704 F.2d 86 (1983)

HUNTER, Circuit Judge:

   In a prosecution for mailing threatening letters in violation of 18 U.S.C. §§ 2, 876 (1976), the United States District Court for the Western District of Pennsylvania ruled that the government could not introduce correspondence in the defendant's handwriting in order to show stylistic similarities between that correspondence and the threatening letters at issue.  United States v. Clifford, 543 F. Supp. 424 (W.D. Pa. 1982). The government appeals from the pretrial order excluding that evidence.  We have jurisdiction to hear the appeal pursuant to 18 U.S.C. § 3731 (1976). We will reverse.


   On March 9, 1982, a federal grand jury returned a two-count indictment against appellee Russell Clifford.  The indictment charged Clifford with mailing threatening letters to Charles Sharon, Clifford's successor as police chief of Saltsburg, Pennsylvania.  Those letters were printed in block style.

   Through its investigation the government acquired various pieces of cursive correspondence apparently written and signed by Clifford.  As one part of its case against Clifford, the government wanted to point out to the jury certain similarities in spelling, abbreviation, syntax, and paragraph structure between the block-style threatening letters and the cursive correspondence. Those similarities included misspellings of "figure" as "figuar" and "explosives" as "explodsives," the inconsistent use of either circles or dots for the letter "i," the lack of indentation or space between paragraphs, the use of certain abbreviations such as "Chas." for "Charles," and the use of long sentences in both sets of letters.

   The government, however, recognized that there was no direct proof that Clifford had written the cursive letters.  The government had circumstantial evidence that Clifford had written the letters because they bore the signature "Russell Clifford" and because they were addressed to city officials with whom Clifford regularly had business contacts as chief of police. Clifford, however, refused to stipulate that he had written the letters and refused voluntarily to provide the cursive handwriting exemplars necessary to establish his authorship.

   On March 26, 1982, the government filed a motion to compel Clifford to provide handwriting exemplars written in cursive style.  The government intended to use those cursive exemplars to prove that Clifford actually had written the cursive correspondence bearing his signature before it began pointing out the similarities between that cursive correspondence and the block-style threatening letters.  

   On April 14, 1982, the district judge conducted a hearing on the government's motion.  In an attempt to explain its need for the exemplars, the government showed to the judge an FBI report which summarized the FBI's forensic linguistic analysis of the threatening letters and of the cursive correspondence bearing Clifford's signature.  The FBI report stated that, based on a comparison between the threatening letters and the cursive correspondence, there was a "strong likelihood" that Clifford had written the threatening letters.  App. at 200A. The FBI report also stated that "you are reminded that since application of the linguistic method is not considered a positive means of identification, results of such examinations are provided for investigative assistance only, and are not intended for testimony." Id.

   The district judge focused on the statement that the results of forensic linguistic analysis "are not intended for testimony," id. In light of that statement, he questioned the admissibility at trial of the cursive correspondence for the purpose of comparing it with the threatening letters.  He further stated that if that evidence were inadmissible, there would be no need to compel Clifford to provide cursive exemplars.  The judge requested additional briefing and took the government's motion under advisement.

   On April 26, 1982, the government filed a Motion in Limine for a ruling that the cursive correspondence would be admissible at trial subject to proper authentication by the government.  On May 17, 1982, the district judge ordered a hearing to examine the state of the art of forensic linguistic analysis before ruling on either of the pending motions.  He ordered the government to produce the author of the FBI report and another forensic linguistics expert to testify at the hearing.  The government filed a response to that order emphasizing that it did not intend to call any expert witnesses at trial.  The government indicated that pursuant to the judge's order, however,  it would present the following three witnesses at the hearing: Dr. Murray S. Miron, professor of psycholinguistics at Syracuse University and consultant in linguistics to the FBI; Special Agent Ronald M. Furgerson, chief of the FBI laboratory unit responsible for forensic linguistics; and Penelope O. Pickett, aural analyst for the FBI laboratory and author of the FBI report at issue.

   The judge conducted the hearing on June 8, 1982.  Dr. Miron testified that forensic linguistic analysis is the process of matching stylistic similarities in different documents and then of assigning weight to those similarities according to their distinctiveness and frequency of occurrence.  He further stated that such an analysis could not provide a positive means of identifying the author of an anonymous document.  He indicated that the results of forensic linguistic analysis could be probative in establishing authorship but could not prove that one person, to the exclusion of all other possible authors, had written a document.

