RANDALL, Circuit Judge:

Claimant-appellant, Pfizer, Inc., appeals the district court's grant of summary judgment in this seizure action instituted by the United States under section 304 of the Federal Food, Drug, and Cosmetic Act, 21 U.S.C. § 301 et seq. (1982). The district court held that Neo-Terramycin Soluble Powder Concentrate was adulterated within the meaning of 21 U.S.C. § 351(a)(5) (1982). We affirm.

I.

It will first be useful briefly to outline the somewhat complicated statutory and regulatory scheme involved in this case. The Federal Food, Drug, and Cosmetic Act, 21 U.S.C. §§ 301- 392 (1982) (the "Act"), provides a comprehensive scheme to protect the public from drugs that may be unsafe or ineffective for their intended uses. As part of this scheme, the Act regulates the marketing of drugs used in treating animals which are raised for human consumption. 21 U.S.C. § 360b. The Act establishes a system of pre-marketing clearance for animal drugs by prohibiting the introduction into interstate commerce of any "new animal drug" unless a Food and Drug Administration ("FDA") approved New Animal Drug Application ("NADA") is in effect for that drug. 21 U.S.C. § 360b(a)(1)(A). The Act establishes procedures for the filing of an NADA, 21 U.S.C. § 360b, and provides that the Secretary of Health and Human Services (the "Secretary") may reject an NADA if he or she finds that the drug is not proven safe, 21 U.S.C. § 360b(d)(1)(A) (B), or that "there is a lack of substantial evidence that the drug will have the effect it purports or is represented to have under the conditions of use prescribed, recommended, or suggested in the proposed labeling thereof." 21 U.S.C. § 360b(d)(1)(E). "Substantial evidence" is defined to mean:

[E]vidence consisting of adequate and well-controlled investigations, including field investigation, by experts qualified by scientific training and experience to evaluate the effectiveness of the drug involved, on the basis of which it could fairly and reasonably be concluded by such experts that the drug will have the effect it purports or is represented to have under the conditions of use prescribed, recommended, or suggested in the labeling or proposed labeling thereof.

21 U.S.C. § 360b(d)(3). The FDA has required, by regulation, that for a fixed combination new animal drug, substantial evidence must be provided both that the drug is effective for its intended use and that each ingredient designated as active contributes to the claimed effects. 21 C.F.R. § 514.1(b)(8)(v) (1983).

A drug that is not a "new animal drug" can be marketed without FDA approval. For a drug not to be considered a new animal drug, it must be "generally recognized" by qualified experts as safe and effective for each of its intended uses. 21 U.S.C. § 321(w).

If a "new animal drug" is marketed without FDA approval, it is deemed to be "unsafe." 21 U.S.C. § 360b(a)(1)(A). An unsafe animal drug is deemed "adulterated." 21 U.S.C. § 351(a)(5). The Act prohibits the shipment of any adulterated animal drug in interstate commerce. 21 U.S.C. § 331(a). The Secretary may sue to enjoin violations, 21 U.S.C. § 332, prosecute the violator criminally, 21 U.S.C. § 333, or seek to have the district court seize and condemn the shipment of the adulterated animal drug, 21 U.S.C. § 334.

Thus, drug manufacturers have two alternative routes in marketing their animal drugs. The first is to seek FDA approval of an NADA. The second is to market the drug without FDA approval on the assumption that the drug is not a "new animal drug," thus leaving the initiative to the Secretary. As such, the Act is somewhat incongruous; it provides a comprehensive scheme for both pre-marketing clearance and post-marketing regulation of new animal drugs, yet leaves it up to the manufacturer of the animal drug to decide whether its product is subject to the scheme in the first place. "The in terrorem effect of potential judicial enforcement actions, the possibility of penalties, and the extremely broad definition of a new drug as any one not generally recognized by experts as safe and effective are apparently designed to encourage voluntary compliance and to render the scheme self-policing." Cutler v. Kennedy, 475 F. Supp. 838, 842, 1979 Food Drug Cosm.Rep. (CCH) ¶ 38,257, at 39,042 (D.C.D.C. 1979).

