TORRUELLA, Circuit Judge.

L.S. Starrett Company ("Starrett") appeals from two orders by the Federal Energy Regulatory Commission ("FERC" or "Commission"). The Commission concluded that Starrett would be required to seek licensing pursuant to Section 23(b) of the Federal Power Act ("FPA")1 if it proceeded with certain proposed changes to its Crescent Street Dam Project ("the Project"), a hydroelectric generating facility. Under Section 23(b), Starrett must seek licensing if (1) its facility is located on a stream over which Congress has Commerce Clause jurisdiction, (2) its proposed changes constitute "post-1935 construction" within the meaning of the FPA, and (3) the proposed modifications will affect the interests of interstate or foreign commerce. See 16 U.S.C. § 817(1). For the reasons below, we conclude that we have no choice2 but to affirm.

I. Background

The Project is located on the non-navigable Millers River in Athol, Massachusetts, on property that belongs to Starrett. It consists of (1) an 87-acre-foot reservoir; (2) a 20-foot-high, 127-foot-long concrete gravity dam; (3) two powerhouses, one at each end of the dam; and (4) various appurtenant facilities. The turbine generator in the powerhouse on the right side of the dam ("the right-side generator") currently has an installed capacity3 of 250 kW and an actual capacity of 80 kW. The turbine generator in the powerhouse on the dam's left side ("the left-side generator") had an installed and actual capacity of 112 kW until 2006, when it failed. Prior to the failure of the left-side generator, the combined installed capacity for the Project was 362 kW. This installed capacity was memorialized in a 1992 FERC order, which concluded that the Project did not require FERC licensing because there had been no post-1935 construction. See L.S. Starrett Co., 61 FERC ¶ 62,200 (1992) ("Starrett I"). The facility, however, could only actually produce 192 kW of electricity, even prior to the failure of the left-side generator, because of the physical limitations of the site.

After the left-side generator failed, Starrett began to investigate its options for replacement or repair. In early 2007, Starrett retained GZA GeoEnvironmental, Inc. ("GZA"), which prepared a feasibility study that examined the financial costs and benefits of repairing the left-side generator. The study concluded that it would be cost effective to use hydropower generated by a new left-side turbine generator ("the new left-side generator"). The new left-side generator's installed and actual capacity would be 198 kW. Installing the new left-side generator would increase the Project's combined installed capacity approximately 24%, to 448 kW, and its total actual capacity approximately 45%, to 278 kW. The total actual capacity of the Project, however, would remain less than the previous total installed capacity (i.e., 362 kW, the capacity documented in Starrett I), a point that Starrett now emphasizes.

In September 2008, believing that it did not require FERC licensing in order to proceed with its proposed changes,4 Starrett ordered a new cross-flow turbine generator5 and began the preparatory work for its installation inside the left powerhouse. According to Starrett, replacing the left-side generator required the following: (1) lowering the floor inside the powerhouse by approximately 5 feet;6 (2)(a) improving the plunge pool7 within the building footprint by mounting the draft tube8 below the turbine generator, and (b) widening the outlet portal from under the powerhouse to the river from approximately 4 to 17 feet, both in order to improve outlet hydraulics and project efficiency; (3) excavating approximately 10 cubic yards of bedrock from the bottom of the Millers River, again in order to improve outlet hydraulics and project efficiency; and (4) installing a transition piece to connect the existing penstock9 to the new turbine.

In March 2009, as Starrett was working to replace its left-side generator, the U.S. Fish and Wildlife Service ("USF & WS") wrote to the Commission to request that the Commission investigate the work occurring at the Project. The USF & WS had become aware that Starrett planned to install higher capacity machinery, and was concerned that the increased capacity at the Project would negatively impact migratory fish.

