Most Americans picture the Clean Air Act as a law that walks up to a factory and hands its owner a limit on what the smokestack may release. That picture fails the One Test that governs this article, and the test is simple: a reader who finishes this guide should be able to state, in one sentence, the design choice that makes this statute unlike every other regulatory law in this series, and should be able to take any question about air regulation and place it in the right section of the law. For six widespread pollutants, the Clean Air Act does almost the reverse of the pictured scene. It does not begin with the factory at all. It begins with the atmosphere people breathe, asks what condition of that atmosphere human health requires, writes that condition down as a legal target, and then orders the technology, the state governments, and the permits to catch up.

The namable claim of this article is the health-first inversion. Almost every other regulatory statute in this series asks what industry can achieve and sets the requirement where the achievable answer lands. The Clean Air Act asks what human health requires and orders achievement to follow. Congress did not ask how clean the air could be made at an acceptable price and then enshrine the answer. It asked how clean the air must be for people, including children, the elderly, and people with asthma, to breathe it without harm, commanded the federal agency to write that answer into law with a margin of safety, and only then turned to the questions of who would cut emissions, by what means, and on what schedule. The inversion is not a slogan. It is a sequence written into the structure of Title I, and once a reader sees the sequence, the whole statute stops looking like a pile of programs and starts looking like a machine with an order of operations. The inversion buys something no technology-based regime can promise: a definition of acceptable air written by medicine rather than by economics. What it costs is perpetual argument about implementation, because a target set without regard to price must still be met with money, machinery, and political will. That argument is not a defect in the statute; it is the statute working as designed. Congress moved the fight from the question nobody should have to relitigate, whether the air should be safe to breathe, to the questions a democracy can legitimately debate: who pays, how fast, and by what means. Every later battle in this article’s machines, over state plans, technology floors, and permits, is a battle inside that settlement, not about it.

The statute as it operates is the Clean Air Amendments of 1970, Public Law 91-604, 84 Stat. 1676, signed by President Richard Nixon on December 31, 1970, amending the Clean Air Act of 1963 (Public Law 88-206, signed by President Lyndon B. Johnson on December 17, 1963, weeks after the assassination of President Kennedy, who had supported the bill). Congress rewrote the law again in the Clean Air Act Amendments of 1977, Public Law 95-95, and the Clean Air Act Amendments of 1990, Public Law 101-549, and the whole structure is codified at 42 U.S.C. sections 7401 and following. The 1970 rewrite was so complete that lawyers treat the modern act as a 1970 statute wearing a 1963 name.

The guide that follows walks through six machines that convert the health target into enforceable duties, and each machine answers a different question. The ambient standards machine, built from sections 108 and 109, turns medical and scientific evidence into the national ambient air quality standards, the legal description of air clean enough to breathe. The state planning machine, built from section 110, hands every state the job of writing an implementation plan that reaches and holds those levels within its borders. The new-source machine, built from section 111 and the preconstruction review programs, imposes technology floors on new and heavily modified facilities so economic growth does not erase the gains the plans produce. The toxics machine, built from section 112, sets emission limits for hazardous air pollutants substance by substance. The mobile-source machine, built from Title II, governs cars, trucks, engines, and fuels, the one domain where Washington regulates the polluting product directly and largely preempts the states from doing so. The permit machine, built from Title V, gathers every applicable duty for each major source into a single operating permit so the entire system can be read, monitored, and enforced in one document.

A word on how to use this guide. Each machine below is presented the same way: the statutory provisions that build it, the job it does in the sequence, and the questions it answers. When a new air-quality question arises, the reader’s first move should be classification, not research: is this about what the air must be, who must cut what, what a new facility must install, which chemicals face direct limits, what vehicles and fuels must meet, or where the resulting duties are collected? Six questions, six machines, and the statute’s table of contents maps onto them with surprising fidelity. That mapping is the practical payoff of the One Test, and it is worth more than any memorized provision.

The complication this article must dissolve is the belief that the Clean Air Act regulates polluters directly. For the six criteria pollutants, it mostly does not. It sets ambient targets and directs the states to write the plans that regulate sources. A factory owner hunting through the federal statute for the rule that governs a particular plant will mostly find the target rather than the rule, because the rule lives in the state implementation plan. Direct federal commands on sources appear in the later machines, and the skill this article teaches is knowing which machine a given question belongs to. Ask what the air must be like, and the answer sits in the ambient standards. Ask who must cut what, and the answer sits in a state plan. Ask what a new plant must install, and the answer sits in the new-source programs. The statute’s architecture, not its slogan, decides who has to do what, and that architecture is the thread that runs through every article in this cluster.

The Health-First Inversion: The Clean Air Act’s Defining Design Choice

To feel the weight of the inversion, consider how a technology-based statute normally works, described here in generic terms rather than through any one law’s provisions. The usual sequence runs like this: survey the control technologies that industry has demonstrated, weigh what they cost and what they achieve, and set the legal limit at the point where the achievable answer lands. Cost and feasibility sit at the table from the first minute, because the limit is defined by what can be done. This is a reasonable way to write law, and Congress has used it in many places. It has one structural consequence worth noticing: the protected resource gets whatever level of protection the available technology happens to make affordable, and if the technology is thin or the price is high, the protection is thin too.

Congress did not choose the inversion in a vacuum. The 1963 act had relied on research grants, technical assistance, and a cumbersome conference procedure in which federal officials could cajole but rarely compel action against interstate pollution, and by the end of the decade the approach had produced studies rather than results. The 1970 amendments replaced persuasion with architecture: a federal health target with a deadline, state plans to meet it, and federal consequences for plans that failed. The inversion was thus born of a specific diagnosis, that leaving both the goal and the means to diffuse goodwill had failed, and its remedy was to nationalize the goal while keeping the means local.

