The wife had received anonymous letters about that affair.
Then candy arrived from California.
Now detectives had a name.
Cordelia Botkin.
But suspicion was not proof.
The Problem With John Dunning
Dunning was useful to investigators.
He was also dangerous to the prosecution.
His private life was messy.
His credibility could be attacked.
He had been involved with women other than his wife.
His career had suffered.
His drinking and gambling were known.
At trial, defense attorneys would exploit the possibility that other women might have motives.
Dunning himself eventually admitted relationships beyond Cordelia.
That created an obvious question.
If he had multiple lovers, why assume one specific woman sent the chocolates?
Investigators therefore needed evidence independent of Dunning's intuition.
They needed to trace the box.
And tracing a mailed parcel in 1898 would require something much more old-fashioned than technology.
Memory.
Detective McVey Goes West
Evidence from Delaware was transported to California.
The box.
Remaining candy.
The note.
The handkerchief.
The wrapping.
Earlier letters.
A Delaware investigator traveled west, where San Francisco police became deeply involved.
Chief of Police Isaiah Lees took charge of the California investigation.
Now the case effectively had two geographical halves.
Dover held the victims and the original package.
San Francisco was believed to contain the sender.
Investigators needed to transform a theory into a chain.
Where was the candy made?
Where was the box sold?
Where was the handkerchief purchased?
Where was the arsenic obtained?
Who mailed the parcel?
Who wrote the note?
If all of those trails converged on Cordelia Botkin, distance would stop protecting her.
Start With the Box
A candy box might seem generic.
But commercial objects carry identities.
Packaging style.
Decorative seals.
Paper.
Brand characteristics.
The investigators examined what remained.
The chocolates appeared connected with a San Francisco confectioner, George Haas & Sons.
The store sold fine candies.
Clerks saw many customers.
That was another problem.
Weeks had passed.
Why would anyone remember one woman buying one box of sweets among hundreds of transactions?
The answer was that the transaction had apparently been slightly unusual.
The customer did not simply buy a standard sealed box and leave.
According to later testimony, a woman asked for candy in packaging without the ordinary store identification or made requests that allowed room for something additional to be placed in the box.
Clerks remembered aspects of the purchase.
And their description began pointing toward the woman police were already watching.
Cordelia Botkin.
A Clerk's Memory
This is where nineteenth-century detective work becomes fascinating.
Today, police would request surveillance footage.
Credit card records.
Digital receipts.
Phone-location data.
In 1898, they asked:
Do you remember her?
The prosecution would eventually depend heavily on human memory.
Candy-store employees were shown the suspect.
Could they identify her?
Had she asked for unusual packaging?
Had she discussed mailing the candy?
The defense would later attack these memories.
Understandably.
Witness identification after publicity can be contaminated.
Newspapers had printed Botkin's image.
The case was famous.
A clerk who originally remembered a general customer might grow more confident after seeing the suspect everywhere.
That is why the investigation needed multiple independent trails.
The candy-store identification could not stand alone.
Then investigators looked at the handkerchief.
The Handkerchief
Inside the package had been a small woman's handkerchief.
It seemed decorative.
Almost sentimental.
Perhaps it was meant to strengthen the illusion that the box came from a friendly woman.
But the handkerchief carried something the sender apparently had not removed completely.
A retail tag.
That was another mistake.
Police traced it to the City of Paris department store in San Francisco.
Again, detectives interviewed employees.
Again, memory became evidence.
A clerk named Grace Harris was associated in later accounts with the sale and said she remembered the woman because the customer's appearance reminded her strongly of her own mother.
That peculiar personal association made the encounter easier to remember.
Police now had a second commercial path.
Candy.
Handkerchief.
Two separate purchases.
Both in San Francisco.
Both potentially connected to Botkin.
A murderer working at enormous distance had avoided leaving footprints at the victim's house.
But she may have left a trail through stores.
The Wrapping Paper
Then there was the paper.
Packaging can be surprisingly revealing.
A manufacturer may use a particular material.
A store may have distinctive seals.
The parcel contained evidence of where it had been prepared.
Investigators searching rooms previously occupied by Botkin reportedly found paper or a fragment of a gilded seal consistent with material associated with the Haas candy store.
Later newspaper accounts emphasized the apparent relationship between material in her former room and packaging from the poisoned box.
Again, no single piece was enough.
A woman living near a candy shop could possess candy-shop packaging innocently.
But the prosecution was building something more dangerous than one dramatic clue.
It was building convergence.
The Arsenic
Now detectives needed the poison.
Arsenic was legally obtainable.
That helped the killer before the crime.
It helped police afterward.
Because pharmacists sometimes recorded the sale of dangerous substances.
Investigators went through drugstore information.
At the Owl Drug Store on Market Street, a record emerged.
On June 1, 1898, a woman had purchased approximately two ounces of arsenic.
The clerk was Frank S. Grey.
He had recorded information associated with the transaction.
The stated reason, according to trial accounts, was unusual.
The arsenic was supposedly wanted for bleaching or treating a straw hat.
Grey reportedly told the purchaser there were more appropriate products for that purpose.
She insisted.
When later confronted with Cordelia Botkin, he identified her as the woman he remembered serving.
That was a major development.
The prosecution now had a person who allegedly:
had motive,
had access to arsenic,
was linked by witnesses to the candy purchase,
was connected with the handkerchief,
and was suspected of writing both the anonymous letters and the note.
But the defense had an answer.
Timing.
The arsenic purchase was in June.
The chocolates were mailed in August.
Buying poison two months earlier did not prove it was used in this crime.
And there was another technical dispute.
The physical form of the arsenic purchased, the defense argued, did not perfectly match the material recovered from the candy.
The case remained circumstantial.
A Conversation About Poison
Another witness remembered something even more disturbing.
A woman named Almura Ruoff testified about a conversation with Botkin in late July.
According to accounts of the case, Botkin asked questions about poisons.