On May 4, 2009, the Commission notified Starrett that its proposed work would increase the capacity of the Project and would be considered post-1935 construction, thus triggering the Commission's licensing jurisdiction. The Commission asked Starrett to submit various details about its dam and the proposed changes. Starrett provided the requested details but maintained that its proposed work would not lead to an increase in capacity above the 362 kW total memorialized in Starrett I because only the installed capacity, not the actual capacity, would be over 362 kW. After reviewing these materials, the Commission issued an order finding that licensing of the Project was required. See L.S. Starrett Co., 129 FERC ¶ 62,053 (2009) ("Starrett II"). Following Starrett's request for rehearing, the Commission issued an order denying rehearing. See L.S. Starrett Co., 130 FERC ¶ 61,112 (2010) ("Starrett III").

II. Discussion

We are now required to review the Commission's determination that the Project fell within its jurisdiction under Section 23(b) of the FPA. Under that section, a hydroelectric project "without a valid pre-1920 permit" is subject to the Commission's licensing jurisdiction if it

Starrett II, 129 FERC ¶ 62,053, at 64,160; see also 16 U.S.C. § 817(1). The Commission concluded that Starrett's dam was subject to licensing under the fourth criterion. We lay out the governing standard of review, and then address each of the three prongs of the fourth criterion.

A. Standard of Review

Reviewing the Commission's orders under the Administrative Procedures Act, we "must reverse an agency action that is `arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.'" Knott v. FERC, 386 F.3d 368, 372 (1st Cir.2004) (quoting Wis. Valley Improvement Co. v. FERC, 236 F.3d 738, 742 (D.C.Cir.2001)); see also 5 U.S.C. § 706.

"We review FERC's findings of fact for `substantial evidence,' and if so supported, such findings are conclusive." Knott, 386 F.3d at 371 (quoting Thomas Hodgson & Sons v. FERC, 49 F.3d 822, 825 (1st Cir. 1995)) (internal quotation marks omitted). "We `defer to the agency's expertise . . . so long as its decision is supported by "substantial evidence" in the record and reached by "reasoned decisionmaking," including an examination of the relevant data and a reasoned explanation supported by a stated connection between the facts found and the choice made.'" Id. (quoting Ne. Utils. Serv. Co. v. FERC, 993 F.2d 937, 944 (1st Cir.1993) (citation omitted)).

"`Pure' legal errors require no deference to agency expertise, and are reviewed de novo." Id. at 372 (quoting Ne. Utils. Serv. Co., 993 F.2d at 944) (internal quotation marks omitted). "Questions involving an interpretation of the FPA involve a de novo determination by the court of congressional intent; if that intent is ambiguous, FERC's conclusion will only be rejected if it is unreasonable." Id. (quoting Ne. Utils. Serv. Co., 993 F.2d at 944) (internal quotation marks omitted). When determining congressional intent, courts must first ask whether Congress has "directly addressed the precise question at issue." Chevron, U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 843, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984); see also Massachusetts v. Sebelius, 638 F.3d 24, 30 (1st Cir.2011). If Congress has indeed addressed the "precise question at issue" and "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." Chevron, 467 U.S. at 842-43, 104 S.Ct. 2778.

B. A "Commerce Clause Stream"

Section 23(b) requires us to ask whether L.S. Starrett's dam is on "a stream over which Congress has Commerce Clause jurisdiction," see 16 U.S.C. § 817(1), i.e., a "Commerce Clause stream." The Commission concluded, and L.S. Starrett does not dispute on appeal, that the Millers River is a "Commerce Clause stream." See Starrett III, 130 FERC ¶ 61,112, at 61,521 n. 6. We agree with the Commission's assessment because "the headwaters and tributaries of navigable waters are [C]ommerce [C]lause streams," id. (citing Fed. Power Comm'n v. Union Elec. Co., 381 U.S. 90, 94-96, 85 S.Ct. 1253, 14 L.Ed.2d 239 (1965)), and the Millers River is a tributary of the Connecticut River, which is navigable, see Starrett II, 129 FERC ¶ 62,053, at 64,161 n. 6.

