dissenting:
The majority’s decision in this case is unfortunate. It scuttles a carefully conceived Fish and Wildlife Service (“FWS”) regulation and creates a split in the circuits on an important statutory question. Despite the significance of the issue presented in this case, the majority insists on granting rehearing and reversing its earlier opinion without the benefit of additional oral argument, and without the benefit of additional briefing tailored to the court’s concerns. What was rightly considered good law in the opinion in this case issued last year, published at 1 F.3d 1 (D.C.Cir.1993), is now “altered” on the basis of a confusing and misguided legal analysis that creates a needless conflict among the circuits. I dissent.
A. Standard of Review
The majority’s first, and biggest, mistake is to jettison the Chevron standard. Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). Only once does the majority mention Chevron^ conceding that it is supposed to govern our review of this case. But the majority concludes that the FWS’s regulation is infirm under Chevron because “the Service’s definition of ‘harm’ was neither clearly authorized by Congress nor a ‘reasonable interpretation’ of the statute.” Maj. op. at 1464.
But' the standard of review that we are bound to apply under Chevron actually reads as follows:
When a court reviews an agency’s construction of the statute which it administers, it is confronted with two questions. First, always, is the question whether Congress has directly spoken to the precise question at issue. If the intent of Congress is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress. If, however, the court determines Congress has not directly addressed the precise question at issue, the court does not simply impose its own construction on the statute, as would be necessary in the absence of an administrative interpretation. Rather, if the statute is silent or ambiguous with respect to the specific issue, the question for the court is whether the agency’s, answer is based on a permissible construction of the statute.
Id. at 842-43, 104 S.Ct. at 2781-82 (footnotes omitted).
Plainly, Chevron does not place the burden on the responsible agency to show that its interpretation is clearly authorized or reasonable. On the contrary, the burden is on the party seeking to overturn such an interpretation to show that Congress has clearly spoken to the contrary, or that the agency’s interpretation is un reasonable. The whole point of Chevron deference is that when Congress has not given a clear command, we presume that it has accorded discretion to the agency to clarify any ambiguities in the statute it administers. In requiring the agency to justify its regulation by reference to such a clear command, the majority confounds its role. Ties are supposed to go to the dealer under Chevron.
Along with (and perhaps because of) this backwards reading, the majority also fails to note which step of Chevron is dispositive of the case. If it is step one, the majority must show that Congress has “directly spoken to the precise question at issue”: whether the word “harm” includes “significant habitat modification [that] actually kills or injures wildlife.” Surely the statute is silent, or at *1474best ambiguous, on this question. The Endangered Species Act (“ESA”) nowhere defines “harm,” and nowhere specifically prohibits the extension of that term to habitat modification. Under step two, the only question is whether the FWS’s interpretation of the word “harm” constitutes a “permissible” reading of the ambiguous language. The question is not — however much the majority might like it to be — whether we think it constitutes the best reading. Under step two, “[t]he court need not conclude that the agency construction was the only one it permissibly could have adopted to uphold the construction, or even the reading the court would have reached if the question initially had arisen in a judicial proceeding.” Chevron, 467 U.S. at 843 n. 11, 104 S.Ct. at 2782 n. 11 (emphasis added).
B. Applying Chevron
Despite the command of Chevron, the majority substitutes its own favorite reading of the Endangered Species Act for that of the agency. But Chevron makes it irrelevant whether the majority’s preferred interpretation is better than the agency’s; the only question is the reasonableness of the agency’s interpretation. And a fair reading allows for no other conclusion than that the agency’s interpretation is reasonable.
The place to start is the statutory language. The ESA makes it unlawful to “take” a member of a species of fish or wildlife designated as “endangered” by the FWS. 16 U.S.C. § 1538(a)(1)(B). “Take” is a defined term, meaning “to harass, harm, pursue, hunt, shoot, wound, kill, trap, capture, or collect, or to attempt to engage in any such conduct.” 16 U.S.C. § 1532(19). The regulation at issue in this case purports to define the term “harm” as “an act which actually kills or injures wildlife. Such act may include significant habitat modification or degradation where it actually kills or injures wildlife....” 50 C.F.R. § 17.3. It is this definition that the majority finds unreasonable.
