UNITED STATES COURT OF APPEALS
                         For the Fifth Circuit

                     ___________________________

                             No. 92-7030
                     ___________________________


                      UNITED STATES OF AMERICA,

                                                   Plaintiff-Appellee,

                                VERSUS


                         JAMES GORDON KELLER,

                                                  Defendant-Appellant.

         ____________________________________________________

             Appeal from the United States District Court
                  For the Southern District of Texas
         ____________________________________________________


Before REAVLEY and DAVIS, Circuit Judges, and TRIMBLE,1 District
Judge.

DAVIS, Circuit Judge:

     James Gordon Keller was convicted in September 1991, of one

count of conspiracy to commit wire fraud, in violation of 18 U.S.C.

§ 371, and ten counts of aiding and abetting the commission of wire

fraud, in violation of 18 U.S.C. §§ 1343 and 2.        The government

presented evidence that Keller and several associates created a

scheme to obtain money from cancer patients and their families by

promising them, by way of interstate telephone conversations, an

effective treatment and cure for their cancer. The principal issue

on appeal is whether Keller was prejudiced by the government's


     1
          District Judge of the Western District of Louisiana,
sitting by designation.
failure    to   disclose   the   grand       jury    testimony    of    one    of    its

witnesses at the time of trial.          We conclude that the government's

nondisclosure was harmless error and affirm Keller's conviction.

                                     I.

     In March 1983, Keller formed the Universal Health Center

("UHC") in Matamoros, Mexico.                Keller advertised that the UHC

offered "an effective therapeutic approach to the treatment of

cancer."    Keller's treatment included injections of an amino acid

solution called       "Tumorex,"   "lymphatic         massages,"       and    "colonic

irrigations."     Although Keller was a college graduate, he had no

medical    or   scientific   training;         his    degree     was   in     business

administration and he spent much of his life working in the water

treatment business.

     One of Keller's co-defendants, Maxine Lowder, operated Western

Health Research, which was a referral and reservation center for

the UHC.    In telephone conversations, personal conversations, and

speeches, Keller and Lowder represented that their "success or cure

rate" for cancer patients who were treated with Tumorex and who had

not undergone prior conventional medical treatment was between 80

and 100%, and that their success rate for patients who had received

conventional medical care was between 40 and 65%.                        Keller and

Lowder also told patients that the United States Government and the

American    Medical    Association   ("AMA")         were   conspiring        to    keep

Tumorex out of the country because it would bankrupt the social

security system and the medical profession.




                                         2
     Keller told his patients that, with a device called the

"Digitron D Spectrometer," he could detect the location and density

of their cancer simply by having them hold a plastic plate attached

to the machine.       Keller also told his patients that he could

diagnose and treat cancer by placing the same plastic plate over a

Polaroid photograph of a person.           In addition, Keller informed his

patients that, by holding a plastic pendulum over a patient's body,

he could determine whether they were suffering from cancer and

whether they would benefit from certain medications or foods.

     Keller    was   tried    before   a    jury   in    August    1991.2   The

government presented testimony from two of the eleven patients

named in the indictment, as well as testimony from the relatives of

the other nine patients who had died.              All eleven patients had

terminal or incurable cancer, and all had been unsuccessfully

treated   by   conventional    medicine      before     seeing    Keller.   The

patients and relatives testified that they had contacted Lowder,

who told them about Keller's treatment methods and his "cure"

rates.

     The relatives described how Keller examined their family

members with the Digitron D Spectrometer and the pendulum.                  They

described the daily injections of Tumorex, the use of Polaroid

photographs to diagnose and treat patients, the prescribed diets,


     2
         Keller co-defendants were tried in 1985. Keller was in
Mexico until March 1991, when American agents allegedly abducted
him and transported him to the United States. On appeal, Keller
has abandoned his argument that the district court lacked personal
jurisdiction in light of the Supreme Court's decision in United
States v. Alvarez-Machain, 

(1992).