C. Post-1935 "Construction"

In 1935, Congress amended the FPA "to require that persons `intending to construct a dam or other project works' on nonnavigable streams obtain a license." Thomas Hodgson, 49 F.3d 822, 826 (second emphasis added) (quoting 16 U.S.C. § 817(1)); see also Public Utility Act of 1935, ch. 687, sec. 210, § 23(b), 49 Stat. 803, 846 (1935). Thus, if post-1935 work constitutes "construction" within the meaning of the FPA and the other two prongs of Section 23(b) are met, a facility will be subject to the Commission's licensing jurisdiction.

The Commission argues that, because Congress has not spoken on the precise question at issue here—which it frames as "what constitutes `construction'"10—we need only determine if its interpretation of Section 23(b) was unreasonable. See Section II.A., supra; Chevron, 467 U.S. at 844, 104 S.Ct. 2778 (explaining that when Congress implicitly delegates a question to an administrative agency, "a court may not substitute its own construction of a statutory provision for a reasonable interpretation made by the administrator of an agency"). The Commission contends that it was reasonable for it to determine that Starrett's proposed work would constitute post-1935 construction because the work would increase the Project's installed capacity. In addition, it notes that the Commission also found that the installation of the new generator would increase actual capacity, another indication that the proposed work constituted post-1935 construction. Alternatively, the Commission argues that it was reasonable for it to conclude that the proposal would result in post-1935 construction because it would increase the Project's "head."11

Starrett, on the other hand, argues that Congress has addressed the precise question at issue here by limiting the Commission's jurisdiction to post-1935 "construction." Starrett contends that because the proposed work was merely a repair, and would not increase actual capacity beyond the 1992 installed capacity, it was not post-1935 construction.

1. Standard of Review

We must first determine whether Congress unambiguously expressed an intent about the precise question at issue here. If we conclude that Congress did unambiguously express such an intent, our analysis ends there. If we conclude that Congress did not unambiguously express an intent on the precise question here, we must analyze whether the Commission's conclusion that the work here constituted jurisdictional construction was unreasonable.

"In determining congressional intent, we employ the traditional tools of statutory construction, including a consideration of the language, structure, purpose, and history of the statute." In re Hill, 562 F.3d 29, 34 (1st Cir.2009) (quoting McKenna v. First Horizon Home Loan Corp., 475 F.3d 418, 423 (1st Cir. 2007)) (internal quotation marks omitted). Our research has not uncovered, and the parties do not call to our attention, any legislative history that sheds light on where Congress would draw the line between jurisdictional construction and other work. Where "[t]here is no legislative history that illuminates the purpose" of a particular statutory term, we "are left with language, structure, and evident purpose." Id.; see also Robinson v. Shell Oil Co., 519 U.S. 337, 341, 117 S.Ct. 843, 136 L.Ed.2d 808 (1997) ("The plainness or ambiguity of statutory language is determined by reference to the language itself, the specific context in which that language is used, and the broader context of the statute as a whole."); Sebelius, 638 F.3d at 31 ("To determine `whether a statute exhibits Chevron-type ambiguity . . . courts look at both the most natural reading of the language and the consistency of the "interpretive clues" Congress provided.'" (quoting Succar v. Ashcroft, 394 F.3d 8, 22 (1st Cir.2005) (quoting Gen. Dynamics Land Sys., Inc. v. Cline, 540 U.S. 581, 586, 124 S.Ct. 1236, 157 L.Ed.2d 1094 (2004)))).

a. Language of the Statute

"We begin with the actual language of the statute, and ask whether . . . ["construction"] has a `plain and unambiguous meaning with regard to the particular dispute in [this] case.'" Perez-Olivo v. Chavez, 394 F.3d 45, 49 (1st Cir.2005) (second alteration in original) (quoting Duckworth v. Pratt & Whitney, Inc., 152 F.3d 1, 5 (1st Cir.1998) (quoting Robinson, 519 U.S. at 340, 117 S.Ct. 843)). When Congress chooses "not to define [a] phrase . . . in the statute itself, we can look to the dictionary for clarification of the plain meaning of the words selected by Congress." Id.