II.

On May 15, 1979, the United States filed a complaint for forfeiture, under section 304 of the Act, praying for seizure and condemnation of a specified quantity of the animal drug Neo-Terramycin Soluble Powder Concentrate ("Neo-Terra Powder") that was shipped in interstate commerce by Pfizer, Inc. ("Pfizer"). The complaint alleged that the drug was a "new animal drug" within the meaning of section 321(w), for which an approved NADA was not in effect. After seizure pursuant to a warrant for arrest in rem, Pfizer intervened as claimant and filed an answer, alleging that Neo-Terra Powder is "generally recognized" as safe and effective, and is not, therefore, a "new animal drug" within the meaning of the Act.

Neo-Terra Powder is a water soluble combination drug, administered in drinking water, labeled for the treatment of various ailments common to chickens, turkeys, swine and calves. The active ingredients in Neo-Terra Powder are neomycin and oxytetracycline. Although Pfizer has marketed Neo-Terra Powder since the mid-1960s, it has not obtained an approved NADA from the FDA.

A jury trial was held from October 14, 1980 to October 23, 1980, at the conclusion of which the jury found that Neo-Terra Powder is "generally recognized" as safe and effective. On November 4, 1980, the district court issued a final judgment based upon the jury's verdict and ordered that the United States Marshal return the seized Neo-Terra Powder to Pfizer.

The next day, November 5, 1980, the United States Marshal released the drug, in violation of the automatic ten-day stay provision of Fed.R.Civ.P. 62(a). The marshal's return shows that the marshal's office called Pfizer on November 5 to notify Pfizer that the drug had been released, and then mailed Pfizer a copy of the written order. The United States was not notified of the release, although it learned informally of the release sometime during the ten-day automatic stay period. The United States did not request or obtain an order staying execution of the judgment releasing the Neo-Terra Powder to Pfizer, nor did it request or obtain an order requiring its return once improperly released.

On November 7, 1980, the United States filed a timely motion for a new trial, pursuant to Fed.R.Civ.P. 59(a). This motion was granted on May 8, 1981.

On June 8, 1981, Pfizer renewed a motion it had made at trial that the proceeding should be dismissed because the status of the drug is governed by the FDA's interim marketing rules developed for certain animal drugs, 21 C.F.R. § 558.15 (1983). On November 9, 1981, Pfizer moved to dismiss the complaint alleging that the court lacked in rem jurisdiction following the release and removal of the seized drug from the district. The district court denied both motions.

The United States, on December 17, 1981, moved for summary judgment, which the district court granted on May 6, 1982. United States v. An Article of Drug. . . Neoterramycin Soluble Powder Concentrate, 540 F. Supp. 363 (N.D.Tex. 1982). Pfizer filed a timely notice of appeal.

III.

The first issue we must address is whether the release by the United States Marshal of the 4,680 pails of Neo-Terra Powder destroyed the jurisdiction of the district court; Pfizer contends that it did. Pfizer's argument rests on the admiralty rule that in rem actions generally require, as a prerequisite to a court's jurisdiction, the presence of the vessel or other res within the territorial confines of the court. United States v. Mack, 295 U.S. 480, 484, 55 S.Ct. 813, 815, 79 L.Ed. 1559 (1935). This rule is predicated upon admiralty's fiction of convenience that a ship is a person against whom suits can be filed and judgments entered, allowing actions to be brought against the vessel when her owner cannot be reached. Continental Grain Co. v. Barge FBL-585, 364 U.S. 19, 22-23, 80 S.Ct. 1470, 1472-1473, 4 L.Ed.2d 1540 (1960). However, in recent years this court has shied away from a strict construction of this in rem rule and has allowed, in certain circumstances, an in rem action to continue despite the absence of the res. Specifically, we have dispensed with strict application of this in rem rule when a legal fiction which exists solely to effectuate the adjudication of disputes is invoked for the opposite purpose, and we have found a substitute basis of in personam jurisdiction. See, e.g., Treasure Salvors, Inc. v. Unidentified Wrecked and Abandoned Sailing Vessel, 569 F.2d 330 (5th Cir. 1978); Inland Credit Corp. v. M/T Bow Egret, 552 F.2d 1148 (5th Cir. 1977).