The majority relies primarily upon nosci-tur a sociis, a seldom-invoked principle of statutory construction. According to this canon, when a potentially broad word appears in a definition along with a list of narrow words, the broad word should be read narrowly to conform with its companions. This the majority purports to do here, finding that all the words in the definition of “take,” other than the crucial word “harm,” involve a “direct application of force.” Ipso facto, the word “harm” must also be read to require a direct application of force, and habitat modification is not sufficiently direct to fall within this definition of “harm.”
To this display of reasoning by the majority, one is tempted to respond, “Watch that first step; it’s a big one.” The definition of “take” does not conform to the paradigm of one broad word alongside many narrow ones, and thus the majority’s decision to apply noscitur a sociis is fatally flawed. Instead, “take” includes several words that might be read as broadly, or nearly as broadly, as “harm”: especially “harass,” “wound,” and “kill.” Indeed, as the regulation at issue defines “harm” as an act (including habitat modification) that “actually kills or injures wildlife,” the FWS might as easily have derived a proscription of habitat modification from the words “kill,” “wound,” and “harass,” as from the word “harm.”
In fact, the House Report specifically comments upon the breadth of the ESA’s prohibition of “harassment,” stating:
[“Take”] includes harassment, whether intentional or not. This would allow, for example, the Secretary to regulate or prohibit the activities of birdwatchers where the effect of those activities might disturb the birds and make it difficult for them to hatch or raise their young.
H.R.Rep. No. 93-412, 93rd Cong. 1st Sess. 11 (1973). Accordingly, FWS has defined the term “harass” nearly as broadly as the term “harm”: *147550 C.F.R. § 17.3. Appellants have not challenged this definition.
*1474Harass in the definition of “take” in the Act means an intentional or negligent act or omission which creates the likelihood of injury to wildlife by annoying it to such an extent as to significantly disrupt normal behavioral patterns which include, but are not limited to, breeding, feeding or sheltering.
*1475A more limited definition of “harass” was required (if at all) in United States v. Hayashi, 5 F.3d 1278 (9th Cir.1993), only because the word “harm” does not appear alongside “harass” in the Marine Mammal Protection Act; the only accompanying terms are “hunt,” “capture,” and “kill.” 16 U.S.C. § 1362(13). By contrast, the ESA also includes the more expansive “harm,” “wound,” and “pursue,” and is clearly meant to proscribe a broader range of activity. The 9th Circuit, the same court that decided Haya-shi has itself reached that conclusion, Palila v. Hawaii Dep’t of Land and Natural Resources, 852 F.2d 1106, 1107-09 (9th Cir.1988) (agency’s interpretation of “harm” to include significant habitat modification is consistent with the language, purpose, and legislative history of the ESA), and today’s contrary decision thus creates a foolish circuit conflict.
Noscitur a sociis means that a word is known by the company it keeps. Jarecki v. G.D. Searle & Co., 367 U.S. 303, 307, 81 S.Ct. 1579, 1582, 6 L.Ed.2d 859 (1961). In the definition of “take,” the term “harm” is accompanied by an assortment of words ranging from the precise and narrow “shoot” to the vague and expansive “harass.” Gratuitous references to our nation’s foreign policy aside, see Maj. Op. at 1464, “harm” is not a single elastic word among many ironclad ones but an ambiguous term surrounded by other ambiguous terms. Consequently, even if it is ever appropriate to measure an agency’s construction of a statute against a seldom-used and indeterminate principle of statutory construction, this is not the place for noscitur a sociis.