                                       3
and the "colonic irrigations."        The relatives recalled how Keller

told their family members that their "cancer was gone" or that they

had been "cured."       Finally, the relatives reported that their

family members had paid Keller between $2500 and $3000 for his

"treatment" and "cure" of their cancer.

      The physicians who diagnosed and treated the eleven patients

also testified.      Dr. Bruce Storrs testified that it was generally

understood that tumors, once removed, frequently recurred. He also

asserted that although a diuretic, such as Tumorex, would reduce

the swelling of a tumor temporarily, it would not cure the patient

of cancer.      Dr. Rand Allen Hock testified that the clinical

definition of a "cure" was "five years of disease free survival."

Both physicians testified that they were unfamiliar with the

Digitron D Spectrometer.

      The government also called Dr. Thomas Dorr, a pharmacologist,

who had tested Tumorex to determine whether it had any anti-cancer

effect. Dr. Dorr testified that his testing had shown that Tumorex

had no anti-cancer effect in a dosage administrable to humans.             He

testified, however, that studies had been published showing that L-

arginine, the amino acid in Tumorex, might inhibit the formation of

tumors, but that no published studies indicated that Tumorex had

any therapeutic effect once a tumor developed.

      The government concluded its case by presenting the testimony

of   FBI   Special   Agent   Robert   Nixon,    who   introduced   the   death

certificates of Keller's patients in an effort to demonstrate the

falsity of his alleged "cure" rates.           Nixon testified that he had


                                      4
conducted a survey of 103 patients that Keller had treated in 1983,

and that by 1985, 78 or 79 of these patients had died, and that by

1991, 91 had died.

      In his defense, Keller contended that he had a good faith

belief in the effectiveness of his treatment.          He presented two

expert witnesses who explained the theory underlying the Digitron

D Spectrometer, as well as four doctors who testified that Tumorex

was   an   effective   anti-cancer   agent.   Keller    also   presented

testimony on the salutary effects of diet in the prevention and

treatment of cancer.       In addition, twenty of Keller's former

patients testified that neither Keller nor any of his associates

promised them a cure or told them that they were cured at the end

of their treatment.     Finally, Keller testified in his own behalf,

stating that he had treated between 180 and 200 patients in 1983.

Keller also testified that he did not believe that cancer was

"curable," but rather, that it could be controlled for long periods

of time if a strict diet was maintained and other parts of his

program were followed.

      The jury found Keller guilty of one count of conspiracy to

commit wire fraud and ten counts of aiding and abetting the

commission of wire fraud.       The district court denied Keller's

motion for "judgment n.o.v." and for new trial.

      After filing his notice of appeal, Keller moved this court to

compel the government to transcribe and produce Agent Nixon's grand

jury testimony which the government had not produced at trial.        We

granted Keller's motion and remanded the case to the district court


                                     5
to determine whether the government's failure to produce this

alleged Jencks Act material warranted a new trial.                  Finding that

Keller had not been prejudiced by the government's nondisclosure,

the district court denied his motion for new trial.                       Keller

challenges this ruling on appeal and contends also that the record

evidence is insufficient to support the verdict.             We consider both

arguments below.

                                      II.

                                       A.

        Keller argues first that his conviction should be vacated and

a new trial ordered because the government failed to produce Agent

Nixon's grand jury testimony as required by Fed. R. Cr. P. 26.2.3

Although Keller moved for the production of all Jencks Act material

prior to trial, the district court rejected his argument that

untranscribed grand jury testimony was subject to disclosure under

Rule 26.2.

        In our earlier remand order, we instructed the district court

to follow the procedure outlined in United States v. Rivero, 

,   459-61   (5th   Cir.   1976),   to     determine     whether   the

nondisclosure of Nixon's grand jury testimony warranted vacating

Keller's conviction and ordering a new trial. The district court's

task under Rivero was to determine:          "whether there is or is not a

reasonable     possibility     that   the   absence    of   [the]    grand    jury

testimony affected the outcome of the case or handicapped [the


    3
          In 1977, Congress placed the substance of the Jencks Act,
18 U.S.C. § 3500, in the criminal rules as Rule 26.2.