The Clean Air Act inverts the sequence for its core program. Sections 108 and 109 order the Administrator of the Environmental Protection Agency to list widespread pollutants that endanger public health or welfare, to assemble the scientific evidence about what concentrations harm people, and then to set the national ambient air quality standards at the levels requisite to protect public health, with an adequate margin of safety. The word requisite does the heavy lifting: the standard must be sufficient for the health purpose, neither more nor less than what the purpose demands. Cost of compliance does not enter the calculation at this stage, because the calculation is not about compliance at all. It is about the human body and what the human body requires. The Supreme Court confirmed the exclusion in Whitman v. American Trucking Associations in 2001, holding that the agency may not consider the economic cost of implementation when it sets these standards. Whether a factory can afford the cuts, and whether the technology exists to make them, are questions for later machines.

Cost’s legitimate reentry points are written into the statute with care. Section 111 directs the agency to set standards of performance for new sources based on the best system of emission reduction that has been adequately demonstrated, taking into account the cost of achieving the reduction and any non-air-quality health and environmental impacts. Section 112’s technology standards for hazardous pollutants begin from the maximum achievable control technology already demonstrated by the best-performing sources, then allow a second look in which cost, feasibility, and health effects inform how far beyond that floor to go. Title II’s vehicle and fuel standards weigh emission cuts against cost, lead time, safety, and energy considerations. In each of these later machines, feasibility and price sit at the table from the first minute, exactly as they do in an ordinary technology-based statute. The inversion is therefore not a war on economics but a sequencing rule: economics may shape the means, never the health target.

The word requisite carries the whole doctrine. Courts have read it to mean sufficient but not more than necessary: the standard must protect health with the margin of safety, and it may not be set tighter than the health purpose demands simply because tighter is achievable. The agency thus walks a defined corridor. It may not plead cost to set the bar lower than health requires, and it may not invoke abundant technology to set the bar higher than health requires. The corridor keeps the target honest in both directions, and it explains why standard-setting proceedings turn on epidemiology and exposure modeling rather than on industry balance sheets.

That ordering is the design choice the One Test asks every reader to carry away. Health first, means second. The agency determines what the atmosphere must be like; the states, the sources, and the technologists then figure out how to get there. The inversion does not mean Congress ignored feasibility or cost. It means Congress relocated them. They reappear, explicitly and legitimately, in the implementation machinery: in the state plans that choose among control strategies, in the section 111 standards of performance that weigh the best system of emission reduction against cost, in the section 112 technology standards for hazardous pollutants, and in the mobile-source programs that balance emission cuts against vehicle cost and fuel economy. The statute is not hostile to economics. It simply refuses to let economics set the health target, reserving that judgment for science and medicine.

The margin of safety sharpens the point. Congress did not merely ask what level protects health on the best available estimate; it demanded a cushion against what the estimate might miss. Scientific knowledge about low-dose, long-term exposure is always incomplete, and the people most exposed to uncertainty are the most vulnerable: children whose lungs are still developing, older adults, people with respiratory disease. The margin of safety is the statute’s instruction to err on the side of the breather rather than the emitter when the evidence has gaps. A technology-based regime would treat uncertainty as a reason to wait, since waiting costs industry nothing. The health-first regime treats uncertainty as a reason to build in protection, and the difference between those two defaults explains more about the act’s real-world bite than any single provision.

None of this means the target is set by whim or by politics wearing a lab coat. Section 108 requires the agency to issue air quality criteria documents that reflect the full body of scientific knowledge, and the statute created the Clean Air Scientific Advisory Committee, an independent body of outside scientists, to review the evidence and the staff’s conclusions. The Administrator must explain decisions in light of that record, and courts review them. The inversion is therefore disciplined rather than arbitrary: the health judgment must survive scientific scrutiny and judicial review, but it must not be diluted by cost arguments smuggled into the health analysis. That discipline is part of what makes the architecture, rather than any slogan about clean air, the decisive feature of the law.

Readers who grasp the inversion will also grasp why the act generates the particular kind of conflict it does. Because the target is set without regard to cost, the target sometimes demands more than current practice can deliver cheaply, and the argument then moves downstream to the state plans, the technology standards, and the courts, where feasibility and price are proper subjects. The fights are real, and they are fought in the open, but they are fought about implementation, not about whether the air should be safe to breathe. Congress settled that question at the front of the sequence, in 1970, and every machine that follows is an instrument for honoring the settlement.

The Ambient Standards Machine: How Section 108 and Section 109 Turn Health Science into Law

The ambient standards machine is where the health-first inversion becomes a legal duty with a deadline. It runs on two sections that do two different jobs. Section 108 decides which substances get standards at all, and section 109 decides what the standards say. Together they form a pipeline: listing, scientific assessment, standard-setting, and periodic review. Nothing in the pipeline regulates a single source. Everything in it describes the condition of the outdoor air that the rest of the statute must then produce.

Listing under section 108 begins with a judgment about danger and ubiquity. The Administrator must publish, and revise as appropriate, a list of each air pollutant whose emissions cause or contribute to air pollution that may reasonably be anticipated to endanger public health or welfare, whose presence in the ambient air results from numerous or diverse mobile or stationary sources, and for which the agency plans to issue air quality criteria. The two qualifiers matter. Endangerment screens out the trivial; the numerous-or-diverse-sources test screens out the local oddity and the single-facility problem, reserving this machine for the contaminants that drift across jurisdictions and come from everywhere at once. A substance that poisons one neighborhood from one stack belongs to a different machine, the toxics program of section 112, and the statute’s architecture keeps the two tracks from blurring.

The ubiquity test has practical bite. A pollutant emitted by millions of tailpipes and thousands of boilers across every state qualifies; a pollutant traceable to a handful of facilities does not, however dangerous it may be locally. Congress thus sorted the air-pollution problem by its geography before sorting it by its chemistry, assigning the widespread to the ambient machine and the localized-but-lethal to the toxics machine.

For each listed substance, section 108 then requires the agency to issue air quality criteria: a document reflecting the full body of scientific knowledge on the effects on public health and welfare of the pollutant in varying quantities in the atmosphere. These criteria documents are the evidentiary foundation of everything downstream. They are long, technical, and periodically rewritten as epidemiology, toxicology, and exposure science advance. The statute also requires the agency to issue information on pollution control techniques, so that the record contains both the health science and the engineering knowledge, kept in separate compartments for separate uses. The separation is deliberate: the health evidence will drive the target, and the control information will inform the plans, and the two are not to be mixed at the listing stage.

