United States v. Connolly

United States Court of Appeals
For the First Circuit

No. 05-2772

UNITED STATES OF AMERICA,

Appellee,

v.

JOHN J. CONNOLLY, JR.,

Defendant, Appellant.

APPEAL FROM THE UNITED STATES DISTRICT COURT

FOR THE DISTRICT OF MASSACHUSETTS

[Hon. Joseph L. Tauro, U.S. District Judge]

Before

Lipez, Circuit Judge,
Selya, Senior Circuit Judge,
and Howard, Circuit Judge.

Terrance J. McCarthy, by appointment of the court, with whom
Edward J. Lonergan and E. Peter Mullane were on brief, for
appellant.
William J. Nardini, Special Attorney, with whom Michael J.
Sullivan, United States Attorney, and John H. Durham, Special
Attorney, were on brief, for appellee.

October 16, 2007
 SELYA, Senior Circuit Judge. Defendant-appellant John J.

Connolly, Jr., a disgraced former agent of the Federal Bureau of

Investigation (FBI), appeals from the denial of his motion for a

new trial in a celebrated criminal case. Connolly musters multiple

theories in support of his appeal, all of which emanate from the

same trove of newly discovered evidence. Unimpressed by the

quality of this evidence and constrained by a deferential standard

of review, we conclude that the district court acted within the

encincture of its discretion in denying the motion.

I. BACKGROUND

The federal courts are by now painfully familiar with the

Winter Hill Gang and its corrupt relationship with the Boston

office of the FBI. See, e.g., United States v. Flemmi, 225 F.3d 78

(1st Cir. 2000); United States v. Salemme, 91 F. Supp. 2d 141 (D.

Mass. 1999); United States v. Salemme, 978 F. Supp. 343 (D. Mass.

1997). The appellant has been prominently featured in several such

opinions. See, e.g., McIntyre v. United States, 367 F.3d 38 (1st

Cir. 2004) (addressing civil suit filed by putative victims' heirs

against the appellant and others). To flesh out the background,

there is a published opinion in this case, upholding the

appellant's conviction on direct review. See United States v.

Connolly, 341 F.3d 16 (1st Cir. 2003) (Connolly I). We assume the

reader's familiarity with these opinions and rehearse here only

those facts most directly relevant to this appeal.

-2-
 Beginning in the early 1970s, the appellant was an FBI

agent in Boston. During his tenure, he served as the "handler" of

two "top echelon" informants: James ("Whitey") Bulger and Steven

("the Rifleman") Flemmi. Although nominally part of the Winter

Hill Gang, Bulger and Flemmi frequently consorted with the Boston

branch of La Cosa Nostra and purported to transmit inside

information to the FBI concerning organized crime activities in New

England. This relationship persisted until 1990 — the year of the

appellant's retirement — when the FBI cut the umbilical cord and

"closed" Bulger and Flemmi as informants.

Notwithstanding their former alliance, the FBI knew (or

at least suspected) that Bulger and Flemmi had been hip-deep in

criminal activity all along. An intensive federal probe ensued.

In December of 1994, the FBI's case against Bulger, Flemmi, and

their cohorts was poised to precipitate an indictment. The

indictment, scheduled to be rolled out on January 10, 1995,

targeted the two quondam informants as well as several other

mobsters. Despite the veil of secrecy attached to grand jury

proceedings, two of the targets — Bulger and Francis ("Cadillac

Frank") Salemme — fled before the indictment was unsealed.

In short order, the indictment was made public; several

defendants, including Flemmi, were arrested; and Bulger and Salemme

became fugitives. Nearly eight months later, the authorities

apprehended Salemme in Florida. Bulger remains at large.

-3-
 Four years after the initial indictment, a different

grand jury handed up a related indictment. This second indictment

named both Flemmi and the appellant as defendants. Pertinently, it

charged the appellant with racketeering, obstruction of justice,

and conspiracy. A superseding indictment, fashioned in October of

2000, added a charge of making a false statement as well as

additional counts of obstruction of justice.

The superseding indictment sketched a corrupt

relationship between the appellant and the Winter Hill Gang. In

the course of that relationship, the appellant allegedly sold

protection, the identities of FBI informants, and the like to

Bulger and Flemmi in exchange for a googol of bribes and favors.

The relationship allegedly continued even after the appellant

retired; one of his final acts was said to be the tip to Bulger and

Salemme that allowed them to abscond before the looming indictment

materialized.

The appellant proclaimed his innocence and stood trial on

the superseding indictment. The jury found him guilty on charges

of racketeering, obstruction of justice, and making a false

statement. See 18 U.S.C. §§ 1962(c), 1503, 1001. On September 16,

2002, the district court sentenced him to 121 months in prison

followed by two years of supervised release. We affirmed both his

conviction and his sentence. See Connolly I, 341 F.3d at 35.

