Opinion for the Court filed by Circuit Judge SILBERMAN.
SILBERMAN, Circuit Judge:Pursuant to his powers under the Endangered Species Act (“Act”), 87 Stat. 884, 16 U.S.C. §§ 1531-1544, the Secretary of Interior issued an emergency regulation on August 4, 1989 listing the Mojave Desert population of the desert tortoise (Gopherus agassizii) as an endangered species. See Emergency Rule, 54 Fed.Reg. 32326 (to be codified at 50 C.F.R. § 17.11) (Aug. 4, 1989). Appellants — the City of Las Vegas, the Nevada Development Authority, several private Nevada real estate developers, and the Státe of Nevada — brought suit in district court alleging that the emergency listing violated both the Administrative Procedure Act and the Endangered Species Act. Appellants sought a preliminary injunction barring the implementation and enforcement of the regulation pending resolution of the litigation. The district court denied appellants’ motion and we affirm.
I.
The Act charges the Secretary of Interior with the responsibility for identifying “endangered” species and provides him with the authority to protect them from further deterioration attributable to human activities — whether industrial, recreational, or even scientific. An endangered species is defined as one “in danger of extinction throughout all or a significant portion of its range....” 16 U.S.C. § 1532(6). The Act, inter alia, generally prohibits the “taking” of an endangered species, which includes harming or capturing it. See 16 U.S.C. §§ 1538(a)(1)(B), 1532(19). The relocation of a listed species or the alteration of its habitat during construction activities constitutes an “incidental taking” that is prohibited by the Act unless the Secretary grants a special permit. See 16 U.S.C. § 1539(a). This restriction is of particular relevance to appellants, who allege that the listing at issue here will bring construction activity in southern Nevada to a standstill.
Normally, the Secretary lists a species that he determines to be endangered1 by promulgating a regulation after undertak*60ing formal rulemaking pursuant to APA procedures. See 16 U.S.C. § 1533(b)(4). In case of an “emergency posing a significant risk to the well-being of any species of fish or wildlife or plants,” however, the Secretary may bypass those procedures and use provisions in the Act which empower him to issue regulations, including a listing, that take effect immediately upon publication in the Federal Register. Emergency regulations remain in force only for 240 days, although formal rulemaking procedures may be followed to enact normal regulations — including listings — within that time. See 16 U.S.C. § 1533(b)(7). To issue an emergency regulation, the Secretary need only publish in the Federal Register “detailed reasons why such regulation is necessary,” 16 U.S.C. § 1533(b)(7)(A), and notify the appropriate state agency in each state where the species is believed to reside, 16 U.S.C. § 1533(b)(7)(B).
In 1985, the Secretary determined that the desert tortoise warranted endangered species protection because of high mortality rates and habitat losses, see 50 Fed.Reg. 49868, 49869 (Dec. 5, 1985), but that such a listing was precluded by the Secretary’s limited resources and the need to list other, more severely imperiled, species — a situation contemplated by the Act. See 16 U.S.C. § 1533(b)(3)(B)(iii). The Secretary subsequently reaffirmed the desert tortoise’s “warranted, but precluded” status on two separate occasions. See 52 Fed. Reg. 24485 (July 1, 1987) and 53 Fed.Reg. 25511 (July 7, 1988). In 1988, Congress amended the Endangered Species Act, and directed the Secretary to “implement a system to monitor effectively the status of all [warranted but precluded] species ... and ... [to] make prompt use of the authority [to issue emergency regulations] to prevent a significant risk to the well being of any such species.” 16 U.S.C. § 1533(b)(3)(C)(iii).
Meanwhile, the desert tortoise population continued to decline. In May of 1989, several environmental groups petitioned the Fish and Wildlife Service of the United States Department of the Interior to list the desert tortoise as an endangered species, calling attention to a previously unknown danger to the tortoise — a fatal, apparently contagious, and incurable respiratory disease (“Respiratory Disease Syndrome”) raging within certain wild tortoise populations, primarily in California. On August 4, 1989, after study, the Secretary published an emergency regulation listing as endangered the desert tortoise population located north and west of the Colorado River (the “Mojave” population), including desert tortoises in parts of California, southern Nevada, southwestern Utah, and northwestern Arizona. See 54 Fed.Reg. 32326 (Aug. 4,1989).2 The Secretary found that, with the newly discovered threat of disease, the already at-risk desert tortoise faced cumulative dangers that compelled emergency listing. The Secretary pointed to the degradation of tortoise habitat resulting from increased cattle and sheep grazing, off-road vehicle recreation and land development — all of which destroy the forage, vegetative cover, and burrow sites upon which the tortoises depend for food and protection from predators. See id. at 32327. That degradation, combined with a burgeoning raven population, has resulted in a severe increase in raven predation, especially on young tortoises. See id. at 32327. In addition, the Secretary noted that many tortoises had been destroyed by hunting and crushed by recreational vehicles. See id. at 32327, 32329. Not only were these dangers previously found to be significant enough in themselves to warrant an endangered species listing — they apparently also exacerbate the tortoises’ susceptibility to the Respiratory Disease Syndrome. See id. at 32328.
