Case No. 10-20763-1-CR-LENARD(GRAHAM)/GOODMAN.


United States District Court, S.D. Florida, Miami Division.

January 10, 2018.

Attorney(s) appearing for the Case

Henry De Jesus Lopez Londono, also known as Mi Sangre also known as Salvador also known as Carlos Mario also known as Brother also known as Krackin also known as Federico, Defendant, represented by Arturo V. Hernandez , Arturo V. Hernandez & Ramon Arturo Hernandez , Courthouse Center.

USA, Plaintiff, represented by Michael Brian Nadler , Assistant United States Attorney, Jared Edward Dwyer , GREENBERG TRAURIG, P.A. & Robert J. Emery , United States Attorney's Office.


Defendant Henry De Jesus Lopez Londono has filed several motions to dismiss the superseding indictment. [ECF Nos. 334; 341; 437-38]. The Undersigned previously issued a Report and Recommendations on Londono's first two motions to dismiss the superseding indictment. [ECF No. 403]. That Report and Recommendations was filed under seal because the motions were filed under seal. Since then, however, Londono submitted a "re-filed" motion to dismiss the superseding indictment and a "supplemental" motion to dismiss the same indictment. [ECF Nos. 437-38]. Londono did not file the motions under seal, although he did redact parts of the re-filed motion.

For the reasons set forth below, the Undersigned respectfully recommends that the District Court deny Londono's motions to dismiss the indictment. As part of the recommendation, the Undersigned rejects Londono's request that I consider polygraph evidence in connection with a pre-trial evidentiary hearing held on his motions because I find it inadmissible. I likewise recommend that the District Court not permit polygraph evidence at trial (should Londono seek to admit it).

I. Procedural and Factual Background (Before the Two Evidentiary Hearings)

This Report and Recommendations will encompass all the motions to dismiss the indictment that have been referred to the Undersigned (and which have not yet been the subject of a Report and Recommendations). [ECF Nos. 334; 341; 437-38].

The arguments in all the motions to dismiss are similar. They concern, in broad terms, Londono's public authority defense. More specifically, they address Londono's arguments that (a) the United States "destroyed" the BlackBerry phones federal law enforcement agents used to communicate with him during his cooperation, and (b) the United States has falsely denied that Special ICE Agent Frank Burrola communicated with him by email (and that the United States destroyed the emails themselves).

Londono contends that the Government's "destruction" of five BlackBerry phones undermines his ability to prove his public authority defense. He likewise argues that the unavailability of the emails (caused by what he says is the Government's bad faith and false representation that Agent Burrola never communicated with him using email) also hinders his ability to adequately establish the public authority defense.1

Because I have already evaluated and ruled upon many of Londono's arguments, this Supplemental Report concerns only the additional information provided by the United States (in a Court-ordered submission) and new (or purportedly new) arguments asserted in Londono's supplemental motion to dismiss.

In the earlier-issued 21-page Report and Recommendations, I made several recommendations, but I also required the United States to submit additional information. Specifically, I required the United States to disclose whether any of the agencies involved instructed any agents to not submit BlackBerry phones for destruction or discarding pending the completion of the Londono investigation and prosecution, despite the agency-wide updating of BlackBerry phone equipment. [ECF No. 403, p. 19]. And I required the United States to explain whether anyone discussed the concept of a "litigation hold" for the five BlackBerry phones at issue. [ECF No. 403, p. 19]. That Report also required both sides to submit supplemental memoranda on spoliation arguments concerning the BlackBerry phones. [ECF No. 403, p. 20]. Finally, I recommended that the District Court postpone a ruling on these aspects of the dismissal motion until I issued a follow-up recommendation. [ECF No. 403, pp. 20-21].

The parties did not follow through on these assignments because they agreed that Londono would be entering a guilty plea. That did not occur, however. At the October 10, 2017 change of plea hearing, the District Court did not accept the Rule 11(c) plea negotiated by the parties. [ECF No. 439]

The next day, Londono submitted his "re-filed" motion to dismiss and his supplemental dismissal motion. [ECF Nos. 437-38]. The supplemental motion focused on what Londono described as "[t]he issue of the lost emails," which he explains is "among the most hotly litigated [issues] in this case." [ECF No. 438, p. 2].

Londono's supplemental motion emphasized three items of so-called "new evidence": (1) a business card which he says ICE Special Agent Burrola gave to him during their first meeting in Cartagena, Colombia in the Fall of 2009; (2) the results of a private polygraph examination he took on the issue of whether he communicated with Agent Burrola via email; and (3) a "Blackberry analysis" performed by a Colombian tech company, which, according to Londono, shows that Agent Burrola's assertion about receiving a BlackBerry message from Londono on January 15, 2010 "cannot possibly be true" because Londono did not even "download the BlackBerry Messenger app for another 26 days." [ECF No. 438, pp. 4-8].

Agent Burrola wrote a live.com email address on the business card and a password, "Panchito," which Londono says is a common nickname for Francisco. [ECF Nos. 438, p. 4; 438-1]. His supplemental motion also contends that the emails were never technically "sent" in the traditional sense because he and Agent Burrola "would compose drafts and then each could access the saved drafts when logging into the email account with the password." [ECF No. 438, p. 4]. The supplemental motion represents that Londono and Agent Burrola would exchange messages through this email account "almost daily" and that defense counsel's analysis of the email account "reveals that it has been permanently disabled." [ECF No. 438, p. 5]. Therefore, the motion argues, "the act of disabling the email account destroyed all the communications that were ever associated with this email address." [ECF No. 438, p. 5].

Londono's supplemental motion also attached a copy of a polygraph examiner's report. [ECF No. 438-2]. It concluded that Londono's answers (i.e., that he communicated with Agent Burrola using the email account provided by the agent on the business card) had a "resounding 99.42% probability for truthfulness." [ECF No. 438, p. 6]. Thus, Londono contends that "[i]n view of these test results and the emergence of the business card, Agent Burrola simply cannot be believed on the question of whether he communicated with Defendant via email." [ECF No. 438, p. 6].

Concerning the "Blackberry analysis," Londono's supplemental motion argues that "Agent Burrola has provided a version of events that cannot possibly be true in order to further cast doubt on the fact that federal agents communicated with Defendant via email so that they could avoid having to admit that a second class of critical exculpatory evidence had been destroyed." [ECF No. 438, pp. 7-8].

Following Londono's submission of his re-filed and supplemental dismissal motions, the Undersigned then ordered the United States to complete the assignments it was given in the earlier Report and Recommendations and also required both sides to submit separate memoranda of law on the issues specified in the Report and additional issues raised in the supplemental motion to dismiss the indictment. [ECF No. 450, pp. 1-2]. That Order also required the parties to discuss their position on Londono's request for an evidentiary hearing. [ECF No. 450, p. 2].

The United States submitted a supplemental response, which responded to the follow-up information required by the initial Report and Recommendations. [ECF No. 464]. In that supplemental response, the United States unequivocally advised that "[n]o agents communicated with the defendant via email." [ECF No. 464, p. 2]. It also represented that "no government officials, e.g. special agents, supervisory special agents, or prosecutors[,] discussed the need to implement a `litigation hold' for the five BlackBerry phones at issue." [ECF No. 464, p. 7].

Londono and the United States submitted the additional memoranda. [ECF Nos. 474-75]. The Undersigned scheduled an evidentiary hearing, which took place on October 30, 2017. [ECF No. 493]. The evidentiary hearing was scheduled to obtain additional information on one issue: whether Londono and Agent Burrola communicated through an email account. Londono contends that he certainly did, and the United States represents that he absolutely did not.

II. The BlackBerry Telephones: Facts and Legal Analysis

Londono argues that the United States had a duty to preserve the Blackberry messages and emails used to communicate with him while he served as an undercover informant in Colombia. Of course, the United States argues that it could not have been under a duty to preserve non-existent emails because agents never communicated with Londono by email.

In its memorandum, the United States asks the Court to reject Londono's claim that the Government purposefully and in bad faith destroyed five Blackberry phones. The United States argues that Londono's presentation is missing "the most obvious fact": that he had "equal access, custody, and control of all these communications and at least at one point had an exact copy of every conversation in his possession." [ECF No. 475, p. 2]. The Government says that the law concerning the Government's purposeful destruction of evidence in bad faith is inapplicable when the evidence was "never in the sole custody and control of the Government." [ECF No. 475, p. 2]. In addition, the United States argues that the notion of its intentional destruction of the five phones to prevent a review of BlackBerry messages is "absurd on its face" because Londono would have equal custody and control of all of those messages and "no one from the Government could have known the defendant destroyed his copies of all the messages." [ECF No. 475, pp. 2-3].

Thus, the Government reasons, "it is bewildering how the defendant can argue that although I destroyed my copy of all the evidence without any government official ever knowing that, the government acted in bad faith when it destroyed their copy of the messages in the normal course of business." [ECF No. 475, p. 3]. The United States notes that Londono admits to purposefully and intentionally destroying his BlackBerry messages "to protect his safety, which may or may not be true." [ECF No. 475, p. 4]. And the United States also points out that Londono "did not disclose this little tidbit of information until the defendant's motion to dismiss was filed." [ECF No. 475, p. 4].

To establish that the government's loss of evidence constituted a denial of due process, Londono must show (1) that the evidence was likely to significantly contribute to his defense and (2) bad faith on the part of the government. United States v. Revolorio-Ramo, 468 F.3d 771, 774-75 (11th Cir. 2006).

Londono disagrees with this legal standard. He contends that there is no bad faith requirement here. He argues that the Eleventh Circuit in Revolorio-Ramo incorrectly conflated a rule from California v. Trombetta, 467 U.S. 479 (1984) and mistakenly included a bad-faith-by-law-enforcement requirement even when the destroyed evidence meets the standard for constitutional materiality. [ECF No. 545]. But Revolorio-Ramo is the law in this Circuit.

If Londono wants to prevail on this argument, then he will need to convince the Eleventh Circuit Court of Appeals or the United States Supreme Court. The Undersigned cannot overrule or reverse or even ignore binding Eleventh Circuit precedent merely because a party contends that the Eleventh Circuit reached the wrong result in a published opinion. Londono argues that "this Court should reject" Revolorio-Ramo [ECF No. 545, p. 3], but cites no authority to support this extraordinary request to unilaterally establish a new legal rule in contravention of binding circuit law.

So the Undersigned declines Londono's invitation and will instead follow Revolorio-Ramo.

Under the first prong, the defendant must demonstrate that the evidence possessed an exculpatory value that was apparent before the evidence was destroyed and that the evidence was of such a nature that he would have been unable to obtain comparable evidence by other reasonably available means. Revolorio-Ramo, 468 F.3d at 774. The mere possibility that the lost evidence might have been favorable to the defense does not satisfy the constitutional materiality standard; where the record reveals that "the chances are extremely low that preserved [evidence] would have been exculpatory," destruction of that evidence does not render a conviction fundamentally unfair. California v. Trombetta, 467 U.S. 479, 489 (1984) (holding that the due process clause did not require law enforcement agencies to preserve breath samples of suspected drunk drivers for the results of the breath-analysis tests to be admissible in criminal prosecutions).