-4-
 On May 27, 2005 — almost three years to the day after the

jury returned its verdict — the appellant moved for a new trial.

See Fed. R. Crim. P. 33. The motion rested on four categories of

newly discovered evidence, each of which (according to the

appellant) undermined the case against him and contributed to a

showing that the government had acted unlawfully.

The first of these proffers constituted an FBI 302

report. That report, promptly disclosed to the appellant's counsel

and the trial judge by the government, memorialized the accusations

of a confidential source (CS). The CS, himself a mobster of some

repute, recounted purported jailhouse conversations with Salemme

(who had been a prominent witness at the appellant's trial).

According to the CS, Salemme, while incarcerated, had recanted

almost the entirety of his testimony, confessing that he had

perjured himself and describing how the government had urged him

down the path of prevarication.

The appellant's second evidentiary proffer consisted of

a 2004 report by the Committee on Government Reform of the United

States House of Representatives. H.R. Rep. No. 108-414 (2004),

available at http://www.gpoaccess.gov/serialset/creports/everything-

secret.html. That report, titled in part "The FBI's Use of

Murderers as Informants," is a caustic chronicle of the dark side

of the FBI's relationship with organized crime spanning three

decades (from the mid-1960s through the mid-1990s). The appellant

-5-
offered the report in an apparent effort to show that the United

States Attorney's Office in Massachusetts was no stranger to

internal misconduct.

The appellant's third proffer comprised a list of alleged

discrepancies between the testimony of witnesses who had testified

at his trial and things that those same individuals subsequently

said in depositions taken in a gallimaufry of civil actions.

Fourth, and finally, the appellant pointed to Salemme's recent

indictment for making false statements to federal agents.

The government opposed the motion. As part of its

opposition, it submitted affidavits from five members of the

prosecution team, three of whom were lawyers and all of whom flatly

denied any knowledge of misconduct or perjury.

On November 15, 2005, the district court summarily denied

the motion. The court acted without holding a hearing and saw no

need to write a rescript. This timely appeal followed.

II. ANALYSIS

The appellant's briefs present four different, but

interrelated, theoretical bases for relief. These rest on (i)

newly discovered evidence; (ii) withholding of evidence by the

government; (iii) prosecutorial misconduct (related to perjury and

the withholding of evidence); and (iv) manifest injustice. We

address the first two theories as an ensemble. We then comment

upon the absence of an evidentiary hearing. Finally, we explain,

-6-
albeit briefly, why we need not consider the appellant's third and

fourth theories.

A. The Preserved Theories.

The appellant's first two theories implicate newly

discovered evidence and the withholding of evidence, respectively.

Both of these closely related theories were squarely raised below

and, thus, are properly preserved. Consequently, we consider them

on the merits. We start, however, with the applicable legal

standards.

A district court's disposition of a Rule 33 motion for

a new trial in a criminal case is ordinarily a "judgment call."

United States v. Maldonado-Rivera, 489 F.3d 60, 65 (1st Cir. 2007).

Hence — at least where the trial judge revisits the case to pass

upon the new trial motion — an appreciable measure of respect is

due to the "presider's sense of the ebb and flow of the recently

concluded trial." United States v. Natanel, 938 F.2d 302, 313 (1st

Cir. 1991). We therefore review such rulings solely for abuse of

discretion. See United States v. Alicea, 205 F.3d 480, 486 (1st

Cir. 2000). Of course, a district court abuses its discretion

whenever it predicates its ruling on an erroneous view of the law,

see United States v. Snyder, 136 F.3d 65, 67 (1st Cir. 1998), and

abstract questions of law engender de novo review, see United

States v. Josleyn, 206 F.3d 144, 151 (1st Cir. 2000). A claim that

-7-
the district court applied an incorrect standard falls within this

purview.

The appellant attempts to complicate this rather

straightforward set of rules. First, he claims that different

standards of review apply to motions for new trials in criminal

cases depending upon whether governmental withholding of evidence

is alleged. Insofar as the standard of review is concerned, we

implicitly have rejected that distinction in earlier cases, see,

e.g., Maldonado-Rivera, 489 F.3d at 65; Josleyn, 206 F.3d at 151,

and we explicitly reject it here.

Second, the appellant contends that the order appealed

from should be reviewed de novo because the district court

summarily disposed of his motion without composing an opinion.