The scientific review is institutionalized rather than left to the agency’s discretion. The Clean Air Scientific Advisory Committee, a standing body of outside experts, reviews the criteria documents and advises the Administrator on the standards, and its letters and recommendations become part of the public record that courts examine when the standards are challenged. The agency must also publish its reasoning and respond to significant comments, so the path from study to standard is traceable. This procedural armor matters because the standards are among the most litigated rules in administrative law; nearly every review cycle ends in court, and the record built under sections 108 and 109 is what survives or falls.

Section 109 converts the criteria into law. The Administrator must promulgate national primary and secondary ambient air quality standards for each listed pollutant, and the distinction between the two kinds of standards is the hinge of the whole machine.

How do primary and secondary air quality standards differ?

Primary standards are set at levels requisite to protect public health, including sensitive groups, with an adequate margin of safety. Secondary standards are set at levels requisite to protect public welfare, which covers visibility, crops, forests, soils, water, wildlife, buildings, and materials. Both kinds are health-and-science driven, but each answers to a different statutory purpose.

The primary standard is the human-health instrument, and its beneficiaries are named with unusual moral specificity in the case law and the legislative history: the asthmatic child, the elderly person with heart disease, the outdoor worker. The adequate margin of safety protects these groups against the unknowns in the science, not merely the knowns. The secondary standard is the welfare instrument, and its beneficiaries are the things people value beyond their own lungs: clear views in national parks, undamaged harvests, uncorroded stone and paint, forests and waters unharmed by deposition. Congress understood that air foul enough to haze a vista or wither a crop might still sit below the threshold of direct human injury, and it refused to leave that harm unaddressed. The two standards are set through the same scientific process but answer different questions, and a state plan must attain both, which means the welfare target sometimes drives controls that the health target alone would not have required.

The statute sets different clocks for the two. Primary standards carry attainment deadlines that press toward timely health protection; secondary standards are to be attained as expeditiously as practicable. The difference reflects a legislative judgment about urgency rather than importance. Both are mandatory, but the health target moves on a tighter schedule, a further expression of the health-first ordering that shapes the entire title.

The welfare side of the machine has produced some of the program’s most visible work. Congress wrote visibility protection into the statute with unusual care, directing protection of views in mandatory class I federal areas, the great national parks and wilderness areas, and the agency’s regional haze program traces directly to that command. Agricultural science documents yield losses from ozone exposure; materials science documents the soiling and corrosion of buildings, stone, and paint; and the acid deposition title added in 1990 addressed the long-range transport of sulfur and nitrogen compounds that acidified lakes and forests far downwind of their sources. None of these harms would necessarily trigger a primary standard, yet each represents the kind of public welfare Congress charged the secondary standards with defending.

What forces the standards to be revisited every few years?

Section 109 requires the Administrator to review the criteria and the standards every five years and revise them as appropriate. Each cycle reopens the scientific record, weighs new exposure and health evidence, considers the independent science advisers’ advice, and ends in a decision to retain or tighten the standards, so the legal definition of safe air tracks advancing knowledge.

This review duty is one of the machine’s most consequential features, because it makes the standards self-updating. A level that reflected the epidemiology of one decade can be tightened when the next decade’s studies show harm at lower concentrations, and the history of the program bears this out across several pollutants and several administrations. The five-year clock also creates a rhythm of litigation and rulemaking that keeps the agency, the states, and the regulated community in a continuous posture of reassessment. Readers who want the full history of how individual standards moved through these cycles will find it in the later article on the national ambient air quality standards, which traces each pollutant’s standard-by-standard story.

Six pollutants have traveled the full pipeline and carry the label criteria pollutants: particulate matter, ground-level ozone, carbon monoxide, sulfur dioxide, nitrogen dioxide, and lead. Each earned the label by satisfying the section 108 tests of endangerment and ubiquity, and each is described here by what it is and where it comes from rather than by any numerical level, since the levels are revised on the five-year cycle and the mechanism matters more than any snapshot.

Particulate matter is the mixture of tiny solid particles and liquid droplets suspended in the air, produced by combustion, industrial processes, construction, agriculture, and natural sources, and sorted by size because size determines how deeply particles lodge in the lungs. Ground-level ozone is not emitted directly; it forms when sunlight cooks precursor gases, principally nitrogen oxides and volatile organic compounds, which is why ozone peaks on hot, still afternoons and why controlling it means controlling its ingredients. Carbon monoxide is the colorless, odorless product of incomplete combustion, most concentrated near heavy traffic. Sulfur dioxide comes largely from burning sulfur-containing fuels such as coal and oil at power plants and industrial boilers. Nitrogen dioxide comes from high-temperature combustion in vehicles, power plants, and industrial equipment, and it doubles as an ozone precursor. Lead, the oldest recognized of the group, entered the air principally through leaded gasoline and through smelters and battery plants, and its ambient standard became one of the great public-health victories of the program as leaded fuel was phased out.

The health evidence behind the six listings reads like a catalog of modern epidemiology. Fine particles are associated in the literature with premature mortality and cardiovascular harm; ozone with aggravated asthma and reduced lung function; carbon monoxide with stress on the cardiovascular system; sulfur dioxide and nitrogen dioxide with respiratory effects, especially in children and people with asthma; lead with neurological injury, most cruelly in developing children. The criteria documents weigh thousands of studies to draw these links, and the margin of safety exists precisely because the links at low doses are the hardest to pin down and the most important to respect.

One more structural point completes the ambient machine: the national standards are a floor, not a ceiling. Section 116 preserves the authority of states and local governments to adopt and enforce standards more stringent than the federal ones, and several have done so. The act nationalizes the minimum definition of acceptable air; it does not forbid any jurisdiction from demanding more. That one-way ratchet reflects the same health-first logic as the rest of the title: uniformity protects everyone at least this much, and local ambition may protect some people more.

Once promulgated, the standards start statutory clocks. Primary standards must be attained as expeditiously as practicable, and the amendments of 1977 and 1990 attached specific outer deadlines and classification schemes that vary by pollutant and by the severity of an area’s violation. The 1990 amendments were especially detailed, sorting ozone nonattainment areas into classifications from marginal to extreme, each with its own attainment date and its own escalating menu of required controls. The deadlines are not aspirational; areas that miss them face reclassification to a more serious category, which tightens both the timetable and the required measures. The machine thus converts a scientific judgment into a dated legal obligation, and the date is what gives the state planning machine its urgency.