Under the second prong, a showing of bad faith requires the government's conduct to reach the level of an "intentional bad faith act." Revolorio-Ramo, 468 F.3d at 775.2 A showing of negligence, even gross negligence, is insufficient to establish bad faith. See United States v. Gayle, 608 F. App'x 783, 790 (11th Cir. 2015) (affirming denial of defendant's motion to dismiss absent "evidence that the police purposely destroyed" evidence); see also United States v. Toro-Barboza, 673 F.3d 1136, 1150 (9th Cir. 2012) (holding that although "better practice" would have dictated that the government preserve certain evidence for the defendants to test, "[b]ad faith requires more than mere negligence or recklessness."); United States v. Spalding, 438 F. App'x 464, 466 (6th Cir. 2011) (holding that destruction of evidence did not constitute a deprivation of due process where "conduct of the government agents was, at most, negligence, but it was not bad faith or reckless."); United States v. Femia, 9 F.3d 990, 995 (1st Cir. 1993) (holding the same where tapes were destroyed due to gross negligence, and not bad faith).

Rather, bad faith is present "if the officer destroyed the evidence `in a calculated effort to circumvent the disclosure requirements established by Brady v. Maryland.'" United States v. Cruz, 508 F. App'x 890, 901 (11th Cir. 2013) (quoting Trombetta, 467 U.S. at 488).

In addition, the requirement of bad faith applies even if evidence is destroyed after the defense files a formal discovery request seeking access to the evidence for testing purposes. Illinois v. Fisher, 540 U.S. 544, 548 (2004) ("We have never held or suggested that the existence of a pending discovery request eliminates the necessity of showing bad faith on the part of police.").

The bad faith necessary for Londono to demonstrate why the indictment should be dismissed cannot be presumed. Revolorio-Ramo, 468 F.3d at 774-75. Establishing bad faith requires a defendant to prove "official animus" or "a conscious effort to suppress exculpatory evidence." United States v. Wright, 260 F.3d 568, 571 (6th Cir. 2001) ("Because we find no bad faith on the part of fire investigators who failed to preserve the potentially useful electrical evidence, we . . . reject [the defendant's] due process claim without further analysis.").

The United States conceded that it has been unable to locate any legal authority pinpointing the nature of Londono's burden of proof for his motion to dismiss the indictment based on the Government's destruction of evidence (in general) and based on the Government's admitted discarding of the BlackBerry telephones. [ECF No. 544]. Nevertheless, the United States analogizes this case to other scenarios and argues that Londono must meet the "clear and convincing" standard, rather than the less-burdensome "preponderance of the evidence" standard. [ECF No. 544, pp. 2-3]. Londono did not brief the issue even though the Undersigned specially directed the parties to address it in the post-hearing memoranda they submitted. [ECF No. 537]. Rather, Londono's memorandum in response to the Undersigned's Order focused on other issues. [ECF No. 545].

Many of this Circuit's decisions on spoliation of evidence arise in civil cases, where there also does not appear to be a binding appellate ruling on the applicable burden of proof. The Eleventh Circuit has not yet decided the appropriate evidentiary standard to use when the requested sanctions are based upon the Court's inherent powers. Nevertheless, the Undersigned finds persuasive a decision by U.S. Magistrate Judge Andrea Simonton in In re Brican American LLC Equipment Lease Litigation, 977 F.Supp.2d 1287 (S.D. Fla. 2013), report and recommendation adopted, 2013 WL 12092311 (S.D. Fla. Nov. 15, 2013).

In Brican, the Court adopted two different evidentiary burdens, depending on the nature of the sanction imposed. For "issue-related" sanctions, defined as "those that are fundamentally remedial rather than punitive and do not preclude a trial on the merits," the proof must be by a preponderance of the evidence. Id. at 1293 n.6 (quoting Compton v. Alpha Kappa Alpha Sorority Inc., 938 F.Supp.2d 103, 104-05 (D.D.C. 2013)). But for "fundamentally penal" sanctions, defined as "dismissals and default judgments, as well as contempt orders, awards of attorneys' fees, and the imposition of fines," the clear and convincing standard is used. Id. (quoting Compton, 938 F. Supp. 2d at 104-05). Accordingly, in evaluating witness-tampering allegations, Judge Simonton applied the preponderance of the evidence standard where plaintiffs sought non-dispositive sanctions and applied the more-exacting clear and convincing standard to their request for dispositive sanctions.

The Undersigned, however, need not wrestle with the thorny standard-of-proof issue because I conclude that Londono failed to show that the BlackBerry phones were destroyed in bad faith even under the less-burdensome preponderance of the evidence standard. Londono has not established the requisite bad faith because the five BlackBerry phones were discarded as part of a standard operating procedure — when new phones are issued to agents or when an agent retires. It might have been especially prudent for the phones to have been preserved. It might have even been careless, cavalier, or negligent to have not preserved the phones. Either way, the Undersigned does not find bad faith.

The Undersigned finds persuasive the Government's practical point that bad faith intent cannot be established when the agents knew or presumed that Londono had his own cellphone and its stored text messages. See United States v. Webster, 625 F.3d 439, 447 (8th Cir. 2010) (declining "invitation to impose a lesser standard of culpability on the part of the government" because "a showing of bad faith [is needed] to sustain a due process claim based on the failure to preserve evidence."); see also United States v. Garza, 435 F.3d 73, 75-76 (1st Cir. 2006) (holding that authorization to destroy evidence to free up laboratory space, although "short-sighted and even negligent," did "not satisfy the requirement of bad faith" because there was no improper motivation); United States v. Brown, 9 F.3d 907, 910 (11th Cir. 1993) (holding that destruction of firearm that was not classified as evidence did not deprive defendant of due process because "officer's failure to reclassify the gun as evidence amounted to negligence, not bad faith."); United States v. Hinton, No. 2:07CR14-SRW, 2007 WL 2670038, at *4 (M.D. Ala. Sept. 10, 2007) (denying motion to dismiss an information in a misdemeanor theft case because the government's decision to not retain the allegedly stolen articles was in good faith and in accordance with its normal practice of disposing of the allegedly shoplifted items).

The Undersigned is also not convinced that Londono's "Blackberry analysis" unequivocally demonstrates that the Government's version of what happened is inaccurate. The parties have not comprehensively briefed this fact-based theory, and the Undersigned concludes that it would be more appropriate for Londono to seek to introduce this evidence at trial (assuming that Judge Graham permits him to assert the public authority defense at the trial).

Londono, moreover, is not entitled to an adverse inference jury instruction concerning the lost BlackBerry phones because their destruction was part of a standard procedure to update the agents' cellphones. Londono has not submitted any legal authority holding that destroying evidence as part of a routine policy indicates bad faith. The case law is to the contrary. See Garza, 435 F.3d at 75-76 (holding that evidence destroyed in the course of implementing routine procedures militates against finding bad faith); United States v. Gomez, 191 F.3d 1214, 1219 (10th Cir. 1999) (holding that destruction of most of the marijuana on which defendant's prosecution was based, done per ordinary statutory procedures and prior to superseding indictment, did not violate discovery rule or due process).

By way of summary concerning the BlackBerry phones, the Undersigned recommends that the District Court deny his motion to dismiss the indictment, disallow any type of adverse inference jury instruction, but permit the parties to submit evidence about the use of the BlackBerry phones and their destruction (including testimony from Londono's IT expert from Colombia).

III. The Alleged Emails Between Londono and Agent Burrola: Factual Background

A. The First Evidentiary Hearing: Londono and Agent Burrola

Two witnesses testified at the first evidentiary hearing: Londono and Special Agent Francisco Burrola. The Undersigned will outline the significant highlights of their testimony.

i. Londono

Londono explained that he was in Argentina during the first half of 2009 and in Colombia for the second half of that year. He said he met federal law enforcement agents in Cartagena, Colombia in November 2009. According to Londono, the meeting's purpose was to inform the agents about his infiltration of a criminal gang called Banda de Uraba, otherwise known as the "Urabenos." Londono said that the meetings occurred in a Sofitel Hotel called the Santa Clara Hotel.

Londono said that several others attended the meetings: two Colombian attorneys who represented him; an attorney from the United States (who also represented him); Burrola; Special Agent Steve Monks; an ICE special agent he knew only by the name of Agent "Eric" (from ICE, and who was assigned to Cartagena); other agents from the DEA (including Jorge Rodriguez); and Sergio Adame (a man he said was providing some type of service to U.S. law enforcement). Londono said that he participated in three days of meetings in November 2009.

Londono said that his attorneys were not permitted to participate in the internal debriefing meetings, even though they were at the hotel and accompanied him to the meetings. According to Londono, Burrola told him that the topics were "hot" and "complicated" and that the infiltration-related tasks he was being given were "high risk" and that his life would be jeopardized if his cooperation or the assignments were to become known. [ECF No. 507, p. 11].

As Londono explained during his direct examination, Burrola told him to be "at ease" during the meetings because he and Londono's attorneys "had already communicated regarding everything which is public authority." [ECF No. 507, p. 11].

Londono testified that Burrola "identified himself with a business card" on two occasions and gave Londono a business card on the third day of the meetings. [ECF No. 507, pp. 11-12]. Londono said that Burrola wrote several things on the pre-printed business card. First, Burrola wrote the word "Panchito" and told Londono that it was his code or password to access an email account. Second, Burrola wrote a "live.com" email address: frankb_1811@live.com. Third, Burrola wrote a number "1" and number "2" on the card, the number "1" next to the handwritten email address and the number "2" next to Burrola's official, pre-printed government email address. Burrola said that his government address was the alternate email if the first one was not available. Fourth, Burrola's business card also contained the numbers "4028D9BB," which Londono said was the BlackBerry pin number.

Londono said that he was supposed to use the number "2" email address (the one already printed on the business card) once things were "normalized." [ECF No. 507, p. 15]. According to Londono, there was at the time an arrest warrant pending for him in Colombia, and Burrola told him that he needed to use the handwritten email address because of the arrest warrant. Thus, the first email address would be used once the arrest warrant was lifted, quashed, or dismissed. Londono said that the agents expressed concern about using an official email address when an arrest warrant was outstanding and about the increased possibility of a leak because the tasks that would be given to him were delicate.

Londono said that he could be anonymous by using the non-official email address handwritten on Burrola's business card. Londono said that he used this unofficial email address to communicate with Burrola approximately 50 to 70 times.

As explained by Londono, the unofficial email address would be "secure" when he used the system explained by Burrola. [ECF No. 507, p. 17]. Burrola instructed Londono to save an email message as a draft in the email system. That way, it would be impossible for someone to intercept an email because Londono would not actually send any emails to Burrola. Londono said that he was told the system was secure because no one would have access to the saved drafts or be able to create a draft without the password.

Londono testified that he could access the email from anywhere in the world, as long as he had the password.

Londono said that Burrola would respond to his saved draft emails by also creating and saving draft emails, to which Londono would then reply by using the same method. Londono said that he used this unofficial email system until he and Burrola switched from email to another form of communication in March 2010. Londono testified that from mid-2009 until March 2010, he used only this unofficial email system, where drafts would be saved to the handwritten email address with the password. During this period, he infiltrated the Urabenos. After March 2010, he used BlackBerry messenger to communicate with Burrola and the other agents.

Londono also said that Burrola's draft emails to him gave him either the specific tasks he was expected to perform for United States law enforcement or the general concepts the agents wanted him to pursue. Londono gave examples of some of the tasks, such as infiltrating "El Chapo" Guzman's well-known drug trafficking organization by establishing drug trafficking deals with Guzman and getting close to his employees.