This contention is unpersuasive. While we always value an

elaboration of the district court's reasoning, it has long been

recognized that a written rescript is not a sine qua non for the

disposition of a new trial motion in a criminal case. See, e.g.,

United States v. Mangieri, 694 F.2d 1270, 1284-87 (D.C. Cir. 1982);

United States v. Wright, 625 F.2d 1017, 1019 (1st Cir. 1980);

United States v. Holy Bear, 624 F.2d 853, 856 (8th Cir. 1980). Nor

does the lack of such a rescript automatically trigger a more

searching standard of appellate review. When the trial court has

not expounded its rationale, the court of appeals will peruse the

record, identify the issues and the controlling legal rules, and

-8-
review the denial of the motion accordingly. Cf. United States v.

Malpica-Garcia, 489 F.3d 393, 395 (1st Cir. 2007) ("[W]e may affirm

a district court judgment on any ground supported by the record.").

We turn from the standard of review to the standards that

attend the resolution of a criminal defendant's motion for a new

trial on the basis of freshly discovered evidence. Typically, the

defendant must make a four-part showing: (i) that the evidence was

unknown or unavailable to him at the time of trial; (ii) that his

failure to learn of it did not result from a lack of due diligence;

(iii) that the evidence is material, not merely cumulative or

impeaching; and (iv) that its availability is likely to bring about

an acquittal upon retrial. United States v. Huddleston, 194 F.3d

214, 218 (1st Cir. 1999); Wright, 625 F.2d at 1019. Every element

of this test (known in this circuit as the "Wright test") is

essential, and a failure to establish any one element will defeat

the motion. Maldonado-Rivera, 489 F.3d at 66.

A somewhat different paradigm applies when the defendant

makes a colorable claim that he would have had access to the newly

discovered evidence but for the government's failure to disclose it

in accordance with the imperatives of Brady v. Maryland, 373 U.S.

83, 87 (1963). See United States v. González-González, 258 F.3d

16, 20 (1st Cir. 2001). This more defendant-friendly standard

applies, then, to what are colloquially known as "Brady

violations."

-9-
 There are three components of an authentic Brady

violation. The evidence at issue (whether exculpatory or

impeaching) must be favorable to the accused; that evidence must

have been either willfully or inadvertently suppressed by the

government; and prejudice must have ensued. Strickler v. Greene,

527 U.S. 263, 281-82 (1999). For purposes of a new trial motion,

the same standard applies to claims that the government knowingly

used perjured testimony. González-González, 258 F.3d at 21. That

standard, however, does not extend to the unwitting use of perjured

testimony. Huddleston, 194 F.3d at 221.

Although we have said that this standard (which we shall

sometimes call the "Brady standard") is defendant-friendly, that

does not mean that proof of a Brady violation automatically entitles

a defendant to a new trial. Even a defendant who moves for a new

trial on the basis of one or more Brady violations still must

establish the first two elements of the Wright test. The difference

affects only the latter two elements; in a Brady scenario, those

elements are replaced with the unitary requirement that the

defendant establish "a reasonable probability that, had the evidence

been disclosed to the defense, the result of the proceeding would

have been different." United States v. Bagley, 473 U.S. 667, 682

(1985); see Kyles v. Whitley, 514 U.S. 419, 434 (1995)

(characterizing that "reasonable probability" as sufficient to

"undermine[] confidence in the outcome of the trial").

-10-
 At least two other nuanced differences distinguish the

two standards. First, the Wright test demands an actual probability

that the result would have differed, whereas the Brady test speaks

in terms of something less — a merely theoretical (but still

reasonable) probability. See United States v. Sepulveda, 15 F.3d

1216, 1220 (1st Cir. 1993). Second, the Wright test categorically

discounts "merely impeaching" evidence as "immaterial." Wright, 625

F.2d at 1019. When a Brady violation is involved, however, the

Wright materiality inquiry disappears as a distinct element, with

the result that undisclosed impeachment evidence, if it suffices to

undermine confidence in the outcome of the trial, may carry the day.

See, e.g., Bagley, 473 U.S. at 682 (applying Brady test to

impeachment evidence); United States v. Dumas, 207 F.3d 11, 16 (1st

Cir. 2000) ("[W]e recognize that impeachment evidence, if powerful

enough, could constitute grounds for a new trial . . . .").

Against this backdrop, we consider the "new" evidence

proffered by the appellant to see how that evidence fares. As a

preliminary matter, we note the appellant's allegation that all of

this evidence is prejudicial and that almost all of it would have

been available to him at trial but for the government's wrongful

withholding.1

1
One prominent exception consists of Salemme's alleged perjury
with respect to his involvement in the murder of Stephen DiSarro.
We discuss that proffer separately. See text infra.

