In this appeal we consider the admissibility of evidence of an out-of-court experiment conducted in order to rebut a criminally accused's version of events. Johnny Jackson was tried and convicted of being a felon in possession of a firearm. One of the key components of his defense was that at the time he was alleged to have committed a shooting and to have possessed a gun, he was actually picking up his girlfriend from work. To rebut the alibi, the government sent a deputy United States Marshal to see how long it would take to drive from the scene of the shooting to the girlfriend's place of employment. The Marshal testified that his drive time was short, which provided the basis for the prosecutor to argue in closing that Jackson could have committed the shooting and still had time to pick up his girlfriend. On appeal, Jackson challenges the admission of this evidence, but since the experiment was offered in rebuttal and was conducted under substantially similar circumstances as Jackson's drive, we find that the evidence was properly admitted. Jackson also contends that the district court gave an improper jury instruction and that the firearm he possessed did not affect interstate commerce, as the statute requires, but we reject these arguments as well and affirm on all grounds.

The indictment charges Jackson with possessing a firearm "[o]n or about July 12, 2004, through July 15, 2004," but to understand what took place at that time we must first travel back a month earlier, to June 19, 2004. On that date Jackson and his nephew, Jarvis Jackson, confronted Tarus Watkins and Watkins's brother, Donny Richardson, in Peoria, Illinois, demanding money owed from a dice game. A shootout between Jarvis and Donny occurred, leaving Jarvis dead, and both Donny and the defendant wounded.1 The government contends that this incident, in which the defendant's nephew was shot to death by Tarus Watkins's brother, created a motive for the defendant to shoot at Tarus Watkins a month later.

That shooting occurred on July 12, 2004. Watkins was the only person at trial to testify about the incident. He stated that "shortly after 2:00 p.m." he left a store on his way home, and that after several blocks of walking the defendant peeked out from behind a church and then ran into the street firing a gun at him. Watkins was not hit, but he heard three gunshots and saw a black, short-nosed revolver in the defendant's hand. He fled and called the police; a dispatch call from police headquarters was placed at 2:34 p.m. sending officers to the scene. Jackson was arrested in his car three days later on July 15; authorities recovered a similar gun on the ground outside the car.2

The defense tried to cast doubt on Watkins's account of the July 12 shooting by introducing testimony from Jackson's girlfriend, Unity Nelms, who stated that Jackson usually picked her up from work. That day, she called him at his house at 1:40 or 1:45 p.m. and asked him to pick her up at 2:30 p.m. Her timecard shows that she punched out at 2:43 p.m., and she testified remembering that Jackson was there waiting for her at that time. She also testified that in her own experience, the drive from Jackson's house to her work took between 20 and 30 minutes. The defense's theory was that Jackson could not have shot at Watkins shortly after 2:00 p.m., because he would have been on his way to pick up Nelms.

That evening, after hearing Nelms's trial testimony, the government sent a deputy U.S. Marshal to drive from the scene of the shooting—which is five blocks away from Jackson's house, in the opposite direction from Nelms's work3—to the restaurant where Nelms worked. The next morning as part of the government's rebuttal case, the marshal testified that the drive, which was 7.4 miles long, took twelve minutes. He testified that he selected what he knew from familiarity with the area to be the shortest route, that he was stopped by red lights at over half of the stoplights, that he conducted the experiment shortly after 7:00 p.m.,4 and that for one portion of the drive he exceeded the speed limit in order to keep up with the flow of traffic. The defense, which objected to the introduction of the evidence, cross-examined the marshal, attacking his experiment because it was not based on the specific route that Jackson used that day, because traffic patterns would vary between night and day, and because his speeding would have affected the overall drive time.

On appeal, Jackson first contends that the district court abused its discretion by admitting the marshal's testimony about the driving experiment. He concedes that he is a "felon" under 18 U.S.C. § 922(g)(1), but contends that he did not possess a firearm, and the evidentiary challenge goes to that element of the offense. Although he does not point to this provision, our review is governed by Federal Rule of Evidence 403, which renders inadmissible evidence whose probative value is outweighed by the danger of unfair prejudice. See United States v. Russell, 971 F.2d 1098, 1106 (4th Cir.1992); Kenneth S. Broun, McCormick on Evidence § 202 (6th ed.2006).