Londono said that he used different computers, including laptops, to access the unofficial email system which Burrola instructed him to use. He said that the computers were calibrated for Spanish (because he does not speak English). He also explained that his computer calibrates accents for Spanish words. He referenced an exhibit received in discovery as a download of a message he sent to Burrola. Specifically, this exhibit is a January 15, 2010 email from Burrola to other federal law enforcement agents, saying, "[b]elow is the response I received" from "the source" (i.e., Londono) on questions from Bogota-based DEA agents. [ECF No. 494-2].

Londono testified that he left this email, which contained four paragraphs in Spanish, in the email draft system established by Burrola.

Londono said that he was found not guilty after a trial in Colombia and that the arrest warrant was then "lifted." [ECF No. 507, p. 27]. The lifting of the warrant impacted the agents' decisions about the methodology used to communicate with him. In March 2010, Londono said, he told Burrola that the Colombian arrest warrant had been lifted, and Burrola told him that their communications could then be "normalized" by using the BlackBerrys. [ECF No. 507, p. 27].

Although the email system he used with Burrola was, in a way, unofficial, Londono explained that he viewed the communications as official because he considered the methodology to be an official way to communicate with Burrola and other agents. Nevertheless, he considered this communication system to be irregular and not normal.

As explained by Londono, he could not preserve the emails because he left them in the email system as drafts. He stopped using the draft system after he and Burrola changed to the BlackBerry BBM system.

Londono said that he and Burrola agreed that Burrola would safeguard the emails because he, Londono, was unable to safely do so due to security concerns.

During cross-examination, Londono's demeanor changed, and he did not articulate his testimony in the same way. He appeared more fidgety and less relaxed. He said that the prosecutor incorrectly summarized Londono's testimony as saying that he had been in Argentina for the first half of 2009 — but that is exactly what Londono said. In any event, Londono said that he had two BlackBerry devices but denied activating one of them shortly after he met Burrola and the other agents in Cartagena in 2009. Londono said that BBM messages were common in 2009 and that he understood that his telephone communications were downloaded.

Londono also acknowledged that his BlackBerry phone could access emails via the internet and conceded that there was not one email he sent to anybody from his BlackBerry. Londono said that he did not use the web browser on his BlackBerry to access emails to and from Burrola. Instead, he said, he used several computers to check on the email account that Burrola established. Londono said that he did not necessarily use only his own computers, but also used computers in hotels and internet cafes. He said that he no longer has the laptop he used at the time.

During cross-examination, Londono said that he did not save his emails between himself and Burrola because he was concerned for his own safety. Londono conceded that he had two attorneys from Colombia, one attorney from Argentina and one attorney from the United States, but did not forward to any of them his emails so that they may save them. In addition, Londono said that his concern for his own security was also why he deleted his BBM messages. Londono, however, also conceded that he saved at least two of the communications he had with U.S. law enforcement agents and may have saved more than two.

Londono also said that he turned over two telephones he said he had been using to communicate with agents.

On redirect, Londono said that he did not send emails to his attorneys because they used cellphones to communicate.

ii. Agent Burrola

Agent Burrola confirmed the existence of a response from Microsoft in connection with a subpoena for records for the frankb_1811@live.com email address. The response was that the account does not exist now and that there is no user at that location. Agent Burrola said that he does not know how that email account, which he created, was disabled or who disabled it. He denied disabling it himself.

In November 2009, Burrola was a field intelligence director, with nine analysts working for him. He said it was not common for him to work directly with sources in 2009, but explained that he had done so before 2009.

Burrola said that the type of communications used with sources varies with the nature of the source. For example, he would not use a home telephone or a cellphone with a source located in Mexico. He has used the draft-email system with other sources. He said that he focuses on security and the country where the source resides as the primary factors when deciding on the method of communication.

Burrola said that he met with Londono in November 2009 because Londono wanted to cooperate as a confidential informant for ICE. He did not know before he met Londono why he wanted to cooperate. Their first meeting was in Cartagena, and he had no communication with Londono before that first meeting. Burrola wanted to debrief Londono to learn the reasons for his desire to cooperate, what Londono was looking for in exchange, and what information he had to offer.

According to Burrola, he met with Londono and another informant nicknamed "Superman" and two DEA agents, who joined them on the second day. [ECF No. 507, p. 75]. He thinks the meetings in Cartagena with Londono lasted two days, not three days.

Burrola said that he decided to use BlackBerry "Messenger" as the form of communication with Londono. [ECF No. 507, p. 76]. He said the decision was based, in part, on a recommendation from Londono, who said it was easier for him. Burrola said that BlackBerry Messenger would be easier for him, as well. He also explained that Londono moved around a lot, which would make it difficult for him to use an internetbased system. Burrola said that he gave his business card to Londono. He also confirmed all the handwriting on the pre-printed business card as his.

He explained that he and Londono could resort to the draft email model if there was a breakdown in the BlackBerry method. Burrola explained that both email addresses were intended to be used as backups for the BlackBerry system.

Confronted with Londono's claim that he used the email draft system with Londono during his cooperation, Burrola said that, based on his recollection and review of the file, he did not think he ever used the email draft system with Londono. And he said that he certainly would not have used it 70 times because he surely would have recalled that level of usage.

Moreover, Burrola said that he would have saved the draft emails and forwarded them to his goverment email account. He also explained that he could not, because of technical limitations in 2009, have forwarded BlackBerry BBM messages to himself. Therefore, he explained, he had to pull out language and paste it into an email if he wanted to forward a BlackBerry BBM message.

Burrola said that he never used the frankb_1811@live.com email address. He also said that he never told Londono that he was not to safeguard messages, nor did he ever tell Londono that he (Burrola) would safeguard the messages. And he also explained that he was not aware of Londono's claim that he destroyed messages exchanged with Burrola. From 2009 through October 2010, Burrola testified, he used BlackBerry BBM to communicate with Londono.

Under the prosecutor's questioning, Burrola denied the claim that he told Londono to destroy the BlackBerry BBM messages because of concerns over safety. In fact, Burrola said that the morning of the hearing was the first time he ever heard of this allegation (when one of the lead Assistant U.S. Attorneys mentioned it to him).

Burrola said that Londono gave him information, and he then passed it on in emails to others in law enforcement. He explained that he sometimes used the word "internet" in emails to describe communications with Londono because he did not want others to know that they were using BlackBerry BBM as their method of communication. [ECF No. 507, p. 82]. According to Burrola, he did not want to reveal that Londono was using BlackBerry BBM because Londono was in Colombia and Burrola was concerned for his safety. Therefore, Burrola explained, he "buried" Londono's use of BlackBerry BBM by strategically using the incorrect term "internet." [ECF No. 507, p. 82].

Burrola said that he would periodically receive requests for information from other agents and he would then place them in a BBM message and send the request on to Londono, who would respond in Spanish via BlackBerry BBM.

Burrola explained that he is not denying that he communicated with Londono; he said that if he did use the email draft function, then he would readily admit it.

On cross-examination, Burrola acknowledged that his email response to a November 4, 2009 email said, "We will have the CIs communicating via internet" [ECF No. 494-3] — a term he said he used to "compartmentalize the information." [ECF No. 507, p. 95]. He said that he used the word "internet" imprecisely because he wanted to protect Londono in case someone intercepted the communication.

According to Burrola, the email draft function methodology would also be used to protect information. He admitted that someone could still capture a message put on the saved draft system.

Burrola said that the email address he wrote on the business card was created for the meetings with Londono, but he was not sure if the email address was created beforehand. He said that "Panchito" is short for Francisco, but it is not his nickname. Regardless, he said it was the password for the frankb_1811@live.com email address. Burrola explained that Londono could use the password if he needed to use the email system in case the BlackBerry BBM system failed.

Burrola said that he did not remember if he knew about Londono's Colombian arrest warrant at the time. He said the crime charged was a financial one and that the charges against Londono were dropped several months later. He denied being worried about being embarrassed if it were known that he and other agents were communicating with a man named in a Colombian arrest warrant.

Burrola repeated his answer that, based on his "recollection and the information that I have at hand," email was never used to communicate with Londono. [ECF No. 507, p. 101].

Londono's attorney questioned Burrola about a May 3, 2010 document in which he said, "I would need to broach him with the testifying questions, which may be something we personally do and not on email." [ECF No. 494-7]. According to Burrola, when he wrote "email," he was actually referring to BlackBerry BBM messenger. [ECF No. 507, pp. 103-04]. He said that he was always concerned for the safety of an informant, like Londono, and he purposefully used the word "email" incorrectly when referring to BBM.

Londono's attorney also questioned Burrola about an October 4, 2010 email that he sent to another federal law enforcement official. In that email, Burrola said, "he [referring to Londono] was provided direction (emails on file)." [ECF No. 494-6]. Burrola said that the phrase "emails on file" refers to emails he received from other agencies. [ECF No. 507, p. 105]. He specifically denied that the directions were sent to Londono by email. He said that the agencies would contact him by email with requests for information to obtain from Londono. Burrola said that he would then provide those requests to Londono in a BBM message.

Burrola was also shown a March 24, 2010 email that he wrote to James Woosley, the ICE director of Intelligence. In that email, Burrola wrote, "Also, I am now in direct communication with the tentative CI via Blackberry Pin. I informed the tentative CI to begin normal business operations [.]" [ECF No. 494-5 (emphasis added)]. Londono's counsel suggested that this email meant that Burrola's earlier communications with Londono were by email, but Burrola rejected that interpretation. Similarly, Burrola denied that the phrase "normal business operations" meant using the official email address. [ECF No. 507, p. 112].

In addition, Burrola explained that he was returning from Haiti in March 2010 and had not been working on the Londono cooperation project much while he was gone. He said Londono had not been his focus while he had been in Haiti for three months. Therefore, Burrola said, "normal business operations" simply referred to the reality that he would be back in country after three months in Haiti. [ECF No. 507, p. 112].

Burrola explained that his goal with Londono had run awry and he had received a direct order from the director to resume the cooperation immediately.

Londono's counsel spent a significant amount of time questioning Burrola about a January 15, 2010 email he had sent to other law enforcement agents in Colombia. The email first said, in English, "Below is the response I received [from] the source on the questions posed by the DEA in Bogota. If you would please pass on or as you deem appropriate." [ECF No. 494-2].

Below that English language introduction is a four-paragraph narrative in Spanish — but in a font different from the font used in his first two introductory sentences in English. [ECF No. 494-2]. Burrola denied the suggestion that he copied and pasted these four paragraphs from an email sent to him by Londono. Instead, he said that he himself retyped them and referred to the fact that quotation marks appeared before the first word and after the last word as evidence that he retyped it.

Londono's counsel pressed him on the fact that in other emails he wrote in which he relayed information received from Londono, he wrote the email in English and provide a summary (not a verbatim retyping in Spanish). Burrola was adamant that he did not cut and copy a Spanish-language email from Londono and then paste it into an email sent to other agents. The email in question was sent to two agents (Luis Sierra and Fernando Plascencia) who speak Spanish.

Referring to the business card he gave Londono during the Cartagena meetings, Burrola said that he created the email address for that specific meeting but did not disable it himself. He conceded that the email account would not have disabled itself and said that he does not recall what happened to the email account.

Burrola said that he never checked the email account (to see if Londono had saved a draft email there for him to access). He also said that he does not know who disabled the account. He said that he never looked into the issue to see who disabled that email account he created solely for Londono's use.