We first dispensed with strict construction of this in rem rule in Inland Credit Corp. v. M/T Bow Egret, supra. In that case, a ship mortgagee brought suit in rem against the vessel and in personam against the vessel's owner seeking to foreclose on its mortgage; creditors, including persons who had advanced funds for an operating fund maintained under the mortgage agreement, intervened. Pursuant to court order, the vessel was sold and the proceeds distributed to the mortgagee and certain creditors. Two creditors who were not permitted to share in the distribution sought to appeal the district court's decision not to include them. The mortgagee, however, argued that, because the res had been distributed and was therefore no longer within the control of the court, this court did not have jurisdiction over the issue of whether the two creditors should have been granted a share of the res funds. While recognizing the general rule that removal of the res destroys a court's jurisdictional base, we found that, because the mortgagee had brought an action both in rem and in personam, we retained jurisdiction.

We conclude that it is proper for us to exercise our jurisdiction in personam to correct any errors in the lower court's in rem adjudication. We recognize, to be sure, that a strict construction of the limits of in rem actions might bar any appeal of an in rem action unless supported by continued in rem jurisdiction. This narrow view of our jurisdiction, however, is difficult to square with the modern view, now almost a cliche, that in personam jurisdiction can be asserted whenever the defendant has those minimum contacts with the forum state that will satisfy "traditional notions of fair play and substantial justice." We need not decide in the present case whether the philosophical underpinnings of the system of in rem jurisdiction in admiralty have been critically shaken. We do decide, however, that where, as here, we find ourselves at the interface of in rem and in personam jurisdiction, we can properly exercise a broad in personam power.

552 F.2d at 1152 (citations omitted).

We were next presented with such an interface of in rem and in personam jurisdiction in Treasure Salvors, Inc. v. Unidentified Wrecked and Abandoned Sailing Vessel, supra. In that case, a Florida corporation, Treasure Salvors, sued for possession of and confirmation of title to an unidentified wrecked and abandoned vessel thought to be the Nuestra Senora de Atocha, a Spanish vessel that sank off the Marquesas Keys in 1622. The United States intervened, answered and counterclaimed, asserting title to the vessel. Because the wreck of the vessel rested on the continental shelf, outside the territorial waters of the United States, the district court entered summary judgment for Treasure Salvors. The United States appealed, asserting that the court lacked in rem jurisdiction to determine the rights of the parties to that portion of the res situated beyond the territorial jurisdiction of the court. We noted that "when a legal fiction which exists solely to effectuate the adjudication of disputes is invoked for the opposite purpose, we have no hesitation in declining to employ it." 569 F.2d at 334 (footnote omitted). Thus, despite the absence of the res from the territorial jurisdiction of the court, we, evidencing a concern with finding the most practical and efficacious means of resolving the dispute before the court and an interest in rendering justice rather than an automatistic reliance upon rigid legalisms, refused to apply a strict construction of the in rem rule. Instead, we again, as we had in Inland Credit, allowed the court's in personam jurisdiction to cure any defects in the court's in rem jurisdiction. Although neither party had brought an in personam action against the other, we held that the court, nevertheless, had in personam jurisdiction over the claimants, citing with approval The Fairisle, 76 F. Supp. 27, 34 (D.Md. 1947), in which the court held that owners who appeared in an in rem action to contest the plaintiffs' claims may equitably be treated as if they had been brought into court by personal process. Thus, we found the vessel's arrest nonessential to the resolution of the action.