   The judge then asked Dr. Miron whether a jury, without the aid of an expert, could analyze the various similarities and "draw valid and certain conclusions beyond a reasonable doubt." App. at 123A.  Dr. Miron stated that an expert could help a jury to decide just how unusual or distinctive certain similarities are and thus how much probative weight to assign to those similarities. App. at 124A.  Dr. Miron further indicated that the jurors "might not draw the proper conclusions if not assisted in how they are to interpret the evidence." App. at 125A.

   Both Special Agent Furgerson and Ms. Pickett testified that forensic linguistic analysis does not provide a positive means of identification. Special Agent Furgerson testified that the purpose of forensic linguistic reports, such as the one prepared by Ms. Pickett in this case, is to provide guidance for investigators.  He testified that those reports are prepared for use in investigations with "no idea that eventually that information would be used for subsequent expert testimony." App. at 142A.  He stated that for that reason, the FBI has never attempted to have Ms. Pickett certified as an expert witness. App. at 135A-37A.

   Following the hearing the district judge orally denied the government's motion to compel handwriting exemplars.  The judge also ruled that the government could not introduce at trial the cursive correspondence listed in its Motion in Limine.  In his supplemental opinion the trial judge stated that he was exercising his discretion under Federal Rule of Evidence 403 and holding that the cursive correspondence was    inadmissible with or without the use of expert testimony.  The judge concluded that admitting such evidence would lead to a jury verdict "based on surprise and speculation, and not on proof beyond a reasonable doubt." 543 F. Supp. at 429.

   Relying on the testimony of Dr. Miron, the judge ruled that the correspondence was inadmissible without the aid of expert witnesses because such evidence was "inherently untrustworthy," id. at 431, and because of the strong possibility that a jury would be misled in evaluating the evidence on its own. He stated that "it is not the similarities or dissimilarities which are controlling, but it is the weight to be given to them which is the crucial factor in performing forensic linguistic analysis." Id. at 429. The judge concluded that the weighing process is "beyond the realm of an ordinary juror." Id.

   He then discussed the admissibility of the evidence with the aid of expert witnesses.  He concluded that the use of experts would not enhance the reliability of the evidence because forensic linguistic analysis "has not been shown to be trustworthy, reliable, accurate, or conforming to a generally accepted scientific theory." Id. at 431. The judge relied primarily on the testimony of the FBI analysts that forensic linguistic reports are used for investigative purposes only.  The judge also stated that he was unable to find any cases in which "testimony pertaining to the linguistic method has been permitted." Id. at 430.  


   On appeal the government contends that the district court erred in refusing to allow the jury to make its own comparisons between the two sets of documents. Appellee Clifford argues that the district court's exclusion of the evidence was a proper exercise of his discretion under Rule 403.    We note at the outset that a trial judge has broad discretion under Rule 403 to exclude relevant evidence when its "probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury." Fed. R. Evid. 403; see Hamling v. United States, 418 U.S. 87, 124-25, 41 L. Ed. 2d 590, 94 S. Ct. 2887 (1974). A trial judge's rulings under Rule 403 are rarely disturbed on appeal unless there is a clear showing that the judge abused his discretion.  United States v. Long, 574 F.2d 761, 766-68 (3d Cir.), cert. denied, 439 U.S. 985, 58 L. Ed. 2d 657, 99 S. Ct. 577 (1978).    In excluding the evidence in this case, however, the district judge misconstrued the nature of the government's request.  The government never intended to offer expert witnesses in forensic linguistic analysis nor did it ever intend to offer the FBI report into evidence as positive proof of Clifford's authorship.  Thus the judge's concern with allowing expert testimony in a relatively new scientific field was misplaced given the nature of the government's request.  The judge's concern with using the FBI report for in-court testimony was equally misplaced.

   The district judge excluded the cursive correspondence, rejecting the government's argument that the jury could make its own stylistic comparisons between the cursive correspondence and the threatening letters.  In excluding that cursive correspondence, however, the judge meshed the question of the sufficiency of the government's overall evidence with the question of the admissibility of the cursive correspondence.  He also considered factors going to the weight which the jury may attach to the evidence as factors affecting its admissibility.