However, on redirect, Burrola explained that Microsoft owns "live.com" and that he is not familiar with Microsoft's policies on disabling email accounts. [ECF No. 507, p. 121]. He emphasized that if he had communicated with Londono, then he would have forwarded those communications to his government email account and catalogued them. Burrola explained that he could not at the time forward BlackBerry BBM messages to his email because of a technology-based limitation.

As explained by Burrola on redirect, his concern over the confidentiality and security of emails varied for emails sent domestically, as opposed to emails sent internationally. He said that he had received information that criminal organizations had obtained equipment to capture communications sent on internet-based email accounts. This answer appears to be the Government's explanation about why Burrola would have been willing to forward communications to his email account even though he was concerned about exchanging emails with Londono in Colombia.

Burrola said that there was no need to use the email draft system with Londono because they were always able to use BBM Messenger. He noted that as of the November 2009 meeting in Cartagena, Londono was already using BBM to communicate with a source named "Superman." [ECF No. 507, p. 125].

B. The Second Evidentiary Hearing: The Daubert3 Hearing on Polygraph Evidence

Londono wanted to have his polygraph expert testify at the first evidentiary hearing, but the United States argued that the notice was inadequate and it objected. Londono had filed a memorandum of law in support of his request to admit polygraph evidence less than four hours before the hearing. [ECF No. 492].

In his eleventh-hour, pre-hearing memorandum, Londono acknowledged that courts historically excluded polygraph results in jury trials, but he cited United States v. Piccinonna (Piccinonna I), 885 F.2d 1529 (11th Cir. 1989) for his argument that "polygraph evidence is not per se inadmissible" in the Eleventh Circuit. [ECF No. 492, p. 3]. Londono argued that "advancements in polygraph technology" lead to the conclusion that his polygraph examination should be admitted "for limited purposes only at this pretrial evidentiary hearing." [ECF No. 492, p. 3]. Londono's memorandum conceded that "numerous courts in this Circuit have found polygraph evidence insufficiently reliable and unscientific" [ECF No. 492, p. 8], but emphasized that the request is for a narrow, non-trial use. Specifically, Londono argued that "because this is a pretrial hearing before a Magistrate Judge as the fact finder there is no risk that the polygraph evidence will be given undue weight or influence." [ECF No. 492, p. 13].

I did not permit the polygraph expert, George Harper, to testify at that first evidentiary hearing. But I gave the United States the opportunity to arrange for Londono to be examined again by a Government-selected polygrapher. [ECF No. 495]. The United States declined the opportunity but asked for a Daubert hearing on the admissibility of Harper's expert witness report and opinions concerning Londono's performance on the defense-arranged polygraph. [ECF No. 510].

The Court held a Daubert hearing on December 14, 2017. [ECF No. 526]. Before the hearing, I required Londono to produce to the United States all the data and information from Harper's polygraph examination. [ECF No. 495].

At the start of the Daubert hearing, the United States announced two challenges to Harper's report and opinions. First, it challenged polygraph evidence in general, contending that it is too unreliable to be presented as evidence in federal court. Second, regardless of whether polygraph evidence could ever be admissible in federal court for any purpose, the Government took the position that this specific test, administered by this particular polygraph examiner, under the precise circumstances surrounding the defense polygraph, was inadmissible — even if it were to be used only at an evidentiary hearing before a federal magistrate judge, as opposed to being presented to a jury.

Two witnesses testified: Harper and DEA Special Agent Brent Barnes.

i. George Harper

Before the hearing, Londono filed Harper's report and resume. [ECF No. 525]. Harper has a B.S. degree in criminal justice from Florida International University, which he earned in 1976. He attended the Metro-Dade County Police Department-sponsored polygraph training in 1978 and attended continuing studies through the Florida Polygraph Association (where he became a member in 1978) from 1987 to 2016.

Harper has been the owner and president of a polygraph company since 1978. His resume does not reflect the publication of any books, treatises, articles, or other publications concerning polygraphs. It does not reveal any lectures or presentations that Harper has given about polygraphs, though it does show that he attended seminars where others lectured.

According to Harper's report, the polygraph was designed to determine the truthfulness of Londono's statements regarding "whether or not he actually sent email correspondence to agents of the federal government on more than fifty occasions while he was a cooperating witness between November 2009 and February 2010." [ECF No. 525-1, p. 2].

The report explains that Harper asked Londono four relevant questions4 and that he answered "yes" to all of them: (1) "Between the dates of November 2009 and February 2010, did you communicate with ICE Agent Francisco Burrola via email using the account frankb1811@live.com?"; (2) "At your first meeting with U.SA. [sic] law enforcement in Cartagena, Colombia, did ICE Agent Francisco Burrola provide you his business card with the email address frankb 1811@live.com?"; (3) "Between the dates of November 2009 and February 2010, did you and ICE Agent Francisco Burrola exchange more than fifty messages using the email account frankb 1811@livecom?"; (4) "Regarding the message included on Document 000343, did you send that message in an email?" [ECF No. 525-1, p. 4].

The "Test Results" section of Harper's polygraph report says "it is the opinion of this examiner that the subject was truthful in his responses." [ECF No. 525-1, p. 4]. Specifically, the report concluded that "the subject's probability of truthfulness is 99.42 percent and probability of deceptiveness is 00.58 percent." [ECF No. 525-1, p. 4]. Harper's report further stated that the "the subject scored a plus (+) 7" on the Utah Empirical Scoring System, which he "concludes that there was no deception indicated." [ECF No. 525-1, p. 4].

During opening statements, Londono's attorney explained that Harper did not himself score the polygraph exam. Instead, he emphasized that it was scored and graded by the computer, a procedure which eliminates the examiner's discretion and supposedly enhances accuracy. On the other hand, the Government argued that electronic scoring is a new methodology and that the algorithm used by the computer has not been adequately tested.

1. Testimony on Direct Examination

Harper said that he has conducted 30,000 polygraphs and began conducting polygraphs when he was with the Miami-Dade County Police Department. Harper described the polygraph as "an emotional stress monitor." [ECF No. 529, p. 87]. He said that a polygraph is more likely to find an innocent person guilty than to find a guilty person innocent. (The Court construes this answer as meaning that the polygraph is more likely to find a person who gave truthful answers to be deceptive than to conclude that a person who gave false or deceptive answers to be truthful. That is because the question asked of Harper used the terms "innocent" and "guilty," rather than "deceptive" or "truthful." [ECF No. 529, p. 88]. Polygraph examiners usually use the terms "truthful" and "deceptive," rather than "guilty" or "innocent.")

Harper conducted the polygraph exam on Londono on August 17, 2017, in a conference room at the Federal Detention Center (the "FDC"). Harper said that he has conducted 40 to 50 polygraphs at the FDC. He was paid $800 for the exam and received an additional $1,000 for post-exam meetings. He was not asked whether he would be receiving additional compensation for providing testimony and opinions at the evidentiary hearing.

In Harper's opinion, the polygraph is 90 percent effective, based on research he has reviewed. Harper briefly discussed United States v. Mosquera, No. 8:14-cr-379-T-36TGW, 2015 U.S. Dist. LEXIS 177631 (M.D. Fla., April 9, 2015), where the magistrate judge mentioned Dr. David Raskin's testimony that laboratory research studies and field studies indicate that "a properly performed polygraph examination has a 90% accuracy rate." Id. at *8. As summarized in Mosquera, Dr. Raskin conducted laboratory and field research on polygraph techniques for the detection of deception for 44 years, taught university courses about polygraph techniques, trained government and law enforcement polygraph examiners, and published extensively on polygraph techniques. Dr. Raskin's testimony in Mosquera concerned both polygraph examinations in general and the specific examination at issue in that specific criminal case.

Harper does not have anywhere near the experience of Dr. Raskin. His resume does not reflect any testing and research and likewise does not list any publications. Neither his testimony at the hearing nor his resume mention any teaching he has done about polygraph exams and techniques.

According to Harper, all 30 police departments in Miami-Dade County use polygraphs, and a similar number of Broward County police departments use it. In fact, Harper said that the federal government is the largest user of polygraph exams in the country. Londono points to this circumstance as a ground to permit his polygraph report to be used for the Undersigned's evaluation of the motions to dismiss the indictment based on, among other reasons, the alleged destruction of the email communications he said he had with Burrola.

Harper explained that Dr. Raskin developed the specific system he now uses — the Utah System. Harper learned another system called the Backster technique (approximately 40 years ago) but switched to the new system developed at the University of Utah. Harper said that the Utah system is often recommended at polygraph association meetings.

Harper discussed the concept of control questions on polygraph exams. As outlined by Harper, a control question has to yield a false answer from an honest person so that the subject's physiological reactions on this false answer can be compared to the reactions to the relevant test questions. Harper emphasized that the "fear of detection" is the key to a control question or relevant question. [ECF No. 529, p. 35].

The control question — also called the comparative question — which Harper used is: "During the first 41 years of your life, did you remember lying to someone who trusted you?" [ECF No. 529, p. 44]. Londono answered "yes." Id. (Judicial note: This answer, which Harper believes to be true, is problematic. Most subjects falsely answer "no," which yields the baseline response for a false answer. But Londono did not provide the expected response. Basically, he answered the "probable lie" question affirmatively with a truthful answer. More on this later, during the discussion of Harper's testimony during cross-examination.)

In any event, Harper estimated that only 2 of the 100 polygraphs he conducted in the past year have generated a "passing" score. [ECF No. 529, p. 45]. So 2 subjects (including Londono) were truthful, and 98 were either untruthful or generated inconclusive results.

Harper said that Londono's polygraph results were "surprising" and "off-thecharts good" and "really as high as you can get." [ECF No. 529, p. 45]. He said that Londono generated an "excellent" score. [ECF No. 529, p. 46]. He also explained that it was "pretty amazing" that Londono scored more than 99 percent on an exam with four relevant questions and three comparison questions. [ECF No. 529, p. 84].

Harper explained that the computer's software program uses an empirical scoring system, which yielded an "NDI" score, meaning "no deception indicated." [ECF No. 529, pp. 50-51]. Specifically, Harper explained, a score of "6" is the cutoff for no deception and Londono surpassed that by receiving a score of "7." [ECF No. 529, pp. 50-51].

Harper also said that he (as opposed to the computer) decided to discard Londono's first polygraph exam chart. Harper said that he did this out of fairness to Londono, who was trembling and had shaking hands. He described Londono as being "overstimulated." [ECF No. 529, p. 51]. Harper said that he used common sense to reach the decision to discard the first chart. Harper said that Londono had calmed down enough to not cause the next three charts to be discarded. Harper, however, testified that the fourth chart was "inconclusive." [ECF No. 529, p. 55].

Harper explained that he has never used the "stimulation test" or "acquaintance test" method because he views them as techniques designed to trick the subject, a procedure he disfavors. [ECF No. 529, pp. 56-57]. Many experts recommend that examiners use this procedure to improve accuracy.

Harper explained that the computer produced different "Probability Truthful" scores for each of the four relevant questions. [ECF No. 529, p. 240]. It allocated a score of 96.22 for the first relevant question, 97.49 for the second relevant question, 90.79 for the third relevant question, and 96.09 for the fourth relevant question. [ECF No. 529, pp. 42-43].

