dissenting:
I concur in parts III and IV of the majority opinion.1 The majority errs, however, in holding that it was harmless error to admit as a business record a laboratory report that indicated, simply by the presence of a check mark, that Bentley’s urine tested positive for marihuana use.2 The majority concludes that this evidence was inconsequential, suggesting that Bentley is unquestionably guilty of importing and possessing marihuana because he was a passenger in a plane transporting odiferous drugs. I cannot join in this misapplication of the harmless error analysis. Moreover, I conclude that the constitutional right to confront adverse witnesses is violated by admitting the results of a urine drug test as a business record when the laboratory report contained no indication of how the test was conducted and the defendant had no opportunity to elicit this information on cross-examination.
I. Harmless Error
The majority holds that any error in admitting the positive results of Bentley's urine drug test was harmless. They correctly cite the test for harmless error: error is harmless where, after reviewing the
facts of the case and surrounding circumstances, the evidence unrelated to the alleged constitutional violation “remains not only sufficient to support the verdict but so overwhelming as to establish the guilt of the accused beyond a reasonable doubt.” Germany v. Estelle, 639 F.2d 1301, 1303 (5th Cir.), cert. denied, 454 U.S. 850, 102 S.Ct. 290, 70 L.Ed.2d 140 (1981). In applying this test, however, the majority simply concludes that “it is obvious” that Bentley was a passenger in the plane, that marihuana was discharged from the plane during flight (this evidence is not obvious although it is a distinct possibility), and that the odor of marihuana was present when the plane crashed. Supra, p. 1118. Some of this may be obvious, but it is not evidence of Bentley’s guilt that is so overwhelming that we can be certain beyond a reasonable doubt that the laboratory test indicating Bentley's marihuana use did not contribute to the guilty verdict. See Chapman v. California, 386 U.S. 18, 23-24, 87 S.Ct. 824, 827-28, 17 L.Ed.2d 705 (1967).
Harmless error “is an exacting standard that must be uncompromisingly applied.” Harryman v. Estelle, 616 F.2d 870, 876 (5th Cir.), cert. denied, 449 U.S. 860, 101 S.Ct. 161, 66 L.Ed.2d 76 (1980). Error is harmless only when the evidence is unusually strong and convincing, so that no reasonable doubt remains as to the disputed issue. See Spears v. Circuit Court, Ninth Judicial Dist., 517 F.2d 360, 366-67 (5th Cir.1975) (In prosecution of midwife for performing illegal abortion, it was harmless error for the state to have failed to call as witnesses the nurse and receptionist at the laboratory that performed the pregnancy test on the woman who received the abortion. The error was harmless because the testimony of two physicians, the lab technician who performed the pregnancy test, and the woman who received the abortion conclusively established her preg*1121nancy).3 We simply do not encounter this type of overwhelming evidence in the present case.
The majority concludes that the fact “that [Bentley] was in the airplane, especially when unexplained by any evidence of the case, is thoroughly damning.” Supra, p. 1118. The clear implication of this statement is that an individual’s presence when a drug crime is being committed is conclusive evidence of his guilt. This Court, however, has consistently held to the contrary. Mere presence at the site of criminal activity is not sufficient evidence to convict an individual of participating in a drug conspiracy or possessing a controlled substance. United States v. Espinoza-Seanez, 862 F.2d 526, 537 (5th Cir.1988). See also United States v. Sneed, 705 F.2d 745, 749-50 (5th Cir.1983) (Son was present when twenty men congregated on his father’s property to unload approximately 35,000 pounds of marihuana from a boat during the night. This evidence was found insufficient to sustain the son’s conviction for possessing marihuana with intent to distribute). The fact that the smell of marihuana pervaded the plane and the crash site is also not overwhelming evidence of Bentley’s guilt. In words particularly applicable to this case, we have noted that “it is not enough for [the evidence] merely to establish a climate of activity that reeks of something foul.” United States v. Galvan, 693 F.2d 417, 419 (5th Cir.1982).
