Shannon Maurice Boyd appeals his sentence for cocaine distribution. Because we conclude that the district court, in calculating Mr. Boyd's sentence, improperly relied on facts outside the record, we vacate Mr. Boyd's sentence and remand for re-sentencing.1

The United States charged Mr. Boyd with three counts of distributing crack cocaine in violation of 21 U.S.C. § 841(a)(1). Counts One, Two, and Three of the indictment charged Mr. Boyd with distributing 9.48 grams, 17.37 grams, and 26.30 grams, respectively, of "a mixture or substance containing a detectable quantity of cocaine base, commonly known as crack cocaine..." Rec. vol. I, doc. 1, at 1-2 (Sealed Indictment). Mr. Boyd pleaded guilty to Count One of the indictment (distribution of 9.48 grams); in exchange, the government successfully petitioned the district court for dismissal of the remaining two counts.

The district court sentenced Mr. Boyd in accordance with United States Sentencing Guidelines ("USSG") § 2D1.1. Under § 2D1.1(c), Mr. Boyd's sentence depended upon the weight of the narcotics distributed by Mr. Boyd. The district court, apparently in conformance with USSG § 1B1.3(a)(2), aggregated the total weight of the narcotics identified in Mr. Boyd's indictment. See United States v. Richards, 27 F.3d 465, 468 (10th Cir.1994) ("Under the guidelines, drug quantities are aggregated for drug offenses which are `part of the same course of conduct or common scheme or plan as the offense of conviction.'") (quoting USSG § 1B1.3(a)(2); some quotation marks omitted). Thus, the district court sentenced Mr. Boyd as if Mr. Boyd had distributed 53.15 grams of "a mixture or substance containing a detectable quantity of cocaine base." Rec. vol. I, doc. 1, at 1-2 (Sealed Indictment). Accordingly, given Mr. Boyd's criminal history level and other uncontested adjustments, the district court faced a sentencing guideline range of 108-135 months of imprisonment and four to five years of supervised release. The district court sentenced Mr. Boyd to 108 months of imprisonment and four years of supervised release.2

In appealing only his sentence, Mr. Boyd now raises a single issue. According to Mr. Boyd, the district court erred in calculating the cocaine quantity attributable to Mr. Boyd. Mr. Boyd notes that § 2D1.1(c) requires that the government prove drug quantity by a preponderance of the evidence. See Richards, 27 F.3d at 468-69 ("The government must prove the existence of the additional quantities by a preponderance of the evidence.... Evidence which does not preponderate or is in equipoise simply fails to meet the required burden of proof."). At his sentencing hearing, Mr. Boyd presented expert testimony, in the form of a report, suggesting that the weight of the cocaine distributed by Mr. Boyd totaled only 49.81 grams, a weight that would have reduced the applicable guideline range to that of eighty-seven to 108 months. The government countered with a lab report prepared by the Kansas Bureau of Investigation (the "KBI"); in the KBI report the narcotic quantity totaled 53.15 grams, as charged (in aggregate) in Mr. Boyd's indictment. Given the admission of these expert findings, Mr. Boyd characterizes the record evidence as resting "in equipoise" as to the relevant drug quantity. Aplt's Br. at 6. And, given such a state of equipoise, the United States, according to Mr. Boyd, cannot have satisfied the preponderance of the evidence standard.

Because the district court never reached the issue, we decline to speculate whether the record evidence actually rested in equipoise. We do, however, conclude that the district court committed a legal error (incorrect application of Federal Rule of Evidence 201, governing judicial notice) that leaves us with a clearly erroneous finding of fact (no record evidence supporting the court's finding that the quantity of a given cocaine sample will (1) inevitably decrease, (2) by as much as 3.34 grams over eight months for a roughly fifty gram sample); thus we vacate Mr. Boyd's sentence and remand for re-sentencing.

The district court concluded that 53.15 grams, the value claimed in the report offered by the government's expert, represented the narcotic weight relevant to Mr. Boyd's sentencing. The district court did not, however, base this conclusion on a balancing of the credibility of the apparently conflicting reports. Instead, the district court carefully explained: "The court need not choose ... between the credibility of the government's evidence and the defendant's evidence.... [T]he court finds no fault in the evidence offered by the defendant." Aplt's Br.App. 3, at 5 (Findings on the Defendant's Objection to the Amount of Cocaine and the Resulting Base Offense Level, undated).