2. Cross-Examination (and to the Court's Questions)

Harper's estimate of administering 30,000 polygraphs in 39 years would mean that, on average, he had conducted 769 tests per year. In addition, before 1990, he had to run and grade the tests manually. Harper said that he now runs approximately 25 polygraphs per week.5 Most of them are pre-employment screening tests conducted outside the United States.

Harper did not know the names of the few cases where federal courts permitted him to offer expert opinion testimony about the results of his polygraph examination. (The cases are not listed on his resume.) He is also unfamiliar with the specific opinions of recognized polygraph experts. "Basically, I work alone," he explained. [ECF No. 529, p. 195]. He repeated a similar point later, when he said, "I work alone, so I don't really know what other people are doing." [ECF No. 529, p. 208].

Harper said that he "believes" that he is a member of the American Polygraph Association ("APA") but never looked into it. [ECF No. 529, pp. 215-16]. He said that the Florida Polygraph Association follows the APA standards. But he does not know the APA standards for when it is appropriate to discard a chart and does not know of any reports or literature accepted in the established polygraph community pinpointing when a chart must be discarded, when it may be discarded, or when it can't properly be discarded.

Harper said that he does not know if the trend in federal courts over the past 20 years is to approve or reject polygraph testimony. Other than Mosquera, which defense counsel questioned him about, he is not aware of any federal cases where polygraph evidence was admitted. By the same token, he is not aware of any federal case where polygraph testimony was rejected.

Although Harper's report represented that Londono answered "yes" to all four relevant test questions concerning his alleged email usage, the actual underlying documents show a "no" answer to question 2 on all the charts.

At that point, Harper explained that Londono did in fact say "yes" to all the questions. He initially testified that he could not explain the discrepancy but then said "[t]his is, obviously, a computer problem." [ECF No. 529, pp. 96-97]. Harper then said that he enters the responses into the computer and that he suspected that the formatting was incorrect. He theorized that he left a "no" response on the template he used for a different exam (with another subject, on an earlier date) and forgot to erase it for the Londono exam.

Harper then explained that the audiotape of the polygraph would confirm that Londono had in fact actually answered "yes" to all relevant questions on all the charts. But he had not turned over the audiotape to defense counsel who retained him, which of course meant that the United States had not been provided with the audiotape, even though the Undersigned required all data and material to be turned over to the United States. The Undersigned required Harper to provide the audiotape after the hearing and also directed the United States to advise if it would seek any further relief after listening to the audiotape. [ECF No. 527]. The United States later filed a formal notice, advising that it would not seek a remedy after listening to the audiotape. [ECF No. 532].

Harper conceded that he discarded the only chart that showed deception. The chart he discarded produced a negative 9 score. The computer did not include the negative 9 score when it ran the algorithm to score the charts. Harper said that he does not know what the overall score would have been had the computer included the deceptive chart. He conceded that the score would be lower but nevertheless predicted that the score would likely still be in excess of 70 percent, which he said would still be a passing score.

According to Harper, the computer does not have the ability on its own to discard a chart or to recommend that a chart be discarded. Harper said that polygraphers have the discretion to include or exclude a chart based on external factors, but he does not know of any applicable standards or guidelines explaining when the discarding of a polygraph chart is required, permissible, or prohibited.

Harper conceded that it is "[n]ot routine" to discard a chart and noted that the decision is selective. [ECF No. 529, p. 259]. He said that the applicable standards do not prevent him from discarding a chart at his discretion.

Harper did not recall whether the relevant questions were first drafted in English and then translated into Spanish or if they were drafted in Spanish in the first instance. One of Londono's attorneys was with him in the FDC conference room for the Londono polygraph.

At first, Harper said that the defense attorney who was with him helped create the relevant questions after discussing the background facts and determining the specific issues to test. [ECF No. 529, p. 231 ("It was just Ray Hernandez [co-counsel] and myself. The subject had no involvement. The defendant had no involvement in that [i.e., drafting and creating the relevant test questions].")]. Later, when asked about this by defense counsel, Harper said that he, but not the attorney, "exclusively" created the relevant test questions. [ECF No. 529, pp. 262-63 (testifying that" I crafted the questions according to the format that's generally accepted" and agreeing that he was "the one who crafted those questions, not Ray Hernandez") (emphasis added)].

Harper acknowledged that Donald J. Krapohl's article "Test Question Construction" [ECF No. 528-2] advised to "[a]void using test questions offered by the examinee or the examinee's attorney," but he said that he could have created the questions on his own after speaking with the attorney. [ECF No. 529, p. 235]. Krapohl is a nationally recognized polygraph expert.

The APA standards require that polygraph exams be videotaped, but Harper said that he did not do it with Londono and explained that he never videotapes the exams he conducts.

Harper said that he thinks the combined number assigned by the software is an average number, but he could not explain how the average for Londono's entire exam was actually higher than the scores for the individual questions. Specifically, the combined score stated a "probably truthful" average score of 0.9942, but the individual scores were 0.9622, 0.9749, 0.9079, and 0.9609. He did not have an explanation for this apparent anomaly but repeated his understanding that the score is an "average" calculation. [ECF No. 529, pp. 240-42]. He said that he was "told" the number is an average. [ECF No. 529, p. 258]. He further explained that he "definitely questioned" it when he saw it and also said, "it's just the way it is." [ECF No. 529, p. 258].

Harper was questioned about inconsistencies in the report results. For example, on chart 1, which he discarded, one page lists the total score at negative 9 and another page lists the total score as negative 8. Similar discrepancies were noted, such as a difference between a 10 score and an 11 score on chart 2 and a difference between a 7 score and a 6 score on chart 3.

According to Harper, the computer "reconfigured" the total scores after evaluating them. [ECF No. 529, p. 126]. Indeed, Harper then said that the computer might reconfigure the data and change the test scores again if he were to ask it to run the algorithm again. At bottom, he explained that the numerical values for the same test could change every time his software reevaluated the physiological data.

"I have no idea how the hell it works," he said, "I'm not a computer expert." [ECF No. 529, p. 126].

Harper acknowledged that he is familiar with Krapohl, who he described as one of the top five polygraph experts in the country. He further acknowledged that Krapohl's understanding of polygraph examinations applies to all polygraph examiners. Harper said that Krapohl's opinions "set the standard" in the polygraph community. [ECF No. 529, p. 191].

According to Krapohl, the same comparison question should not be repeatedly asked. But Harper asked the same control question — designed to elicit a lie from an honest person to create a measure for relevant questions against which one can determine if those questions are eliciting a lie — twelve times (i.e., three times on four separate charts). Harper said that Krapohl's view conflicts with Backster's perspective, whose system he learned in 1978. Harper did not have an explanation for his still-current use of Backster's "repeat-the-control-question" approach even though he abandoned Backster's method for the Utah method approximately ten years ago.

Harper explained that Backster developed his system before computers were in use — which seems to be a reason why he would not follow the Backster approach now, when he uses a computer for the Utah approach.

Furthermore, Harper said that he does not know any polygraph examiner still alive who still uses the same control questions multiple times in the same exam. He does not know the name of a recognized polygraph expert who disagrees with Krapohl and agrees with him about whether it is problematic to repeat the same control question. "Basically, I work alone," he said. [ECF No. 529, p. 195]. On redirect, however, Harper said that the use of repetitive control questions did not affect the final scores.

Concerning the control questions, Harper explained that they were in fact the "special procedures" that his report says were used "to monitor the subject's response capability." [ECF No. 529, p. 223]. He advised that he always uses a control question protocol, which (as noted above) is designed to elicit a deceptive response to the question. Most subjects falsely answer "no," which creates a comparison base level for the responses to the relevant test questions.

Harper was pressed about the fact that Londono answered "yes" to the control question designed to trigger a false "no" in an otherwise honest person. The Government suggested that because this control question is used to yield a base for comparison purposes, Londono's "yes" answer, which appears to be a truthful one, cannot be used as a base for comparison purposes. Given that most subjects provide a false "no" answer to the control question, Harper was asked about the consequences arising from Londono's atypical response. Harper's explanation was that Londono's truthful "yes" response to the control question would "overwhelm a `yes' or `no' answer or a threat to well-being would override that." [ECF No. 529, p. 227].

Because the Undersigned views Harper's answers to questions about the control question — which Londono surprisingly answered "yes," rather than the false, expected response of "no"— as problematic and unconvincing, it is helpful to excerpt Harper's specific responses:

If you knew, for instance, you robbed a bank — I'm sorry, if you knew you did lie to someone before, it wouldn't be that much of a threat to your well-being if he said "yes" or "no." I mean, if he said "yes," and you had, it would come out as lying most times, because that would override his response. It's just a "no." But the body, this physiology is taking over, "Oh, yes, we have," and you would say an arousal. So "no" is not necessarily or "yes" is not necessarily a barrier to the question, and it's going to show that that response of — if you're lying, you're lying. You could say "no" all you want or "yes" all you want. But that's going to be overridden. * * * Okay. You can answer whatever you want on a test question, on a control question. You can answer any question you want any way. * * * The fear of detection overrides that answer. In other words, you could say "maybe" and you would still react if the fear of detection overwhelmed you. So you'll see it in the control. If you do say "no" and there's no reaction there but reaction in the relevant which follows it, you're on the right track. The person's not being truthful. But as far as a control question is concerned, it's all about threat to well-being. You could say "yes," "no" or "maybe." But if you're really concerned about that, you will react to it regardless of what you say.

[ECF No. 529, pp. 227-29 (emphasis added)].

During the cross-examination of Harper, the United States introduced into evidence an article written by Krapohl entitled "Utah Probable-Lie Test." [ECF No. 528-3]. In the article, Krapohl explained that "[t]he first test will be a demonstration test." [ECF No. 528-3, p. 1]. In fact, Krapohl's article stressed that "[t]he first test should always be some form of demonstration test" because those tests "improve polygraph accuracy." [ECF No. 528-3, p. 4]. Harper does not use the demonstration or acquaintance test, though, and did not use it for Londono's charts either. Harper transferred his decision to follow the no-demonstration-question approach from the Backster methodology to the Utah approach. He was not aware of polygraph experts who recognized this hybrid methodology as being acceptable.

Other than saying that he does not like the demonstration question protocol because he does not want to use trickery, Harper did not explain why he chose to retain this portion of the Backster method. And Krapohl's article clarifies that "trickery or appearance of trickery" should not be used in the demonstration questions. [ECF No. 528-3, p. 5]. Even so, Harper said that the omission of a demonstration question would not invalidate an otherwise good polygraph examination.

Harper said that he is not sure if there is any accepted literature recommending the elimination of the demonstration question from the Utah test. He has not read anything on that specific topic and conceded that it "could be" that the contemporary literature consistently notes that the demonstration or acquaintance test is an integral part of the Utah test. [ECF No. 529, p. 149].

Harper also conceded that it is best to alter the order of the relevant questions when a chart is run more than once. But Harper used the same order for three of the four charts and changed the order of only two questions on one chart.

Harper said that the literature does not suggest crafting relevant questions, if possible, to yield a correct "no" response. But Krapohl wrote a published article in which he recommended that questions should be drafted so that "no" is the truthful answer. [ECF No. 528-2, p. 3]. Harper said that, for the Londono exam, he could not create relevant questions designed to generate a truthful answer of "no." But the AUSA questioning Harper came up with two relevant questions to which Londono could have answered "no" (assuming that his answers were truthful in the first place). For example, the prosecutor suggested the following question: "Have you lied about communicating through e-mail with Special Agent [Burrola]?" [ECF No. 529, p. 266].