Equally inappropriate is Judge Sentelle’s use of the presumption against surplusage. There is no reasonable definition of the word “harm” (or, for that matter, the word “harass”) that would not render superfluous some of the other defined terms. For example, one cannot “kill” or “wound” an animal without also “harming” it, even under the narrowest conceivable interpretation of “harm.” Does that mean we must read “harm” out of the statute altogether? That would hardly be faithful to Congress’s intent, which was to define takings “in the broadest possible manner to include any conceivable way in which a person can ‘take’ or attempt to ‘take’ any fish or wildlife.” S.Rep. No. 93-307, 93rd Cong., 1st Sess. 7 (1973), 1973 U.S.C.C.A.N. pp. 2989, 2995. No, instead it means we must uphold the agency’s reasonable accommodation of all the statutory terms. Defining “harm” to include “significant habitat modification” renders no more terms superfluous than would a definition that did not include habitat modification but did include “direct” forms of killing and wounding. And indeed, the majority’s holding that “harm” cannot include indirect means of injuring wildlife may render “harm” itself superfluous, or nearly so, as “direct” means of injury are well covered by the other terms. The plain language of the statute, then, lends no support to the majority’s contention that the FWS’s interpretation of “harm” is impermissible.
The statute’s purpose and legislative history provide equally insufficient grounds for rejecting the agency’s interpretation. The majority, adopting the view of the appellants in this case, acknowledges that Congress intended to halt injurious habitat modification when it passed the ESA. It contends that Congress intended to combat this problem solely through § 1534’s provision for federal land acquisition, and not through § 1538’s prohibition of private takings.
According to the majority, the legislative history of the “take” provision establishes that Congress did not mean for that term to encompass habitat modification. It notes that the original bill that was referred to the Senate Committee on Commerce, S. 1983, defined “take” to include “destruction, modification, or curtailment of [an endangered species’) habitat or range.” S. 1983, 93rd Cong., 1st Sess. § 3(6) (1973). The bill reported out of committee, however, did not refer to habitat modification in the definition of “take.” This omission, the majority finds, evinces Congress’s intent not to include habitat modification within the scope of protected “takings.” The majority also points to several statements from the floor suggesting that some members of Congress may have want*1476ed land acquisition, not the prohibition of land uses, to be the ESA’s sole weapon against habitat modification on private lands.
I find the legislative history to be most ambiguous regarding whether Congress intended to include habitat modification within the meaning of “take.” It is true that the Senate Committee chose not to use the S. 1983 definition of “take,” which specifically encompassed habitat modification. Instead, the Committee adopted a definition from the other bill under consideration, S. 1592, which did not explicitly include habitat modification. But as the district court noted, there is no indication in the legislative history as to why the Committee selected one definition over the other. And, in any event, the crucial word “harm” was never voted on by the Committee but was added later on the floor of the Senate. It might well have been intended to cover the entire landscape originally contemplated by the S. 1983 definition. See 119 Cong.Rec. 25,683 (July 24, 1973) (Statement of Sen. Tunney) (“The amendments will help to achieve the purposes of the bill and will clarify some confusion caused by language remaining in the bill from earlier drafts or omitted from earlier drafts which went unnoticed during the final committee markup.”).
Most importantly for our Chevron inquiry, there is nothing to suggest that Congress chose the definition it did in order to exclude habitat modification. The Committee may have rejected the S. 1983 definition only because it apparently would have made habitat modification a per se violation of the ESA, as opposed to leaving such determinations to the discretion of the responsible agency whose judgment this court today casts aside. Surely there is nothing to indicate that the Committee intended to foreclose an administrative regulation prohibiting habitat modification — particularly a prohibition accompanied by limitations, such as those contained in the FWS regulation under review, requiring that there be actual injury or death to wildlife.
In any case, Congress manifested no clear intent to exclude habitat modification from the “take” definition. Indeed, the Senate Committee Report states that “ ‘Take’ is defined ... in the broadest possible manner to include every conceivable way in which a person can ‘take’ or attempt to ‘take’ any fish or wildlife.” S.Rep. No. 93-307, 93rd Cong., 1st Sess. 7 (1973), 1973 U.S.C.C.A.N. p. 2995. And, as I have noted above, the bill reported out of Committee did not even include the word “harm” in the definition of “take.” That was added later on the floor, in a “technical and clarifying amendment ].” 119 Cong.Rec. 25,682-83 (July 24, 1973) (Statement of Sen. Tunney). As the amendment added the word “harm,” while subtracting nothing, that amendment can only have broadened the definition from the bill reported out of Committee — “clarifying” that “take” should be defined “in the broadest possible manner.”