None of the machinery can run without measurements, so the statute builds a monitoring obligation into the system. States must operate air monitoring networks under federal reference methods, and the resulting data determine area designations, track progress toward attainment, and supply the evidence for enforcement. A standard without monitors would be a wish; the network of stations sampling the air day after day is what turns the legal target into an observed fact, and disputes about whether an area has attained a standard are, at bottom, disputes about what the monitors show.

The machine’s output, then, is a set of legal descriptions of air clean enough to breathe and clean enough to protect welfare, written from science, insulated from cost, cushioned by a margin of safety, and refreshed on a statutory schedule. Nothing has yet told any source to do anything. That instruction arrives with the next machine, and the handoff between the two is the most misunderstood joint in the statute.

The State Planning Machine: Why Washington Sets the Target and the States Write the Rules

If the ambient standards machine describes the destination, the state planning machine assigns the journey. Section 110 provides that each state shall adopt and submit to the Administrator a plan, the state implementation plan, that provides for implementation, maintenance, and enforcement of the national ambient air quality standards within that state. The sentence is short and the consequences are vast. Congress nationalized the target and localized the means, creating the cooperative federalism engine that drives the entire criteria-pollutant program.

The plan is not a wish or a white paper. The statute requires it to contain enforceable emission limitations and other control measures, means, or techniques, along with the machinery of a working regulatory program: an inventory of emissions from sources in the state, a monitoring network that measures whether the air actually meets the standards, an enforcement program with teeth, and a demonstration that the chosen measures will bring any violating area into attainment by the statutory deadline and keep it there afterward. The state chooses the mix. It may tighten limits on power plants, require cleaner paints and solvents, restrict wood burning on stagnant winter days, redesign transit, or impose inspection programs on vehicles, and the federal role is to judge whether the mix adds up to attainment, not to dictate the recipe.

The statute gives the states a deadline measured in years from each new or revised standard to submit their plans, and it gives the agency a parallel duty to determine whether submissions are complete and to approve or disapprove them. The completeness finding matters more than it sounds: an incomplete submission triggers the same federal backstop as no submission at all, which is why states treat the administrative requirements of section 110, the inventories, the monitoring commitments, the enforcement authority demonstrations, as seriously as the control strategies themselves. A plan that promises the right reductions without the legal machinery to deliver them is, in the statute’s eyes, not a plan.

Practitioners distinguish two kinds of state submissions, and the distinction helps readers locate documents. The infrastructure plan addresses the standing elements every state needs regardless of its air quality: the monitoring network, the enforcement authority, the permit programs, the consultation procedures. The attainment or control-strategy plan addresses the specific job of bringing a violating area into compliance, with the inventories, the modeling demonstration, and the control measures the area needs. Both travel under the name state implementation plan, and both are subject to federal approval, but they answer different questions, and a state with clean air still owes the first even when it owes nothing under the second.

Before a plan can be written, the air must be graded. After standards are set or revised, the agency designates areas as attainment, nonattainment, or unclassifiable, and the designation sets the planning burden. An attainment area’s plan must prevent significant deterioration of its clean air, holding the gains rather than coasting. A nonattainment area’s plan must do the harder work of reaching the standard on a statutory timetable, with progressively stronger measures for areas graded into more serious classifications. The classification ladder, with its subcategories from marginal to extreme for ozone, calibrates the medicine to the severity of the disease, and it is one reason the same national standard can demand far more from one metropolitan region than from another.

The 1990 amendments built the calibration into the statute itself. Ozone nonattainment areas were sorted into five classifications, from marginal to extreme, with more serious classifications receiving later attainment dates but facing stricter mandatory controls: reasonably available control technology on major sources, inspection and maintenance programs for vehicles, clean-fuel requirements, and steeper offset ratios for new construction. Particulate matter and carbon monoxide areas received their own, simpler classification schemes. The centerpiece control concept for nonattainment areas is reasonably available control measures, including reasonably available control technology, the set of controls that are both technologically feasible and economically reasonable for existing sources in a dirty-air area. The worse the air, the more of these measures the plan must contain, and the statute leaves the agency little discretion to waive them.

Attainment is not the end of the story. An area that meets the standard may be redesignated to attainment, but only if the state submits a maintenance plan showing how it will keep the standard for at least ten years, with a second ten-year plan to follow. The maintenance plan must include contingency measures that snap into place if the area slips, so redesignation is a supervised graduation rather than a release. The requirement reflects the statute’s long memory: the conditions that fouled the air once can return with growth, and the law insists the plan account for the return before it celebrates the departure.

The plans reach beyond stationary sources into the decisions of government itself. Under the conformity provisions, transportation plans and programs funded or approved with federal money must conform to the state plan’s emissions budgets, and federal agencies must ensure their own actions conform as well. A highway project that would bust the region’s motor-vehicle emissions budget cannot proceed as drawn, which makes the state plan a constraint on public infrastructure choices, not only on private factories. The provision closes a loophole the drafters foresaw: it would be perverse for the law to squeeze emissions from industry while public spending underwrote new sources of the same pollutants.

The federal backstop is what makes the state duty real rather than aspirational. If a state fails to submit a plan, submits an incomplete one, or submits one the Administrator disapproves in whole or in part, section 110 authorizes the agency to impose a federal implementation plan that does the job the state left undone. The backstop is supplemented by sanctions: the statute provides for the cutoff of certain federal highway funds and for a two-to-one offset requirement on new sources in areas where the state defaults on its planning obligations. These are not punishments for dirty air. They are consequences for failing to plan, and the distinction matters, because the statute’s theory is that a state with a serious, enforceable plan will produce clean air, while a state without one will not. The federal government rarely wants to run a state’s air program, and the sanctions exist to make sure it rarely has to.

Why does the same national target produce different rules in different states?

Because Congress assigned the states the choice of means while reserving to Washington the judgment of ends. Two states face different source mixes, different weather, different terrain, and different starting levels of pollution, so each writes a plan fitted to its own conditions and the agency approves each plan on whether it attains the uniform target.