In short, the fact that Bentley was a passenger (and not the pilot) of an airplane that contained marihuana, while certainly evidence of his guilt, is not so overwhelming as to render any other improperly admitted evidence harmless beyond a reasonable doubt. Moreover, in holding that any error in admitting the laboratory report was harmless, the majority fails to consider the impact this evidence had on Bentley’s defense. Throughout the trial, Bentley’s counsel argued that the government had failed to prove that Bentley had knowledge of the criminal activity and possession of the marihuana. In support of this defense, the trial testimony indicated that Bentley ran a business selling Mexican folk art, and thus had a legitimate reason to travel to and from Mexico. The evidence also established that it was the pilot who purchased, fueled, and flew the plane; Bentley was not licensed to fly.
The laboratory report indicating that Bentley’s urine tested positive for marihuana after the crash severely undermines Bentley’s contention that he did not possess the marihuana. It is for this reason, of course, that the government specifically mentioned the positive urine test in its closing argument.4 In view of Bentley’s defense and the circumstantial evidence upon which the government relied, the admission of the laboratory report, which provided the only physical evidence linking Bentley to marihuana, was not harmless error.
II. Admission of the Laboratory Report
Turning to the merits of Bentley’s claim, I conclude that the admission of the labo*1122ratory report, under the particular circumstances of this case, violated Bentley’s right to confront adverse witnesses guaranteed by the Sixth Amendment.5 The laboratory report was admitted pursuant to the business record exception to the hearsay rule. Fed.R.Evid. 803(6).6 The fact that evidence was properly admitted under the Federal Rules of Evidence does not, however, automatically satisfy the dictates of the Constitution. Indeed, the Supreme Court has rejected the view that the requirements of the Federal Rules of Evidence and the Confrontation Clause are completely congruous. California v. Green, 399 U.S. 149, 155-56, 90 S.Ct. 1930, 1933-34, 26 L.Ed.2d 489 (1970). See also United States v. Washington, 688 F.2d 953, 959 (5th Cir.1982).
The Confrontation Clause protects “the accuracy of the truth-determining process in criminal trials by assuring that the trier of fact has a satisfactory basis for evaluating the truth of [a] prior [out-of-court] statement.” Dutton v. Evans, 400 U.S. 74, 89, 91 S.Ct. 210, 220, 27 L.Ed.2d 213 (1970). In determining whether hearsay can be admitted consistent with this constitutional guarantee, the focus is on the “indicia of reliability” accompanying a statement. Mancusi v. Stubbs, 408 U.S. 204, 213, 92 S.Ct. 2308, 2313, 33 L.Ed.2d 293 (1972). Evidence admitted pursuant to a “firmly rooted” exception to the hearsay rule is usually, but not always, found to be sufficiently reliable to comport with constitutional requirements. See Ohio v. Roberts, 448 U.S. 56, 66, 100 S.Ct. 2531, 2539, 65 L.Ed.2d 597 (1980).7
*1123The government concedes that Bentley’s contention that the admission of the laboratory report violated his right of confrontation is “troubling.” The government argues, however, that this information is sufficiently reliable so that no constitutional problem arises. But the laboratory report in this case indicated by the presence of a simple check mark on an extensive form that Bentley’s urine tested positive for can-nibinoids. The report summarized the results of an unidentified test conducted by an anonymous technician. At trial, the government called only the medical records custodian to testify that the laboratory report was kept as a part of Bentley’s hospital records. Thus, Bentley had no opportunity to learn the type of test conducted, the chain of custody of the substance tested, and the qualifications of the laboratory personnel.
In short, none of the indicia of reliability that are usually present when the results of a laboratory drug test are admitted into evidence accompanied the report in this case.8 Other courts have concluded that medical reports admitted “without a detailed explication of either the facts or reasoning processes on which they were based” violate a defendant’s right to confront adverse witnesses. Phillips v. Neil, 452 F.2d 337, 347 (6th Cir.1971), cert. denied, 409 U.S. 884, 93 S.Ct. 96, 34 L.Ed.2d 141 (1972); Pickett v. Bowen, 626 F.Supp. 81, 85 (M.D.Ala.1985), aff'd per curiam, 798 F.2d 1385 (1986). This is a sound rule to apply in the drug testing context, where some of the common laboratory procedures produce a high rate of false positive results and are considered adequate only for screening purposes.9
This Court has expressly recognized that the reliability of urine drug tests varies greatly according to the type of laboratory procedure employed.10 When a laboratory report omits this crucial information, and no witness is available at trial to explain the testing procedure, there is no way to determine if the test results are reliable. The laboratory report in this case, which contained a check mark and nothing more, is simply hearsay without any indicia of reliability. As such, its admission into evidence violated Bentley’s constitutional right to confront adverse witnesses.