-11-
 This is a dubious premise. The law is settled that "[n]o

inference of government knowledge of perjury arises from the mere

fact of a convict's hearsay report that a material witness recanted

testimony." González-González, 258 F.3d at 23. In an abundance of

caution, however, we elect to examine this evidence under the more

defendant-friendly Brady standard. Should the appellant's

challenges fail under this standard, then a fortiori they would fail

under the more stringent standard applicable to non-Brady claims.

See Josleyn, 206 F.3d at 148 n.2.

The appellant places most of his emphasis upon Salemme's

jailhouse recantation. To recapitulate, Salemme had been a witness

at the appellant's trial. His testimony was helpful to the

prosecution; he testified, among other things, that on several

occasions the appellant had given assurances that he would keep

Salemme apprised of developments with respect to the ongoing federal

investigation. Salemme also testified that, on January 5, 1995 —

two weeks after the appellant allegedly had leaked information about

the nascent indictment — he had met with Flemmi and other mobsters

at the home of Salemme's ex-wife; that Flemmi had informed him that

the indictment was about to be unsealed; and that Flemmi knew that

fact thanks to the appellant.

Importantly, however, Salemme's testimony did not occur

in a vacuum. Several other witnesses corroborated aspects of it.

Most notably, Kevin Weeks, Whitey Bulger's former aide-de-camp,

-12-
testified extensively about the appellant's corrupt relationship

with the Winter Hill Gang. For example, Weeks described an episode

that transpired on December 23, 1994, when he met with the appellant

at a local Winter Hill hangout. The appellant summoned him into a

cold-storage freezer and informed him of the impending indictment.

Others — such as the appellant's secretary, Kathleen Orrick, and his

high-school chum, John Ford — testified to Salemme being in the

appellant's company at roughly the times that Salemme claimed to

have been. Salemme's ex-wife, Alice McLaughlin, also confirmed that

the January 5 meeting took place at her abode. Thus, although

Salemme testified to a number of things as to which no other witness

had personal knowledge, much of his testimony received substantial

circumstantial corroboration.

As a counterpoint to this testimony, the appellant

proffers the FBI 302 report. That report describes the CS

befriending Salemme while both men were in the witness protection

program. According to the CS, Salemme recounted a litany of

prosecutorial misconduct, interwoven with tales of his own

mendacity. The CS attributed the following kinds of statements to

Salemme:2

C The government had shaped his
narrative by asking questions such as,

2
Neither party disputes that the CS's statement to the FBI was
made during the fall of 2004. Thus, the statement was not
available at the time of trial and the appellant could not have
obtained it earlier with even the utmost diligence.

-13-
 "are you sure it didn't happen like
this?" and then describing alternate
versions of events.

C Salemme came to realize that the
prosecutors wanted him to testify that
he had paid the appellant for
information, so he obliged.

C Prior to the unsealing of the
indictment, Salemme had never met the
appellant and was able to recognize
him only because the FBI had furnished
him with a photograph.

C Salemme ascribed his willingness to
help convict the appellant to profound
hatred stemming from what the
appellant had done to Salemme and his
family.

C When shown an article about someone
sentenced to several years in prison
for perjury, Salemme stated that if
the article was accurate, he himself
should receive "a thousand years."

This is an illustrative list, not a complete compendium — but it

suffices to set the stage for our discussion.

The appellant's rationale is that the report demonstrates

that Salemme's trial testimony was made up out of whole cloth. But

even if we assume, favorably to the appellant, that the CS

faithfully recounted Salemme's rodomontade, the district court

easily could have found that recantation unworthy of credence and

insufficient to shake its confidence in the jury's verdict. We

explain briefly.

"It is well established that recantations are generally

viewed with considerable skepticism." United States v. Carbone, 880

-14-
F.2d 1500, 1502 (1st Cir. 1989). Here, the wisdom of that insight

is evident. As a mafioso turned government witness, Salemme had

called both his honor and his safety into question. It would seem

only natural, then, that he would try to rationalize his actions and

insulate himself from their consequences. See, e.g., González-

González, 258 F.3d at 22 (taking into consideration in Brady inquiry

the witness's incentives for recantation given the risk that she

might meet defendant in the future); United States v. Badger, 983

F.2d 1443, 1457 (7th Cir. 1993) (discounting probative value of

affidavit in light of threats of jailhouse retaliation against

affiant); United States v. Leibowitz, 919 F.2d 482, 483 (7th Cir.

1990) (describing affidavit as "not worthy of belief" in light of

affiant's psychological frailty and the threats likely communicated

by defendant).

The mixture of braggadocio and self-serving excuses

contained in the recantation fits this pattern. And Salemme likely

would have thought that he had little to lose by boasting about his

ostensible perjury in the prison yard to a fellow mobster; there

would have seemed to be little chance that the statements could come

back to haunt him. In short, Salemme's recantation, like many

jailhouse recantations, lacked any meaningful indicia of reliability

and, therefore, was "properly regarded as 'highly suspicious.'"