If the district court had reached the credibility issue, certainly the court might have found one expert's report more credible than the other. The district court might have rested such a determination on, perhaps, the identification of certain internal inconsistencies in one of the reports or, perhaps, on the greater experience of one of the experts. On the other hand, however, the district court, having examined the reports, may well have concluded that the court had no means of crediting one report over the other: that the evidence rested in actual `equipoise.' In any event, such possibilities represent pure speculation and we will leave to the district court the initial determination of such matters.

What we can determine is that the district court, in (understandably) endeavoring to avoid the seeming conflict in drug quantity evidence, looked outside the record to discover a factual basis for an implicit finding that narcotic quantity decreases over time. The district court discovered in the case law four reasons why narcotic quantity may, in fact, decrease over time: (1) the possibility of "static cling" by which some quantity of the narcotic will be lost during an initial measurement of the given narcotic sample; (2) the possibility that government purity testing will consume some further quantity of the narcotic; (3) the possibility of the evaporation of some of the moisture initially within the narcotic; and (4) the possibility that atmospheric conditions, such as humidity, will affect the measurements of narcotic quantity. Aplt's Br.App. 3, at 3-4 (Findings on the Defendant's Objection to the Amount of Cocaine and the Resulting Base Offense Level, undated).3

Having uncovered these reasons why an earlier-in-time measurement of narcotic quantity may prove more accurate than a measurement performed later-in-time, the district court proceeded to conclude: (1) that, because Mr. Boyd's expert weighed the given narcotic sample some eight months after the KBI, the measurement of narcotic quantity submitted by Mr. Boyd's expert did not conflict with that submitted by the government's expert (thereby allowing the district court to avoid having to reach the difficult credibility determination just discussed) and (2) that the government's measurement should be credited on the basis of having been recorded earlier than that of Mr. Boyd's expert.

B. Whether the District Court's Determination Regarding the Decrease in Narcotic Quantity Has a Basis in the Record (Via Admitted Evidence and/or Judicial Notice)

A court is ordinarily limited by the record before that court; thus, a district court's finding of fact is clearly erroneous where not based on record evidence. See United States v. Contreras, 59 F.3d 1038, 1041 (10th Cir.1995) (determining that a district court's finding of fact "had a basis in the record" and thus "was not clearly erroneous"); United States v. Mahaffey, 53 F.3d 128, 133 (6th Cir.1995) ("Because the district court's finding ... totally lacks an evidentiary basis in the record, that finding is clearly erroneous."); see also United States v. Flowal, 163 F.3d 956, 961 (6th Cir.1998) (remanding for re-sentencing upon concluding that "the sentencing court's decision to choose an arbitrary amount that differs from the weights submitted by both DEA agents amounts to clear error").

Federal Rule of Evidence ("FRE") 201 provides an exception to this ordinary requirement that record evidence support any judicial finding of fact. FRE 201(b) permits a court to take "judicial notice" of a particular fact where that fact is "not subject to reasonable dispute in that [the fact] is either (1) generally known within the territorial jurisdiction of the trial court or (2) capable of accurate and ready determination by resort to sources whose accuracy cannot reasonably be questioned." In applying FRE 201, we are mindful that "[i]f a court takes judicial notice of a fact whose application is in dispute, the court removes the[] weapons [of rebuttal evidence, cross-examination, and argument] from the parties and raises doubt as to whether the parties received a fair hearing." General Elec. Capital Corp. v. Lease Resolution Corp., 128 F.3d 1074, 1083 (7th Cir.1997). "[T]he effect of taking judicial notice under Rule 201 is to preclude a party from introducing contrary evidence and, in effect, directing a verdict against him as to the fact noticed." Id.

1. Mr. Boyd's Case is Devoid of Record Evidence Supporting the District Court's Finding that the Quantity of a Given Narcotic Sample will Decrease over Time

Here, Mr. Boyd notes, and the United States cannot dispute, that the United States failed to introduce any evidence suggesting that an earlier-in-time measurement is any more likely to be accurate than one performed later-in-time. More specifically, the United States failed to introduce any evidence that static cling, consumptive testing, evaporation, and/or particular atmospheric conditions accounted for any, much less all, of the relevant narcotic quantity differential in this particular case. Mr. Boyd's case features no record evidence that the KBI utilized a measurement technique that produces the `static cling' phenomenon and no record evidence that any static cling would here fully account for the difference between the measurement results. Mr. Boyd's case features no record evidence that the KBI performed purity testing after initially weighing the narcotic sample, no record evidence that the KBI employed a method of purity testing that actually consumes a quantity of the original narcotic, and no record evidence that any narcotic quantity actually consumed for purity testing fully explains the difference between the measurement results. Mr. Boyd's case features no record evidence as to the type of container and/or atmosphere in which the KBI stored the narcotic sample and no record evidence as to whether, given the conditions of the KBI's storage of the given narcotic sample, the possibility of evaporation fully explains the difference between the measurement results. And, finally, Mr. Boyd's case contains no record evidence that any particular atmospheric conditions may be responsible for any decrease in narcotic weight, much less the full difference between the measurement results.