III. Conclusion
As the majority concedes, Bentley’s conviction rests on circumstantial evidence. The results of a laboratory test indicating marihuana use was the only physical evidence linking Bentley to marihuana. The admission of the laboratory report into evidence was error and was not harmless.
Bentley had no opportunity to question the accuracy of the laboratory report indicating that his urine tested positive for marihuana use. A check mark alone was deemed sufficiently reliable to overcome *1124his right to confront adverse witnesses. Bentley was implicated by the results of an unidentified test conducted by an anonymous technician. This surely violates the core value protected by the Confrontation Clause: the opportunity to face one’s accuser face-to-face, with the concomitant ability to cross-examine the witness.
In an era where urine drug testing is increasingly common, reports of positive tests will undoubtedly appear, not infrequently, in criminal trials. Under the Confrontation Clause, a defendant must be afforded the opportunity to ascertain how this devastating information was obtained. The Constitution does not allow a simple check mark on a piece of paper to be used to establish conclusively the drug use of a criminal defendant. I must dissent. The case should be reversed for a new trial.
. In part IV, I agree with the majority that Bentley failed to preserve his objection to the contested jury charge, and that the alleged error was not plain error. I do not join in the majority's determination that the error was also harmless.
. Congress has identified the controlled substance in this case as “marihuana.” 21 U.S.C. § 812(c), Schedule I(c)(10) (1982). I follow the spelling of the name of the drug that is used in the official text of the federal statute.
. The majority cites the Spears case for the proposition that “alleged constitutional error involving a urine test has been found harmless.” Supra, p. 1118. In that case, however, the evidence supporting the results of the test was indeed overwhelming — two physicians, the lab technician, and the formerly pregnant woman all confirmed that she was pregnant when the abortion was performed. In the present case, there is absolutely no confirmation of the results of Bentleys urine test, and the rest of the evidence supporting the conviction is circumstantial.
The majority’s citation to Spears is not only inapposite, it is also ironic given the prior history of that case. The original conviction of the midwife was reversed after the Mississippi Supreme Court concluded that the failure of the lab technician to testify in the first trial violated the defendant’s right to confront adverse witnesses guaranteed by the Mississippi constitution. Spears v. State, 241 So.2d 148 (Miss.1970).
. The prosecutor stated:
The records of Lubbock General Hospital ... the third page here shows that while [Bentley] was in the hospital on the 17th of September, 1987, at 4:45 AM, it’s marked urine and bladder, urine drug screen, and you do down here and you get to cannibinoids. You remember what the cannabinoids are? You notice that there’s a check mark right there, positive. Again indicating the knowledge of this man with respect to marihuana.
. “In all criminal prosecutions, the accused shall enjoy the right ... to be confronted with the witnesses against him.” U.S. Const.Amend. VI.
. In addition to his constitutional claim, Bentley argues that the report was improperly admitted as a business record under Fed.R.Evid. 803(6) because it was inadmissible under Fed.R.Evid. 803(8), which provides that certain public records cannot be admitted against the accused in a criminal case. See United States v. Cain, 615 F.2d 380, 382 (5th Cir.1980). The record in this case, however, does not support Bentley’s contention that the laboratory report would be inadmissible as a public record. We have no information regarding the public affiliation of the hospital or lab where the test was conducted. Nothing in the record indicates that the laboratory test was conducted under the auspices of the police or other law enforcement officials. In short, Bentley has failed to point to any evidence which suggests that the exception he relies upon applies in this case.
