Skip to main content


June 2024
16min read

A newly discovered document casts a disturbing light on exactly how Frank’s prosecutor won his case

FEW CRIMINAL TRIALS IN AMERICAN HISTORY have been so carefully studied as Leo Frank’s, and with all the principals of the case now deceased and the written record generally available, it may come as a surprise that there is something new to be said about the case. But there is.

About three months after the murder of Leo Frank, a case was tried in the Fulton County Superior Court, of which Atlanta is the county seat. Unfortunately the record of this trial is not available; the case was appealed, however, and papers associated with that appeal provide an accurate, if less than full, account of the trial’s proceedings, a trial that reveals much about the one that doomed Frank.

The original case, Pinkertoris National Detective Agency v. National Pencil Company , was heard before Judge W. D. Ellis on November 17, 18, 19, and 22, 1915. At issue was an unpaid bill in the amount of $1,286.09 for detective services rendered by Pinkerton in 1913 for the investigation of Mary Phagan’s murder. Pinkerton prevailed in the superior court; National Pencil appealed the decision to the Georgia Court of Appeals, and Pinkerton, now the appellee, won that decision as well.

The reason that the National Pencil Company refused to pay Pinkerton’s bill can be found in the Amended Motion for the New Trial, in which National Pencil claimed that Pinkerton “did not seek honestly and in good faith to ascertain the truth, but, on the contrary, endeavored dishonestly and in bad faith to suppress and distort the truth and to bring about the conviction of Frank regardless of guilt or innocence.” Even though the court found against the National Pencil Company, a fair-minded reading of the Brief of Evidence—a 134-page document that summarizes the trial and whose accuracy was ratified by lawyers for both sides—and of supporting papers strongly suggests that Pinkerton wanted Frank to be found guilty and worked toward that end.

The National Pencil Company’s case against Pinkerton turned on the actions of the agency’s employee Harry Scott. As assistant superintendent of Pinkerton’s Atlanta office, Scott was in charge of the investigation of the Phagan murder from the day after her body was discovered in late April until sometime in August, the month Frank was convicted. Ironically it was Leo Frank himself, as manager of the pencil factory, who arranged with Scott to hire Pinkerton. Not only did Scott take an active part in the investigation, he also supervised all the numerous Pinkerton employees looking into the crime.

From the beginning the pencil company’s lawyers should have been wary of the way Scott handled the investigation, for in his testimony Scott explained that “the established policy of the plaintiff [i.e., Pinkerton] in dealing with [a] crime in its relations with the local police is to cooperate with the State and County authorities to the fullest extent and work with them in the interest of public justice.” When Frank’s lawyers requested that Scott inform them first about any new development, he was adamant that “any new facts that we unraveled in a criminal case, we go right to the police about it.” The Atlanta police, in turn, assigned Detective John Black to work with Scott.

The record indicates that Scott may have been more than merely ineffective as an investigator; there is ample evidence that he actually conspired with the prosecutor, Hugh Dorsey. After Scott was subpoenaed by the prosecution in the Frank trial, Frank’s lawyers asked Scott to discuss with Dorsey beforehand the nature and content of what the prosecutor planned to ask him on the stand and then report back to them. But Scott never spoke to Frank’s lawyer about his forthcoming testimony, although it would prove a devastating setback for the defense.

The reason was that Scott differed with Frank about many things—what Frank had done on the day of the murder, Frank’s knowledge of the principals in the crime, Frank’s behavior after the discovery of the victim’s body—and the accumulation of differences undermined the veracity of the defendant’s testimony. Worse still, Scott posed not only as a disinterested third party whose only concern was the truth but as an employee of the defense.

Early in the trial Scott contradicted what he had said at the coroner’s inquest and written in a report he had made as a Pinkerton employee to the National Pencil Company’s lawyers. At issue was what Frank had told Scott soon after the murder, when he explained how Mary Phagan had come by the pencil factory for her pay and that before leaving she had asked whether the metal had arrived. This was important to her because her job was to attach the metal sleeves to the pencils, and no metal meant no work. Both in his report and at the coroner’s inquest, Scott had said that Frank told him he had answered no to Phagan’s question. However, at the trial Scott changed the answer to “I don’t know,” which let the prosecution contend that Frank had an excuse to go with Mary Phagan to another part of the factory where he could check on the metal’s availability and that it was there that the murder had taken place. Though caught off guard, Frank’s attorney, Luther Z. Rosser, was able to question Scott on why he had changed his testimony from what he had written in his report and thus blunt, but not entirely remove, doubts about Frank’s actions on the day of the murder.