United States v. Walker, 25 F.3d 540, 549 (7th Cir. 1994) (citation

omitted); see United States v. Goodwin, 496 F.3d 636, 641 (7th Cir.

-15-
2007) (noting that inmate composed the recanting affidavit under

threat of coercion and later withdrew it); United States v. Baker,

479 F.3d 574, 577-78 (8th Cir. 2007) (refusing to credit hearsay

jailhouse recantation by trial witness); González-González, 258 F.3d

at 22 (describing evidence as weak when it depended upon the

credibility of convicted felons and those under threat of

retaliation for their prior adverse testimony).

To be sure, the inference of unreliability that arises

from the circumstances of the recantation is merely permissive.

Here, however, that inference is reinforced by the record. When one

examines the individual assertions that the CS attributes to

Salemme, a worrying pattern emerges: many of Salemme's readily

verifiable assertions are demonstrably false. We offer a few

examples.

First, Salemme supposedly told the CS that his prison

term was to be commuted in exchange for his perjured testimony. But

the record belies that assertion: it shows unequivocally that the

terms of Salemme's reduced sentence were left to the determination

of the district court, not the prosecution. For his part, the

prosecutor did not advocate for Salemme's release from incarceration

but, rather, recommended a 16-month sentence reduction (from 136 to

120 months).

Second, Salemme supposedly told the CS that he did not

know the appellant before his indictment and arrest. Yet two

-16-
independent witnesses — Orrick and Ford, each of whom was favorably

disposed to the appellant — testified very clearly to the contrary.

And to add frosting to this particular piece of cake, the record

reflects that the appellant personally arrested Salemme in 1972. The

district court certainly did not have to accept hook, line, and

sinker the totally implausible notion that Salemme could so

perfectly turn the other cheek as to expunge from his memory the law

enforcement officer who placed him under arrest and set in motion

a chain of events that led to him spending 16 years in a federal

penitentiary.

Third, and finally, Salemme supposedly told the CS that

he had been pressured by the prosecutors. However, the trial record

contains nothing that would suggest that this was so. The

government filed affidavits from no fewer than five members of the

prosecution team flatly denying that any such misconduct had

occurred. These affidavits have a patina of plausibility since

virtually all of Salemme's initial debriefings were conducted in the

presence of his attorney.3 These debriefings were consistent with

the testimony that Salemme later gave at the appellant's trial, and

one could just as easily believe in the tooth fairy as believe that

3
Salemme was represented by Anthony M. Cardinale, a veteran
criminal lawyer. In an earlier, unrelated opinion, we had occasion
to quote from an article in Boston magazine chronicling Cardinale's
storied career and noting that "[e]very troubled mobster in
[Boston] kisses Cardinale's ring sooner or later." United States
v. Boylan, 898 F.2d 230, 258 n.17 (1st Cir. 1990) (quoting "All in
the Family," Boston magazine, Aug. 1988, at 18).

-17-
a lawyer of Cardinale's stature was complicit in the most brazen

sort of prosecutorial misconduct.

These are neither minor nor collateral matters.4 They go

to the very heart of the issues upon which the recantation is meant

to shed light. Although the fact that a recantation is peppered

with apparent falsehoods does not compel a court to discredit it,

that fact certainly gives the court sufficient reason to invoke the

hoary doctrine of falsus in uno, falsus in omnibus.5 A court may

reject a recantation on that basis.

4
The government also argues that Salemme lied when, in the
course of his recanting statements, he boasted of having perjured
himself by testifying as to giving money to the appellant. The
government claims that this boast is clearly belied by the trial
transcript, according to which Salemme testified to no more than
having given money to Flemmi for the purpose of paying it to the
appellant.
In this regard, we think that the government demands too much
precision, especially given the casual context of the alleged
recantation. Salemme plausibly could claim having testified to
paying the appellant, even if that testimony did not indicate in
haec verba that Salemme personally handed over the cash. Indeed,
the government itself appears to have adopted this interpretation
of Salemme's testimony; in his summation to the jury, the
prosecutor argued that Salemme had testified that "he and Flemmi
had been paying money to John Connolly" and that "they had paid
$5,000 . . . to Mr. Connolly." We should not place greater demands
for precision upon a prison-yard conversation than upon an account
provided by a prosecutor in open court.