. The Roberts court set out a general test for determining when the admission of hearsay violates a defendant’s right to confront adverse witnesses:
In sum, when a hearsay declarant is not present for cross-examination at trial, the Confrontation Clause normally requires a showing that he is unavailable. Even then, his statement is admissible only if it bears adequate ‘indicia of reliability’. Reliability can be inferred without more in a case where the evidence falls within a firmly rooted hearsay exception. In other cases, the evidence must be excluded, at least absent a showing of particularized guarantee of trustworthiness.
448 U.S. at 66, 100 S.Ct. at 2539. Later cases make clear, however, that this test is only a guideline that “should not be read as an abstract answer to questions not presented in that case.” United States v. Inadi, 475 U.S. 387, 392, 106 S.Ct. 1121, 1125, 89 L.Ed.2d 390 (1986). Recently, the Supreme Court has referred to the Roberts analysis in more qualified terms, stating that "the Court has, as a general matter only, required the prosecution to demonstrate both the unavailability of the declarant and the 'indi-cia of reliability’ surrounding the out-of-court declaration.” Bourjaily v. United States, 483 U.S. 171, 182, 107 S.Ct. 2775, 2782, 97 L.Ed.2d 144 (1987).
The language in Roberts suggesting that reliability can be inferred "without more” when evidence falls under a well-settled hearsay exception has not been construed as an absolute rule. Indeed, elsewhere in the opinion, the Supreme Court stated that “certain hearsay exceptions rest upon such solid foundations that admission of virtually any evidence within them comports with the substance of the constitutional protection." 448 U.S. at 66, 100 S.Ct. at 2539 (emphasis added). See also Puleio v. Vose, 830 F.2d 1197, 1207 (1st Cir.1987), cert. denied, — U.S. -, 108 S.Ct. 1297, 99 L.Ed.2d 506 (1988) ("[T]he mere fact that a ... court, in admitting evidence, tucks it into a pigeonhole which bears the label of a time-honored hearsay exception cannot be entirely dispositive”). Courts have found that evidence admitted pursuant to the business record exception to the hearsay rule can, in certain situations, violate a defendant’s right of confrontation. See e.g., United States v. McClintock, 748 F.2d 1278, 1291-92 (9th Cir.1984), cert. denied, 474 U.S. 822, 106 S.Ct. 75, 88 L.Ed.2d 61 (1985). See also United States v. Washington, 668 F.2d at 959 (This Court found that evidence was properly admitted as a business record, but remanded the case for the district court to determine whether the Confrontation Clause was violated).
. Significantly, the lab technician who tested the substance found at the crash site was called to testify at trial in order to establish that the substance was marihuana. A second technician who conducted tests on a substance that was presumably dropped from the plane also testified. Both lab technicians gave a complete account of the chain of custody of the substance, the type of tests conducted, and the bases for their conclusion.
. Many of the tests for identifying marihuana are not specific. This means that other drugs and chemicals present in the urine can lead to a false positive result. Giannelli and Imwinkel-ried, Scientific Evidence, § 23-2, pp. 933-34 (1986). Two of the common tests for detecting marihuana use, the EMIT and RIA tests, produce inconclusive results that require confirmation by other testing methods. Note, Admissibility of Biochemical Urinalysis Testing Results for the Purpose of Detecting Marihuana Use, 20 Wake Forest L.Rev. 391, 393-94, 409 (1984). The fact that the laboratory report of Bentley’s drug test was labeled “Urine Drug Screen” might suggest that one of these inconclusive screening procedures was used. Unfortunately we, like Bentley, have no way of knowing if this was the case.
.In Nat'l Treasury Employees Union v. Von-Raab, 816 F.2d 170 (5th Cir.1987), aff'd in part, - U.S. -, 109 S.Ct. 1384, 103 L.Ed.2d 685 (1989), we considered the reliability of the Customs Service drug-testing program. We noted that “the initial screening test, EMIT, may have too high a rate of false-positive results for the presence of drugs.” 816 F.2d at 181. We concluded, however, that the Customs Service drug testing program was not so unreliable as to violate due process because “the follow-up test, GC/MS, is almost always accurate, assuming proper storage, handling, and measurement techniques." Id.