Scott may been more than merely ineffective as a defense investigator; he may have actually conspired with the prosecutor.

THIS IS ALSO TRUE OF ANOTHER CHANGE in Scott’s testimony. Scott reported to the pencil-company lawyers that in an early interview Frank had said he had left the pencil factory to go home for lunch at 1:00 P.M. on the day of the murder. At Frank’s trial Scott changed this to 1:10 P.M. , suggesting that the defendant had more time to move Mary Phagan’s body to the basement. Again, Rosser pressed Scott for the reason why he changed his testimony from 1:00 P.M. to 1:10 P.M. Scott claimed that “it must be a typographical error” in the coroner’s report and that his notes would show that Frank had said 1:10 P.M. From what is available of the trial record, it appears that these notes were never entered into evidence.

On another point there was no written evidence to help Rosser contradict Scott’s damaging testimony. Scott claimed that Frank had told him that J. M. Gantt, a former employee whom Frank had fired only a couple of weeks before the murder, “was very familiar and intimate with Mary Phagan.” This testimony suggested that Frank knew the dead girl by name, although he had testified to the contrary at the coroner’s inquest; indeed, he’d had to check his payroll books to verify that it was Mary Phagan who had been paid around noon that day.

With no written record on this point, all Rosser could obtain from Scott was a lame admission that his failure to mention it at the coroner’s inquest “was an oversight, if anything at all.” Under cross-examination Scott made a contradictory statement: “I did not consider it material at all to mention in the report to the Pencil Co. that statement of Leo Frank regarding Gantt’s intimacy with Mary Phagan.…I knew that Frank had stated that he did not know Mary Phagan and that he had to look into the books to tell her name, but it wasn’t a material fact against Frank at that time that he said to me that Gantt was familiar with her.…The first time I saw the materiality of it was when the Solicitor [Dorsey] asked me the question [during Frank’s murder trial].” This is quite an admission for a man who considered himself a savvy and seasoned detective.

The record, however, suggests more than a mere oversight. The following exchange during Frank’s murder trial between Rosser and Scott implies collusion between Dorsey and Scott:

Q. [Rosser] Was it an oversight before the coroner’s inquest too[?] Look at it [i.e., the transcript of the coroner’s inquest], and see if you said anything about that before the coroner’s inquest; your mind was fresher then about a verbal conversation [between you and Frank] than it is now, wasn’t it[?]

A. [Scott] Well, it was fresher on my mind at the time, certainly, but you will understand the coroner asked me certain questions, and I gave him answers to the questions, but he did not cross examine me like Mr. Dorsey has.…

There is more. Scott’s testimony at Frank’s trial on how Frank behaved when confronted with the night watchman Newt Lee, who had found the body and was for a time a prime suspect, went far beyond what Scott had said at the coroner’s inquest. Two days after the discovery of the body, Scott and Detective John Black asked Frank to talk with his employee Lee and persuade him to be more forthcoming. Black and Scott left the two men alone in a room for about ten minutes, then returned. Scott, at Frank’s trial, described what happened next: ”…we took seats alongside of both of them; Newt Lee was handcuffed to the chair, and he says: ‘Mr Frank, it is awful hard for me to remain handcuffed to this chair,[’] he says: ‘It is awful hard, awful hard, Mr Frank.’ Frank hung his head the entire time the negro was talking to him. Finally in about 30 seconds, he says: ‘Well, they have got me, too.’”

Dorsey continued the questioning:

Q. [Dorsey] Now, describe if you can the appearance and deportment and manner in which Frank talked and carried himself at the conference set forth on that occasion.

A. [Scott] Well, he was extremely nervous at that time.…very squirmy in his chair, crossing one leg and then with the otherf,] he didn’t know how to put his hands, he was moving them up and down on his face, and he hung his head a great deal of the time while the negro was talking to him, that is, in my presence.

Q. How did he talk[?]

A. Well, as I say, he hesitated some…

Q. How did he breath[e][?]

A. Well, he just took a long sigh that [illustrating], more of a sigh than a breath.

Q. Did you notice his eyes[?]

A. Yes sir, I judged their insecure condition all the way through, yes.

On cross-examination Rosser was not able to refute much of Scott’s testimony. However, he did establish that at the coroner’s inquest Scott had testified he heard nothing of the conversation between Lee and Frank and then at the murder trial claimed that Frank had declared: “Well, they have got me, too.” Rosser forced Scott to admit that this was a change in his testimony.