As noted, Harper discarded the first chart (which showed significant deception — negative 9) and classified the final and fourth chart as "not usable." He said that this fourth chart was after Londono was comfortable with the exam and the examiner. He described this phenomenon as "dampening" but did not otherwise specifically and adequately explain why this familiarity would lead to an unusable chart, as opposed to a better, less-deceptive result. [ECF No. 529, pp. 250-51, 270-71].

Harper testified that an imprecise question by a polygrapher could cause a false positive physiological response. When formulating his test questions, he was never told that Londono allegedly used the draft function of email to converse with Agent Burrola, instead of sending emails back and forth. Harper also used a slightly different email address then the one at issue when questioning Londono, and he testified that an examinee could get an incorrect physiological response if the wrong email address was used.

He testified that publicly available information existed on the internet regarding physical countermeasures — e.g. suppression of breathing, clenching the sphincter, manipulation of toes — which an examinee could employ to produce a false physiological response. He did know that detainees at FDC have access to the internet. During the exam, Harper did not use a foot sensor or examine Londono's shoes to identify possible counter-measures; rather, he only used a seat sensor.

In certain charts, Harper made the sole determination to remove data based on movement. As a result, this removed data did not figure into the scoring. If the data was not removed, then it could have affected the scoring. He did not know if other polygraph examiners remove data based on a movement.

The Undersigned also questioned Harper, giving both sides an opportunity to ask follow-up questions. He testified that an examiner wants an examinee to lie to the control question because the examiner uses the response as a benchmark to compare the responses to the answers to the relevant questions. But he later said that even if a person does not give a lie — in order words, answers "yes" to the probable lie control question — then a "fear of detection" overrides the response. [ECF No. 529, p. 228].

In response to the Court's questioning, Harper said that he disregarded chart 1 based on common sense. He could not identify any literature accepted by the polygraph community for when an examiner should or should not disregard a chart. If chart 1 had been included, then it would have lowered the overall scoring and percentage. He disregarded chart 1 because he perceived that Londono was agitated.

The prosecutor also questioned Harper about the wording of relevant question four, which asked Londono regarding a certain message, "did you send that message in an email." [ECF No. 525-1, p. 4]. Specifically, Harper was questioned about the fact that Londono's testimony was not that he sent emails to Burrola, but that he saved them as drafts, which the agent would later view after login into the same email account. Harper said that the difference in wording in the question is inconsequential and branded the distinction as "splitting hairs." [ECF No. 529, p. 158]. Nevertheless, Harper acknowledged that an incorrectly worded question could affect the physiological response.

Harper was also questioned about "detrending," which he explained as a technique to even out the responses. [ECF No. 529, pp. 160-61]. Harper said that detrending makes the responses easier to view. Harper used detrending to view some of Londono's results. Harper said that the computer does not require or suggest detrending.

Harper explained that he used an activity monitor on Londono's chair to detect any countermeasures which Londono might have used in an effort to trick or beat the test. Harper, however, removed a movement from the chart even though no movement was shown on the activity monitor. In addition, although (as noted above) Harper discarded the first chart because he said Londono's entire body was shaking, the body movement monitor did not confirm that. Harper theorized that the omission of movement on the movement monitor might be because only Londono's hands and arms were shaking (as opposed to his entire torso). His other theory is that Londono's body was "just trembling," which he said is "not movement." [ECF No. 529, p. 268].

The prosecutor focused Harper's attention to a specific page of his chart, where the computer scored the response as a positive 2. But Harper said that he would have scored it as a negative 2, reflecting deception. So Harper would have scored this answer differently (and lower) than the computer (which used an algorithm).

Although Harper now uses a computer to conduct polygraph exams and switched from the Backster method to the Utah approach approximately ten years ago, he also said that the techniques and technology for polygraphs have not significantly changed over the past 50 years.

ii. DEA Special Agent Brent Barnes

Special Agent Barnes is the unit chief for the DEA's polygraph unit. He is a supervisor and oversees 21 polygraphers. The National Center for Credibility Assessment (the "NCCA") certified him as a polygrapher in April 2012. According to Agent Barnes, the NCCA is the gold standard in the polygraph field. Only federal employees are selected to attend. Krapohl was an NCCA instructor. Agent Barnes has administered more than 340 polygraph exams and has reviewed thousands of exams.

Agent Barnes's testimony at the evidentiary hearing was, for the most part, focused only on the control/comparative question. He explained that it is a probable-lie question — designed to elicit a lie — and that it is "absolutely imperative to the exam." [ECF No. 529, p. 299].

Agent Barnes explained that he has never used the same control/comparative question multiple times in the same polygraph exam. Moreover, he said that the different approach of using different control questions is applicable for all polygraphs.

Agent Barnes explained that the control question becomes innocuous if the answer is not a lie — and it would invalidate the entire test. He also said that he never heard of a theory where the question itself, rather than the answer, overwhelms the subject.

C. Post-Hearing Developments

Londono's counsel provided the United States with the audio of the polygraph exam, and the United States advised that it did not seek any further relief. Presumably, the audio confirmed that Londono answered "yes" to question 2 on all four charts even though the charts technically contained a "no" response.

The Undersigned required the parties to submit proposed Reports and Recommendations (on the admissibility of Harper's report) and a brief (i.e., no more than five pages) memorandum on certain legal issues. Both parties timely filed both submissions. The United States, however, also filed a "Notice Regarding Late Production of Raw Physiological-Response Data." [ECF No. 543]. In that Notice, the United States flagged this Court's December 15, 2017 directive that required Harper to contact the Stoelting Company (which created the software he used to polygraph Londono) to determine if he could download the raw physiological data to a flash drive or other storage device so that the Government could review it. If it was doable, then Londono was required to produce it by December 20, 2017. [ECF No. 527].

On December 19, 2017, Londono gave notice that Harper had difficulty obtaining could review it. [ECF No. 531]. Londono later advised that Stoelting refused to provide an answer despite several requests. [ECF No. 536]. On January 3, 2018, one of the two primary AUSAs handling this case contacted Stoelting and spoke to a technical expert, who explained that the raw data could be downloaded and viewed on another computer with the software installed. The next day, Londono's counsel delivered to the AUSA a thumb drive containing the raw physiological data I ordered turned over by December 20, 2017.

According to the Government's Notice, the DEA needs one to two weeks to purchase and install the software and interpret the data and produce a report, which would then need to be peer-reviewed. [ECF No. 543]. The United States says that it therefore does not have the time to reference the raw data in its proposed Report and Recommendation, which it filed on January 5, 2018.

The Government's suggested remedy for this tardy disclosure is that the Undersigned not consider the data if Londono mentions it in his filings. In addition, the United States also contends that "the Court should draw yet another conclusion about the unreliability of Harper's report. Mr. Harper clearly does not understand how to operate his own software and equipment." [ECF No. 543, p. 3].

IV. Legal Analysis on the Admissibility of Londono's Polygraph Exam

The Undersigned must tackle the issue of whether I can or should consider the polygraph examiner's report for the limited purposes of the pre-trial evidentiary hearing concerning Londono's motions to dismiss the indictment. The answer is no. First, Londono has not met his burden of establishing that polygraph exams in general pass muster under Daubert. Second, even if polygraph evidence could in theory be admitted at an evidentiary hearing under Daubert,6 this specific test is not admissible.

There are two circumstances in the Eleventh Circuit when a court, in its discretion, may admit or consider polygraph evidence: (1) where the parties stipulate in advance, or (2) where the evidence is used to impeach or corroborate the testimony of a witness. Piccinonna I, 885 F.2d at 1536. The United States did not stipulate and advised that it will not stipulate, so the only other possibility is for impeaching or corroborating a witness' testimony.

A court may admit such corroboration evidence only after the following three prerequisites are satisfied: (1) "the party planning to use the evidence at trial must provide adequate notice to the opposing party that the expert testimony will be offered;" (2) the opposing party must be given "reasonable opportunity to have its own polygraph expert administer a test covering substantially the same questions;" and (3) the polygraph examiner's testimony must be admissible under the Federal Rules of Evidence governing the admissibility of corroboration testimony. Id. Even if a party satisfies these prerequisites, a district court can exercise its discretion to exclude the polygraph evidence under other applicable rules of evidence. Id. at 1536-37. In fact, Piccinonna I holds that admission of polygraph evidence for impeachment or corroboration purposes "is left entirely to the discretion of the trial judge." Id. at 1536 (emphasis added).

The Undersigned provided the Government with the opportunity to have its own polygraph expert examine Londono. It declined, opting instead to challenge the admissibility of polygraph evidence in general and the admissibility of this specific exam.

District courts in this district hold that a single polygraph testing session represents an inadequate basis upon which an expert can base an opinion on a defendant's character for truthfulness. United States v. Piccinonna (Piccinonna II), 729 F.Supp. 1336, 1338 (S.D. Fla. 1990), aff'd, 925 F.2d 1474 (11th Cir. 1991) ("It is inconceivable that anyone, expert or not, could form a valid, reliable, and admissible opinion as to the `character' of a witness based on nothing more than one single polygraph examination.")

Based on the specific circumstances surrounding Londono's submission of a polygraph report, the Undersigned will not consider it. See United States v. Gilliard, 133 F.3d 809, 815 (11th Cir. 1998) (holding that the district court did not abuse its discretion in excluding under Daubert polygraph examination that used hybrid technique that had not gained general acceptance within the relevant scientific community); see also Norelus v. Denny's, Inc., 628 F.3d 1270, 1284 (11th Cir. 2010) (explaining that "[o]pinions based on polygraph examinations are seldom, if ever, admissible into evidence.").7

The proponent of the polygraph evidence must "offer evidence to establish the polygraph evidence at issue here is reliable." Pavlenko, 845 F. Supp. 2d at 1325. To be admissible under Federal Rule of Evidence 702 (which is entitled "Testimony by Expert Witnesses"), as interpreted by Daubert, the expert polygraph evidence must constitute scientific knowledge and assist the trier of fact to understand the evidence or to determine a fact at issue. See Daubert, 509 U.S. at 589-91. To determine whether the scientific knowledge prong is satisfied, this Court must consider polygraph theory or technique in light of at least five factors: (1) whether the theory or technique underlying polygraphy can be, and has been, tested; (2) whether it has been subjected to peer review and publication; (3) the known or potential rate of error for polygraphy; (4) the existence and maintenance of standards controlling the operation of polygraphy; and (5) whether polygraphy has attained general acceptance within the relevant scientific community. See id. at 593-94; see also Gilliard, 133 F.3d at 812.

On the other hand, a court is not required to consider each of these factors, "and a federal court should consider any additional factors that may advance its Rule 702 analysis." Quiet Tech. DC-8, Inc. v. Hurel-Dubois UK Ltd., 326 F.3d 1333, 1341 (11th Cir. 2003).

The Undersigned finds that Londono has failed to meet his burden here. At bottom, Londono presented only the testimony of Harper. Harper is an experienced polygraph examiner, to be sure. But he did not offer much about the five factors. For all practical purposes, he simply referenced (and only briefly, at that) Mosquera's summary of the testimony of a nationally-known expert who has conducted research and who has published literature. By contrast, Harper said that he has merely read about and heard about the purported 90 percent reliability factor.