5
Although the maxim "falsus in uno, falsus in omnibus" may be
fallacious when understood as a mandatory rule of inference, see
Dicenso v. Michaels, 668 F.2d 633, 634 (1st Cir. 1982), it retains
validity as a basis for a permissive inference, see, e.g., Kanawha
& M.R. Co. v. Kerse, 239 U.S. 576, 581 (1916); Castañeda-Castillo
v. Gonzales, 488 F.3d 17, 23 (1st Cir. 2007) (en banc); Yongo v.
INS, 355 F.3d 27, 33 (1st Cir. 2004).

-18-
 Two final considerations cement our intuition that the

appellant has failed to demonstrate an abuse of the district court's

discretion. For one thing, no evidence has been presented

suggesting that Salemme himself would be willing, under oath, to

admit to perjury. Relatedly, there has not been a credible argument

put forth as to how the hearsay account contained in the FBI 302

report could be introduced into evidence. Cf. Fed. R. Evid.

801(d)(1) (defining as non-hearsay prior inconsistent statements

"given under oath subject to the penalty of perjury at a trial,

hearing, or other proceeding"); 804(b)(3) (providing exception for

statements against interest when declarant is unavailable). These

deficiencies weigh in the balance against reversing the district

court's exercise of its discretion. See United States v. Rosario-

Diaz, 202 F.3d 54, 66 (1st Cir. 2000) (discounting the force of new

evidence in a Brady inquiry in light of the evidence's likely

inadmissibility).

For another thing, even assuming that the recantation

were true, it would not prove very much. In his musings to the CS,

Salemme gave no indication that the appellant was innocent of the

charged crimes. In this sense, his recantation, if believed, would

merely be impeaching and, consequently, would have a limited effect

-19-
upon the outcome of a new trial in which substantial corroborating

testimony existed.6

Along the same lines, the recantation would to some

extent be cumulative. During cross-examination, the jurors learned

much about Salemme's unsavory past. They knew, for example, that

he was a career criminal who had blood on his hands in consequence

of numerous murders and other violent crimes. The jurors also knew

the details of his plea bargain with the government.

Given Salemme's extensive criminal history, it would not

have been an abuse of discretion for the district court to find that

the absence of additional cross-examination on essentially the same

well-developed theme would not undermine confidence in the jury's

verdict. See United States v. Cruz-Kuilan, 75 F.3d 59, 63 (1st Cir.

1996); see also United States v. Sanchez, 917 F.2d 607, 618-19 (1st

Cir. 1990) ("Impeachment evidence, even that which tends to further

undermine the credibility of the key Government witness whose

credibility has already been shaken due to extensive cross-

examination, does not create a reasonable doubt that did not

otherwise exist when that evidence is cumulative or collateral.")

6
Though the Brady test does not treat impeachment evidence
with the same categorical wariness as does the Wright test, this
Court in applying Brady has recognized that the force of
impeachment evidence is diminished when the witness's testimony is
supported by substantial corroborating evidence, see González-
González, 258 F.3d at 22-23, or when the impeachment evidence is
"cumulative or collateral." United States v. Sanchez, 917 F.2d
607, 619 (1st Cir. 1990) (citation and internal quotation marks
omitted).

-20-
(quoting United States v. Shelton, 588 F.2d 1242, 1248 (9th Cir.

1978)).

In this instance, all roads lead to Rome. For the

reasons limned above, the jailhouse recantation, even if it

occurred, did not require the granting of a new trial.

This conclusion does not end our odyssey. The

appellant's motion for a new trial rested on more than Salemme's

alleged recantation, so we need to inspect those additional items

of "new" evidence.

We start with the appellant's frequent and tantalizing

references to a recently issued congressional report and a series

of depositions taken over the course of several civil suits arising

out of the maraudings of the Winter Hill Gang. Given the passing

nature of these references, it is an open question whether the

appellant has waived any right to have us consider this evidence.

See, e.g., United States v. Zannino, 895 F.2d 1, 17 (1st Cir. 1990)

(explaining that "issues adverted to in a perfunctory manner" are

deemed abandoned). For ease in exposition, however, we assume that

the answer to this threshold question is favorable to the appellant

and consider the impact of the evidence. As we shall explain, it

contributes very little to the appellant's cause.

The congressional report falls to the wayside on

relevancy grounds. The report, issued by the Committee on

Government Reform of the United States House of Representatives,

-21-
communicates the Committee's low opinion of the FBI's Boston field

office in regard to its handling of organized crime matters from the

mid-1960s to the mid-1990s. The report also raises a suspicion

that, during a 1997 Department of Justice investigation, the United

States Attorney's Office withheld evidence that Bulger and Flemmi

had been the beneficiaries of federal prosecutorial discretion.