We are presented with a similar interface of in rem and in personam jurisdiction in the instant case. Pfizer entered a general appearance in the district court. It did not, as is allowed under Admiralty Rule E(8), enter a limited appearance. Moreover, even after the release of the res, Pfizer voluntarily appeared before the district court. Although the res had been released on November 5, 1980, Pfizer submitted a memorandum in opposition to the government's new trial motion on November 26, 1980 and a response to the government's reply on December 31, 1980. On neither occasion did Pfizer make mention of the release of the drug. On June 8, 1981, Pfizer invoked the jurisdiction of the district court in moving for disposition under 21 C.F.R. § 558.15 (1983), again without any mention of the release. On September 22, 1981, Pfizer submitted interrogatories to the United States, and on November 3, 1981, it filed answers to the United States' interrogatories. Therefore, we find, as we did in Treasure Salvors, supra, that the district court had in personam jurisdiction over the parties, thus rendering the res' arrest non-essential to the court's jurisdiction.

Pfizer contends, however, that the district court did not retain jurisdiction because the release of the res rendered the case moot. Because the relief sought by the United States' complaint was forfeiture and destruction of the res, Pfizer argues, once the res was beyond the court's control, it could no longer grant the relief requested; accordingly, the case is moot. We do not agree. While, in form, the relief requested by the United States is the forfeiture and destruction of a specifically identified lot of Neo-Terra Powder, the action, in substance, seeks much more. The substantive character of the remedy sought in this case is not forfeiture and destruction of a specific lot of Neo-Terra Powder, but a declaration that Neo-Terra Powder is or is not a new animal drug. The effect of this finding would be res judicata against Pfizer in a subsequent seizure, United States v. 4 Cans . . . Master Liquid, 127 F. Supp. 243, 246 (N.D.Iowa 1955), as well as being res judicata in a subsequent injunction action. United States v. Nysco Laboratories, Inc., 318 F.2d 817, 817-18 (2d Cir. 1963). Conversely, a judgment in this case for Pfizer would be res judicata against the United States. United States v. One Dozen Bottles . . . Boncquet Tablets, 146 F.2d 361, 363 (4th Cir. 1944). Thus, the ultimate issue for each of the parties is whether or not Pfizer must submit Neo-Terra Powder to the FDA for pre-marketing approval; the effect of the court's holding goes far beyond a single lot of the drug. The fiction of bringing suit against the drug itself is invoked simply to get the issues and the parties before the court; this purpose was accomplished. With both parties before the court, the action, for all practical purposes, becomes one for a declaratory judgment to determine whether Neo-Terra Powder is a new animal drug. Therefore, we refuse, in this case, to elevate form over substance. Both the controversy and the parties being before the court, the action is not moot and need not be dismissed. Accordingly, we turn to an examination of the merits of Pfizer's appeal.

IV.

In reviewing the district court's grant of summary judgment in favor of the United States, we are mindful that summary judgment may be granted only if it appears from pleadings, depositions, admissions and affidavits, considered in the light most favorable to the non-moving party, that there is no genuine issue as to any material fact and that the moving party is entitled to judgment as a matter of law. Poller v. Columbia Broadcasting System, Inc., 368 U.S. 464, 468, 82 S.Ct. 486, 488, 7 L.Ed.2d 458 (1962); Mahfood v. Continental Grain Co., 718 F.2d 779, 781 (5th Cir. 1983). The moving party carries the burden of showing that no genuine issue of material fact exists. Once the movant makes this showing, the burden shifts to the opposing party to show that summary judgment is not appropriate; the opposing party must counter the moving party's affidavits with opposing affidavits or other competent evidence setting forth specific facts to show that there is a genuine issue of material fact for trial. Mere allegations are insufficient. Fed.R.Civ.P. 56(e).