   In assessing the probative value of the evidence the judge asked Dr. Miron whether a jury could examine the correspondence and "draw valid and certain conclusions beyond a reasonable doubt." App. at 123A (emphasis added).  A piece of evidence, however, need not conclusively prove a fact beyond a reasonable doubt in order to be admissible.  Fed. R. Evid. 401 advisory committee notes. The test of relevancy used by the trial judge to determine the admissibility of a single piece of evidence is less stringent than the standard the judge must later use under Fed. R. Crim. P. 29 to determine whether all the government's evidence is sufficient to allow the case to go to the jury. C. McCormick, McCormick on Evidence § 185 (2d.ed.1972). Rule 401 defines relevant evidence merely as evidence "having any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable." Fed. R. Evid. 401; see United States v. Steele, 685 F.2d 793, 808 (3d Cir.), cert. denied, 459 U.S. 908, 103 S. Ct. 213, 74 L. Ed. 2d 170 (1982); Carter v. Hewitt, 617 F.2d 961, 966 (3d Cir. 1980).
   The correspondence which the government wanted to present to the jury in this case is relevant.  The similarities between the cursive correspondence and the threatening letters, particularly the unusual misspellings, clearly have some tendency to make Clifford's authorship of the threatening letters more probable. The evidence is thus admissible unless "its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury." Fed. R. Evid. 403.

   We reject the district court's conclusion that the "jury will be seriously misled" by the evidence. 543 F. Supp. at 428. In reaching that conclusion the district judge relied on the testimony of Dr. Miron that the jurors "might not draw the proper conclusions if not assisted in how they are to interpret the evidence." App. at 125A.  Given that testimony the judge concluded that "the determination of the weight and significance [of the similarities] is something which is beyond the realm of an ordinary juror." 543 F. Supp. at 429.

   We disagree.  The jury's function in every case is "to weigh the credibility of witnesses, resolve evidentiary conflicts and draw reasonable inferences from proven facts." United States v. Young, 573 F.2d 1137, 1139 (9th Cir. 1978); see United States v. Barber, 442 F.2d 517, 522 (3d Cir.), cert. denied, 404 U.S. 958, 30 L. Ed. 2d 275, 92 S. Ct. 327 (1971). Clearly the weighing process which the jury will perform in this case is not the same scientific process which Dr. Miron performs.  In his work Dr. Miron relies on computer programs in order to identify similarities in writings.  He then uses standard "counts" of present American English, listings of the frequencies of occurrence of words, in order to assign weight to those similarities.  App. at 92A-97A.  We see no reason, however, why jurors cannot merely examine the documents for themselves and consider the similarities between the documents along with the rest of the evidence presented by the government.  

   Weinstein notes that "the technique [of identifying a writer by the internal patterns of the writing] -- without the aid of experts or computers -- is one long used in the courts." 5 J. Weinstein & M. Berger, Weinstein's Evidence para. 902(b)(4)[03] (1982) (citing Magnuson v. State, 187 Wis. 122, 203 N.W. 749 (1925)). The government cites numerous cases where a jury has been allowed to compare similarities in misspellings in two sets of documents as one part of the government's overall case against a criminal defendant.  See, e.g., United States v. Larson, 596 F.2d 759, 765 n. 10 (8th Cir. 1979) (jury considered misspelling of "approach" as "approuch" in kidnapper's ransom note and in defendant's letter to parole board); United States v. Pheaster, 544 F.2d 353, 371-72 (9th Cir. 1976) (jury considered similar misspellings in kidnapper's notes and in defendant's exemplars), cert. denied, 429 U.S. 1099, 97 S. Ct. 1118, 51 L. Ed. 2d 546 (1977); Hughes v. United States, 320 F.2d 459, 461 (10th Cir. 1963) (jury considered similar misspellings in materials illegally sent through the mails and in defendant's known writing), cert. denied, 375 U.S. 966, 11 L. Ed. 2d 415, 84 S. Ct. 483 (1964); see also Fed. R. Evid. 901(b)(4) (authentication of a document through examination of internal patterns and distinctive characteristics is proper).

   Effective rebuttal by the defense attorney will lessen any danger that the jurors might assign improper weight to their comparisons.  Any evidentiary arguments, however, are properly addressed to the weight and not to the admissibility of the cursive correspondence.  See J. Wigmore, Wigmore on Evidence § 29 (1940); cf. United States v. Flenory, 619 F.2d 301, 304-05 (3d Cir. 1980) (involving admissibility of in-court identifications).


We therefore conclude that the district judge abused his discretion in excluding the cursive correspondence offered to show stylistic similarities between that correspondence and the threatening letters.  We hold that the documents listed in the government's Motion in Limine are admissible subject to proper authentication by the government.  See United States v. Goichman, 547 F.2d 778, 783-84 (3d Cir. 1976) (per curiam); Fed. R. Evid. 901.   We will reverse the district judge's pretrial order and will remand for proceedings consistent with this opinion.