Evidence of experiments is most commonly used in the context of products liability law, where recreations of accidents, explosions, and product malfunctions are now common. See, e.g., Buscaglia v. United States, 25 F.3d 530, 533 (7th Cir.1994); Carey ex rel. Carey v. Hy-Temp Mfg., 929 F.2d 1229, 1235 n. 2 (7th Cir.1991). Because this type of evidence can be quite persuasive, in order to avoid unfair prejudice, the conditions under which an experiment is performed must be "substantially similar" to those surrounding the simulated event. Mihailovich v. Laatsch, 359 F.3d 892, 908 (7th Cir.2004). This is a flexible requirement: "substantially similar" does not mean "identical," and dissimilarities can be explored on cross-examination. See Buscaglia, 25 F.3d at 533. In other words, as a general matter, "dissimilarities between experimental and actual conditions affect the weight, not the admissibility of the evidence." 33A Fed. Proc., L.Ed. § 80:254 (2006).

Although there is little circuit precedent on the subject, the substantially similar requirement also applies in the criminal context. See United States v. Baldwin, 418 F.3d 575, 579-81 (6th Cir. 2005); United States v. Birch, 39 F.3d 1089, 1092-93 (10th Cir.1994); Russell, 971 F.2d at 1105-06. Whether in a criminal case or a civil one, however, the requirement's application always depends on the purpose for which the experiment is introduced. See Jones v. Ralls, 187 F.3d 848, 853 (8th Cir.1999). So if the purpose is to recreate an event, the timing and physics of which are critical, courts will only admit evidence of experiments that are conducted under nearly identical conditions as the actual event. Broun, McCormick § 202. For instance, in Jackson v. Fletcher, 647 F.2d 1020, 1026-28 (10th Cir.1981), the district court erred by admitting evidence of an experiment purporting to recreate an accident between a car and a truck in order to determine the precise speed of the truck at the time of collision. The court held that since the simulation truck was empty whereas the actual truck carried a full load (creating a weight differential of 37,000 pounds), and the two trucks were different model years, the experimental conditions were not substantially similar to the actual ones. Id.

By contrast, where the purpose of the experiment is not to recreate events but simply to rebut or falsify the opposing party's sweeping hypothesis, the substantial similarity requirement is relaxed. Broun, McCormick § 202. In Osborne v. United States, 542 F.2d 1015, 1019-20 (8th Cir.1976), the most similar federal circuit case on point, the defendant, accused of robbery, contended that he could not have committed the crime in the early afternoon because he applied for unemployment benefits at 2:00 p.m. in an office 23 miles away from the scene of the crime. In rebuttal, an FBI agent testified that he had driven from the crime scene to the unemployment office and showed, based on the drive time, that it was possible for the defendant to have committed the robbery and driven to the office, still arriving by 2:00. The appellate court affirmed the district court's admission of the evidence, holding that "the experiment did not purport to recreate events as they actually occurred on the day of the robbery but rather established whether it was physically possible for [the defendant] to have traversed the distance between the stash point and the railway office in the time allowed." Id. at 1120. Other circuits have also concluded that dissimilarities of the type Jackson points out can be identified in cross-examination to weaken the evidence's impact, but they do not bar its admission in the first place. See Birch, 39 F.3d at 1092-93; Russell, 971 F.2d at 1105-06; United States v. Metzger, 778 F.2d 1195, 1204-05 (6th Cir. 1985); cf. Baldwin, 418 F.3d at 581 (although district court did not abuse discretion in excluding evidence of experiment, it would not have abused discretion had it admitted the evidence).5

The case here is quite similar to Osborne. The government was attempting to cast doubt upon the defense's implication that Jackson could not have committed the shooting sometime after 2:00 p.m. and still had time to pick up his girlfriend at 2:43 p.m. (Since Nelms testified that the drive from Jackson's home to her work takes between 20 and 30 minutes, this is not the strongest alibi to begin with.) The government was not trying to recreate Jackson's actual drive. If, as the evidence tended to show, the drive could be completed in as little as 12 minutes, then Jackson could potentially have fled the scene of the shooting as late as 2:31 p.m., and could therefore have committed the shooting (and possessed the handgun) shortly after 2:00 p.m. The jury was permitted to weigh this evidence and consider the weaknesses elicited on cross-examination. If jurors thought the conditions of the experiment too dissimilar, they could give the evidence little weight. Although an ideal test would have recreated the traffic conditions, route, and time of day under which Jackson performed the drive, given the limited purpose for which the experiment was offered, the experiment was substantially similar to actual conditions. Had the testimony been offered in the government's case-in-chief, rather than in rebuttal of the defendant's story, a different result might be warranted. However, on the facts of this case, the district court did not abuse its discretion in admitting the evidence.

