A police colonel walks into the office of the minister of internal affairs carrying a tray of little bottles, names each poison and explains how it works, and feeds a measured dose to one of Rasputin's cats to prove the stuff is lethal. Every word of it is theatre. The bottles are ordinary medicine vials, their contents harmless, and the poison names were copied straight out of a pharmacology textbook the night before.1 The minister, Alexei Khvostov, believes he is arming an assassination. The colonel, Mikhail Komissarov, and the man who sent him, Stepan Beletsky, are building evidence — not against Rasputin, but against Khvostov himself. In the winter of 1916 a plot to poison Rasputin's Madeira was real on paper and fake in fact, and the fakery was the whole point: the deputy minister was using his own boss's murder scheme as the rope to hang him with.2
This is tradecraft turned inward. The Department of Police was built to surveil and entrap revolutionaries; here its full apparatus — agents, provocations, planted statements, the careful manufacture of self-incriminating paper — is aimed at a colleague one office down the hall. The target of the operation is not the peasant everyone claims to be killing. The target is the man running the operation.
The Madeira plot runs on a specific stock of raw material, and none of it is poison. It ingests Khvostov's humiliation first. He had already tried to have Rasputin beaten — agents in disguise waiting by a car outside the reporter Snarsky's building, Khvostov himself pacing the frost incognito to watch — and Rasputin had simply failed to show, spending the night drinking up the very money meant to pay for his beating in a private room at the Palace Theatre.3 Made ridiculous, Khvostov moved from beating to murder, and it was after the Snarsky fiasco that he first told Komissarov, "Rasputin must be killed."4
Second, it ingests money — startling, untraceable money. When Komissarov feigned hesitation, Khvostov promised him 100,000, then physically showed him two packets of 50,000 each, then raised it to 200,000.5 Komissarov was amazed precisely because no minister could pull that from a ministry budget; Beletsky understood that the proposal "was backed by very powerful forces" — a "significant private allocation."6 The murder fund itself becomes intelligence: its size tells the plotters who is really behind the throne's enemies.
Third, it ingests a cover story already half-written: the banker Rubinstein. Khvostov's plan was to send a case of poisoned Madeira "as if from the banker Rubinstein," so the death would read as someone else's crime.7 Beletsky's job, assigned by no one but himself, was to feed this appetite while starving the result — to "criticize those plans, postponing their execution," or simply to demolish them.8
The mechanism has two layers running in opposite directions at once, and the genius is that the lower layer wears the costume of the upper.
On the surface, Beletsky is the loyal lieutenant procuring poison for his minister's scheme. He "immediately sent Komissarov to get poisons."9 Komissarov returns with the bottles, the explanations, the cat. Everything an assassin's quartermaster would do, done convincingly.
Underneath, every gesture is sabotage dressed as service. The poisons are textbook names on medicine vials.10 The cat experiment is the tell — you do not need to prove a real poison works; you stage a demonstration only when the demonstration is the deliverable, when the point is to be seen complying. Komissarov is not building a murder weapon. He is building a witness's memory: a colonel who can later testify, accurately, that the minister handed him money and ordered a killing. The harmless bottles are the alibi inside the frame — if it ever surfaces, Beletsky's men poisoned no one, because there was never poison.
Beletsky never believed the murder would work and never wanted it to. He told the Extraordinary Commission he "had no faith in its success," having been "convinced of Rasputin's cunning and intuition and of Khvostov's incompetence in organizing secret police actions."11 More coldly, he calculated that removing Rasputin would only summon "a strange new person in the spirit of Misha Kozelsky" into the palace — a devil he did not know, replacing one he did.12 What he did not say in his testimony was the real engine: he "had decided to dupe Khvostov, to let him organize the murder and then prevent it, and hand 'Fat Belly' over to the tsars," so that Tsarskoe Selo would conclude there was no better minister of internal affairs than Beletsky himself.13 The plot's true product was a job opening.