These portions of the report, along with the Committee's

complaints about the Department of Justice's ready resort to

executive privilege, make interesting reading — but they have

nothing to do with the appellant. This trial concerned the

lawfulness vel non of the appellant's actions — a subject as to

which the report does not speak. Material of such dubious relevance

cannot be the basis for the granting of a new trial. See Rosario-

Diaz, 202 F.3d at 67 (finding newly discovered evidence purporting

to establish victim's involvement in drug operations irrelevant to

defendants' conviction for murdering her); United States v. West,

672 F.2d 796, 799-801 (10th Cir. 1982) (holding that newly

discovered Department of Justice inquiry into alleged prosecutorial

misconduct was not relevant to defendant's guilt).

The deposition testimony is newly discovered in the sense

that the civil proceedings for the most part post-dated the

appellant's indictment and trial. Many of those excerpts are at

least arguably relevant, even though they touch upon predicate acts

of which the appellant was not ultimately convicted. As we

-22-
concluded on direct review of the appellant's conviction, the jury

could have relied upon evidence of those acts in finding him guilty

of participating in a continuing criminal enterprise. See Connolly

I, 341 F.3d at 27. Relevancy aside, however, the deposition

testimony suffers from a different infirmity: the testimonial

inconsistencies noted by the appellant are of meager weight. To

illustrate, we offer a representative sampling.

The appellant observes that Weeks's trial testimony and

Flemmi's deposition testimony conflict as to the length of time that

Weeks controlled a particular Winter Hill fund. Unrebutted, though,

is the more salient fact, testified to by both men, that the bribes

paid to the appellant came out of that fund.

Next, the appellant notes that Weeks's trial testimony

about John McIntyre's murder is arguably inconsistent with his

deposition testimony in a civil case brought by McIntyre's heirs.

At trial, Weeks testified that Bulger murdered McIntyre after

learning from the appellant that McIntyre had become a snitch; in

his deposition, Weeks seemed to imply that Bulger killed McIntyre

because he found him "weak and [not to] be trusted." But these

answers are not necessarily inconsistent and, at any rate, the

appellant's responsibility for McIntyre's demise was not in issue

at trial (that is, the government never charged the appellant with

criminal conduct with respect to McIntyre's death). Thus, this

"new" evidence could not logically have affected the jury's verdict.

-23-
 The appellant also envisions an inconsistency between

Weeks's testimony that an FBI informant, Halloran, was outed by the

appellant and the deposition testimony of a mysterious individual

named Ronald Costello that Halloran's informant status was public

knowledge. Here, again, there is no necessary contradiction between

the two statements (and, thus, no basis for believing that the jury

verdict hung in the balance).

Wholly apart from these bits and pieces of evidence, the

appellant proffers yet a fourth category of newly discovered

evidence. He notes that, on October 27, 2004, a federal grand jury

indicted Salemme, alleging that Salemme had deceived federal

prosecutors in plea negotiations by falsely denying his involvement

in the gangland murder of Stephen DiSarro.

This proffer spotlights an apparent falsehood in

Salemme's testimony during the appellant's trial. On cross-

examination, Salemme denied having participated in any murders while

at the head of La Cosa Nostra in New England. That testimony

reiterated a statement that he had made to prosecutors in

negotiating his plea bargain. The government cannot square the

circle. If it believes that facts exist that prove Salemme's

involvement in DiSarro's slaying — and the recent indictment

indicates that it does — it cannot be allowed to argue that Salemme

was wholly truthful in his trial testimony in this case.

-24-
 As it happens, this possible perjury is, for present

purposes, of minor moment. DiSarro's murder occurred in the course

of a criminal enterprise separate from the Winter Hill Gang and

remote from the appellant. Furthermore, the jury already was aware

of Salemme's nefarious past and had good reason to believe that his

every word was not deserving of unqualified respect. In these

circumstances, the district court was fully entitled to conclude

that neither this alleged perjury nor the attribution of one more

murder to Salemme was likely to taint his other testimony.

This situation closely approximates the one before us in

Sanchez. There, in the context of a Brady inquiry, we considered

the materiality of undisclosed evidence of payments from the

Commonwealth of Massachusetts to a witness, made in exchange for the

witness's services as an informant. See 917 F.2d at 618. In

concluding that the payments were immaterial, we noted that those

payments were dwarfed by similar payments made by the federal

government — these latter payments having in fact been disclosed —

and that the Commonwealth's payments had been for services unrelated

to Sanchez's case. Id.

That completes our canvass of the appellant's proffers of

newly discovered evidence. None of them creates any plausible basis

upon which to question the integrity of the verdict. And here, the

whole does not exceed the sum of the parts: it cannot fairly be said

that these proffers, each of which is asthenic, gain strength by

-25-
aggregation. Taken together, they do not impugn the verdict. See,

e.g., Dumas, 207 F.3d at 18.