Black's Law Dictionary defines "construction" as "[t]he act of building by combining or arranging parts or elements." Black's Law Dictionary 355 (9th ed.2009). Another dictionary defines "construction" as "the act of putting parts together to form a complete integrated object," and the verb "construct" as "to form, make, or create by combining parts or elements." Webster's Third New International Dictionary (Philip Babcock Gove et al. eds., 1971). These definitions do not suggest that "construction" has "a plain and unambiguous meaning with regard to the particular dispute" here. Duckworth, 152 F.3d at 5 (quoting Robinson, 519 U.S. at 340, 117 S.Ct. 843). Rather, "construction" "is a chameleon, capable of taking on different meanings, and shades of meaning, depending on the subject matter and the circumstances of each particular usage." Strickland v. Comm'r, Me. Dep't of Human Servs., 48 F.3d 12, 19 (1st Cir.1995).

b. Statute as a Whole

If we conclude that the "plain language of the statute, standing alone, is ambiguous," the next step is to "ask whether this ambiguity can be resolved by looking to the `specific context in which [the] language is used, and the broader context of the statute as a whole.'" Perez-Olivo, 394 F.3d at 49 (quoting Robinson, 519 U.S. at 341, 117 S.Ct. 843) (alteration in original). The parties have not called to our attention, and we have not found in our review of the FPA, any clues about the meaning of the word "construction" as it applies here. Therefore, we move on to the second stage of the Chevron analysis.

2. Reasonableness

The Commission argues that its conclusion here—i.e., that Starrett's proposed work would constitute jurisdictional construction—was reasonable because the proposed changes involved (1) an increase in installed capacity and (2) an increase in head. Responding to the Commission's argument about increased capacity, Starrett contends that the Project's new actual capacity (278 kW, up from 192 kW) would remain below the 1992 installed capacity (362 kW), and thus the Commission should not have exercised its jurisdiction. We conclude that the Commission's determination was reasonable because there is no doubt that, under Starrett's plan, there would be an increase in capacity no matter how the capacity was measured; both the actual and the installed capacities would be greater than their respective 1992 values.12 For this reason, we need not analyze the head issue or resolve any of the factual disputes related to that issue.13

Given the state of the law, we must conclude that the Commission's interpretation of "construction" as including the work here was reasonable. In 1965, the Supreme Court explained that

Union Elec. Co., 381 U.S. at 98, 85 S.Ct. 1253. In 1986, Congress amended the FPA and made clear that when the Commission decides whether to grant a license under Section 23(b) and various other sections of the FPA, it should "give equal consideration to" (1) "the power and development purposes for which licenses are issued" and (2) "the purposes of energy conservation, the protection, mitigation of damage to, and enhancement of, fish and wildlife (including related spawning grounds and habitat), the protection of recreational opportunities, and the preservation of other aspects of environmental quality." 16 U.S.C. § 797(e); Electric Consumers Protection Act of 1986, Pub.L. No. 99-495, 100 Stat. 1243. We cannot say, as a matter of law, that it was unreasonable for the Commission to conclude that in order to ensure that the Nation's waterways be used in a "harmonious" fashion, and to ensure, among other things, that fish and wildlife were protected, it could interpret "construction" as including all increases in capacity.

The cases that Starrett cites to support its position are distinguishable because they did not involve increases in capacity. In Thomas Hodgson, the court relied in part upon the fact that there was no increase in capacity when concluding that no post-1935 construction had occurred. See Thomas Hodgson, 49 F.3d at 828 (no post-1935 construction where dam owners restarted operation of inactive dam after twelve years but there was "no project enlargement . . . in capacity, diversion, or physical plant" (quoting Puget Sound Power & Light Co. v. Fed. Power Comm'n, 557 F.2d 1311, 1316 (9th Cir.1977)) (internal quotation marks omitted)). The same can be said of Puget Sound. See 557 F.2d at 1316 (noting that the work there "merely restored the . . . project to its original specifications and configuration" and that there was no increase in the project's electrical generating capacity). Similarly, in Aquenergy Systems, Inc. v. Federal Energy Regulatory Commission, 857 F.2d 227 (4th Cir.1988), although the court ruled on another ground that the dam was subject to the Commission's licensing jurisdiction, it acknowledged that where a new project was "carefully planned . . . to meet the specifications of the original project" and where neither "designed capacity," head, nor the amount of electricity generated was to increase, such work would not ordinarily constitute jurisdictional construction under Section 23(b). Id. at 229-30.