A state dominated by coal-fired power plants will write a different plan from a state dominated by traffic; a valley that traps winter inversions will write a different plan from a windy plain; a region already close to attainment needs lighter measures than a region far from it. The national standard is identical in both places, and that identity is the point: every American is promised air that meets the health-based definition, but the route to that air runs through local facts. The complication dissolves once the division of labor is clear. The target is federal and uniform; the rules are state-written and varied; and a factory owner who wants to know the applicable limit looks to the state plan, not to the federal standard, because the plan is where the target becomes a duty.

This division of labor also explains the act’s distinctive politics. Because the states write the plans, the hardest arguments about cost, feasibility, and economic dislocation happen in state capitals and in state administrative proceedings, where the people who breathe the air and the people who employ its breathers are neighbors. The federal agency sets the destination and enforces the discipline of the timetable, but the state decides whether to reach it by scrubbing power plants, cleaning up solvent use, expanding transit, or some combination the state’s own analysis supports. Cooperative federalism is sometimes described as Washington ordering and the states obeying, but the planning machine is closer to Washington grading and the states composing, with the grade depending entirely on whether the composition performs.

The machine has sharp edges worth naming. A state that writes an approvable plan on time faces no federal second-guessing of its policy choices; a state that stalls faces the federal plan and the sanctions. Air does not respect borders, so section 110 also contains the good-neighbor provision, which bars any state plan from emitting pollutants in amounts that contribute significantly to another state’s nonattainment or interfere with its maintenance of the standards. The provision has generated some of the fiercest disputes in the program’s history, precisely because it forces upwind states to account for downwind consequences, and it makes the planning machine a continental system of interlocking promises rather than fifty isolated exercises. Each state’s attainment depends in part on the fidelity of the others, and the statute’s answer is a legal duty owed across state lines.

Two final features give the plans their democratic and legal bite. First, the statute requires public notice and hearing opportunities in the development of state plans, so the choice among control strategies happens in the open, where affected communities and industries can argue their case. Second, once the agency approves a state plan, its provisions become federally enforceable, which means violations can be pursued not only by the state and the federal government but also by citizens acting under the act’s citizen-suit provision. The plan is therefore not merely a promise a state makes to Washington. It is law that neighbors can enforce, and that enforceability is what turns the cooperative federalism of section 110 from an administrative arrangement into a genuine engine of cleaner air.

Taken together, the two machines built in this article form the core loop of the Clean Air Act’s criteria-pollutant program: science sets the target, states write the plans, monitors measure the air, and the gap between measured and required drives the next round of controls. The remaining machines attach to this loop at defined points. The new-source programs keep growth from widening the gap; the toxics program handles the dangerous substances that ubiquity screening excludes; the mobile-source program attacks the one source category Congress chose to govern directly; and the permit program gathers each source’s resulting duties into a single enforceable document. The loop is the architecture, and the architecture is what decides who has to do what.

Later articles in this series take up the remaining machines in turn: the new-source performance standards and preconstruction review that keep growth from erasing the plans’ gains, the hazardous pollutant program that handles the dangerous substances this machine screens out, the mobile-source title that governs the nation’s vehicles and fuels, and the operating-permit title that gathers each source’s duties into one enforceable document. Each will be read most profitably with the inversion in mind, because every one of them is a means chosen after the health target was set, and the target is the fixed point around which all of them orbit.

The New Source Machine: Section 111 Performance Standards

Section 111 of the statute, codified at 42 U.S.C. 7411, operates on a single organizing judgment: the date a plant is built or rebuilt determines which emission rules it must satisfy. That timing choice shapes the act’s treatment of stationary sources more than any single numerical limit inside it. A factory whose construction began before the relevant standard was proposed faces one set of duties; an identical factory whose construction began a day later faces another, stricter set. The section does not apologize for the asymmetry. It builds the entire new source program around it.

The mechanism starts with a listing duty. EPA must publish, and periodically revise, a catalog of categories of stationary sources that, in the agency’s judgment, cause or contribute significantly to air pollution that may reasonably be anticipated to endanger public health or welfare. Power plants, refineries, cement kilns, steel mills, and dozens of other industrial categories appear on that roster. Once a category is listed, the agency must set standards of performance for new sources within it. The phrase “standard of performance” is a term of art here. It means an emission limit that reflects the degree of emission limitation achievable through the best demonstrated technology, with the agency instructed to take cost into account, along with nonair quality health and environmental impacts and energy requirements. Cost enters the picture at this stage deliberately. The standard-setting question under section 111 is a technology question, and the statute tells EPA to weigh what the technology costs before imposing it.

Notice the contrast with the ambient standards discussed in the first two machines. There, the statute forbids cost from entering the room when the numerical target is chosen. Here, the statute invites cost in. The difference is not an accident of drafting. The ambient program answers the question “how clean must the air be,” and the answer is meant to be health-driven without compromise. The new source program answers the question “what must this particular plant install,” and the answer is meant to be achievable in practice. A reader who keeps those two questions separate will find the statute’s apparent inconsistency dissolving into design.

The timing trigger is precise. A “new source” is a stationary source whose construction or modification commenced after the publication of regulations, or proposed regulations, prescribing a standard of performance for that source category. The proposal date matters, not the final rule date. An owner who breaks ground after the proposal is on notice that a stricter regime is coming, and the statute treats that owner as having planned accordingly. An owner who broke ground before the proposal locked in the older expectations and cannot fairly be told to have anticipated the new ones. That is the grandfathering logic in its cleanest form.

Modification is where the timing logic shows its teeth. The statute defines modification as any physical change in, or change in the method of operation of, a stationary source that increases the amount of any air pollutant emitted or results in the emission of a pollutant not previously emitted. A modification turns an existing source into a new source for purposes of section 111. The rebuilt boiler, the expanded production line, the changed feedstock that raises emissions, each of these can reclassify a grandfathered plant and pull it under the new source standard. The practical consequence is enormous. Plant owners face a continuing choice between maintaining existing equipment in its current configuration and undertaking changes that improve efficiency or expand capacity but trigger stricter emission duties. Regulators face the mirror-image question of how to distinguish routine maintenance, which does not count as a modification, from life-extending reconstruction, which does. Disputes over that boundary generated some of the most heavily litigated enforcement actions under the act.