                                       6
defendant] or his counsel in their presentation or defense."    

  The district court affirmed Keller's conviction after

finding that:

     there is no reasonable possibility that the absence of the
     grand jury testimony at trial affected the outcome of the case
     or otherwise significantly prejudiced the conviction in light
     of the extensive cross-examination at trial on each and every
     particular raised in the Defendant's Motion to Vacate.

     In United States v. Beasley, 

(1978), cert.

denied, 

(1979), we stated that fidelity to the

principle underlying the Jencks Act requires that "'appellate

courts should be hesitant to take it upon themselves to decide that

the defense could not have effectively utilized a producible

statement.'"    

(quoting Rosenberg v. United States, 

, 375-76 (1959) (Brennan, J., dissenting)).   The harmless

error standard in this context is quite strict:

     The inquiry cannot be merely whether there was enough to
     support the result, apart from the phase affected by the
     error. It is rather, even so, whether the error itself had
     substantial influence [on the judgment].

-30 (quoting Kotteakos v. United States, 

,

764-765 (1946)); see also United States v. McKenzie, 

,

609 (5th Cir. 1985), cert. denied, 

(1986).   A failure

to produce Jencks Act material at trial, however, is harmless error

where there is no "substantial inconsistency, contradiction or

variation" between the prior statements and the witness' trial

testimony. United States v. Merlino, 

, 1019 (5th Cir.

1979), cert. denied, 

(1980); see also United States

v. Welch, 

, 274 (5th Cir.), cert. denied, 

(1987); United States v. Sink, 

, 1051 (5th Cir. 1978),

                                 7
cert. denied, 

(1979).         We have carefully reviewed

Nixon's grand jury testimony and his trial testimony and find no

substantial inconsistency.

     In his grand jury testimony given April 30, 1991, Nixon

testified that he had surveyed 93 of the patients that Keller

treated between March and December 1983.      He stated that of the 93,

two did not have cancer before seeing Keller, and two would not

talk to him about Keller.      Of the remaining 89 patients, Nixon

reported that "as of one year later," 50.56% were dead and 41.47%

still had cancer.    We interpret this as an assertion that in 1984,

one year after the subjects of Nixon's study had completed their

treatment at Keller's clinic, 45 patients had died and 44 were

still alive.   Nixon concluded his remarks by noting that: "We have

40-something people that were still alive the last time we checked

[presumably in 1984].     We're going to have to get out there and

find out if they're still alive."

     Four months later, in August 1991, Nixon testified at trial

that he had surveyed 103 of Keller's patients, that 91 of the 103

were dead as of the time of trial, and that of the 12 still alive,

only three claimed that they had been cured.              During cross-

examination, the district court asked Nixon: "How many of those 93

[sic] had already died by 1985?"       Nixon responded:   "There was ---

I think it was like 78 or 79 of them, in fact, back in 1985 that

had already died."

     Keller maintains that Nixon's trial testimony that 78 or 79

people had died by 1985 conflicts with his grand jury testimony


                                   8
that 45 people had died "the last time we checked."                We do not

agree.   When Nixon testified before the grand jury in April 1991,

he explained that he had not updated his research since 1984, and

that he would have to do further research to determine how many of

the "40-something" patients that were still alive in 1984, had died

by   1991.     Before   testifying    at   trial   in   August   1991,   Nixon

obviously had performed this additional research. At trial, he was

able to report that 91 of the 103 patients that he surveyed had

died by 1991, a statistic which had not been available when he

testified before the grand jury four months earlier.

      Having performed the follow-up research as he promised, Nixon

undoubtedly knew how many patients had died since he had performed

his initial research in 1984.        It therefore is reasonable that, at

trial, he would have known how many patients had died by 1985.