In his listing, however, the Secretary excluded those tortoises located south and east of the Colorado River (the “Sonoran” population). Although he expressed concern about the few instances of a respirato*61ry disease (which was not shown to be Respiratory Disease Syndrome) within the Sonoran population, he explained that those tortoises occur in patchy or disjunct groups, thereby reducing the danger that the disease would spread throughout the population. See id. at 32327.
Appellants argue that the Secretary’s emergency listing, as it applies to Nevada, was arbitrary and capricious because: (1) it was based on inferior scientific evidence, (2) it did not adequately take into account Nevada’s programs for protecting the tortoise, (3) it will have no ameliorative effect on the respiratory disease that inspired it, (4) it is an unexplained departure from prior agency practice of using emergency regulations only to provide an immediate solution to a specific problem, and (5) it excluded the Sonoran population which has exhibited symptoms of the disease while including the tortoises of Nevada where, appellants claim, no disease has yet been detected. Appellants moved for a preliminary injunction against the implementation and enforcement of the rule. The district court concluded that the Secretary relied on inconclusive scientific data in “finding respiratory disease among wild tortoise populations in Nevada.” See City of Las Vegas v. Lujan, Memorandum Order, No. 89-2216 (D.D.C. August 24, 1989) at 5 (“Memorandum Order”). It further agreed with appellants that “listing the desert tortoise as endangered will have no specific effect on respiratory disease in any of the populations.” Id. Nevertheless, the district judge refused to grant the preliminary relief, essentially because he believed that appellants were unlikely to prevail on the merits of their challenge to the emergency rule. Although not entirely clear, it seems that the district judge believed that the promulgation of emergency listings is entirely within the Secretary’s discretion once he identifies an emergency affecting a portion of a species population — provided only that he offer the detailed reasons and notice to the states required by 16 U.S.C. § 1533(b)(7). We agree, for somewhat different reasons, that appellants have little chance of succeeding on the merits of them claim, and we therefore affirm.
II.
Ordinarily, our review of district court decisions to grant or deny preliminary relief is conducted under the extremely deferential clear error or abuse of discretion standard. See Friends For All Children v. Lockheed, 746 F.2d 816, 834-35 (D.C.Cir.1984). As we explained in Friends, that deference is justified by the latitude the district court properly enjoys in balancing the four factors that traditionally constitute the preliminary injunction calculus: the movant’s likelihood of success on the merits, the movant’s potential irreparable injury absent preliminary relief, the hardship a preliminary injunction will inflict on the non-moving parties, and the public interest. See Friends, 746 F.2d at 834-35 & n. 32; Washington Metropolitan Area Transit Comm’n v. Holiday Tours, Inc., 559 F.2d 841, 842 (D.C.Cir.1977); Virginia Petroleum Jobbers Ass’n v. FPC, 259 F.2d 921, 925 (D.C.Cir.1958). Furthermore, we owe special deference to the district court’s factfinding, which is always reviewed under a clearly erroneous standard, whether on appeal of a preliminary injunction motion or a final order. See Friends, 746 F.2d at 835 n. 32. On the other hand, we engage in “essentially de novo review” when the district court’s decision hinges on a question of law. See Friends, 746 F.2d 835 n. 32.
Here, we properly review the denial of the preliminary injunction de novo. The district judge did not make any factual determinations to which we owe deference since he was sitting in appellate review of agency action, and did not compile an independent record. Neither did the district judge explicitly undertake the traditional four-factor balancing that is normally associated with motions for preliminary injunctions. And to the extent that he implicitly balanced any equities, he found them all skewed toward the appellants.3 The dis*62trict judge denied the preliminary injunction because, and only because, he believed the Secretary was likely to succeed on the merits of his defense of the emergency listing. Our review of that conclusion is properly de novo since we owe no deference to the district court’s assessment of the legal merit of the Secretary’s position.