Why does the Clean Air Act treat new and existing factories differently?

Congress reasoned that requiring expensive retrofits on every existing plant would impose uneven costs and invite litigation over feasibility, while applying strict limits only to new and modified units lets the standard rise over time as the capital stock turns over. That timing split is the engine of section 111.

The category listing under section 111(b)(1)(A) carries its own endangerment finding, and it operates differently from the ambient program’s pollutant listings. Before listing a source category, EPA must find that the category causes or contributes significantly to air pollution that may reasonably be anticipated to endanger public health or welfare. The finding is category-specific: the agency asks whether this industry’s emissions, in the aggregate, meaningfully contribute to the pollution problem. The listed categories number in the dozens and span the industrial economy, from electric utility steam generating units and petroleum refineries to Portland cement plants, iron and steel mills, and municipal waste combustors. Each listing decision reflects a judgment about significance that the agency must defend, and each one triggers the mandatory duty to set new source standards for the category. The listing power is thus the gateway to the entire machine: no listing, no standard. The eight-year review of the standards, and the periodic revision of the category list itself, keep the gateway from closing behind the first generation of industries.

The conservative side of the timing logic deserves equal attention. Grandfathering has a cost that the statute accepts openly. An older plant operating under weaker limits can remain in service for decades, and because its compliance costs are lower than a new entrant’s, its owners have an economic incentive to keep it running rather than replace it. Members of Congress who supported the 1970 and 1977 frameworks understood this dynamic and accepted it as the price of political feasibility. Members who wanted faster turnover argued for tighter modification triggers and shorter compliance horizons. The statute as enacted holds the middle: strict for the new, patient with the old, and watchful at the boundary between them through the modification rule.

Section 111(d) addresses the other side of the ledger: existing sources of certain pollutants. The provision directs EPA to establish a procedure, similar in shape to the state implementation plan process, under which each state submits a plan establishing standards of performance for existing sources. This is the existing-source companion to the new source program, and its scope is carefully bounded. It applies to pollutants that are neither listed as criteria pollutants under section 108 nor emitted from a source category already regulated under section 112 for that pollutant. In other words, section 111(d) fills the gaps left by the other two machines rather than overlapping them. The state plan must set emission limits for the designated existing sources, and EPA reviews the plan for adequacy, with a federal backstop available if a state fails to submit a satisfactory one. Historically the provision saw limited use, applied to pollutants such as fluoride emissions from phosphate fertilizer plants, and it functioned quietly as a gap-filler rather than as a centerpiece of the program.

The relationship between the two halves of section 111 rewards careful reading. For new sources, EPA writes the standard directly and the limit is uniform nationally within the category. For existing sources of designated pollutants, EPA writes guidelines and the states write the standards, which introduces the same state-by-state variation found in the implementation plan program. The new source half is technology-forcing in the sense that it can require technology not yet in widespread use, so long as EPA finds it adequately demonstrated and considers the cost. The existing source half is more deferential to state judgment about what the covered facilities can achieve. Both halves share the same vocabulary of performance standards, but they allocate the decision differently.

A final feature of the new source machine is its periodic review duty. EPA must review and, if appropriate, revise the standards of performance at least every eight years. The review asks whether the best demonstrated technology has advanced since the last round and whether the existing limit should be tightened to reflect the advance. This keeps the program from freezing in place. A standard set in one decade reflects the technology of that decade; the eight-year review is the statute’s way of ratcheting expectations upward as control equipment improves. Combined with the modification rule, the review duty means the program tightens from two directions at once: new plants face better technology at construction, and rebuilt plants get reclassified into the newer tier.

Section 111(j) provides a waiver mechanism for innovative technology that deserves mention because it shows the statute’s attitude toward experimentation. An owner or operator proposing to use an innovative technological system of continuous emission reduction may obtain a waiver from the applicable standard of performance, subject to conditions EPA imposes to assure that the system will not cause unreasonable risk to public health or the environment. The waiver lets a source serve as a full-scale demonstration of a technology that might become the basis for a future standard, trading short-term forbearance for long-term learning. Few waivers have been sought, but the provision’s existence confirms the program’s orientation: the standard reflects demonstrated technology, and the statute creates a pathway for the next generation of technology to become demonstrated.

Readers sometimes ask why the statute bothers with this elaborate timing apparatus instead of simply requiring every plant to meet the best available limit. The answer is that section 111 is a compromise between environmental ambition and the economics of sunk capital, written into law. The compromise is visible, legible, and enforceable. Anyone who can determine when a source was built, when it was last modified, and which category it belongs to can determine which duties apply. That legibility is the point. The timing logic turns a political settlement into an administrable rule.

The phrase “adequately demonstrated” has generated its own body of interpretation, and the leading early decision remains instructive. In Portland Cement Association v. Ruckelshaus, decided in 1973, the court of appeals upheld EPA’s authority to base a new source standard on technology that had been demonstrated in practice even if it was not yet in widespread routine use across the industry. The decision drew the line the statute implies: the technology must be more than a laboratory concept or a paper design, but it need not be the industry’s common practice. That reading gives the technology-forcing dimension of section 111 its legal footing. EPA may look to the best-performing plant, the pilot installation operating at scale, or the control system proven in a comparable application, and set the national standard at what that technology achieves. The cost consideration then operates as the check: the agency must take cost into account and explain why the required expenditure is reasonable in relation to the emission reduction gained.

Enforcement of the new source standards runs through a straightforward prohibition. Section 111(e) makes it unlawful for an owner or operator of a new source to operate in violation of an applicable standard of performance. The duty is direct and federal, which distinguishes it from the implementation plan program where the state’s plan supplies the operative limit. A violation of a new source standard is a violation of the act itself, enforceable by EPA, by the state where the state has been delegated authority, and by citizens under the citizen suit provision. Delegation deserves a brief note: EPA may delegate implementation and enforcement of the standards to a state whose procedures are adequate, and most states have accepted delegation for most categories. Delegation does not change the content of the standard; the federal limit remains the federal limit, administered locally.