The basic issue in this case, whether Neo-Terra Powder is a "new animal drug," involves a determination of whether it is "generally recognized" among qualified experts as safe and effective. 21 U.S.C. § 321(w). The Act is silent as to the kind of evidence required to establish general recognition. The Supreme Court faced this issue in Weinberger v. Hynson, Westcott Dunning, 412 U.S. 609, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973), and determined that "the statutory scheme and overriding purpose of the 1962 amendments [to the Act] compel the conclusion that the hurdle of `general recognition' of effectiveness requires at least `substantial evidence' of effectiveness for approval of [a new drug application]." 412 U.S. at 629, 93 S.Ct. at 2483. Thus, "general recognition" requires a two-step showing: first, that there is general recognition in fact, i.e., that there is an expert consensus that the product is effective; and second, that the expert consensus is based upon "substantial evidence" as defined in the Act and in FDA regulations. "Substantial evidence" itself embodies two elements — one quantitative and one qualitative. Quantitatively, the studies concerning the product need not be unanimous, or even overwhelming, in their support of the product's effectiveness. The legislative history of the Act makes this clear:

The proposed committee amendment clarifies and strengthens the previously reported bill by restating and carefully defining the quality and quantum of evidence which the Secretary must find to exist as a basis for clearance of the drug or for withdrawal of a previously approved new-drug application. In the course of committee deliberations a distinction evolved, in this connection, between two tests — the "preponderant evidence" test and the "substantial evidence" test as now specifically defined. Under the former a claim would not be accepted under the new-drug section unless it represented the preponderant view of experts qualified by training and experience in the subject that the claim was supported. The committee recognizes that in the difficult area of drug testing and evaluation there will frequently, if not usually, be a difference of responsible opinion. The committee feels that the existence of such a difference should not result in disapproval of a claim of effectiveness if it is supported by substantial evidence defined in the manner set forth below and evaluated by the Secretary in the light of all the information available to him at the time.

S.Rep. No. 1744, Part 2, 87th Cong., 2d Sess. 6, 1962, U.S. Code Cong. Admin.News, p. 2884. As noted by the Fourth Circuit, "substantial in this connection does not mean `preponderant evidence' or `conclusive evidence.' Congress specifically discarded those terms for the milder term `substantial,' which was understood to embrace the idea, not of a preponderance, but rather of a responsible body of qualified opinion." Hynson, Westcott Dunning, Inc. v. Richardson, 461 F.2d 215, 220 (4th Cir. 1972), aff'd sub nom. Weinberger v. Hynson, Westcott Dunning, Inc., 412 U.S. 609, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973) (footnote omitted). Qualitatively, the studies must be adequate and well-controlled, upon which qualified experts can reasonably conclude that the product is effective for its labeled indications. 21 U.S.C. § 360b(d)(3).

While general recognition is normally a factual question of whether there is an expert consensus that is based upon substantial evidence, should we find the studies relied upon by Pfizer and its experts as substantial evidence to be conclusively deficient with respect to the statutory and regulatory standards, then summary judgment is proper. Accord Masti-Kure Products Co. v. Califano, 587 F.2d 1099 (D.C. Cir. 1978); United States v. An Article of Drug . . . Entrol-C Medicated, 513 F.2d 1127 (9th Cir. 1975). The district court found Pfizer's studies to be legally deficient and accordingly granted summary judgment for the United States. The district court found Pfizer's studies legally deficient in three respects: (1) Pfizer's studies were not performed on a drug with the same form, dosage and application as Neo-Terra Powder; (2) none of Pfizer's studies complied with the FDA's "combination policy;" and (3) Pfizer's studies did not establish that Neo-Terra Powder is generally recognized as effective for all of its labeled indications. Because we agree that Pfizer's studies fail to comply with the requirements of the FDA's "combination policy," we affirm the district court's award of summary judgment.

The FDA's combination policy provides:

Each ingredient designated as active in any new animal drug combination must make a contribution to the effect in the manner claimed or suggested in the labeling, and, if in the absence of express labeling claims of advantages for the combination such a product purports to be better than either component alone, it must be established that the new animal drug has that purported effectiveness.

