*119OPINION
PER CURIAM.Appellant, the United States, appeals from the judgment of the district court, contending that the court erred in several of its evidentiary rulings made below in the course of a bench trial. The United States requests a new trial. As we determine that the district court either did not err or committed harmless error in its rulings, we affirm.
I.
James Howard and his corporation Down East Farms, Inc. (the “defendants”) farmed approximately 2,400 acres of land during the 1992 crop year. In September and October of 1992, Howard sold approximately 45,000 bushels of corn owned by Down East Farms under a fictitious name. Also in 1992, the defendants signed or caused to be signed various documents in connection with filing for and obtaining crop insurance indemnity payments and disaster relief payments from several federal agencies and programs, including the Risk Management Agency and the Farm Service Agency’s Disaster Assistance Program. The defendants received $85,893 for crop losses from these programs. The defendants did not report the 45,000 bushels of corn they sold in 1992 on the documents filed with the federal agencies. If the bushels were from the 1992 harvest, then evidently the defendants would not have been entitled to the insurance and disaster relief payments received for crop losses in 1992.
The United States sued the defendants under several legal theories, including violation of the False Claims Act, common law fraud, unjust enrichment, and payment under mistake of fact. All of these claims turned on the allegation that the defendants gave false information on the forms submitted to the federal agencies in order to defraud the United States. This allegation in turn depended on the United States’ contention that the 45,000 bushels sold under a false name were part of the 1992 harvest.
The basis for the United States’ contention that the corn was harvested in 1992 was as follows (according to the testimony of the expert witnesses presented by the United States, or admitted by the defendants). Corn can be stored safely for long periods of time only at or below a moisture level of 15.5-16%. Corn stored at a higher moisture level runs the risk of disease, mold, or rot. When sold, the moisture level of corn is measured by machine. The type of machine used to measure the disputed bushels’ moisture was known as a Dickey John. When samples of the disputed 45,000 bushels of corn were measured by a Dickey John, the readings returned by the machine placed the moisture level of the corn between 15.5% and 17.8%. The buyer of the disputed corn reduced the price paid for drying charges, but not for poor quality, as one would presumably expect if the corn were stored for a long period of time above a 15.5-16% moisture level. Hence, the United States argued, this evidence, although circumstantial, established that the disputed 45,000 bushels sold by the defendants in 1992 were harvested in 1992.
The defendants claimed that the 45,000 bushels sold in 1992 were from the defendants’ stored corn, harvested in 1991 or earlier. Therefore, they argued, they did not give false information on the various documents relating to crop losses in 1992. To explain the high moisture level of the corn, the defendants claimed that they had sprayed water on the corn either during storage or during the transfer of the corn from the storage bins into the trucks used to ship the corn.
*120To support this explanation, at trial the defendants called a witness by the name of Harold Smith, an operator of a Dickey John, and the individual who purchased the disputed bushels of grain from the defendants. After an objection from the United States (overruled by the district court), Mr. Smith testified and then performed an in-court experiment on the Dickey John machine. He took a small sample of corn and measured it in the Dickey John machine. The machine returned a reading of 13.8% moisture. Then, Mr. Smith sprayed a small amount of water on the corn, and ran it through the machine. After approximately 15 minutes (during which time Mr. Smith was vigorously cross-examined by counsel for the United States as to his observations and his expertise in the Dickey John and in corn), the machine returned a reading of 19.5% moisture.
At the close of the trial, defendants moved for a judgment on partial findings pursuant to Fed.R.Civ.P. 52(c). The district court, relying in part on the result of the in-court experiment in its factual findings, granted the defendants’ motion and entered a judgment against the United States on all of its claims.
The United States now appeals, challenging the admission of opinion testimony from Mr. Smith, as well as the admission of the in-court experiment.
II.
The United States argues that the district court erred in allowing the in-court experiment into evidence. Although it is not clear as to which of the Federal Rules of Evidence the United States contends was violated, two seem possible: Fed. R.Evid. 402 (stating that evidence “which is not relevant is not admissible”) and Fed. R.Evid. 403 (stating that “evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury.... ”).
First, we conclude that the in-court experiment was relevant. “‘Relevant evidence’ means evidence having any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence.” Fed. R.Evid. 401. The fact that the spraying of water on the exterior of a sample of corn could raise the moisture reading of the corn sample on the Dickey John does make the defendants’ explanation for the high moisture readings of the disputed bushels at least slightly more probable. Hence, the admission of this evidence did not violate Fed.R.Evid. 402 (“All relevant evidence is admissible, except as otherwise provided.... ”).
Second, the United States’ objection fares no better under Fed.R.Evid. 403. “[I]n the context of a bench trial, evidence should not be excluded under 403 on the ground that it is unfairly prejudicial.” Schultz v. Butcher, 24 F.3d 626, 632 (4th Cir.1994). As the instant case involved a bench trial, rather than a jury trial, Fed. R.Evid. 403 was not violated by the admission of the in-court experiment.
The United States also raises numerous objections to aspects of Mr. Smith’s testimony, characterizing portions of the testimony as opinion testimony, admitted in violation of Fed.R.Evid. 701 and 702. But, the only factual findings in the district court’s order relating to Mr. Smith’s testimony are the following: “13. The court finds, through the testimony of grain buyer Harold Smith and his demonstration with the actual Dickey John used to measure the moisture content of the disputed 45,000 bushels of corn, that it is possible to spray water onto corn and thereby raise *121the moisture content measurement on the Dickey John. 14. The aforementioned demonstration in the court showed that spraying water on corn with an initial content of 13.8 could raise the moisture level of the corn to approximately 19.5, an increase of 5.7.” J.A. 553. These factual findings are seemingly based entirely on the result of the in-court experiment, which we have determined to be admissible. The district court does not rely elsewhere on any of the objected — to opinion testimony in its order.* Hence, to the extent that the district court entered opinion testimony in violation of Fed.R.Evid. 701 and 702, these errors were harmless. See Taylor v. Virginia Union Univ., 193 F.3d 219, 235 (4th Cir.1999).
The judgment of the district court is affirmed.
AFFIRMED.
In its order, the district court did reference one piece of opinion testimony by Mr. Smith. J.A. 550. But, the opinion testimony was mentioned in a section labeled "Preamble.” This section appeared to be simply a short description of what occurred during the trial, rather than a discussion of the district court's reasoning. Also, the opinion testimony mentioned by the district court was given in response to direct questions from opposing counsel on cross-examination, and thus any error was invited by the United States’ counsel and its admission cannot be challenged by the United States on appeal. See United States v. Neal, 78 F.3d 901, 904 (4th Cir. 1996).