Perhaps most importantly, Harper candidly explained that he works alone, does not speak much with other polygraph experts, and is not familiar with the recent decisions in the federal courts about the admissibility of polygraph evidence. Although he did note that the federal government often uses polygraphs, this alone is unconvincing and does not establish any of the five factors. Moreover, the Undersigned does not equate a pre-employment screening polygraph with a case-specific polygraph given to a criminal defendant or suspect in a criminal investigation.

The Undersigned is not convinced that the Government's use of polygraph exams in other settings means that the technique passes muster under Daubert. There are vast differences between criminal investigations and employment screenings and evidence submitted in court. As the Supreme Court has already explained: "Governmental use of polygraph tests . . . is primarily in the field of personnel screening, and to a lesser extent as a tool in criminal and intelligence investigations, but not as evidence at trials." United States v. Scheffer, 523 U.S. 303, 312 n.8 (1998). "Such limited, out of court uses of polygraph techniques obviously differ in character from, and carry less severe consequences than, the use of polygraphs as evidence in a criminal trial." Id.

Significantly, Harper did not testify about the fifth factor, but the Undersigned observes (without making a specific finding about whether this factor could ever be met in another case) that "a review of the case law in the federal circuits and the Supreme Court show substantial disagreements over the effectiveness of polygraphs" and "there is no consensus within the scientific and legal community over the efficacy of polygraph evidence." Pavlenko, 845 F. Supp. 2d at 1327 (citing Henderson, 409 F.3d at 1303).

Assuming that Londono had established the reliability and admissibility of polygraph evidence in general so that a Court might consider admitting the results and opinions in a specific case (which he most assuredly has not), the specific test here is not admissible, even for the limited purpose of an evidentiary hearing before a judicial officer.

Because the Undersigned also finds that this particular exam is inadmissible, it is not necessary to exhaustively analyze the factors underlying an assessment of the overall reliability of polygraphs. It would, however, be extremely difficult, if not impossible, for Londono to have adequately cleared the first hurdle — whether the theory underlying polygraphs can be, and has been, tested. That is because there is no "Pinocchio" factor associated with polygraphs.8 In other words, "[t]here is no known physiological response to lying, and thus no scientific theory in that regard." Pavlenko, 845 F. Supp. 2d at 1326. In the end, "the only thing the polygraph examination `proves' is that the examinee believes her own story." Norelus, 628 F.3d at 1284. Indeed, "physiological responses consistent with deception may also be the product of various conditions that have nothing to do with the truth or falsity of a response." Pavlenko, 845 F. Supp. 2d at 1326.

Moving beyond the issue of whether polygraph evidence in general is admissible, the Undersigned now analyzes whether this specific exam is admissible in a pre-trial setting generated by Londono's efforts to dismiss the indictment because of destruction of evidence.

Federal Rule of Evidence 702 provides:

A witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or otherwise if: (a) the expert's scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and (d) the expert has reliably applied the principles and methods to the facts of the case.

Fed. R. Evid. 702.

In Kumho Tire Co. v. Carmichael, 526 U.S. 137, 147 (1999), the Court held that the trial judge's role as gatekeeper applies not only to "scientific" testimony, but to all expert testimony. Rule 702, as amended, is consistent with Kumho and provides that all types of expert testimony presents issues of reliability and relevance for a judge. "The burden of establishing qualification, reliability, and helpfulness rests on the proponent of the expert opinion, whether the proponent is the plaintiff or the defendant in a civil suit, or the government or the accused in a criminal case." United States v. Frazier, 387 F.3d 1244, 1260 (11th Cir. 2004).

Here, the Court finds that myriad reasons compel the conclusion that Harper's opinion about the polygraph exam he gave to Londono is inadmissible. It does not clear any of the three hurdles erected by Rule 702. The Undersigned will list only the most significant reasons for this ruling.

First, the fundamental control-question protocol was not properly implemented, and Harper has not explained how a truthful "yes" answer to a control question designed to trigger a false response of "no" could still produce a valid test. According to Special Agent Barnes, this invalidates the entire test. Second, Harper discarded the first chart — which showed significant deception — and was unable to adequately articulate whether applicable standards permitted this decision. Explaining that "common sense" prompted his decision is not convincing. Third, Harper was unaware of any recognized expert or widely accepted literature which would permit the discarding of a problematic chart. Fourth, Harper did not know how the results would have changed if the first chart had in fact been included in the algorithm.

Fifth, the last of the four exams was "inconclusive," and yet Harper did not sufficiently explain why this result would not change his opinion that Londono's polygraph was excellent. Sixth, Harper repeated the same control question 12 times even though one of the country's top experts recommended that the question be changed. Seventh, Harper did not significantly vary the order of the questions even though the same nationally-recognized expert recommended that such variation take place. Eighth, Harper could not explain how the average score for all four relevant questions could be higher than the individual scores for those same four questions. Ninth, Harper testified that he discarded the first exam because Londono was trembling, but an on-the-chair sensor did not detect movement. And tenth, the questions contained some factual inaccuracies.

Given these deficiencies, the Undersigned declines to consider the polygraph evidence and brands it as inadmissible, even for a non-jury hearing. See generally United States v. Carroll, 450 F. App'x 937, 940 (11th Cir. 2012) (holding that district court did not abuse its discretion when excluding polygraph evidence).

V. The Non-Polygraph Evidence

Setting aside the inadmissible polygraph report, the Undersigned has two other categories of evidence to evaluate when considering Londono's motions to dismiss because of the Government's purported failure to preserve emails (and for supposedly lying about the existence of emails and therefore perpetrating a fraud on the Court): (a) Agent Burrola's business card, and (b) the testimony from the evidentiary hearing (including discussion of the business card).

The Undersigned finds convincing the Government's common sense point that the mere possession of a business card with an email address on it "does not mean that emails were sent." [ECF No. 475, p. 10]. The United States highlights the fact that the agents have repeatedly and unequivocally stated that they communicated with Londono through BlackBerry Messenger, not email. Therefore, according to the Government, Londono is seeking to exploit the logic "that it is incredibly difficult to prove a negative" by "trying to prove the exist[ence] of emails because he knows there is no way to contradict his assertion other than through live testimony." [ECF No. 475, p. 11 (emphasis added)].

The evidentiary hearing, of course, was the opportunity for the parties to present the live testimony flagged in the Government's memorandum. Moreover, although a business card does not in and of itself prove that emails were sent, its existence hardly proves that emails were not exchanged. Indeed, Londono's possession of Agent Burrola's business card when arrested in Argentina would tend to support his claim and would help to undermine (albeit perhaps only slightly) the Government's "noemail-at-all" position.

But the testimony (from Londono and Burrola) at the evidentiary hearing did not convince me one way or the other that emails were (or were not) exchanged between Londono and Burrola. By way of summary, both witnesses provided testimony which a jury could rationally believe or could logically reject. Neither witness provided bullet-proof testimony, free from potential impeachment.

The Government, of course, suggests that Londono is a biased witness with a huge personal stake in the outcome of the case (in general) and in his motion to dismiss (in particular). This point is undoubtedly correct, so the Undersigned will offer some comments on Agent Burrola's testimony. The Undersigned is not concluding that Burrola is not credible or that he provided false testimony. Instead, the Undersigned is simply flagging some potentially problematic points that a jury might rely on to accept some or all of Londono's contentions about his use of email (or to reject some or all of Burrola's testimony).

First, Burrola did not unequivocally testify that emails were in fact never used. Instead, he testified that, to the best of his recollection after reviewing material, he and Londono did not use emails. This is a qualified, equivocal answer, significantly different from a flat-out denial that emails were never used.

Second, Burrola had no explanation for the shutting down of the email account created to provide an alternate method of communication. He said that he personally did not close or terminate the account and did not know who did. A jury could infer that this vague answer militates in favor of a conclusion that emails were in fact used.

Third, a jury could infer that Burrola's use of the word "internet" to describe communications with Londono was not actually a security measure designed to disguise the actual method of communication, but, instead, was an accurate term used to reference the actual use of email communications.

Fourth, the existence of a different font in Spanish in an email Burrola sent to other federal law enforcement agents and officials might be grounds that a jury could use to infer that Burrola pasted Londono's email into his own email, rather than retyping multiple paragraphs into a separate email to his colleagues.

To be sure, the Undersigned is hardly concluding that Londono's testimony was free from challenge. It wasn't. He sometimes contradicted himself and his demeanor was at times far-more fidgety on cross than on direct. And a jury could surely find his testimony to not be credible, especially his explanation about why he no longer has his own emails.

The record at the evidentiary hearing (where Londono and Burrola testified) is far from clear. The Undersigned finds that Londono has not now met his burden (even if governed by the preponderance of the evidence standard) of establishing a due process violation arising from the alleged destruction of emails, so I am recommending that the District Court deny the motion to dismiss the indictment. See generally United States v. Thomas, 62 F.3d 1332, 1339-40 (11th Cir. 1995) (holding that defendant's failure to carry burden on bad faith concerning motion to dismiss for pre-trial delay was fatal to the motion's success).

If Londono's testimony about using emails is correct, even partially, then that likely would mean that Agent Burrola was not credible on that important point. It also could mean that the emails were destroyed or lost by the Government. Conversely, if Londono's testimony about his use of email communication is false, then that would credit Agent Burrol's testimony and render unbelivable Londono's claim that the Government destroyed emails. For purposes of his pre-trial motions to dismiss the indictment, Londono has not persuaded me. But I am not prepared to conclusively and irrefutably find that his testimony is completely false, either.

Resolving this important credibility challenge between Londono and Burrola would best be considered by a jury, in a full trial, rather than by a judicial officer who observed the two testify only in a limited-purpose, relatively brief evidentiary hearing. A jury would be entitled to hear the evidence concerning Londono's destruction-ofevidence theory in connection with his public authority defense because Londono's hypothesis is that the United States is attempting to obliterate all evidence of the public authority defense and that this goal is related to the purported destruction or loss of evidence.

Therefore, if Judge Graham permits Londono to assert the public authority defense at trial and if Londono testifies, then the parties can present evidence about the discarded BlackBerry telephones and the alleged use of email communication for the jury's consideration. Granting Londono's motions to dismiss the indictment is unwarranted (based on the record evidence presented to date) and premature, but determining now that Londono is foreclosed from raising factual issues asserted in the motions at trial would be equally premature.

But, as noted, Londono should not be permitted to introduce polygraph evidence at trial. This ruling on polygraph evidence is an Order for purposes of the pre-trial ruling and evidentiary hearing (i.e., the Undersigned will not credit it) but is only a recommendation as to use at trial, a proceeding over which Judge Graham will preside.

Finally, for reasons explained below, the Undersigned recommends that a decision on the possible use of an adverse inference concerning the alleged destruction of emails (as opposed to the unavailability of the BlackBerry phones) be postponed until trial.

In United States v. Lanzon, 639 F.3d 1293, 1302 (11th Cir. 2011), the Eleventh Circuit stated that the Court had not yet decided whether a criminal defendant who presents evidence of the destruction of evidence is entitled to an adverse-inference instruction. In Lanzon, the Court noted that it had not recognized the "spoliation doctrine in the criminal context" but had done so in the civil context. Id. The Lanzon Court affirmed the denial of such an instruction concerning instant messages not preserved by the detective, finding no evidence to support the requisite bad faith showing. Id.; see also United States v. McCray, 345 F. App'x 498, 501 (11th Cir. 2009) (discussing without deciding which sanctions, such as an adverse-inference instruction, might be available for spoliation involving destruction of firearm at issue).