21 C.F.R. § 514.1(b)(8)(v) (1983). Pfizer argues that the combination policy is not applicable here since the policy is invoked only if there are express label claims of advantages for the combination or the product purports to be better than the ingredients alone. Pfizer contends that no such claims were made. We find, however, that there remains no genuine issue of material fact that such claims were made and therefore, that the combination policy is, indeed, invoked. As the district court noted, "[o]xytetracycline and neomycin have different spectra of antibiotic activity, acting against different organisms, although there is some overlap. The two drugs are also active in different areas, neomycin primarily in the gut and oxytetracycline primarily systemically." 540 F. Supp. at 377. The labels, which were put into evidence at the prior trial and submitted by both parties with their affidavits in this summary judgment motion, clearly make no express claims of the advantages of Neo-Terra Powder as a combination; the United States does not argue otherwise. However, the United States presented affidavits by experts who averred that, as to those uses on the labeling for which oxytetracycline alone would be effective, the dosage levels indicated on the labeling for Neo-Terra Powder were substantially lower than those which would be required for oxytetracycline alone. This was uncontroverted by the affidavits of Pfizer's experts. In the absence of counter documents to oppose the United States' affidavits, the facts attested must be deemed true and established. The district court thus determined that Neo-Terra Powder conclusively purports to be better than its individual components. We agree. One of the rationales for the combination policy is that a single component of a drug may affect the other components, react with them, and enhance or reduce the total effect of the drug. United States v. Article of Food and Drug . . . Coli-trol 80, 518 F.2d 743, 746 (5th Cir. 1975); United States v. X-Otag Plus Tablets, 441 F. Supp. 105, 111 (D.Colo. 1977). Thus any combination drug, the labeling of which indicates substantially lower dosage levels than would be required for one of the active ingredients alone (as to those indicated uses for which that ingredient alone would be effective), must be held to conclusively purport to be better than its individual components, thereby requiring compliance with the FDA's combination policy.

Therefore, in order to substantiate that Neo-Terra Powder is generally recognized as effective under the combination policy, 21 C.F.R. § 514.1(b)(8)(v) (1983), Pfizer must demonstrate that Neo-Terra Powder is more effective than are neomycin and oxytetracycline alone. Pfizer must demonstrate the validity of this assumption through "substantial evidence," i.e., adequate and well-controlled investigations. Adequate and well-controlled investigations required by the combination policy necessitate studies that compare the effect of the individual ingredients with the effect of the combination, and that include a control group. Masti-Kure Products Co., Inc. v. Califano, supra, 587 F.2d at 1103-04, 1105; Record Vol. 3 at 600-01 (Affidavit of FDA employee Howard Meyers). The United States presented the affidavits of experts who averred that none of Pfizer's studies satisfy this legal requirement. Pfizer does not controvert this except to point to the expressed opinions of its experts that Neo-Terra Powder is more effective than its component parts and to one in vitro study concerning the synergistic effects of the combination of neomycin and oxytetracycline.

Substantial evidence does not consist of the expressed opinions of experts hired to testify on behalf of one party or the other. Instead, it consists of adequate and well-controlled studies that must be generally available to the scientific community. In this regard, Pfizer points to only one in vitro study. Such a study cannot alone satisfy the requirements of substantial evidence. First, the very definition of substantial evidence in the Act requires that field investigations be included; an in vitro study is not a field investigation. Further, the regulation that defines "adequate and well-controlled studies," by its tenor, clearly requires in vivos studies. For example, the regulation requires certain standards for "[t]he method of selection of the animals to be studied . . . ." 21 C.F.R. § 514.111(a)(5)(ii) (1983) (emphasis added). Since Pfizer has wholly failed to provide any adequate and well-controlled studies that compare the effect of the individual ingredients with the effect of the combination, there has been presented no substantial evidence which could support a finding of general recognition of effectiveness under the FDA's combination policy. Accordingly, the United States is entitled to judgment as a matter of law and the district court's award of summary judgment must be affirmed.

V.

Pfizer also appeals the district court's denial of its motion for disposition of this action under 21 C.F.R. § 558.15 (1983). During the early 1970s, a concern arose in the scientific community concerning the safety of widespread, long-term subtherapeutic use of antibiotics in food-producing animals. Specifically, this concern involved the possibility that the use of such drugs, over time, might favor the selection and development of antibiotic-resistant bacteria that might, under certain conditions, promote an increase in the prevalence of antibiotic-resistant bacteria in humans.