The eight-year review duty merits a closer look because it is the program’s ratchet. Each review asks a bounded question: has the best demonstrated technology advanced since the last standard was set, and does the advance warrant a tighter limit. The review is not a reexamination of whether the category should be listed, nor an occasion to revisit the health science. It is a technology survey. EPA examines control equipment installed since the last round, vendor guarantees, performance data from operating plants, and cost information, then decides whether the floor of demonstrated performance has risen. Where it has, the standard follows. Where it has not, the standard stays. The reviews have produced some of the program’s most contested rulemakings, because each one redistributes compliance costs across an industry and because the determination of what counts as demonstrated is inevitably judgmental at the margin. But the structure of the inquiry keeps the disputes focused on facts about equipment rather than on first principles about the statute’s purpose.

Section 111(d) state plans follow a procedural pattern borrowed from the implementation plan program but adapted to the provision’s narrower scope. EPA first issues emission guidelines for the designated pollutant and the affected source categories, identifying the best system of emission reduction the agency considers adequately demonstrated for existing sources and the degree of limitation achievable. Each state then submits a plan establishing standards of performance for the existing sources within its borders and providing for their implementation and enforcement. EPA reviews the submission and approves it if it satisfies the statutory criteria; where a state fails to submit a satisfactory plan, the agency must prescribe a federal plan. The guidelines-plus-plan structure gives states room to account for the remaining useful life of older facilities and other source-specific factors, a flexibility the new source side does not offer. The provision’s limited historical use meant that its procedures attracted less attention than the implementation plan machinery, but the architecture is the same cooperative federalism in miniature: federal guidelines set the direction, state plans supply the operative limits, and a federal backstop stands ready.

One more boundary question completes the picture of the new source machine: its relationship to preconstruction review. A new or modified major source must typically obtain a preconstruction permit under the new source review program before it builds, demonstrating that it will meet the applicable emission limits and, in cleaner areas, that it will apply the best available control technology determined case by case. The section 111 standard and the preconstruction determination are separate duties that can point in the same direction or diverge. The performance standard sets a national floor for the category; the preconstruction review can require stricter controls based on the specific facility and its location. A plant must satisfy both. The permit machine of Title V then consolidates both into the operating permit. Readers who keep these three layers distinct, the categorical floor, the case-specific preconstruction determination, and the consolidating operating permit, will avoid the most common confusion in the stationary source program.

The Toxics Machine: The Clean Air Act’s Section 112 After 1990

Section 112 is the statute’s answer to a class of pollutants that the ambient program cannot handle well. The criteria pollutant system works by setting a safe concentration in the outdoor air and directing states to plan their way down to it. That approach assumes a pollutant is emitted widely enough, and measured easily enough, that an ambient concentration is the right regulatory object. Many dangerous substances do not fit that assumption. Benzene, asbestos, mercury, vinyl chloride, and scores of other compounds are emitted by specific industrial processes, often in quantities that are hard to measure in the ambient air but easy to control at the stack. Regulating them through ambient targets would be like regulating a factory’s wastewater by measuring the river miles downstream. The 1990 amendments rebuilt section 112 around that insight, replacing a risk-based program that had produced almost nothing with a technology-based program organized by source category.

The pre-1990 record explains why Congress intervened. The original section 112 directed EPA to list hazardous air pollutants and then set emission standards providing “an ample margin of safety” to protect public health. The listing process was discretionary and slow. Over two decades the agency listed only a handful of substances, asbestos and benzene among them, and each listing invited protracted scientific dispute about how much risk was acceptable and what an ample margin of safety required. Litigation over the vinyl chloride standard stretched across years and produced court decisions that faulted the agency’s reasoning without producing a workable template for the next substance. By the late 1980s the program had generated standards for only a small number of pollutants while hundreds of known or suspected toxic emissions went unaddressed. Members of Congress from both parties described the provision as a failure, and the regulated industries that had lived under the uncertainty were among those asking for a clearer, more predictable system.

The 1990 rewrite took listing out of EPA’s discretion and put it into the statute. Congress itself listed the hazardous air pollutants, naming 189 substances and compound groups in section 112(b), a list later adjusted to 188. The agency retained authority to add pollutants that present a threat of adverse human health effects, or to delete listed pollutants shown not to warrant control, but the starting point was fixed by law. That single change eliminated the listing bottleneck that had paralyzed the program. No one could any longer stall regulation by contesting whether a substance deserved to be listed, because Congress had already answered that question.

With the list fixed, the rewrite organized regulation by source category rather than by pollutant. EPA must list categories of major sources of hazardous air pollutants, the industrial groupings such as chemical plants, refineries, and steel mills that emit the listed substances, and then set emission standards for each category. The standard-setting concept is maximum achievable control technology, universally known by its initials MACT. The statute defines the minimum stringency of a MACT standard through a floor: for existing sources, the floor is the average emission limitation achieved by the best performing twelve percent of sources in the category, or by the best performing five sources where the category contains fewer than thirty. For new sources, the floor is the emission control achieved in practice by the best controlled similar source. EPA may set the standard above the floor, that is, stricter than the floor requires, but only after considering the cost of achieving the additional reduction, nonair quality health and environmental impacts, and energy requirements. The floor itself, however, requires no cost justification. It is derived from what the industry’s best performers already achieve, which means the technology is demonstrated by definition.

What does maximum achievable control technology require of a source category?

The law defines the floor as the average emission limitation achieved by the best performing twelve percent of existing sources in the category, or the best performing five where fewer than thirty sources exist. EPA then may go beyond the floor only by considering cost, non-air health and environmental impacts, and energy requirements.

The floor mechanism deserves a closer look because it embodies a distinctive theory of regulation. Rather than asking scientists to determine a safe level of exposure for each substance, an inquiry that produced endless dispute under the old program, the floor asks a factual question about industry: what are the best-run plants in this category already achieving? The answer becomes the minimum legal duty for everyone else. Laggards must rise to the level of the leaders. The mechanism is self-updating in the sense that each round of standards reflects current practice, and it sidesteps the risk-assessment wars by anchoring the requirement in demonstrated performance rather than in toxicological modeling. Proponents in Congress presented this as the rewrite’s central virtue: it would produce standards quickly, predictably, and on a statutory schedule, replacing two decades of drift with a decade of rulemaking deadlines.

