LOKEN, Circuit Judge.

Coleen Sue Funk, David Vernon Harden, and Lonnie James Sharpe, Jr., appeal the sentences they received after pleading guilty to a charge of conspiracy to manufacture methamphetamine in violation of 21 U.S.C. § 846. Defendants contend that the district court erred in determining their base offense levels under U.S.S.G. § 2D1.4 (1989) based upon the most abundant precursor chemical seized at their methamphetamine laboratory. We affirm.

On July 31, 1989, Funk summoned police to the farm she was renting near Rockwell, Iowa, because Sharpe had shot Harden. While investigating the shooting, police detected a strong chemical odor from a shed and contacted the federal Drug Enforcement Agency. DEA agents obtained a warrant and searched the shed, where they discovered a methamphetamine laboratory, 3.6 grams of methamphetamine, and various quantities of precursor chemicals. After defendants were charged in a three count indictment, each pleaded guilty to Count 3, which alleged that they knowingly conspired to manufacture approximately one kilogram of methamphetamine and thirty grams of phenyl-2-propanone, a controlled substance used in manufacturing methamphetamine.

The sentencing hearing focused primarily on the quantity of methamphetamine attributable to defendants for sentencing purposes. Each side presented an expert who estimated the production capability of defendants' laboratory from the quantities of precursor chemicals seized. The government's expert was DEA agent Scott Masumoto, whose experience included investigating and rendering an opinion as to the capacity of some thirty other methamphetamine laboratories. Defendants' expert was Dr. Peter P. Wickham, a professor of chemistry at Coe College.

Agent Masumoto testified that all materials necessary for the production of methamphetamine were present at defendants' laboratory except aluminum foil and alcohol. The precursor chemicals seized included 1550 grams of phenylacetic acid, 800 milliliters of acetic anhydride, and 700 milliliters of methylamine. Using the methamphetamine manufacturing formulas seized at the farm, both experts then testified as to the quantities of methamphetamine that defendants could have produced from these precursor quantities:

Estimated Methamphetamine Production

On appeal, defendants argue that the district court erred in using the most abundant precursor chemical, phenylacetic acid, to determine their base offense levels. Defendants contend that this was both an error of law and of fact. We conclude it was neither.

1. Application note 2 to § 2D1.4 required that the sentencing judge approximate the quantity of controlled substance for sentencing purposes when the amount seized does not reflect the scale of the offense: "In making this determination, the judge may consider, for example . . . the size or capability of any laboratory involved." We have frequently upheld sentences in which the quantity of methamphetamine was estimated based upon the manufacturing capability of the laboratory using the quantity of precursor chemicals seized. See, e.g., United States v. Stockton, 968 F.2d 715, 719-20 (8th Cir. 1992); United States v. Fulcher, 943 F.2d 824 (8th Cir. 1992); United States v. Evans, 891 F.2d 686 (8th Cir. 1989), cert. denied, 495 U.S. 931, 110 S.Ct. 2170, 109 L.Ed.2d 499 (1990); United States v. Wagner, 884 F.2d 1090, 1097-98 (8th Cir. 1989), cert. denied, 494 U.S. 1088, 110 S.Ct. 1829, 108 L.Ed.2d 958 (1990). Here, the amount of methamphetamine seized (3.6 grams) was far less than the capability of the defendants' laboratory using the precursor chemicals on hand. Therefore, § 2D1.4 required the district court to estimate a larger quantity for sentencing purposes.

Defendants correctly note that the estimate upheld in Fulcher was based upon the least abundant precursor discovered with that methamphetamine laboratory. However, our opinion in Fulcher did not hold that the estimate must be based upon the least abundant precursor, and we decline to adopt such a rule. Estimating the appropriate quantity for sentencing purposes is a fact intensive inquiry that should not be constricted by an inflexible rule as to one relevant factor. See United States v. Olderbak, 961 F.2d 756, 764 (8th Cir.), cert. denied, ___ U.S. ___, 113 S.Ct. 422, 121 L.Ed.2d 344 (1992). Moreover, limiting the district court's estimating discretion to the least abundant precursor would be contrary to the mandate in § 2D1.4(a) that the offense level for conspiracies of this nature "shall be the same as if the object of the conspiracy . . . had been completed."

2. Alternatively, defendants argue that the district court's quantity estimate was clearly erroneous. Defendants contend that the most abundant precursor method yields an unrealistic quantity estimate in this case, first, because they were inexperienced methamphetamine "cooks" who were using primitive equipment in a filthy laboratory, and second, because Sharpe's shooting of Harden effectively ended the conspiracy.

Both Agent Masumoto and Professor Wickham first calculated theoretical methamphetamine yields from the precursors on hand and then reduced those yields substantially in estimating the amount of methamphetamine that defendants would have produced. Their estimates of the "practical yield" of methamphetamine from the phenylacetic acid on hand were substantially similar — 250 grams and 225.5 grams. The evidence showed that Funk had prior experience as a methamphetamine cook and that the laboratory was fully equipped except for readily available household products. Thus, the district court's finding that defendants would have produced 250 grams of methamphetamine from the 1550 grams of phenylacetic acid on hand at this laboratory is not clearly erroneous. Compare United States v. Kingston, 922 F.2d 1234, 1235 (6th Cir. 1990) (500 grams of methamphetamine estimated from 2000 grams of phenylacetic acid), cert. denied, ___ U.S. ___, Ill S.Ct. 2054, 114 L.Ed.2d 460 (1991).

Nor can we accept defendants' argument that the conspiracy ended with Sharpe's shooting of Harden. The district court made no finding that the conspiracy had ended, and defendants requested none. While it is likely that the shooting would have interrupted defendants' criminal enterprise, and might have ended their ability to work together in furtherance of the conspiracy, they had invested substantial resources in renting the farm, setting up their laboratory, and purchasing a large quantity of phenylacetic acid, other precursor chemicals, and laboratory equipment. In these circumstances, the district court's finding that the quantity of methamphetamine attributable to the conspiracy should be based upon full use of the phenylacetic acid on hand was not clearly erroneous. Compare Smallwood, 920 F.2d at 1236-39 (thirty pounds of methamphetamine estimated from 110 gallons of phenylacetic acid despite missing precursors, no immediate intent to manufacture, and a disassembled lab). Defendants' attempt to distinguish the decision in Smallwood is unpersuasive.

3. Finally, Sharpe makes a somewhat narrower factual argument based upon his unique situation. To obtain relief from their mandatory minimum sentences, Funk and Harden must reduce the estimated quantity of methamphetamine below 100 grams. See 21 U.S.C. § 841(b)(1)(A)(viii). However, Sharpe received a higher sentence, at the top of his Guidelines range, that would be affected if the quantity estimate were reduced below 250 grams. Sharpe argues that the district court's finding of 250 grams was clearly erroneous because Professor Wickham's estimate of 225.5 grams from the phenylacetic acid on hand should have been accepted over the "willy-nilly shotgun calculations" of Agent Masumoto.

This is a question of witness credibility and weight of the evidence as to which the district court's findings are entitled to particularly great deference. See United States v. Candie, 974 F.2d 61, 64 (8th Cir. 1992) ("witness credibility is an issue for the sentencing judge that is virtually unreviewable on appeal"). We have carefully reviewed the record of the sentencing hearing and conclude that the district court's finding accepting Agent Masumoto's quantity estimate was not clearly erroneous.

The judgment of the district court is affirmed.