As for Frank’s nervousness, Scott had never mentioned it at the inquest. When Rosser asked him about this, Scott said: “At the time, Frank’s nervousness had no effect whatever on my mind, because I did not consider Frank any suspect at all. Knowing the man was under a strain, I did not suspect him at all at that time, and therefore it was not a material fact at the time. I did not consider him a suspect.” This is impossible to believe, for Scott admitted that on the previous day, when he first interviewed the defendant, he “knew then that Frank was under strong suspicion.”

Though Rosser blundered later in the trial, he was powerful and effective when confronting the state’s witness John Black, the city detective who worked with Scott on the case. Through careful and insistent questioning, Rosser all but destroyed Black’s testimony against Frank. At one point Black was so confused that he took six minutes to answer a question, and near the end of his testimony he declared, “I don’t like to admit that I’m crossed up, Colonel Rosser, but you have got me in that kind of a fix and I don’t know where I’m at.” No surprise then that the headlines of the story of Black’s testimony read: DEFENSE RIDDLES JOHN BLACK’S TESTIMONY/SLEUTH CONFUSED UNDER MERCILESS CROSS-QUESTIONS OF LUTHER ROSSER . The Atlanta Georgian said, “There is a feeling growing more fixed every day,…that the state, if it hopes to win, must set up something more than it has yet made public.”

When Scott took the stand, he proved a more difficult target for Rosser’s cross-examination than Black had been. Dorsey was able to get Scott to give testimony that damaged the defense’s case, and as we have seen, because much of this was a surprise to the defense, Rosser was hard-pressed to get Scott to retract it. Summarizing the day’s testimony, the Atlanta Constitution declared, “Harry Scott…proved a strong witness for the state, although at first it looked as if he would prove of more value to the defense.” Though the trial had been under way for several days, Scott was the first witness who really aided the case against Frank.

Rosser was emphatic about how the changes in Scott’s testimony had damaged the defense. In the trial over the unpaid Pinkerton bill, he declared: “At no time prior to the trial of the Frank case, was I informed verbally by Mr. Scott…that he intended to change the testimony that he gave at the Coroner’s inquest and the information that he gave me in his reports as to the matter of Frank’s saying ‘no’ or ‘I don’t know[.]’ He certainly did not tell me before he went on the stand at the trial that he was going to testify that Frank told him that Gantt was intimate with Mary Phagan. His testimony at the trial on that point certainly surprised me. In my opinion, as an attorney, that was certainly a matter of materiality and consequence in the case.”

On cross-examination Rosser expressed in even stronger terms what he considered Scott’s duplicity: “Mr. Scott never made any effort to get a conference with the attorneys for the defense before the trial. I did not know that Mr. Scott had any opinion about the case that had not been communicated to me…and everything he knew was supposed to have been put in writing in the reports, that he had made to me, and I supposed that those reports were true. If I had thought that Mr. Scott was going to testify anything different from what was in the writings that had been submitted to me, I would have wanted to talk to him.”

SCOTT CLAIMED THAT AFTER HE HAD DISCUSSED his testimony with Dorsey, he attempted to speak to Frank’s attorneys but was refused an interview. But surely, if Frank’s lawyers had had any indication that Scott’s testimony at the murder trial would differ from what he’d said at the coroner’s inquest or in his written reports, they would have found time to talk to him.

More plausible is that even though he was employed by the defense lawyers, Scott had been conspiring with Dorsey for some time to establish Frank’s guilt. This is nowhere more evident than in the role he played in the several statements made by the prosecution’s most devastating witness, Jim Conley.

At first Conley had been overlooked as a possible suspect because the investigators believed his claim that he could not write. But at some point two Pinkerton detectives, L. P. Whitfield and W. D. McWorth, became suspicious. According to Leo Gottheimer, a National Pencil Company salesman, the two detectives visited the factory on May 16, 1913, a little more than two weeks after the murder, and asked “those present if they knew whether Jim Conley could write.” At McWorth’s request Gottheimer went over to the “Tower,” the county jail where Frank was being held, to ask Frank. The answer was definitive: “Yes, I know he can write, I have had notes from him asking me to lend him money…”

Frank directed Herbert Schiff, acting superintendent of the pencil factory, to a drawer in the company safe that contained documents associated with Conley’s purchase of watches from local jewelry stores and pawnshops. These led the Pinkerton detectives to three shops where they were able to secure loan contracts signed by Conley, and when the signatures were compared with the murder notes, Whit field noted in his report for that day, “the hand writing appeared to be identical.”