III.
At the threshold of our evaluation of appellants' likelihood of success on the merits of their challenge to the emergency listing is a determination of the appropriate standard to review the lawfulness of emergency regulations — as opposed to regulations that emerge from normal rulemaking. At first glance, the standard might appear to be the same since all action by the Secretary taken pursuant to the Endangered Species Act is subject to “arbitrary and capricious” style review under the Administrative Procedure Act. Nevertheless, what might constitute arbitrary and capricious action or an unacceptable explanation for a regulation in the normal course of events might well pass muster under the emergency provisions of 16 U.S.C. § 1533(b)(7).
Section 1533(b)(7) provides that an emergency regulation automatically expires after 240 days “unless, during such 240-day period, the rulemaking procedures which would apply to such regulation without regard to this paragraph are complied with.” The statute thus contemplates a somewhat less rigorous process of investigation and explanation for emergency regulations than for normal rulemaking, or else no subsequent rulemaking would be required. Moreover, the statute directs the Secretary to withdraw emergency regulations if, during that 240-day period, he determines that “substantial evidence does not exist,” to warrant the regulation. 16 U.S.C. § 1533(b)(7). This seems to suggest that whatever quantum of data the Secretary must possess to issue an emergency regulation, it need not rise to the level of “substantial evidence” — or at least that Congress contemplated that the Secretary would not inquire as thoroughly at the emergency listing stage. And with respect to emergency listings, Congress amended the Endangered Species Act in 1988, specifically directing the Secretary to “make prompt use of the authority under [16 U.S.C. § 1533(b)(7)] to prevent a significant risk to the well being of any ... species” that had previously been found to warrant an endangered listing (such as the desert tortoise). See 16 U.S.C. § 1533(b)(3)(C)(iii) (emphasis added). At least with respect to “warranted but precluded” species, Congress therefore indicated that the Secretary was to use his emergency powers less cautiously — in a sense to “shoot first and ask [all of the] questions later.” Our scrutiny of such emergency regulations is therefore less exacting on the Secretary than it would be if he enacted precisely the same regulation and gave the same explanation after normal rulemaking. Without expressing any opinion about whether appellants’ arguments would prevail if the Secretary had engaged in normal rulemaking, we conclude that they have not shown a likelihood of prevailing on the claim that the Secretary failed to satisfy the requirements of emergency rulemaking.
Appellants attack, for instance, the quality of the evidence relied upon by the Secretary when he supposedly determined that Respiratory Disease Syndrome exists among tortoises in Nevada.4 They argue, and the district court agreed, that the evi*63dence used was not the “best scientific and commercial data available” as required by 16 U.S.C. § 1533(b)(1)(A) for “determinations” made in the course of normal listings. At least for emergency listings (particularly of warranted but precluded species), however, this provision merely prohibits the Secretary from disregarding available scientific evidence that is in some way better than the evidence he relies on. Even if the available scientific and commercial data were quite inconclusive, he may— indeed .must — still rely on it at that stage. We think that is the only fair inference from the requirement in 16 U.S.C. § 1533(b)(7) (the emergency regulations provision) that, if the Secretary finds after publishing the emergency regulation, “on the basis of the best appropriate data available to him, that substantial evidence does not exist to warrant such regulation, he shall withdraw it.” Since there is no allegation that the Secretary disregarded scientifically superior evidence that was available to him at the time he published, he satisfied his duties under the 16 U.S.C. § 1533(b)(7).
In any event, whether or not the Secretary possessed evidence of the existence of the Respiratory Disease Syndrome — or any respiratory disease — within the borders of Nevada is not determinative as to the propriety of the emergency listing. Appellants do not challenge on appeal the Secretary’s scientific support for the proposition that:
“The potential exists for the Respiratory Disease Syndrome to reach epidemic proportions throughout the Mojave population. There appear to be no natural barriers that would prevent transfer of infectious agents from California subpopulations to Nevada, Utah, and Arizona subpopulations in the Mojave desert.” 54 Fed.Reg. 32326, 32328 (Aug. 4, 1989) (emphasis added).5 Given the unambiguous congressional direction to the Secretary that emergency listings of warranted but precluded species be issued prophylactically, the danger of the epidemic spreading to Nevada is sufficient justification for including Nevada’s tortoises in the emergency listing — he need not wait until the epidemic crosses the Califomia/Nevada border.