This page gives the vault a clean specimen of intelligence tradecraft turned against the institution that owns it. Where most of the Rasputin surveillance material shows the Okhrana pointed outward — at Rasputin, at revolutionaries, at the street — this shows the same toolkit pointed sideways, agent against agent, minister against deputy. It supplies the vault's case study in manufactured self-incrimination: a provocation that does not lure the target into a crime so much as let the target write his own confession and then makes sure the paper survives.
It also emits a precise reading of how a fake operation produces real evidence. The harmless bottles, the copied poison names, the cat — none of these could kill Rasputin, but all of them can be testified to. The page hands the rest of the vault a model where the theatre of compliance is itself the weapon, useful anywhere an inside actor must appear loyal while building the file that destroys the principal. And it connects the murder fund to the broader question of who paid: the 200,000 traces back, in Khvostov's own File testimony, to Princess Zinaida Yusupova and "the views of the whole grand-ducal milieu."14
Reconstruct the operation as Beletsky ran it, in order, from the File. Khvostov, fresh from the Snarsky humiliation, decides on murder and tells Beletsky they are both "burdened by the meetings with Rasputin and the constant fear of having our closeness to him exposed."15 Beletsky agrees out loud and lies — the agreement is the first move of the counter-provocation. The killing is entrusted to "our colonel," Komissarov, who accepts only "after portraying a certain hesitation for the sake of plausibility."16 Already three of the four men in the room are performing.
Then the money appears, and it is too large. Two packets of 50,000, raised to 200,000, shown not described.17 This is the moment the operation's intelligence value spikes: a sum that cannot be ministerial means private backers, and private backers behind a Rasputin murder is a fact worth more to Beletsky than the murder itself. He files it.
Khvostov proposes the poisoned Madeira from Rubinstein. Beletsky sends Komissarov "to get poisons," and Komissarov returns with the little bottles, narrating each one's effect.18 The cat experiment follows — a demonstration staged "for the sake of appearance" on one of Rasputin's own cats. Radzinsky then delivers the reveal flat: "In fact, the bottles were medicine bottles and their contents harmless: Komissarov had merely copied out the names of different poisons from a pharmacology textbook."19 The entire poisoning apparatus is a prop. Khvostov is being shown a play in which he is the only spectator who thinks it is real.
The finale was not poison at all but paper. When the parallel Rzhevsky provocation blew up and Khvostov panicked, Beletsky advised him to commit "political suicide" — go to the tsar, show him the external surveillance log on Rasputin, lay everything out.20 Khvostov went. But Beletsky, "as befitted the chief of the secret police," was not too fastidious to inspect Khvostov's briefcase afterward, found both copies of the report still inside, and understood Khvostov had reported nothing about Rasputin — he had reported on Beletsky instead, proposing to ship him off as a provincial governor-general.21 Each had tried to betray the other to the throne. Beletsky struck back through Simanovich, who gave the Department of Police a statement that an engineer named Heine had named Rzhevsky as organizing the attempt on Khvostov's orders.22 Khvostov "grasped it all, but it was too late." By March 1916 he was retired so coldly the tsar would not even summon him to say so.23 The harmless bottles had done their work — not by killing Rasputin, but by being part of a record that killed a career.
You are reading the File testimony cold, and the tell arrives in a single sentence. It is past midnight in your reconstruction; the colonel has set the bottles on the minister's desk; the cat has been fed. Watch the verbs. Komissarov "experimented on one of Rasputin's cats" — but the phrase that follows is "for the sake of appearance."24 Stop there. An assassin testing a real poison does not test it for appearance. He tests it to know. The instant a demonstration is performed to be seen rather than to learn, you are no longer reading a murder plot; you are reading a man building an alibi inside a frame.
Now run it backward through the other gestures and feel them invert. Komissarov "portraying a certain hesitation for the sake of plausibility" — performance.25 Beletsky saying "he was in agreement" while testifying later that he "was of course lying when he said he agreed" — performance.26 Three confederates acting loyalty toward a principal who believes he is the only one not acting. The diagnostic move is to ask, of each compliant gesture, what record does this leave behind, and who benefits from that record surviving? The cat leaves a memory of a colonel who obeyed an order to procure poison. The shown packets of cash leave a memory of a minister who funded murder. Every act of apparent obedience is depositing testimony into a future commission's hands.