The contention that Congress intended land acquisition to be the exclusive instrument for curbing habitat modification on private lands is totally speculative. Nothing in the language of 16 U.S.C. § 1534 or in the legislative history proves this sweeping assertion. The only evidence the majority can garner in support of its proposition is a few isolated and ambiguous remarks by members of Congress on the floor. See 119 Cong.Rec. 25,669 (July 24, 1973) (Statement of Sen. Tunney); 119 Cong.Rec. 30,162 (Sept. 18, 1973) (Statement of Rep. Sullivan). The general rule is that “debates in Congress expressive of the views and motives of individual members are not a safe guide ... in ascertaining the meaning and purpose of the lawmaking body.” Duplex Printing Press Co. v. Deering, 254 U.S. 443, 474, 41 S.Ct. 172, 179, 65 L.Ed. 349 (1921). In any case, these statements do not establish that even the speakers themselves intended land acquisition to be the exclusive protective mechanism for habitats on private lands.
In sum, the majority has found nothing in the language, structure, purpose, or legislative history that unambiguously shows that “harm” does not encompass habitat modification. Under Chevron, that should dispose of the case: the FWS’s gap-filling measure is a permissible exercise of its discretion as delegated by Congress. Moreover, not only is the majority’s evidence ambiguous, but there is additional evidence supporting the agen*1477cy’s interpretation. I have already cited the Senate and House Committee Reports, which suggest that Congress envisioned a broad interpretation of “take,” even before the crucial word “harm” was added to the definition of that term. S.Rep. No. 93-307, 93rd Cong., 1st Sess. 7 (1973), 1973 U.S.C.C.A.N. p. 2995 (“ ‘Take’ is defined ... in the broadest possible manner.”); H.R.Rep. No. 93-412, 93rd Cong., 1st Sess. 15 (1973) (ESA “includes, in the broadest possible terms, restrictions on the taking, importation and exportation, and transportation of [endangered] species, as well as other specified acts”); id. at 11 (“Harass” includes activities of birdwatchers “where the effects of those activities might disturb the birds and make it difficult for them to hatch or raise their young.”). And I have adverted to the floor amendment that added the word “harm,” purportedly to “clarify” language that was “omitted” from the draft that emerged from Committee. 119 Cong.Rec. 25,683 (July 24, 1973) (Statement of Sen. Tunney).
In addition, the agency’s interpretation draws support from a subsequent amendment to the ESA. In 1982, Congress amended the ESA to include a provision authorizing the FWS to issue a permit allowing “any taking otherwise prohibited by section 1538(a)(1)(B) of this title if such taking is incidental to, and not the purpose of, the carrying out of an otherwise lawful activity.” 16 U.S.C. § 1539(a)(1)(B). By negative inference, this provision demonstrates that Congress thought at least some “incidental takings” must be prohibited by § 1538(a)(1)(B) in the first instance. The majority speculates on what Congress could have meant by “incidental takings,” but the evidence quite clearly suggests that Congress meant habitat modification. The House Report states,
This provision is modeled after a habitat conservation plan that has been developed by three Northern California cities, the County of San Mateo, and private landowners and developers to provide for the conservation of the habitat of three endangered species....
This provision will ... provide the institutional framework to permit cooperation between the public and private sectors in the interest of endangered species and habitat conservation.
H.R.Rep. No. 97-835, 97th Cong., 2d Sess. 30-31 (1982), 1982 U.S.C.C.A.N. pp. 2871, 2872.
Moreover, the 1982 amendments came after the Secretary promulgated the present definition of “harm” at issue in this case. But instead of using the amendments as an occasion to overrule the Secretary’s interpretation, Congress chose to allow the definition of “take” to stand, while amending another section of the statute — making clear in the process that it knew habitat modification could be (and was being) prohibited under the ESA. See Lindahl v. Office of Personnel Management, 470 U.S. 768, 782 n. 15, 105 S.Ct. 1620, 1629 n. 15, 84 L.Ed.2d 674 (1985) (“Congress is presumed to be aware of an administrative or judicial interpretation of a statute and to adopt that interpretation when it re-enacts a statute without change.”); Albemarle Paper Co. v. Moody, 422 U.S. 405, 414 n. 8, 95 S.Ct. 2362, 2370 n. 8, 45 L.Ed.2d 280 (1975) (same); Johnson v. Transportation Agency, 480 U.S. 616, 629 n. 7, 107 S.Ct. 1442, 1450 n. 7, 94 L.Ed.2d 615 (1987) (congressional inaction after interpretation may be probative of approval).