Four times in all, and all under oath, Schiff insisted Harry Scott had stated explicitly that Pinkerton wanted Frank to be found guilty.

Though Scott could not prevent the knowledge that Conley was able to write from becoming known, he did try to hide Frank’s role in its discovery, for if Frank were guilty of the murder and Conley were his accomplice, then Frank would be expected to try to shield Conley from questioning.

Luther Rosser had no doubts about the significance of Scott’s failure to inform him that Frank was instrumental in the discovery. At the unpaid-debt trial he stated, “In my opinion, as an attorney, it was material that I should have known before hand the information that the Pinkerton’s [ sic ] had that Leo Frank had said that Conley could write and that information should have been given me by the Pinkertons.” Under questioning by the attorney Harry A. Alexander, acting for the pencil company, Rosser explained how valuable this knowledge could have been for the defense:

Q. [Alexander] When Mr. Scott took his stand at the trial and testified that he ha[d] gotton [ sic ] the information about Conley writing from sources entirely disconnected from the pencil factory, would it or would it not have been material to you?…You could have disproved it by their own reports, couldn’t you[,] if Mr. Scott, had—

A. [Rosser] If they had reported to me, I could have shown it in their reports, of course.

Q. Yes, if it was in that report that they had got it from Leo Frank?

A. If they had given me that information I could have just handed it up to him [Scott], and said: “What did you report that to me for?”

Evidence offered by the National Pencil Company lawyers shows that Scott edited Pinkerton documents to remove any mention of Frank. This evidence consisted of two copies of Whitfield’s report for May 16, 1913—a draft in Whitfield’s hand and the final typed version. In the handwritten version the following words were crossed out: “but that he would sent [ sic ] to the tower and learn from Leo Frank if Conley could write.” These words did not appear in the typed version submitted to Frank’s lawyers, and it was Harry Scott who had edited Whitfield’s draft report.

Once Conley became a prime suspect, he eventually made four different statements, a process that involved Harry Scott more than any person except Conley himself. At Frank’s trial Scott claimed that he coaxed Conley to write by dictating “That long, tall, black negro did by himself,” words similar to those on the murder notes. Scott explained: “We [Scott and John Black] talked very strongly
to him, and tried to make him give a confession[;] we used a little profanity, and cussed him, and he made that statement that he knew that I knew that he could write; we talked for about 2 or 3 hours that day. He made another statement on May 24th, which was put in writing.”

On the basis of the second statement, Scott and Black “questioned him [Conley] very closely for about 3 hours” and again the next day, but Conley stuck to his story. In this statement Conley claimed that Frank had paid him to write the murder notes on Friday, April 25, the day before the killing. Scott continued: “We saw him [Conley] again on May 27th in Chief [Newport] Lanford’s office. Talked to him about 5 or 6 hours. We tried to impress him with the fact that Frank would not have written those notes on Friday, that was not a reasonable story, that showed premeditation and that wouldn’t do.”

On May 28, 1913, Scott, joined this time by Chief Lanford, “grilled” Conley for “5 or 6 hours, endeavoring to make clear several points which were far fetched in his statement; we pointed out to him [Conley] that his statement would not do, and would not fit. He then made us another long statement on May 28th.”

This was Conley’s third statement. On the very next day, May 29, Scott and another person, most likely either John Black or Chief Lanford, spent “almost all day” talking with Conley in an attempt to improve on it. As Scott explained, “we pointed out things in his story that were improbable, and told him he must do better than that, anything in his story that looked to be out of place, we told him wouldn’t do; after he had made his last statement, we did not wish to make any further suggestions to him at that time; he then made his last statement on May 29th.”

MORE THAN TWO YEARS LATER, when he was confronted with these statements by the National Pencil Company attorneys, Scott explained that he was only trying “to make Conley confess that he killed the girl. That was
my idea, and I put most unusual efforts in that line. The affidavits that I took from Conley were taken to make him confess that he committed the crime himself.”

This strains credulity. Scott’s own words give ample evidence that rather than induce Conley to confess that he alone murdered Mary Phagan, Scott was working with Conley to produce a statement that would convict Frank of the killing and portray Conley as a paid accomplice after the fact. Had Scott and the others not pushed Conley on his several statements, it is very likely that Conley rather than Frank would have been found guilty.