In short, we conclude that the Commission's determination that Starrett's facility met the second required prong was not unreasonable.

D. The Effect on Interstate Commerce

Before allowing the FERC to exert its licensing jurisdiction over projects involving post-1935 construction, Section 23(b) requires the Commission to "find that the interests of interstate or foreign commerce would be affected by [the] proposed construction." 16 U.S.C. § 817(1). The Commission argues that Starrett's construction meets the interstate commerce requirement because its dam is a member of a class of small hydroelectric projects that collectively have a substantial impact on interstate commerce because they produce power that would otherwise have to be produced elsewhere on the interstate grid. Starrett responds that (1) it is improper for the Commission to rely on this "cumulative effect" theory because it leaves the Commission's Commerce Clause jurisdiction without boundary; and (2) in any case, the Commission has not shown that Starrett's facility belongs to a class of small hydroelectric projects that collectively affect interstate commerce.

We first address whether it was unreasonable for the Commission to consider the cumulative effect on commerce of many small hydroelectric facilities, and then address whether there was substantial evidence that Starrett's dam, in conjunction with others, actually has a significant impact on interstate commerce. See Habersham Mills v. FERC, 976 F.2d 1381, 1384-85 (11th Cir.1992) (first addressing whether the FERC "misappl[ied] the [FPA] by considering the cumulative effect of a class of small hydroelectric projects that include[d]" the two projects at issue, and then addressing whether there was substantial evidence of an effect on interstate commerce); City of Centralia v. FERC, 661 F.2d 787, 791-93 (9th Cir.1981) (first explaining that if "a local activity belongs to a class of activities having a cumulative effect on interstate commerce, it may fall within the commerce power," and then analyzing whether there was substantial evidence that the hydropower facility at issue, either alone or in conjunction with other facilities, had a real and substantial effect on interstate commerce).

1. The "Cumulative Effect" Theory

The Supreme Court has noted that the language of the FPA "strongly implies that Congress drew upon its full authority under the Commerce Clause" in enacting the statute. Union Elec. Co., 381 U.S. at 96, 85 S.Ct. 1253. "Full authority under the Commerce Clause includes the power to reach a local activity whose effect on commerce, `taken together with that of many others similarly situated, is far from trivial.'" Habersham, 976 F.2d at 1384 (quoting Wickard v. Filburn, 317 U.S. 111, 128, 63 S.Ct. 82, 87 L.Ed. 122 (1942)); see also Gonzales v. Raich, 545 U.S. 1, 17, 125 S.Ct. 2195, 162 L.Ed.2d 1 (2005) (noting that Supreme Court "case law firmly establishes Congress' power to regulate purely local activities that are part of an economic `class of activities' that have a substantial effect on interstate commerce"). Assuming there was substantial evidence supporting the Commission's factual findings, it would not be unreasonable for the Commission to regulate Starrett's dam because "a small hydroelectric project that affects commerce only slightly" can "still be subject to congressional regulation if it is part of a class with a significant cumulative effect." Habersham, 976 F.2d at 1384. We thus turn to the second part of our analysis and ask whether the Commission's conclusion that Starrett's dam is part of a class of projects that, in the aggregate, have the required effect on interstate commerce "is supported by substantial evidence." City of Centralia, 661 F.2d at 792.