Jackson next argues that the jury's verdict may not have been unanimous because the district court provided only a general unanimity instruction. He points out that when the district court explained to the jury the elements of 18 U.S.C. § 922(g)(1), and specifically the element of "possession" of a firearm, the court stated that the jury must find that Jackson possessed a firearm on "July 12 and/or July 15."6 His fear is that some of the jurors might have found that he possessed a gun on July 12th and some on July 15th — which would not, he claims, be a unanimous verdict as to the element of possession. The only way to prevent this possibility, Jackson says, would have been a specific instruction admonishing the jury that it must unanimously find that he possessed the gun either on July 12 or on July 15, rather than a general instruction that the jury must reach a unanimous verdict.

Since Jackson did not object to the district court's instruction or proffer his own, our review is only for plain error. United States v. Peters, 435 F.3d 746, 754 (7th Cir.2006). It is a rare case in which an improper jury instruction will justify upsetting a criminal conviction. Id. When a defendant fails to object to the trial court's general unanimity instruction, that instruction will usually suffice and no specific instruction will be needed. See United States v. Cherif, 943 F.2d 692, 701 (7th Cir.1991); United States v. Williams, 737 F.2d 594, 613-14 (7th Cir.1984); United States v. Pavloski, 574 F.2d 933, 936 (7th Cir.1978).

We find no error here, let alone one that is plain. Unanimity can indeed present a concern where an indictment includes in a single count two offenses that could be charged separately. For instance, in United States v. Payseno, 782 F.2d 832 (9th Cir.1986), a case that Jackson cites, the defendant's indictment for extortion charged, in a single count, three separate extortionate acts against three victims over the course of eight months. The appellate court found that jurors might not have understood that they needed to find unanimously that the defendant was guilty of one of the three acts, and therefore held that the district court's failure to give a specific unanimity instruction was plain error. Id. at 836-37. But see United States v. Anguiano, 873 F.2d 1314, 1319-20 (9th Cir.1989) (emphasizing that a general unanimity instruction is sufficient in the absence of a complex set of facts or a broad and ambiguous indictment that could easily confuse a jury as to the need for unanimity).

But here the indictment charged not multiple offenses under one count of being a felon in possession, but a single course of illegal conduct that spanned three days. See United States v. Fleischli, 305 F.3d 643, 658 (7th Cir.2002) ("Possession of a firearm is a continuing offense which ceases only when the possession stops."); see also United States v. Buchmeier, 255 F.3d 415, 424-25 (7th Cir.2001) (discussing distinction between continuing course of conduct and multiple offenses). If half of the jurors found that Jackson possessed the gun on July 12 and half found that he possessed it on July 15, the jury would still be unanimous that he possessed a gun between July 12 and July 15—which is exactly what the indictment charged. See R.20, at 1 ("On or about July 12, 2004 through July 15, 2004 . . . the defendant herein, did knowingly possess a firearm."); see also Williams, 737 F.2d at 613-14 (emphasizing sufficiency of clear indictment coupled with general unanimity instruction); Anguiano, 873 F.2d at 1320 (same). Indeed, to convict Jackson on two separate counts of being a felon in possession, the government would have had to prove that he lost possession of the gun at some point between the two charged dates. See United States v. Conley, 291 F.3d 464, 470-71 (7th Cir.2002).7

The fact that this case does not present two offenses listed under one count, coupled with our rule that a general unanimity instruction is usually enough in the absence of an objection or a tendered instruction, Pavloski, 574 F.2d at 936, dooms this claim.

Finally, Jackson argues that his possession of the firearm did not affect interstate commerce. He does not dispute the evidence that shows that the gun was manufactured in Brazil, entered the United States in Miami, and then traveled to Illinois, and he acknowledges that this movement in interstate commerce is all the Supreme Court requires under the statute. Scarborough v. United States, 431 U.S. 563, 97 S. Ct. 1963, 52 L. Ed. 2d 582 (1977); see United States v. Williams, 410 F.3d 397, 400 (7th Cir.2005). Rather, he raises this issue to preserve the possibility of further appellate review. Only the Supreme Court can reverse itself, so we need not discuss this contention further.

For these reasons, the judgment of the district court is AFFIRMED.