Finally, watch for the briefcase moment — the instant the operator stops trusting the principal's word and goes to the physical evidence. Beletsky does not believe Khvostov's account of the tsar's reaction; he opens the briefcase and finds both copies of the unread report.27 When you see an inside actor verifying the principal behind his back, the loyalty was never real and the file is already being assembled. The whole episode reads, in retrospect, as a single sustained performance whose audience was one man and whose script ended with his retirement.
The plot "fails" only from Khvostov's chair — and his failure is the diagnostic catalogue. The first sign is that the operation is too easy to staff: Khvostov hands the most sensitive task imaginable to "our colonel," a man Beletsky controls, and never asks whose man the colonel truly is. An operator who lets his enemy's friend run his secret killing has already lost.
The second sign is the money that announces itself. Cash shown rather than disbursed, raised on the spot from 100,000 to 200,000, is not how a careful conspiracy moves — it is how a man buys loyalty he is not sure of, and the buying itself broadcasts desperation and powerful backers in the same gesture.28
The third sign is the willingness to be talked into delay. Beletsky's assignment was to "criticize those plans, postponing their execution."29 Khvostov accepted the postponements. A principal whose subordinate can indefinitely defer the deed without consequence is a principal whose deed will never happen — the subordinate owns the timeline, and whoever owns the timeline owns the plot.
The fourth and fatal sign is the panic confession. When Rzhevsky was arrested and named him, Khvostov ran to Beletsky — the author of his ruin — for advice, and took it.30 An operator who, cornered, seeks counsel from the very man cornering him has lost the ability to tell ally from handler. That is the moment the bottles full of medicine finish their work.
The episode rests almost entirely on the self-interested testimony of the two principals to the 1917 Extraordinary Commission, and they were enemies recounting a mutual betrayal. Beletsky's account is the spine — the harmless bottles, the copied poison names, the cat, the briefcase inspection — but Beletsky was the man who engineered the whole counter-provocation and had every reason to present himself as the clever puppeteer rather than a co-conspirator in a murder plot.31 Radzinsky flags exactly this: Beletsky, testifying, framed his refusal as principled doubt about success and the danger of a "new Misha Kozelsky," while concealing his real motive — duping Khvostov to seize his post.32 The witness is narrating his own innocence.
Khvostov's testimony cuts the other way and is the only thread to the backers: in 1917 he "very carefully named" Princess Zinaida Yusupova and the grand-ducal milieu as the source of the unlimited funds.33 Whether the 200,000 truly came from there, or whether Khvostov was distributing blame from the dock, the File cannot settle. Vyrubova's testimony supplies the outside corroboration that Iliodor's wife arrived with telegrams in which Khvostov offered 60,000 for the murder, and that the sovereign ordered Stürmer to investigate.34 The open question the source leaves wide: did the poison plot ever have a chance of becoming real, or was it stillborn from the first bottle? Radzinsky's reconstruction says it was theatre all the way down — but theatre staged by a man who needed it to look exactly like the real thing.
Radzinsky builds this entire chapter from File interrogation testimony, and the seam in his sourcing is the gap between Beletsky's words and Beletsky's motive. The two converge on the facts of the operation — both Beletsky and Komissarov describe the same bottles, the same cat, the same fake poisons — and that convergence is what lets Radzinsky treat the harmless-Madeira mechanism as solid rather than as one man's boast.35 Komissarov's recollection of the shown cash and Beletsky's confirmation of the "significant private allocation" cross-check each other on the money, the one detail that points beyond the room toward the grand dukes.36
The tension is between Beletsky-as-witness and Beletsky-as-actor. In the File he presents the refusal to murder as reasoned reluctance; Radzinsky, reading past the testimony, names the suppressed motive — the plot to "dupe Khvostov... and hand 'Fat Belly' over to the tsars."37 Radzinsky is doing here what he does throughout the book: trusting the File for events while distrusting the File's self-portraits. The split reveals something the testimony alone would hide — that a 1917 commission witness, performing innocence for investigators, was the same man who in 1916 performed loyalty for Khvostov. The counter-provocation and the testimony are the same skill exercised twice, once to destroy a rival and once to survive a revolution. Where Vyrubova's testimony enters, it is corroboration from a hostile-to-Beletsky angle — she received Khvostov's letter, she relayed the empress's refusal to clear him — and her account independently confirms the blackmail architecture even as she calls it "a Yid provocation arranged by Beletsky," repeating Khvostov's own framing back into the record.38
Here is the plain version: a plot to kill someone with poison and a plot to ruin someone with paperwork are the same plot when the poison is fake — and the only domains that explain that are the history of secret-police tradecraft and the psychology of how blackmail leverage gets built. The Madeira bottles are not chemistry; they are evidence-manufacture, and evidence is a behavioral and historical object, not a pharmacological one.