We similarly find no merit in appellants’ contention that the Secretary acted unlawfully by not adequately “taking into account” Nevada’s efforts to protect the tortoise. See 16 U.S.C. § 1533(b)(1)(A). It is true nothing in the explanation accompanying the emergency listing can be described fairly as “taking into account” Nevada’s regulatory efforts. Nevertheless, the record indicates that the Director of the Fish and Wildlife Service of the Department of the Interior consulted in detail and ultimately disagreed with the Director of the Nevada Department of Wildlife with regard to the desert tortoise. Whatever obligation the Secretary has under the Endangered Species Act and the APA explicitly to discuss the strengths and weaknesses of state regulatory mechanisms in the explanation accompanying a normal listing, the onus is lighter with respect to emergency listings which the Secretary may be obliged to issue even before he has access to all the relevant information.
Appellants also argue that the emergency listing runs afoul of the requirement that an agency decision must bear a “‘rational connection between the facts found and the choice made,’ ” Motor Vehicle Manufacturers Ass’n v. State Farm Mutual Automobile Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 2866, 77 L.Ed.2d 443 (1983) (quoting Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, *64168, 83 S.Ct. 239, 246, 9 L.Ed.2d 207 (1962)), because the protections which accompany a listing will not cure or otherwise alleviate the Respiratory Disease Syndrome that immediately prompted the Secretary’s action. We think it doubtful that this argument would have any force even if this were a normal listing because the logic of appellants’ position would foreclose an endangered listing even if 99% of all desert tortoises throughout their entire range (including in Nevada) had been reliably diagnosed to suffer from Respiratory Disease Syndrome, unless the listing would somehow abate the disease. Whether or not any such nexus must be established when the Secretary enacts a permanent listing through normal rulemaking, it is enough for our purposes to note that the Act does not require the Secretary to demonstrate that invoking his emergency power to list a species as endangered will stave off the demise of that species, let alone that it will address the specific emergency that led to the listing. Cf. State of Louisiana ex rel Guste v. Verity, 853 F.2d 322, 333 (5th Cir.1988) (“regulations aimed at preventing the taking of a protected species cannot be invalidated on the ground that the record fails to demonstrate that the regulatory effort will enhance the species’ chance of survival.”).
Appellants contend, nonetheless, that such a causal nexus is required for emergency listings, if not by the statute, then by virtue of the Secretary’s prior practice of issuing emergency regulations only when the use of emergency procedures (as opposed to normal rulemaking) would provide an immediate solution to a specific problem.6 In their view, the Secretary has impermissibly departed from prior practice and precedent without explanation. See Greater Boston Television Corp. v. FCC, 444 F.2d 841, 852 (D.C.Cir.1970), cert. denied, 403 U.S. 923, 91 S.Ct. 2229, 29 L.Ed.2d 701 (1971). But appellants point to no indication that the Secretary, as a matter of agency policy, limited reliance on emergency procedures to situations where it could be shown that an impending danger would be averted by the listing. That the Secretary has previously used the emergency power cautiously does not constitute an articulated policy that the Secretary will not also use the emergency power when its efficacy is less assured. Moreover, the prior uses to which appellants point largely precede the 1988 amendments to the Act in which, as we explained above, the Secretary was specifically directed to use the emergency power preemptively with regard to “warranted but precluded” species. A change in congressional instructions, of course, absolves the Secretary from explaining why his policy has changed, if indeed it has.
More troubling is appellants’ contention that the Secretary acted irrationally by including the Nevada portion of the Mojave population — where, we assume arguendo, no respiratory disease has been reliably established — within the emergency listing and excluding the Sonoran population — where there is evidence of respiratory disease. As we noted, the Secretary defends this distinction on the grounds that the disease — even if it is Respiratory Disease Syndrome — is unlikely to spread among Sonoran tortoises because they occur in patchy or disjunct groups. See 54 Fed.Reg. at 32327. Appellants claim that the distinction has no rational scientific basis and is actually the result of a bureaucratic dispute between two different sections of the Fish and Wildlife Service that have jurisdiction, respectively, over the Mojave and the Sonoran populations.