In all events, as an appellate tribunal we must cede

appropriate deference to the trial court with respect to the

granting or denial of a new trial motion. See Alicea, 205 F.3d at

486. That deference is magnified where, as here, the judge who

considered the motion presided over the trial itself. Natanel, 938

F.2d at 313. Viewed from this vantage point, there is no principled

way that we can say the district court abused its discretion in

finding this congeries of evidence insufficient to undermine its

confidence in the jury verdict.

B. The Need for an Evidentiary Hearing.

The appellant has a fallback position. He asseverates

that the district court should, at the very least, have granted his

motion for an evidentiary hearing on the new trial motion.

The baseline rule is that a "criminal defendant has no

absolute or presumptive right to insist that the district court take

testimony on every motion." United States v. Panitz, 907 F.2d 1267,

1273 (1st Cir. 1990). Accordingly, when considering the question

of whether an evidentiary hearing should be granted in connection

with a Rule 33 motion, a district court should ask if the defendant

made a threshold showing sufficient to warrant such a hearing.

Alicea, 205 F.3d at 487. In pursuing this inquiry, the court must

make a practical, commonsense evaluation. When, for example, the

-26-
motion is "conclusively refuted . . . by the files and records of

the case," an evidentiary hearing would be supererogatory. Carbone,

880 F.2d at 1502; accord González-González, 258 F.3d at 23.

The short of it is that evidentiary hearings on new trial

motions in criminal cases are the exception rather than the rule.

Such motions ordinarily are decided on the basis of affidavits,

without convening evidentiary hearings. See United States v.

Kearney, 682 F.2d 214, 219 (D.C. Cir. 1982). Even disputed matters

of fact arising from post-trial motions "are often properly decided

on the basis of affidavits." United States v. Abou-Saada, 785 F.2d

1, 7 (1st Cir. 1986) (Breyer, J.).

The granting of an evidentiary hearing lies within the

district court's sound discretion, and the denial of an evidentiary

hearing is reviewed for abuse of that discretion. United States v.

Colón-Muñoz, 318 F.3d 348, 358-59 (1st Cir. 2003). The court of

appeals should defer, within wide margins, to the district court's

assessment — at least when, as in this instance, it is reasonable

to believe that the district court possesses a more finely honed

sense of the situation and a superior feel for the facts. See,

e.g., United States v. Winter, 663 F.2d 1120, 1155 (1st Cir. 1981).

On this record, we see nothing that would demand an

evidentiary hearing. The district judge had presided over the trial

and was steeped in the lore of the case. The appellant's position,

as it appeared from the papers submitted to the court, left few

-27-
avenues for additional corroboration. There was no showing below —

and there has been none here — that an evidentiary hearing was

either necessary or desirable. Consequently, we hold that the

district court acted within the realm of its discretion in denying

the appellant's request.

C. The Unpreserved Theories.

This leaves the appellant's theories of prosecutorial

misconduct and manifest injustice. The appellant failed adequately

to raise either of these theories at the trial-court level. His

original motion for a new trial did not advert to either theory, and

his reply to the government's opposition did not remedy this

omission.

To be sure, the term "prosecutorial misconduct" can be

found scattered throughout the record below. But these sporadic

allusions to prosecutorial misconduct are unaccompanied by any

attempt at developed argumentation and are bereft of any citations

to relevant authority. They are, therefore, waived. See Paterson-

Leitch Co. v. Mass. Mun. Wholesale Elec. Co., 840 F.2d 985, 990 (1st

Cir. 1990) ("One should not be allowed to defeat the system by

seeding the record with mysterious references to unpled claims,

hoping to set the stage for an ambush should the ensuing ruling fail

to suit."); see also Zannino, 895 F.2d at 17.

As for manifest injustice (a shorthand for the theory

that the district court's supervisory powers should have been

-28-
exercised to overturn a verdict procured by egregious falsehoods and

government misconduct), the appellant's filings below contain not

even a glancing mention of this point. A theory that makes its

debut in the court of appeals is perforce unpreserved; "[i]t is a

bedrock rule that when a party has not presented an argument to the

district court, she may not unveil it in the court of appeals."

United States v. Slade, 980 F.2d 27, 30 (1st Cir. 1992).

IV. CONCLUSION

We need go no further. The evidence newly discovered by

the appellant is, in Judge Bownes's phrase, "not the straw that

would have broken the camel's back; it was just more chaff to

scatter in the wind." United States v. Martorano, 663 F.2d 1113,

1119 (1st Cir. 1981). Even under the more defendant-friendly Brady

standard, we cannot say that this evidence undermines our confidence

in the jury's verdict or justifies a finding that the district court

abused its discretion in failing to vacate the judgment and order

a new trial.

Affirmed.

-29-