Judge Williams devotes a large portion of his majority opinion to a refutation of this argument. I presume the reason for this emphasis is that he concurred in the initial panel opinion in this case solely on the ground that the 1982 amendments “support the inference” that the prohibition of takings includes a prohibition of some habitat modification. 1 F.3d at 11 (Williams, J., concurring). Having placed all of his eggs in that basket, he understandably finds it necessary to explain his error at length now that he has changed his mind.
I agree that Judge Williams was wrong the first timé. He was wrong to rely solely on the 1982 amendments for his decision; I agree that they do not alone support its weight. They indicate that Congress in 1982 probably believed that habitat modification was properly covered by the prohibition on takings. Admittedly, the 1982 amendments prove little about Congress’s intent in 1973, *1478and had Congress in 1973 specifically stated that “take” does not include habitat modification the 1982 amendments would not save the FWS regulation. But Congress did no such thing in 1973; it was silent on the question. Consequently, the 1982 amendments do lend some weight to the reasonableness of the agency’s definition — if Congress in 1982 believed the definition was reasonable, and the agency believed it was reasonable, then Chevron demands that we uphold the regulation unless we find solid evidence to the contrary. No such evidence exists.
Thus the court today moves from wrong to more wrong in attempting to parse this statute. Having forsaken the 1982 amendments as dispositive evidence, no effort is made to determine whether the agency could reasonably have relied on such amendments as persuasive evidence supporting its interpretation. Instead, the agency is asked to prove that the best interpretation of “harm” encompasses habitat modification. Beginning from a wrong premise, applying a wrong standard, it is not surprising that the wrong result is achieved.
Overall, there is nothing in the ESA itself, or in its legislative history, that unambiguously demonstrates that the term “harm” in the definition of “take” does not encompass habitat modification. Indeed, there is evidence to the contrary. Chevron commands that we defer to an agency’s interpretation of a statute it is entrusted to administer, unless that interpretation is contrary to Congress’s unambiguous command or an unreasonable exercise of Congress’s vague or ambiguous delegation. The majority has not been so quick to ignore Chevron before. See, e.g., Railway Labor Executives’ Ass’n v. National Mediation Bd., 988 F.2d 133, 144-45 (D.C.Cir.) (Williams, J., dissenting) (“legislative silence [is] precisely the condition that under Chevron is understood to create a gap to be filled by the agency.”), vacated on motion for rehearing, 996 F.2d 1271 (D.C.Cir.1993). Neither should it ignore Chevron today.
Finally, the majority should be even more hesitant to cast aside the agency’s interpretation in light of the circuit split that this decision now creates. The Ninth Circuit determined, in Palila v. Hawaii Dep’t of Land and Natural Resources, 852 F.2d 1106 (9th Cir.1988), that the FWS’s “harm” definition was a permissible interpretation of the statute. “The Secretary’s inclusion of habitat destruction that could result in extinction follows the plain language of the statute because it serves the overall purpose of the Act, which is ‘to provide a means whereby the ecosystems upon which endangered species and threatened species depend may be conserved....’ 16 U.S.C. § 1531(b).” Id. at 1108. The purpose of the Endangered Species Act, lest we forget, is to protect endangered species. In today’s abandonment of our decision of less than a year ago, this court takes a large step backward from that purpose. The majority may believe it is making good policy — but that is not our job. Under Chevron, we may overturn an administrative regulation only if it contradicts the agency’s legislative mandate. Congress does not always speak as plainly as it might in designing administrative missions for the executive branch, so it is not always easy to decipher Congress’s marching orders to an agency. But at least we ought to try. I dissent.