implausible bystander excuse — Purishkevich's wife too "lazy and capricious" to burn a fur coat, sending a grand duke back "like a little boy." 32 When a story needs a silly woman to explain why physical evidence (the unburned coat, the surviving boot found in the car) contradicts the timeline, the timeline is the lie. 33 The fifth and surest sign is the seam: the truthful detail the fabricator can't bear to cut, sitting awkwardly beside the invention it should contradict. Rasputin's "too sweet" refusal is that seam. Find one and pull.
The central evidentiary tension is that the autopsy record — the one document that could have settled the poison and the cause of death — disappeared after the revolution. 34 Radzinsky proceeds by triangulation instead. Beletsky testified that Protopopov told him Rasputin was still alive when thrown into the ice hole, "shown by the autopsy"; if true, neither poison nor gunshot killed him and drowning did. 35 That contradicts both murderers' claim of a clean kill in the basement and supports the "wounded, revived, never finished off" reconstruction. [POPULAR SOURCE, secondhand testimony — flagged]
Radzinsky offers two competing reconstructions and does not fully resolve them. The "erotic version" speculates that something sexual passed between Felix and Rasputin in the basement, explaining why the peasant waited three hours and why his famous intuition was "fast asleep." 36 [SPECULATIVE — Radzinsky's own conjecture, drawn from Nikolai Mikhailovich's imaginings] The "realistic version" is more boring: Rasputin simply refused the food and wine, Felix fetched the pistol and shot him quickly, and the poison drama was invented afterward. 37 These two readings sit side by side; the erotic one explains the lost three hours, the realistic one needs no such explanation but leaves the three hours a puzzle. Radzinsky leans realistic while keeping the erotic possibility open. The open question Radzinsky himself foregrounds and never closes: which of the conspirators fired the two "masterful" final shots — and his answer, Dmitry, rests on inference (Dmitry the Olympic marksman; Purishkevich the philologist who missed at "two or three paces"), not on a confession. 38
The contest here is not between two historians but between a body of testimony and the published myth it sits beneath. On one side stand Yusupov and Purishkevich, the émigré memoirists, whose accounts "have been repeated from book to book" and form the canonical story. 39 On the other stand the witnesses Radzinsky mines from the 1917 File and the police archive: Constable Vlasyuk and Constable Efimov; the building custodian on Gorokhovaya Street; Beletsky, Khvostov, Chikhachev, and Filippov on Rasputin's diet; Vyrubova naming Marianna Derfelden; the police agents tracking Vera Karalli; and Grand Duke Nikolai Mikhailovich, the sceptical "Voltairean" diarist who refused to believe in demons. 40 Where the memoirists converge with these witnesses — the gramophone, the basement layout, the constable conversation, which "in its general features was identically described by them all" — Radzinsky accepts the detail. 41 Where they diverge, he treats the memoir as the suspect.
The revealing split is internal to the conspiracy. Dmitry Pavlovich himself broke ranks. In a 1928 Matin interview and a Russian émigré clipping, he condemned Yusupov for publishing at all — "Yusupov acted quite improperly in publishing his book" — and said the publication ended their friendship. 42 One conspirator wanted the thriller in print; another found "it was hateful to maintain the lie." That fracture is the strongest evidence the published account is a lie: the man it was built to protect refused to endorse it. The convergence that matters most, then, is between the bored constable and the bitter grand duke — the witness with nothing to gain and the beneficiary who couldn't stomach the cover — both, in different registers, contradicting the heroes' own bestseller.
Why would a history of a murder need help from psychology and from the study of how myths are built? Because the document at the centre of this page is not really a record of what happened — it is a constructed object, authored to manage threat and to manufacture a legend, and understanding that requires the tools that study how stories defend their makers and how founding myths get assembled. The murder is history; the making of the account is the manufacture of ideology.
The first handshake is with historiography as ideology. That page argues that historical accounts encode the interests of those who write them; this page is the laboratory specimen. The killers' memoirs are not biased reporting — they are reporting whose every fabrication maps onto a specific interest: the pseudonym serves Irina's honour, the all-male guest list serves Karalli's and Derfelden's, the "Purishkevich fired the fatal shot" insistence serves Dmitry's claim to the throne. 43 The insight neither domain produces alone: ideology in history is usually diffuse, a tilt across a whole narrative, but here it is itemizable — you can list each lie and name the interest it protects, turning "historiography as ideology" from a thesis into an audit. The structural parallel is exact and the tension is sharp: the most cited, most "primary" source on the murder is the most ideologically loaded, which inverts the usual archival hierarchy where first-person accounts outrank later commentary.
The second handshake is with suppressed-testimony epistemology. That page concerns how knowledge gets distorted when some voices are silenced and others amplified. The fabricated narrative is a case of active suppression generating false knowledge: the women are erased not by accident but by design, and the erasure leaves a detectable hole — the missing woman's voice from upstairs that should have been staged and wasn't, which Radzinsky uses to prove women were there. 44 The insight that emerges only at the join: suppression is forensically legible. A well-made cover-up has to delete not just the fact but the trace of the fact, and when it forgets one trace — the unburned fur coat, the surviving boot, the absent female voice in a story that requires one — the deletion itself becomes evidence. Epistemology supplies the principle (silenced testimony warps the record); the murder supplies the proof that the warp leaves a fingerprint.
The third handshake is with founding-myth construction. The killers were not only covering a crime; they were authoring a legend in which patriotic heroes slay a demon to save Russia — Purishkevich announces to the soldiers that he has killed "Grishka, Russia's enemy and the tsar's," and writes through the night of "that great land I call Motherland." 45 The thriller form is the myth's natural vessel: a monster, a trial (poison that fails), a resurrection, a final righteous kill. The insight the join produces: the supernatural elements that look like mystification (the indestructible body) are simultaneously the myth's engine — a hero is only as large as the monster he kills, so inflating Rasputin into a man-devil inflates the killers into saviours. Myth-construction explains why the cover-up reached for the supernatural rather than a flat denial: a flat denial protects you legally, but only a legend makes you a hero, and these men wanted both.
The Sharpest Implication. The destabilizing reading is that the murderers' account is more historically valuable as a fabrication than it would be as a true report — because a true report would tell us only how Rasputin died, while the lie tells us what his killers feared, whom they protected, and what legend they needed Russia to believe. The document fails as evidence of the murder and succeeds completely as evidence of the murderers' politics. Read straight, it is the worst source on the night. Read inverted — as a list of problems each fabrication was built to solve — it becomes the best source on why the killing happened and what the conspirators thought they were buying with it. The fable is a confession in negative.
Generative Questions. (1) If the canonical primary source on a famous death is a deliberate literary construction, how many other "eyewitness" accounts in the historical record are thrillers we've mistaken for depositions — and what genre markers (the repeated disclaimer, the load-bearing miracle, the convenient silly bystander) would let us flag them? (2) Dmitry refused to endorse Yusupov's book yet never told his own version; is a beneficiary's silence a stronger falsification of a cover story than an outsider's contradiction, and how would a historian weight it? (3) The thriller form did real political work — could you write a comparative study of "heroic assassination" memoirs (tyrannicides who publish) and show that the genre conventions are load-bearing across cases, i.e. that murderers reach for the same plot whenever they need to be heroes?