Consequently, when the district judge asked Nixon how many of

Keller's patients had died by 1985, he was able to respond:                 "I

think it was like 78 or 79 of them, in fact, back in 1985 that had

died."

      Keller also maintains that Nixon's grand jury testimony that

he surveyed 93 patients conflicts with his trial testimony that he

surveyed 103 patients.      Again, we disagree.         Nixon performed his

first survey in 1984, and a follow-up survey between April and

August 1991.    That Nixon increased the patient base by 10 when he

performed his supplemental survey in 1991 creates no inconsistency.




                                      9
                                  B.

     Keller argues next that the government presented insufficient

evidence to support the jury's finding that he intended to defraud

his patients.    "In assessing a challenge to the sufficiency of the

evidence, we must consider the evidence in the light most favorable

to the verdict and must afford the government the benefit of all

reasonable inferences and credibility choices."     United States v.

Stouffer, 

, 921-22 (5th Cir.), cert. denied, 

(1993).    The key question is whether a rational jury could have

found the essential elements of the crime beyond a reasonable

doubt.   

a conviction for wire fraud under 18 U.S.C. § 1343,

the government must present evidence of (1) a scheme to defraud,

and (2) the use of, or causing the use of, wire communications in

furtherance of the scheme.     United States v. Dula, 

,

778 (5th Cir.), cert. denied, 

(1993). The government

also must prove that defendant intended to defraud or deceive his

victims.   United States v. St. Gelais, 

, 95 (5th Cir.),

cert. denied, 

(1992).    It is this last element of the

offense that Keller contends is not supported by the evidence.

     The requisite intent to defraud exists if the defendant acts

"knowingly and with the specific intent to deceive, ordinarily for

the purpose of causing some financial loss to another or bringing

about some financial gain to [himself]."    

  Also, "proof

of intent [to defraud] may arise by inference from all of the facts

and circumstances surrounding the transactions."    United States v.


                                  10
Shively, 

, 814 (5th Cir.), cert. denied sub nom.

Johnson v. United States, 

(1991).

     We are persuaded that the facts and circumstances surrounding

Keller's claim that he could cure cancer, when combined with his

bizarre treatment method, amply support the jury's conclusion.

Keller represented that he could "cure" cancer, and told his

patients that they were "cancer free" following their treatments.

Keller   extracted   hundreds   of    thousands   of   dollars   from   his

patients, which he insisted be paid in cash.       Most importantly, the

government demonstrated the outrageous nature of Keller's claim

that he could diagnose cancer with a plastic pendulum, a Polaroid

photograph, and the "Digitron D Spectrometer."

     Based on this evidence, the jury could reasonably infer that

Keller knew his treatment was ineffective and that therefore he

intended to defraud his patients when he told them he could cure

their cancer.   Moreover, the jury was entitled to conclude that

Keller's claims that he could diagnose cancer with a plastic

pendulum and a Polaroid photograph were patently false and that

Keller knew that they were false.4

     4
           Keller also challenges the denial of his motion for new
trial based on the district judge's remarks at sentencing that he
thought that Keller believed in his treatment method.        Keller
contends that the judge's remark indicates that the judge did not
believe that Keller had the requisite intent to defraud. Keller,
however, neglects to point out that the judge also stated that
"from all the evidence before me, I totally and completely concur
in the determination of the jury."        As our discussion above
demonstrates, the evidence presented was sufficient to support the
jury's conclusion, and the district judge obviously did not believe
the verdict was against the great weight of the evidence.       The
district court therefore did not err in denying Keller's motion for
new trial.

                                     11
                              III.

     Because we find that the government's nondisclosure of Agent

Nixon's grand jury testimony was harmless error, and because we

find that the government presented sufficient evidence of Keller's

intent to defraud, we affirm his conviction.

     AFFIRMED.




                               12