dissenting.
Judge Wood for the majority has fairly presented the factual circumstances, as well as most of the relevant policies, bearing on the admissibility of prior convictions’ evidence in this case. Upon reflection, however, I believe that the balance between probative value and prejudice is so lop-sided in the direction of unfair prejudice that I would reverse the conviction and remand for a new trial.
The basic problem both in the argument of the district court here and of the majority is that the special circumstances and characteristics of a stolen shipment of interstate goods that give rise to the possessor’s knowledge that the goods were stolen are not universal but, of course, vary sharply from case to case. In one case, a possessor’s knowledge that the goods were stolen may derive from his active participation in hijacking the goods, while in another it may come from being told to deliver stolen goods to a fence. Proof in the first instance might include testimony that the hijacker was found driving the truck with the original driver bound and gagged in the back. Proof in the second case might in-elude testimony that the driver had no bill of lading and asked for payment in cash. In these examples, therefore, the underlying facts that would allow the trier of fact to infer knowledge in each case are different. One could not reasonably infer that the possessor in the second case knew the goods were stolen from the mere fact that he had knowledge under the facts of the first case, since the respective facts that give rise to the inference of knowledge in each case are totally different.
The problem with introducing only a certificate of a prior conviction under the same criminal statute is that it says nothing about the facts underlying the conviction. In some instances, of course, the facts underlying the conviction may be unnecessary because the mere fact of a prior conviction alone may be highly probative of an element in a later case. For example, if a person were convicted of possession of cocaine and was similarly charged on a later occasion, a defense that the person did not know what the substance was would be undercut by the fact of the prior conviction. Here, however, the contested issue is whether the defendant knew that these particular goods were stolen. Unless the prior convictions involved circumstances similar to those from which defendant’s knowledge could be inferred here, they have little if any probative value on that issue and must be considered of only the most marginal relevance.
Numerous courts have, of course, admitted certifications of prior convictions to establish knowledge, intent, and so on in a subsequent case. Some of these courts, while expressing reluctance to admit such evidence, have felt bound to do so by precedent. See, e.g., United States v. Mehrman-esh, 689 F.2d 822, 831 (9th Cir.1982) (“[Njowhere has [the government] specifically established which facts such evidence is relevant to show, nor has it articulated *668the hypotheses by which the jury could infer such facts from the proffered evidence[;] ... [nevertheless, prior cases in this Circuit compel us to conclude that there is at least some logical connection, however weak, to a central element in this case .... ”). Unfortunately, I believe that this tendency in the cases shows “a reluctance to attempt a logical analysis of other crime evidence, and an overreliance upon the result reached in legal precedent as a complete solution to problems of admissibility.” Trautman, Logical or Legal Relevancy — A Conflict in Theory, 5 VAND.L.REV. 385, 409 (1952). Although I can conceive of instances in which the mere fact of a prior conviction may tend in a significant way to prove an element of the crime and not merely a propensity to commit criminal acts, see, e.g., United States v. Long, 706 F.2d 1044, 1052 (9th Cir.1983), this is not such a case.
I believe the defendant’s four prior convictions fall into that category of evidence that, because it has only slight apparent similarity to or connection with the violation involved here, must be excluded. This case is not unlike United States v. Burkhart, 458 F.2d 201 (10th Cir.1972) (en banc), in which certified copies of the defendant’s two prior convictions for interstate transportation of stolen motor vehicles with knowledge that they were stolen (one conviction occurring 15 years earlier and the other 4 years earlier) were admitted to show that the defendant had knowledge that the vehicle he was charged with transporting was stolen. In reviewing the admissibility of these convictions, the court said:
[The certified copies of the judgments of conviction] were not shown to have been connected with the case on trial either as a part of the transaction on trial or to have been related transactions in terms of sharing a common plan, scheme, design or intent, or in terms of factual or other similarity except that they were all Dyer Act violations. The trial court’s cautionary instruction at the time was that this evidence was to be considered by the jury only for the purpose of determining whether defendant acted willfully and with specific intent and not because of mistake or accident or other innocent reason. We have no criticism of either the accuracy or sufficiency of this instruction. It is the extreme remoteness in time and space, together with the lack of any apparent similarity or connection with the principal charge, which creates the problem.
The judgment in this case must be reversed primarily because of the lack of manifest relevancy of the evidence of other or similar offenses in relationship to the cause on trial. The evidence is neither an inseparable part of the case on trial nor does it serve to supply anything more than irrelevant and invalid support of any element of the charge. The plan, scheme, design or intent was not shown to be a common one in relationship to the offense charged and this is an essential requirement.
Factual similarity is to be considered and weighed in determining the strength attributable to the relevancy factor.
Id. at 203-04 (footnote omitted), 208.
Even though the prior convictions were not lacking completely in probative value and thus could not be considered entirely irrelevant, the probative value of these convictions is so slight in comparison with their obviously weighty prejudicial effect that the district court’s failure to exclude them under Rule 403 was an abuse of discretion. The unfair prejudice flowing from prior convictions — particularly convictions carrying over a span of 20 years — is that they “cast[] a shadow over the remaining evidence so that a jury hearing this is necessarily distracted from considering the particular merits.” Id. at 208.1 Because of this obvious and palpable danger, it is important that the probative value of prior crime evidence be substantial before it is admitted. Here, the district court did an *669excellent job in attempting to minimize the prejudicial effect of the prior convictions. The district court told the jury immediately after the certifications of the prior convictions had been entered into the record and again in the jury instructions that the prior convictions related solely to the issue whether the defendant knew that the goods were stolen. Transcript at 143-44; Record Doc. 50. My concern, however, is not with the steps the district court took to minimize the prejudicial effect of the prior convictions but with the way the district court evaluated the balance between probative value and the prejudice inherent in jury consideration of this “life of crime.” By giving undue probative value to the remote and unconnected convictions and in failing to perceive the extreme prejudice created by them, the district court abused its discretion in conducting the balancing required by Rule 403. Although I am extremely reluctant to disturb a district court’s exercise of discretion, I would intervene to do so here.
I therefore respectfully dissent.
. The government apparently understood that the jury might regard the prior convictions as evidence of Falco’s bad character, rather than as evidence of knowledge. It argued to the *669trial court that failure to admit the convictions would result in “the truth [being] distorted[J for the picture presented will be that of a grey-haired, grandfatherly gentleman who has been himself victimized----” Government’s Memorandum In Support of Motion In Limine at 8. I cannot help but believe that the potential for the jury to convict the defendant because of his propensity to engage in criminal activity is too great to allow these convictions to be introduced.