The sharpest handshake runs to The Cyanide Failure & Sabotage Doubt. Both pages turn on poison that did not poison — the Madeira bottles full of medicine, and the cakes and wine at the Yusupov murder that Rasputin reportedly ate without effect. The structural parallel is exact and the divergence is the insight: in the Madeira plot the poison is known fake, manufactured harmless on purpose by a man sabotaging his own boss; at the Moika the poison's failure is a mystery that breeds the legend of an unkillable holy devil. Set side by side, they expose a continuum of "failed poisoning" running through the whole Rasputin material — and they raise the same forensic suspicion from opposite directions. The Madeira case proves that men close to Rasputin were entirely capable of staging a poisoning that was never meant to work; the cyanide case asks whether the Yusupov poisoning, too, was sabotaged or simply fictional. Neither domain alone gets there: the chemistry says "poison should kill," the history says "but inside-actors had every reason to fake it," and only holding both produces the doubt that the famous poisoning was theatre like the Madeira was theatre.
The second handshake runs to Fixer-Blackmail Architecture. The Madeira plot is not really about murder; it is about leverage, and its true mechanism is the construction of compromising material. Khvostov, by being shown the cash and accepting the colonel and proposing the Rubinstein cover, deposits self-incrimination into hands that will weaponize it. That is precisely the architecture the fixer page maps — the systematic accumulation of holds over powerful men. The insight neither domain produces alone: the fixer page treats blackmail as a tool used on outsiders to the apparatus (ministers, bankers, petitioners), but the Madeira plot shows the same architecture turned inside the apparatus, deputy against minister, with the secret police building a hold on its own chief. The history of the secret service supplies the toolkit; the behavioral logic of leverage supplies the reason it gets aimed at a colleague — when the prize is the colleague's job, the colleague becomes the mark.
The third handshake runs to Okhrana Manufactured-Consent Forgery Ops. Both pages are about the Department of Police manufacturing reality rather than recording it — there, forged documents to shape opinion; here, fake poison and the planted Heine-Simanovich statement to shape an outcome.39 The parallel is the institutional habit of fabrication; the tension is the target. Forgery ops manufacture consent in the public; the Madeira counter-provocation manufactures a self-incriminating record for an audience of the tsar. The insight: the same forgery muscle that the Okhrana built to deceive the population is, in the Madeira plot, flexed to deceive a fellow minister and the throne — proving the apparatus's deception capacity had no fixed direction. It would manufacture against revolutionaries, against the public, or against itself, wherever advantage lay.
The Sharpest Implication. The destabilizing reading is that the Madeira plot makes the real poisoning at the Moika historically suspect. If a colonel of the Department of Police could, in early 1916, stage a flawless fake poisoning — bottles, demonstration, dead-serious cat — purely to build evidence against his own minister, then the December 1916 cyanide that "failed" to kill Rasputin should not be read as proof of supernatural resistance or even of botched chemistry. It should be read as further evidence that men around Rasputin were fluent in staging poisonings that were never meant to poison. The third wire: the famous unkillable-holy-devil legend may rest on a poison plot that, like the Madeira, was theatre — and the vault has the documentary proof that such theatre was a known secret-police craft within the same year, the same city, the same apparatus.
Generative Questions.