The Eleventh Circuit requires civil defendants in suits brought by the Government to establish bad faith before the Government's destruction of evidence could provide the basis for an adverse-inference instruction. See S.E.C. v. Goble, 682 F.3d 934, 947-48 (11th Cir. 2012).

Several appellate courts which have addressed the availability of an adverse-inference instruction for destruction of evidence have found that a criminal defendant must establish bad faith on the part of the government before an instruction is appropriate. See United States v. Tyerman, 701 F.3d 552, 561 (8th Cir. 2012) (noting that the appellate court has not applied the spoliation doctrine in a criminal case but would require bad faith by the government if it were to do so); United States v. Wright, 333 F. App'x 772, 778 (4th Cir. 2009) (holding that "without bad faith there was simply no basis for an inference" instruction to the jury due to destruction of evidence); United States v. Wise, 221 F.3d 140, 156 (5th Cir. 2000) (holding that district court properly declined to give adverse-inference instruction to the jury absent bad faith conduct by the Government); see also United States v. Greco, 734 F.3d 441, 447 (6th Cir. 2013) (requiring that the Government act with a "culpable state of mind" for a spoliation-ofevidence type of presumption to issue).9

If Judge Graham were to determine that Londono presented sufficient trial evidence of the existence and destruction of emails and that the spoliation sanction of a permissible adverse inference is available in a criminal prosecution, then the jury can be instructed that it may infer that the emails could have supported his public authority defense — if it concludes that the emails existed and that the Government destroyed them in bad faith.10

Unlike the BlackBerry-cellphones scenario, where their unavailability is based on a standard practice (and not a bad faith destruction of evidence), the unavailability of the emails (assuming that they existed in the first place) could be attributable to bad faith (because it would mean that the Government had been untruthful about their existence and had no logical, good faith reason for their disappearance).

The potential scenario involving purported emails might not necessarily support Londono's public authority defense. That is because the Undersigned has not seen the emails (assuming that they ever existed) and cannot now determine whether they said what Londono now claims they said. Therefore, the possibility of an adverse-inference instruction must be tailored to reflect this inherent uncertainty. Rather than requiring the jury to make a mandatory inference and or allowing it to permissibly infer that the emails would support the defense, the adverse inference possible here (if warranted by the evidence and if authorized by Judge Graham) would at most permit the jury to infer that the emails might have supported the defense.

This discussion of a possible, permissible adverse inference would become a reality at trial only if Londono sufficiently established that the email communications were used and that the Government destroyed them in bad faith. Wise, 221 F.3d at 156.11

VI. Conclusion

First, the Undersigned finds Harper's expert opinions about Londono's polygraph exam to be inadmissible for purposes of the Daubert hearing and my overall assessment of the two motions to dismiss the indictment.

Second, the Undersigned respectfully recommends that the District Court not permit Harper to testify at trial.

Third, the Undersigned respectfully recommends that the District Court deny the two motions to dismiss the indictment.

Fourth, the Undersigned respectfully recommends that the parties be permitted to present evidence at trial about the communications by BlackBerry, the purported communications by email, and the circumstances surrounding the absence at trial of the BlackBerrys and the emails.

Fifth, the Undersigned respectfully recommends that the Court postpone until trial the decision on whether the jury would be permitted to reach the inference that the missing emails could have supported Londono's public authority defense — but only if there is a finding that the emails existed at one time and that Burrola or other law enforcement officials destroyed or discarded them in bad faith.

VII. Objections

The parties shall have three (3) calendar days from the date of this report to file written objections, if any, with the District Judge. Each party may file a response to the other party's objection within three (3) calendar days from the date of the objection.12 Failure to timely file objections shall bar the parties from a de novo determination by the District Judge of an issue covered in the report, and shall bar the parties from attacking on appeal unobjected-to factual and legal conclusions contained in this report, except upon grounds of plain error and if necessary in the interest of justice. See 28 U.S.C. § 636(b)(1); Thomas v. Arn, 474 U.S. 140, 149 (1985); Henley v. Johnson, 885 F.2d 790, 794 (1989); 11th Cir. R. 3-1 (2016).



1. The United States has filed a motion to exclude or limit Londono's public authority defense. [ECF No. 443]. Earlier today, the District Court denied that motion without prejudice. [ECF No. 557]. The Court reserved ruling "pending trial testimony." [ECF No. 557, p. 7].
2. For instance, in Revolorio-Ramo, the Eleventh Circuit affirmed a conviction for conspiracy to possess cocaine with intent to distribute on a vessel even though fishing vessel was destroyed. 468 F.3d at 775. A U.S. Navy team had boarded a vessel at sea and arrested the crew for cocaine smuggling. Id. at 772. The boarding team decided that the vessel was not seaworthy and could not be towed back to port, so the vessel was destroyed. Id. at 773. Photographs of the boat were taken before it was destroyed, but the photos were of such bad quality that they were useless. Id. at 775. The defendants argued that fishing gear on the vessel proved their innocence. Id. at 774. Even so, the boat's destruction and the useless photographs were not sufficient to dismiss the indictment because these negative results arose from the crew's unsuccessful and incompetent manner of trying to preserve evidence, not an intent to deprive defendants of favorable evidence. Id. at 775.
3. Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993).
4. Relevant polygraph questions are those addressing the factual issue being tested. For Londono, the relevant questions concerned his contention that he and Agent Burrola communicated through emails.
5. Standard 1.7.9 of the American Polygraph Association provides that "[a] member polygraph examiner shall not conduct more than four diagnostic or three evidentiary examinations in one day, and no more than five examinations of any type in one day." APA Standards of Practice (Effective September 1, 2015), http://www.polygraph.org/assets/docs/Misc.Docs/apa%20standards%20of%20practice% 20effective%203-11-16%20final.pdf.
6. The Undersigned is not conclusively holding that polygraphs can never be admitted into evidence in a federal court trial or evidentiary hearing or for other purposes (such as sentencing hearings). Instead, the holding on this first point is far-more limited: the defendant here did not meet his burden.
7. Piccinonna I was decided in 1989, four years before the U.S. Supreme Court's Daubert decision, which is the landmark case on the standards for the admissibility of expert testimony. Because it was decided before Daubert, one might wonder if Piccinonna I is still good law in this Circuit. The Undersigned has not found any Eleventh Circuit case receding from Piccinonna I. See United States v. Henderson, 409 F.3d 1293, 1304 (11th Cir. 2005) (explaining that "[b]ecause Piccinonna's holding concerned only the admissibility of polygraph evidence under the Frye `general acceptance test,' it does not preclude a district court from excluding polygraph evidence for not satisfying the Daubert factors.").

Other district courts have flagged as notable that Piccinonna I predates Daubert and proceed to analyze the proposed admissibility of polygraph evidence under the post-Piccinonna I standards established by Daubert and its progeny. See, e.g., United States v. Pavlenko, 845 F.Supp.2d 1321, 1324 (S.D. Fla. 2012) (finding as notable that Piccinonna I was decided before Daubert and rejecting the need for an evidentiary hearing after analyzing the Daubert factors). Therefore, even if Piccinonna I is still the law in this Circuit, polygraph evidence must still also satisfy Federal Rule of Evidence 702 and Daubert. See United States v. Evans, 469 F.Supp.2d 1112, 1114-15 (M.D. Fla. 2006) (denying defendant's motion in limine to admit the results of polygraph test in crack cocaine prosecution).

8. The story of Pinocchio concerns a wooden boy whose nose grew when he told a lie. Based on Carlo Collodi's "The Adventures of Pinocchio," it was later made into a movie (several times).
9. To be sure, the Ninth Circuit has contradictory precedent on this question. Compare United States v. Romo-Chavez, 681 F.3d 955, 961 (9th Cir. 2012) (requiring bad faith) with United States v. Sivilla, 714 F.3d 1168, 1173 (9th Cir. 2013) (bad faith not necessary, but "quality of the Government's conduct" is a key consideration). And as another appellate court put it, although an adverse-inference instruction "usually makes sense only where the evidence permits a finding of bad faith destruction," on that point "the case law is not uniform in the culpability needed for the instruction." United States v. Laurent, 607 F.3d 895, 902 (1st Cir. 2010).
10. The Undersigned's jurisdiction here is limited to the motions to dismiss the indictment, not possible evidentiary or jury instruction rulings at trial, which Judge Graham will preside over. Therefore, the Undersigned is discussing the potential for an adverse-inference instruction and is not making a recommendation one way or the other because (1) the evidence has not been fully developed and will not be adequately developed until trial and (2) Judge Graham will make decisions about how to conduct the trial.
11. The Wise Court discussed the bad faith requirement for an adverse inference sanction resulting from destruction of evidence in a criminal prosecution by citing to spoliation cases in the civil context.

Given that the Eleventh Circuit has not yet conclusively determined the adverse inference issue in criminal cases and the fact that civil spoliation cases are sometimes used as authority for a similar analysis in criminal cases, it is appropriate to flag decisions in this Circuit in civil cases. See U.S. E.E.O.C. v. Suntrust Bank, No. 8:12-CV-1325-T-33, 2014 WL 1364982, at *11 (M.D. Fla. Apr. 7, 2014) (allowing the EEOC to introduce evidence at trial of defendant's policies regarding preservation of its surveillance videos and its failure to preserve surveillance video evidence); Socas v. The Nw. Mut. Life Ins. Co., No. 07-20336-CIV, 2010 WL 3894142, at *9 (S.D. Fla. Sept. 30, 2010) (denying the defendant's request for dismissal or adverse inference but noting that "this ruling is not intended to preclude Defendant Northwestern from introducing into evidence the facts concerning this failure to preserve relevant documents"); Managed Care Sols., Inc. v. Essent Healthcare, Inc., 736 F.Supp.2d 1317, 1334 (S.D. Fla. 2010) (denying sanctions motion but explaining that "this ruling does not foreclose the possibility that the plaintiff will be able to introduce evidence of the defendant's failure to retain relevant documents at trial."); Floeter v. City of Orlando, No. 605CV-400-ORL-22KRS, 2007 WL 486633, at *7 (M.D. Fla. Feb. 9, 2007) ("Courts have found that loss of evidence may be relevant and admissible for the jury's consideration, and that adverse inferences arising from such destruction can be argued by counsel in closing.").

12. The time for objections and responses is being shortened because of the upcoming trial date. Moreover, more than once the Undersigned advised that the objection and response times would be accelerated. [ECF Nos. 511 (stating that the parties "will be required to submit post-hearing memoranda and that the deadline to do so will likely be comparatively brief" and that "the parties are urged to now begin their legal research and drafting for the sections of the memoranda not dependent on the specific expert witness testimony."); 529, pp. 3-4 (issuing the same warning)].


1000 Characters Remaining

Leagle.com reserves the right to edit or remove comments but is under no obligation to do so, or to explain individual moderation decisions.

User Comments

Listed below are the cases that are cited in this Featured Case. Click the citation to see the full text of the cited case. Citations are also linked in the body of the Featured Case.

Cited Cases

  • No Cases Found

Listed below are those cases in which this Featured Case is cited. Click on the case name to see the full text of the citing case.

Citing Cases