In November 1994, defendant-appellant Hamilton pleaded guilty, pursuant to a plea agreement. She stipulated to having purchased 40,000 tablets of ephedrine and two gallons of acetone to be used in the manufacture of methcathinone, in violation of 21 U.S.C. §§ 846 and 841(a)(1). At sentencing, the district court found that the 40,000 25-mg. tablets of ephedrine (1,000 grams) could be converted into 500 grams of methcathinone. Hamilton appeals, arguing that the district court's use of the 50% conversion ratio was based solely on a general understanding of the methcathinone production process, and that the district court failed to inquire into the particulars of the conversion ability of her and her conspirators. Because we have already held that such conversion calculations must have some individualized component, we remand for further sentencing proceedings consistent with this decision.

* The Sentencing Guidelines provide a framework for sentencing methcathinone offenders to which the district court added a 50% formula for converting the precursor chemicals into methcathinone. The Guidelines equate one gram of methacathinone with 380 grams of marijuana for purposes of calculating an offender's base offense level. Hamilton was sentenced on the basis of 1,000 grams of ephedrine, which the district court concluded could be converted into 500 grams of methcathinone. 500 grams of methcathinone, in turn, equates with 190 kg of marihuana and a base offense level of 26 (which applies for 100-400kg of marihuana).

The district court's 50% conversion ratio was first enunciated by Judge McKeague in United States v. Baker, 852 F. Supp. 609 (W.D.Mich.1994). Expert testimony in Baker suggested that a 50% conversion ratio was reasonable for the type of makeshift labs operating in Michigan's Upper Peninsula. In this case, the district court appears to have accepted the Baker reasoning without any testimony specific to the lab with which Hamilton was associated. It is important to note, however, that Hamilton would have received the same sentence even if the district court had picked a conversion ratio as low as approximately 27%, since that would make Hamilton responsible for approximately 270 g of Methcathinone (equal to 100 kg of marihuana) and base level 26 applies to a range of 100-400 kg of marihuana. Despite the wide margin of error this fact suggests, we must still examine the process whereby the district court arrived at Hamilton's sentence to ensure that the court's process was in accord with the law of the circuit.

We review a district court's factual findings underlying the application of the Sentencing Guidelines for clear error. United States v. Garner, 940 F.2d 172, 174 (6th Cir.1991). At sentencing, the prosecution bears the burden of proving by a preponderance of evidence the quantity of drugs that should be attributed to a defendant for sentencing purposes. United States v. Sims, 975 F.2d 1225, 1242 (6th Cir.1992). When choosing between equally plausible estimates, a court should err on the side of caution. United States v. Meacham, 27 F.3d 214, 216 (6th Cir.1994).

Hamilton relies on United States v. Mahaffey, 53 F.3d 128 (6th Cir.1995), to make her case that the government was required to offer evidence bearing specifically on the yield she and her conspirators could expect from the ephedrine she bought. We did establish in Mahaffey that a district court could not simply rely on a standardized formula from prior caselaw (the standard in Baker, for example), but must instead tailor its findings to the particular situation of individual defendants' abilities to convert raw materials into drugs. Further, in Mahaffey, we noted specifically that "the government produced no witnesses and filed no affidavits of experts who were familiar with the size and capabilities of the laboratories involved." Id. at 132.

The government attempts to distinguish this case from Mahaffey by pointing to specific evidence before the court that supports its 50% ratio. According to the government, "the sentencing judge in the instant case had before him transcripts of the expert testimony on which to base his own independent findings .... [and] the agreement and stipulation by Mr. Clements (another drug manufacturer on whom Hamilton informed) that he converted ephedrine to methacathinone with a 50% yield." Appellee's Brief at 10. The government provides no citation for these facts, though there are exhibits attached to its sentencing memorandum consistent with these claims. JA at 46-59.

The bottom line, however, is that none of this evidence helps the government meet the Mahaffey standard for particularized findings. The transcripts of expert testimony reflect the findings of various chemists as to expected yields, or yields from particular experiments, and clearly do not meet the concern that is articulated in Mahaffey for findings that are particularized to individual laboratories. Indeed, they appear to be the exact same testimony relied upon in Mahaffey. Mahaffey cites a conversion figure of 55.2%, 53 F.3d at 132, identical to the figure in one of the government's exhibits in this case. JA at 58. The evidence regarding the capacity of the lab used by "Mr. Clements" is equally unavailing to the government in meeting the Mahaffey burden--the government admits that "there was no evidence that Defendant Hamilton had provided Mr. Clements with ephedrine." Appellee's Brief at 6. Nor is there any argument by the government as to why Clements's capacity for conversion was similar to that of the lab that Hamilton did supply.

The government argues at length about the legal and policy rationales that favor a deterministic and general conversion factor. Citing consistency benefits and the decisions of other circuits that allow courts to use general conversion formulas, the government apparently hopes to convince this panel to overrule the Mahaffey decision. Since we cannot do so, Salmi v. Secretary of Health and Human Services, 774 F.2d 685, 689 (6th Cir.1985), we must reverse.

Mahaffey articulated the rule that a finding of drug quantities in a methcathinone case must be based on at least some facts related to the individual capacity of a defendant's laboratory. This does not mean that general evidence of scientific processes or of usual capabilities may not be useful. But Mahaffey requires that the sentencing court consider specific information as to the capability of the relevant lab, which was not done here. We therefore VACATE Hamilton's sentence and REMAND this case for further findings regarding the amount of drugs attributable to Hamilton, consistent with this opinion.