In response to this concern, the FDA assembled an expert Antibiotic Task Force. This group of experts studied the problem at some length and recommended to the FDA that extensive studies be undertaken to resolve the question of whether there is in fact any real danger of an increase in antibiotic-resistant, bacterially-induced disease in humans or animals due to the use of antibiotics in food-producing animals.

Acting upon the recommendations of the task force, the FDA, after notice and opportunity for comment, published a final regulation, 21 C.F.R. § 135.109 (since renumbered 558.15), in the Federal Register of April 20, 1973. That regulation states:

The Commissioner of Food and Drugs will propose to revoke currently approved subtherapeutic (increased rate of gain, disease prevention, etc.) uses in animal feed of antibiotic and sulfonamide drugs whether granted by approval of new animal drug applications, master files and/or antibiotic or food additive regulations, by no later than 2 years following effective date of this order, unless data are submitted which resolve conclusively the issues concerning their safety to man and animals and their effectiveness under specific criteria established by the Food and Drug Administration based on the guidelines included in the report in animal feeds. All persons or firms previously marketing identical, related, or similar products not the subject of an approved new animal drug application must submit a new animal drug application by July 19, 1973, if marketing is to continue during the interim. New animal drug entities with antibacterial activity not previously marketed, now pending approval or submitted for approval prior to, or following the effective date of this publication, shall satisfy such criteria prior to approval.

21 C.F.R. § 558.15(a) (1983) (emphasis added).

The FDA interpreted the regulation initially to include water soluble drugs. Pfizer, pursuant to this interpretation of the regulation, submitted a timely NADA for Neo-Terra Powder on July 16, 1973. However, the FDA later proposed not to include water soluble drugs within the scope of the regulation, 39 Fed.Reg. 28382, 28383 (August 6, 1974); after a notice and comment period, the agency formally chose to exclude water soluble drugs from the regulation. 41 Fed.Reg. 8282, 8286 (February 25, 1976). Pfizer argues that the FDA did not change its interpretation of section 558.15 to exclude water soluble products or, if it did, that it did not do so in a proper manner. Accordingly, Pfizer claims that Neo-Terra Powder is entitled to full interim marketing privileges while its NADA is under review by the FDA. The United States contends that it properly changed its interpretation of section 558.15 to exclude water soluble animal drugs and that, therefore, Neo-Terra Powder is not entitled to interim marketing privileges thereunder. For the reasons given in the thorough opinion of the district court, published at 540 F. Supp. 363 (N.D.Tex. 1982), we agree with the United States. We write here only to deal with those of Pfizer's arguments on this appeal that appear not to have been directly addressed below.

Pfizer argues that the exclusion of Neo-Terra Powder from the FDA's lists of all the drugs that were to be covered by the regulation was merely administrative error. This argument ignores, however, the fact that absolutely no water soluble product was included on the lists and the explicit language in the FDA's publication of the final list in 1976, which makes clear that water soluble products were being purposely excluded. The exclusion of Neo-Terra Powder was not merely an administrative error.

Pfizer further argues that the "Human Health Safety Criteria," which are incorporated in the FDA's regulations, and which included the definition of "Feeds" as "any feedstuff including water to which antibacterial drugs are added," has the force of law as set forth in 21 C.F.R. § 10.90 (1983):

A guideline represents the formal position of FDA on a matter and, except as provided in paragraph (b)(3) of this section, obligates the agency to follow it until it is amended or revoked. The Commissioner may not recommend legal action against a person or product with respect to an action taken in conformity with a guideline issued under this section that has not been amended or revoked.

While the original inclusion of water soluble products within the definition of section 558.15 had the force of law at that time, the significance of Pfizer's argument is unclear since we have held that the FDA has lawfully changed its interpretation of the regulation, thereby amending the guidelines. Accordingly, we affirm the district court's dismissal of Pfizer's motion.

VI.