We need not settle this dispute, however, in order to affirm the district court’s denial of appellants’ preliminary injunction motion. Our preceding discussion demonstrates that the Secretary had ample basis for including the Nevada population in his emergency listing. The reasons offered for excluding the Sonoran population are not inconsistent with and therefore do not *65undermine the reasons given for including the Nevada population. If there is any problem with the line drawn by the Secretary (an issue on which we express no view), it could only lie in excluding the Sonoran population rather than in including the Nevada portion of the Mojave population.7
Appellants face a heavy burden in establishing that the Secretary acted irrationally by including Nevada but not including the Arizona Sonoran population in the listing. Since agencies have great discretion to treat a problem partially, we would not strike down the listing if it were a first step toward a complete solution,8 even if we thought it “should” have covered both the Mojave and Sonoran populations. See National Ass’n of Broadcasters v. FCC, 740 F.2d 1190, 1207 (D.C.Cir.1984). Even if this case were before us on direct appellate review of final agency action (rather than on review of a preliminary injunction motion) and we decided that the explanation given by the Secretary for the distinction was unacceptable, we would not be obliged to vacate the entire emergency rule. Instead, we would be free to remand the case to the Secretary for further proceedings while leaving the rule in force, if we believed that to be the more equitable course. See, e.g., Independent U.S. Tanker Owners Comm. v. Dole, 809 F.2d 847, 854 (D.C.Cir.), cert. denied sub nom. Atlantic Richfield Co. v. Independent U.S. Tanker Owners Comm., 484 U.S. 819, 108 S.Ct. 76, 98 L.Ed.2d 39 (1987). Here, where appellants seek appellate review of a denial of a preliminary injunction from a district court reviewing the Secretary of Interior’s emergency listing, they have not nearly satisfied their burden.
Under these circumstances, we do not believe that appellants would be likely to succeed on the merits of their challenge to the emergency listing — and certainly not in achieving the relief they seek. The order of the district court denying appellants motion for a preliminary injunction is therefore
Affirmed.
. 16 U.S.C. § 1533(a)(1) provides that:
The Secretary shall ... determine whether any species is an endangered species or a threatened species because of any of the following factors: (A) the present or threatened destruction, modification, or curtailment of es habitat or range; (B) overutilization for commercial, recreational, scientific, or educational purposes; (C) disease or predation; (D) the inadequacy of existing regulatory mechanisms; or (E) other natural or man-made factors affecting its continued existence.
. This population also includes tortoises located on the Beaver Dam Slope in Utah which were excluded from the emergency rule since they had already been listed as a "threatened" species — a status that provides them with similar protection. See 54 Fed.Reg. 32326 (Aug. 4, 1989).
. The district judge noted that "the [Secretary’s] emergency regulation is broader than necessary, and that the [Secretary] readily could have limited the geographic scope of its rule with no *62detriment to the tortoise____" Memorandum Order at 5. He further asserted that "the Court believes that it would have been significantly preferable for the [Secretary] to have proceeded with normal rulemaking as to those areas in which there are populations unaffected by respiratory disease____’’ Id. at 6.
. It is unclear to us whether the Secretary even made such a finding. In the explanation accompanying the emergency listing, the Secretary asserts merely that: "Interviews of personnel at veterinary hospitals in the Las Vegas, Nevada area by Service personnel have revealed that most cases of Respiratory Disease Syndrome are found in captive tortoises, but that wild tortoises have been brought in with symptoms of respiratory disease.” 54 Fed.Reg. 32326, 32328 (Aug. 4, 1989).
. Appellants did present expert testimony in the district court that physical and genetic barriers would prevent the spread of Respiratory Disease Syndrome from California to Nevada. And while the district judge stated that the Secretary did not rely on the best available scientific evidence in determining that the disease already existed in Nevada, the judge expressed no view about the proposition (or its supporting evidence) that the spread of the disease from California to Nevada is a real danger. On appeal, appellants do not specifically challenge as arbitrary and capricious or clearly erroneous the Secretary’s discrete determination that the disease may spread into Nevada.
. For example, the Secretary promulgated an emergency regulation in 1982 listing the Ash Meadows speckled dace and the Ash Meadows Amargosa pupfish as endangered species thereby halting imminent land development projects that threatened the entirety of the species’ limited habitat.
. We note that the listing is currently being challenged in district court precisely on the grounds that it was irrational to exclude the Sonoran population. See Environmental Defense Fund, et al. v. Manuel Lujan, Jr., Civil No. 89-2034 SSH (D.D.C.).
. In enacting the emergency listing, the Secretary represented that the respiratory disease among the Sonoran population "is currently being addressed by the [Fish and Wildlife] Service.” 54 Fed.Reg. at 32327.