Days later, Mary Elizabeth opened it.
The mechanism completed itself.
That is why California argued it had jurisdiction.
The murderous act had begun long before the victim swallowed poison.
The Jury Retires
On December 30, 1898, after weeks of testimony, the jury withdrew.
The prosecution had never produced an eyewitness to the poisoning.
It had produced something else.
A network of small facts.
The jury deliberated for only a few hours.
Then it returned.
Guilty of murder in the first degree.
But the jurors fixed the punishment at life imprisonment rather than death.
Newspaper reports described Botkin collapsing after the courtroom began to empty.
The woman accused of sending a murder weapon across a continent had avoided the gallows.
But the legal story was nowhere near finished.
February 1899
Botkin was formally sentenced to life imprisonment.
Her husband soon sought divorce.
Yet before she could simply disappear into prison, California law intervened.
Another murder case had produced a ruling concerning how judges instructed juries about circumstantial evidence.
The same type of problematic instruction had been used in Botkin's trial.
The California Supreme Court reversed her conviction.
Importantly, the court did not declare her innocent.
It ordered further proceedings because the trial had contained legal error.
And in the same appellate process, the court addressed her argument that California had no jurisdiction.
It rejected that argument.
She could be tried again.
A Case That Refused to End
Years passed.
The case remained alive.
Witnesses had to be brought again from Delaware to California.
Memories were older.
Some investigators had died.
Evidence had already been handled repeatedly.
The prosecution still had the same basic challenge.
No eyewitness.
Circumstantial chain.
But it also had something powerful.
The first jury had believed the chain.
Could a second?
The retrial began in 1904.
And almost immediately the case produced another scandal.
A Jury Scandal
During the 1904 proceedings, allegations surfaced that jurors had been approached.
One juror reportedly told police that someone had claimed several jurors had already been secured for the defense and offered money for another vote.
Judge Carroll Cook responded dramatically.
The jury was discharged.
A new jury had to be selected.
Botkin's attorneys denied that she or her legal team had any involvement.
There was no reason simply to attribute the alleged bribery scheme directly to her.
But the incident reinforced the atmosphere surrounding the case.
Six years after the chocolates reached Dover, the prosecution of Cordelia Botkin could still produce front-page drama.
The Second Trial
A new jury heard the case.
Again:
Mary Elizabeth Dunning.
The mysterious parcel.
The chocolates.
Arsenic.
The San Francisco connection.
The affair.
The candy shop.
The poison purchase.
The handwriting.
The witnesses from Delaware.
The prosecution's central theory had not changed.
Neither had Botkin's denial.
But the second jury reached the same essential conclusion as the first.
On April 7, 1904, Cordelia Botkin was again found guilty of first-degree murder.
Once again the punishment was fixed at life imprisonment.
Judge Cook reportedly believed the punishment should have been harsher.
But the jury had spoken.
The woman whose alleged murder weapon had traveled through the mail would not be executed.
Why the Case Was So Difficult
Modern retellings sometimes make the investigation sound almost effortless.
Dunning says Cordelia.
Police find arsenic.
Case solved.
That understates what investigators actually faced.
Imagine beginning with the evidence available on August 12, 1898.
Two dead women in Delaware.
Several sick survivors.
A partly eaten box of chocolate.
A note signed only "Mrs. C."
A parcel originating somewhere around San Francisco.
That is all.
The sender could be anyone.
The candy could have changed hands.
The package had crossed a continent.
The victims had acquaintances in multiple states.
The husband himself had multiple romantic relationships.
There was no digital trail.
No laboratory could compare fingerprints from inside the chocolates with a national database.
No DNA could be recovered from saliva on an envelope and searched genealogically.
Investigators solved the case by reconstructing commercial life.
Where was this handkerchief sold?
Who remembers this box?
Who sold arsenic?
Who asked strange questions?
Who writes this way?
Who knew the victim loved candy?
Who benefited if she died?
That is why the case feels unusually like nineteenth-century detective fiction.
Every ordinary object had to become a witness.
The Handkerchief as Witness
The handkerchief could not speak.
Its tag could.
The sender probably intended it to make the package seem more affectionate.
Instead it created an additional retail trail.
A decorative addition became evidence.
Had the box contained only chocolates, one entire branch of the investigation would have disappeared.
The killer had added camouflage.
The camouflage added clues.
That is a recurring phenomenon in planned crime.
The more elaborate the deception, the more material the offender must manipulate.
Every manipulated object creates another opportunity for error.
The Anonymous Letters as Witness
The same principle applies to the earlier correspondence.
If jealousy was the motive, the letters made emotional sense before the murder.
The writer wanted Mary Elizabeth to know about the affair.
Perhaps the goal was to make her abandon Dunning permanently.
If that had worked, murder might have been unnecessary.
But the letters also preserved writing.
Later, when the writer allegedly escalated from psychological interference to poison, the old correspondence allowed investigators to compare the two stages.
The campaign against Mary Elizabeth had left its own archive.
The Drugstore Ledger as Witness
The arsenic transaction demonstrates another difference between fiction and real investigation.
The dramatic clue was not a vial labeled "poison" found in Cordelia's hand.
It was a mundane business entry.
A clerk.
A date.
An address.
A purchase.
Without that record, prosecutors could still argue that she had access to arsenic.
With it, they could put a specific acquisition into the timeline.
Again, not proof alone.
But another coincidence the defense had to explain.
Human Memory as Forensic Technology
One of the most interesting things about the case is how heavily it depended on people remembering ordinary transactions.
The candy clerk had no reason, at the time of sale, to believe she was witnessing preparation for murder.
The handkerchief clerk did not know she might later testify in one of America's most sensational cases.
The pharmacist was performing an ordinary commercial transaction involving a regulated poison.
Postal employees processed countless parcels.
Yet weeks later investigators returned and asked each person to reconstruct a few minutes of an ordinary day.
Before cameras and electronic records, memory itself functioned as surveillance.
It was imperfect.
But sometimes it was all detectives had.
Newspapers Become Part of the Investigation
The press did more than report the Botkin case.
Newspaper publicity brought witnesses forward.
It also created danger.