The government's failure to introduce such evidence stands in marked contrast to the cases relied upon by the district court, cases that each featured the respective court's reference to expert testimony, admitted in that particular case, that one or more of the proffered explanations were applicable on the unique facts of the given case.4 This failure on the part of the government forced the district court in our own case, left without the benefit of any evidence explaining the discrepant measurements of narcotic quantity, to, in effect, take judicial notice of evidence presented in cases other than Mr. Boyd's.

2. The District Court Erred in Taking Judicial Notice of the Alleged Fact that Narcotic Quantity Decreases over Time

Given, then, the district court's reliance on facts outside the record to establish that an approximately fifty gram quantity of cocaine will decrease in weight by 3.34 grams over approximately eight months, the court's determination in that regard is clearly erroneous, unless judicial notice is appropriate for that fact; i.e. unless judicial notice is appropriate for the fact that static cling, consumptive testing, evaporation, and/or particular atmospheric conditions accounted for the full differential between the proffered measurements of narcotic quantity. We conclude that, because the fact that Mr. Boyd's narcotic sample decreased by 3.34 grams over eight months is neither "generally known within the territorial jurisdiction of the trial court" nor "capable of accurate and ready determination by resort to sources whose accuracy cannot be reasonably questioned," that `fact' is not one amenable to judicial notice.

The science and technology of narcotic collection, measurement, and storage are far too uncertain (particularly in light of variances in local practice), and evolving, to permit the conclusion, without the production of any evidence, that the passage of eight months necessarily harmonizes two measurements of a given cocaine sample where one measurement yields a 53.15 gram value and the other yields a 49.81 gram value.5 In this very case, it may well be that the KBI utilized a measurement technique that does not induce the `static cling' phenomena, measured the narcotic purity before measuring the narcotic quantity, stored the narcotic sample in a container that would not permit evaporation, and measured the narcotic quantity under substantially the same atmospheric conditions as did Mr. Boyd's expert. In such a case, the `fact' of which the court purported to take judicial notice (that static cling, consumptive testing, evaporation, and/ or particular atmospheric conditions accounted for the relevant narcotic quantity differential in this particular case, thereby leaving the government's earlier-in-time measurement more reliable than Mr. Boyd's later-in-time measurement) would be no fact at all. Judicial notice is not appropriate where so many variables are in play. Cf. Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 592 n. 11, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993) (describing those scientific theories of which courts may take judicial notice as those theories "so firmly established as to have attained the status of scientific law, such as the laws of thermodynamics").

Indeed, permitting judicial notice of such a fact would effectively prevent a criminal defendant from ever successfully challenging the quantity of narcotics attributed to the defendant. The government will always weigh the quantity first, after which some period of time will elapse before the defendant is able to obtain his or her own expert weighing of the narcotic sample. If a court could take judicial notice of the fact that the narcotic quantity may have decreased over that period of time and, based upon that `finding,' conclude that the government's measurement is more reliable, government errors in narcotic measurement would go uncorrected. Cf. United States v. Richards, 27 F.3d 465, 469 (10th Cir.1994) ("[W]hen choosing between a number of plausible estimates of drug quantity, none of which is more likely than not the correct quantity, a court must err on the side of caution.").

The inappropriateness of judicial notice leaves the record devoid of support for the district court's conclusion that, given the possibility of static cling, consumptive testing, evaporation, and/or particular atmospheric conditions, (1) the proffered narcotic quantity measurements are not in conflict and that (2) the government's measurement is more reliable because conducted earlier in time. The absence of any record support for this finding leaves the finding clearly erroneous.

For the foregoing reasons, we VACATE Mr. Boyd's sentence and REMAND for re-sentencing in consideration of this opinion. At re-sentencing, the district court, of course, retains the discretion to both (1) admit additional evidence as to the relevant narcotic quantity and/ or the credibility of the given expert reports and (2) make additional findings regarding those matters.