2. Substantial Evidence

Starrett contends that its situation is comparable to the one in City of Centralia, where the Ninth Circuit concluded that the record failed to support the Commission's conclusion that a hydroelectric project either (1) itself had a substantial effect on commerce, 661 F.2d at 792; or (2) was part of a class of projects that, cumulatively, had a substantial effect on interstate commerce, id. at 793, We agree with the Commission that this case is more comparable to Habersham, where the Eleventh Circuit concluded that two small dams did meet the interstate commerce requirement because the Commission presented evidence that (1) by supplying power to a factory, the two hydroelectric projects "effectively displace[d] electricity that the factory otherwise would draw from the interstate grid," 976 F.2d at 1384; and (2) the Commission referred to two FERC reports that "indicate[d] that the small [hydroelectric] projects [around the nation] collectively account for a substantial portion of the nation's hydroelectric generating capacity," id. at 1385.14 Here, the Commission (1) explained that Starrett's dam produces power that Starrett would otherwise receive from the interstate grid, a point supported by the record; and (2) cited Habersham—which in turn, as noted above, cited two FERC studies—to support the proposition that small hydroelectric projects that displace power from the national grid can have a significant cumulative effect on interstate commerce. Starrett III, 130 FERC ¶ 61,112, at 61,522-61,523. We believe that there was substantial evidence to support the factual findings underlying the Commission's interstate commerce conclusion.

III. Conclusion

For the reasons stated, we affirm.15

Affirmed.

STAHL, Circuit Judge, concurring, joined by TORRUELLA, Circuit Judge.

I join this opinion with great reluctance. I do so because Chevron deference requires the result reached here, not that the result makes economic or realistic sense.

Here, we have the last full-line precision tool company producing its product within the United States. Although Starrett has several manufacturing locations worldwide, the Athol location produces most of the precision tools and has remained the company's headquarters since its founding in 1880. Starrett is the largest employer in the greater Athol area, and its payroll typically contributes over $2 million per month to the economy.

In order to remain competitive in the global marketplace, Starrett has aggressively sought to lower its cost structures and has instituted many energy conservation measures, which have both saved operating costs and reduced the company's carbon footprint. One of these measures included the replacement of the failed left turbine generator with a new, energy-efficient generator, the source of controversy in this case.

Innovations like those taken by Starrett are a necessary concomitant if we are to reinvigorate the nation's manufacturing base. Our decision today, however, may well mean that this company loses the economic advantage it would have from its low-cost, nonpolluting power structure. Cost-saving measures like those instituted by Starrett are particularly key for companies based in high energy cost states, like Massachusetts, and may well make the difference in keeping the plant open, providing good paying jobs, and maintaining an essential business such as this in our country. Indeed, machine tools are the lifeblood of industry, and when we have lost all of our domestic capacity, we become less secure and less able to compete. It is said by some that American industry has died from a thousand cuts, and many contend that over-regulation bears a share of the responsibility.

Further, it is unfortunate that a small power producer like the Starrett facility falls within the ambit of the Commission's jurisdiction because it is located on a non-navigable stream that is a tributary to a navigable water and affects interstate commerce through its connection to the interstate grid. Although I acknowledge that Wickard v. Filburn, 317 U.S. 111, 63 S.Ct. 82, 87 L.Ed. 122 (1942), and its progeny give the Commission power to reach purely local activity, the result strikes me as ironic. In Wickard, the government was confronted with a surplus of wheat, and it regulated production to avoid dramatically low wheat prices around the country. The market at issue here, however, proves just the opposite. Today, rising energy prices and a diminishing supply of resources pose a real challenge, and our national and state governments are doing all that they can to promote energy efficiency in order to lower energy costs. It would seem that Starrett's Project is a prime example of efficient usage through a nonpolluting power source and is one that we should be encouraging, not stifling.

Perhaps a better argument not advanced by Starrett would have been that, although Chevron applies, the Commission's definition of post-1935 construction was unreasonable in view of the realities presented by this project. Defining construction to include any increase in capacity still less than that originally authorized, without a de minimis exception and without consideration of a project's increased efficiency and economic impact, strikes me as troubling. But Starrett did not make this point, nor was there evidence of the costs it would incur in seeking the Commission's licensing and whether those costs and the necessary delay would take away from the project's economic advantages. We must deal with the record we have.