The sequence of events lends further support to the hypothesis that Scott and the prosecution were working singlemindedly to establish Frank’s guilt. Conley made his second statement on May 24, 1913, the same day that Frank was indicted for the murder. If Scott was really trying to get Conley to confess, the indictment could have been delayed pending the results of his interrogation. Rather, they had no doubt already made up their minds; Scott could not have been ignorant of the proceedings of the grand jury as he testified before it on the very day Frank was indicted.

THE TRIAL OVER THE UNPAID BILL OFFERS further evidence about Scott’s motives. According to H. B. Pierce, superintendent of Pinkerton’s Atlanta office, Scott failed to broaden the investigation to include Conley but “was entirely interested in developing the Frank proposition.” Pierce testified that he and Scott “clashed very often” and “had several discussions on some matters [associated with the investigation] bordering at time[s] on quarrels.” Pierce went on to say that Scott “was influenced by public opinion, he was of the opinion that if so many people saw it that way, that is the way the case was being developed, [and] that, in his opinion, must be right, against all other facts or anything else, regardless of [the] facts.” Pierce, on the other hand, thought that public opinion could be wrong and that if there were differences between the Pinkerton agency and the police, Pinkerton should “go on making [the] investigation for our client, regardless of the theories of the police department or anybody else, or we would quit.”

The reason that Scott could go against the wishes of Pierce, his supervisor, was that the Pinkerton hierarchy was on his side. Pierce testified that “Mr. Scott had the weight of opinion both with his superiors and with himself. By superiors, I mean his general superintendent and his other superior officers. Mr. Scott, was then in correspondence with the officers higher than myself and his course in working on the Frank angle met [with] their approval.”

Pierce queried his immediate superior A. S. Cowerdin. “I went over the case with him in detail and explained my views. He very politely replied that from the revealed facts and the reports that had been submitted, and that were being rendered, it was his opinion that the investigation was being carried on in a proper way. He disagreed with me and agreed with Mr. Scott.”

Pierce was not the only one in the local Atlanta Pinkerton office who differed with Scott over the direction of the investigation. Leo Gottheimer, the National Pencil Company salesman who had been sent to ask Frank whether Conley could write, testified about conversations he had had with Whitfield and McWorth. According to Gottheimer, the two investigators “said as to the relations between themselves and their superiors…there seemed to be friction with their superiors, everything they done…they told me they would not accept their theories, and they seemed to be tickled to death to get this new evidence…[that Conley could write. Gottheimer quotes them as saying] ‘We have got the goods now, they can’t deny this, we can prove this on them in such a way that they can’t deny it.’”

Herbert Schiff, Frank’s successor as factory superintendent, also testified about tensions in the Pinkerton’s Atlanta office. During several of the many visits that Whitfield and McWorth made to the factory, “they told me of the dissention [ sic ] in the office, and of the things that they put up that never seemed to agree with Mr. Scott, and Whitfield told me on one occasion that Mr. Scott called him into the private office and told him that if Leo Frank wasn’t convicted it would be the last of the Pinkerton agency in Atlanta.” Twice under crossexamination and twice more under re-direct examina- don—four times in all, and all under oath—Schiff insisted that Scott had stated explicitly that Pinkerton wanted Frank to be found guilty.

In itself this document, the Brief of Evidence, which has for so long lain dormant, does not prove guilt or innocence. It does, however, add substantially to the evidence that Leo Frank did not receive a fair trial. In fact, the conclusion that he was railroaded is now inescapable.

Whatever his reasons, Harry Scott was a key figure in convicting Frank of murder. Less certain, but still highly suggestive, was the malign role played by the prosecutor, Hugh Dorsey. Here ambition was certainly a motive, and a successful one, for Dorsey was twice elected governor of Georgia. This document strongly suggests that Dorsey urged witnesses to embellish their testimony, even lie under oath, to build a case against Frank.

The picture that emerges from this civil trial over an unpaid bill is of a conspiracy between the prosecutor Hugh Dorsey and Harry Scott of Pinkerton’s National Detective Agency to find Leo Frank guilty of murder. Although we will almost certainly never know just what was said between Dorsey and Scott, their collaboration seems to have assured that Leo Frank would not receive a fair trial for a crime he almost certainly did not commit.

Enjoy our work? Help us keep going.

Now in its 75th year, American Heritage relies on contributions from readers like you to survive. You can support this magazine of trusted historical writing and the volunteers that sustain it by donating today.