Finally, Pfizer appeals the district court's granting of the United States' motion for a new trial following the original jury trial in this case. Following a jury trial in which the jury found that Neo-Terra Powder is "generally recognized" as effective, the district court granted a motion by the United States for a new trial "on the ground that there is insufficient evidence to support the jury's verdict." Record Vol. 2 at 387.

Initially, we note that this case comes to us in a rather unusual procedural posture — a grant of summary judgment following the grant of a new trial. It might seem, at first blush, that our disposition with regard to the district court's grant of summary judgment would be dispositive of the issue of the propriety of the district court's previous grant of a new trial. While this might be true in an appropriate case, we do not think this is such a case. First, our affirmance of the district court's summary judgment is based, at least in part, upon uncontroverted affidavits presented on the motion for summary judgment which were not presented at the original trial. Moreover, the district court granted the motion for a new trial, not upon the ground of the legal deficiency of Pfizer's studies, as was our affirmance of the summary judgment, but upon the weight of the evidence. Therefore, we review the district court's grant of a new trial upon the basis of the evidence presented at the original trial and upon its stated grounds for doing so.

Pfizer first notes that, at the jury trial, it presented the testimony of seven experts who testified that Neo-Terra Powder is generally recognized as effective. Pfizer then argues, relying upon language from cases in which we have examined denials of grants of motions for new trials, that the district court's grant of a new trial can be upheld only where there is an "absolute absence" of evidence to support the jury's verdict; therefore, since it presented evidence that would support the jury's verdict, Pfizer argues, we must reverse the district court's order and reinstate the jury verdict.

Pfizer miscomprehends our standard of review for a grant of a new trial. The trial judge can grant a new trial if he believes the verdict is contrary to the weight of the evidence. Montgomery Ward Co. v. Duncan, 311 U.S. 243, 61 S.Ct. 189, 85 L.Ed. 147 (1940); Bazile v. Bisso Marine Co., 606 F.2d 101 (5th Cir. 1979). In a motion for a new trial, the judge is free to weigh the evidence. Id. As an appellate court, our review of the grant or denial of a new trial is severely limited; we may interfere only when the court abuses its discretion or fails to exercise it. Bazile v. Bisso Marine Co., supra. In light of the trial court's discretion, its freedom to weigh the evidence, and our limited scope of review, we have held in cases reviewing the denial of a motion for new trial that the district court does not abuse its discretion — and, a fortiori, we may not reverse that decision — unless there is an "absolute absence" of evidence to support the jury's verdict. See, e.g., Litherland v. Petrolane Offshore Construction Services, 546 F.2d 129, 132 (5th Cir. 1977).

This in no way intimates, however, that, in reviewing the grant of a new trial, we may sustain the district court only if there is an "absolute absence" of evidence to support the jury's verdict. There is no deference to the district court's discretion in such a standard. Rather, the district court's discretion, its freedom to weigh the evidence, and our limited scope of review, suggest the exact opposite: we may reverse the district court's grant of a new trial only if we find an abuse of discretion, i.e., only where there is an "absolute absence" of evidence contrary to the jury's verdict.

We do not find such an "absolute absence" in this case. The factual issue to be decided at trial was whether Neo-Terra Powder is "generally recognized" as effective. While general recognition does not require unanimous recognition, the existence of a "severe conflict" in the expert testimony precludes a finding of general recognition. United States v. X-Otag Plus Tablets, 441 F. Supp. 105, 113-14 (D.Colo. 1977), aff'd, 602 F.2d 1387 (10th Cir. 1979); United States v. Article of Drug . . . Beuthanasia D Regular, 1979 Food Drug Cosm. L.Rep. (CCH) ¶ 38,265 (D.Neb. 1979). The United States presented the testimony of several experts who testified that Neo-Terra Powder is not generally recognized as effective. The district court, in weighing the evidence, could reasonably have determined that there was a severe conflict in the expert testimony that precluded a finding of general recognition. Accordingly, we do not find that the district court abused its discretion.

The judgment of the district court is AFFIRMED.

. . . . .