The rewrite also distinguished major sources from area sources. A major source is a stationary source that emits, or has the potential to emit, ten tons per year or more of any single listed pollutant or twenty-five tons per year or more of any combination of them. Major sources face MACT standards. Smaller area sources, which collectively can account for significant emissions in categories like dry cleaning or auto body refinishing, face standards based on generally available control technologies, a less demanding concept that the statute leaves largely to EPA’s judgment. The distinction reflects a proportionality judgment: the heaviest regulatory machinery is reserved for the largest emitters, while smaller operations face requirements scaled to their size and resources.

After the technology-based standards are in place, section 112(f) requires a second look. Within eight years of a MACT standard’s promulgation, EPA must assess the residual risk remaining after the technology standard is implemented and, if the remaining risk to public health is unacceptable, impose additional standards to provide an ample margin of safety. This residual risk review preserves a role for health-based analysis, but it sequences the analysis after the technology floor is established rather than before. The ordering matters. Under the old program, risk analysis was the gateway and nothing moved until the analysis was complete. Under the rewritten program, technology standards move first and risk analysis follows as a backstop. The delay that strangled the original program is structurally prevented.

The distinction between criteria pollutants and hazardous air pollutants is one of the most tested concepts in the statute, and the two machines illustrate why the distinction exists. Criteria pollutants are ubiquitous, emitted by many source types and by everyday activities, and the right regulatory object is the concentration in the air people breathe. The regulatory method is therefore ambient and health-based: set the concentration, plan down to it. Hazardous air pollutants are emitted by identifiable industrial categories in ways that are better controlled at the source than measured in the air. The regulatory method is therefore categorical and technology-based: identify the category, find the best performers, require the rest to match. A substance can in principle be addressed by either method, and the statute’s gap-filler in section 111(d) is defined partly by reference to which method already covers a pollutant. But the two methods reflect different theories of the problem, and a reader who confuses them will misread both.

The 1990 toxics rewrite also illustrates the statute’s use of deadlines as an enforcement device against the agency itself. Congress did not merely authorize EPA to write MACT standards; it scheduled them. Categories were to be listed on a timetable, standards were to follow on a timetable, and citizens could sue to enforce the timetable. This hammer of deadlines reflected Congress’s judgment that discretion had been the old program’s undoing. The schedule forced a decade of intensive rulemaking, and while litigation over individual standards continued, the program produced the volume of regulation its predecessor never had. For readers tracing the broader history of the amendments, a companion article in this series examines the 1990 amendments and the political coalition that produced them.

One caution belongs here. The technology-based approach has a characteristic blind spot: it controls what the best performers already do, but it does not ask whether the resulting emissions are safe. The residual risk review is meant to catch that gap, and its effectiveness depends on the quality of the risk assessment and the willingness to impose a second round of controls where the first round leaves unacceptable risk. Members who favored a purely health-based program criticized the rewrite on exactly these grounds, while members who favored predictability for industry defended the sequencing. The statute as enacted holds both ideas: technology first, health as the backstop. Whether the backstop has done its intended work is a question the statute’s own review provisions keep returning to the agency.

Compliance timing under the rewritten program follows a statutory schedule designed to prevent the drift that characterized the old one. An existing major source must comply with a MACT standard no later than three years after the standard’s effective date, with the permitting authority able to grant a one-year extension where necessary for the installation of controls. New sources must comply immediately upon startup. The three-year clock reflects a legislative judgment about how long it reasonably takes to design, procure, and install major control equipment, and the extension provision acknowledges that some installations cannot be completed within that window without shutting down production. The schedule is deliberately tighter than the open-ended implementation timelines of the ambient program, because the technology-based standards specify the control method more precisely and leave less to state discretion.

The rewrite included an early reduction incentive that reveals how Congress thought about the transition. Section 112(i)(5) offered sources that voluntarily achieved deep reductions of hazardous pollutants before the MACT standards took effect an alternative emission limit and a six-year extension of the compliance deadline. The provision rewarded early actors and gave the agency a way to credit reductions already achieved, softening the complaint that the new program punished facilities that had invested in controls ahead of the mandate. The incentive’s structure, a defined percentage reduction from a baseline year in exchange for regulatory relief, became a template for later voluntary programs, but its original function was transitional: to bridge the gap between the old program’s paralysis and the new program’s deadlines.

Beyond the emission standards, section 112 houses two programs that operate on different theories. Section 112(r) addresses accidental releases, the catastrophic failure rather than the routine emission. It requires EPA to list regulated substances and threshold quantities, and it imposes on owners and operators of covered facilities a general duty to identify hazards, design and maintain a safe facility, and minimize the consequences of releases. Covered facilities must prepare risk management plans documenting their hazard assessments, prevention programs, and emergency response measures. The provision was a response to industrial disasters that focused congressional attention on the difference between controlling steady-state emissions and preventing sudden catastrophic ones. Its vocabulary is the vocabulary of process safety rather than emission limitation, and its enforcement runs through plan review and the general duty clause rather than through stack testing.

Section 112(d)(7) addresses a different corner: work practice standards for situations where measurement is impracticable. Where it is not feasible to prescribe or enforce an emission standard, because the pollutant cannot be emitted through a conveyance designed to capture it or because measurement methodology is impracticable, EPA may instead prescribe design, equipment, work practice, or operational standards. Dry cleaning operations and certain coating processes fall into this category. The work practice approach sacrifices the precision of a numerical limit for the enforceability of a required method, and its presence in the section shows the rewrite’s pragmatism: the program prefers numerical limits where they can be measured and enforced, and accepts operational requirements where they cannot.

The delisting and listing petitions keep the statutory list from ossifying. Any person may petition EPA to add a substance that presents a threat of adverse human health effects through inhalation or other routes of exposure, and may petition to delete a substance where the petitioner shows that it does not warrant regulation. The agency must grant or deny the petition with a stated basis, and the denial is subject to judicial review. In practice the list has been far more stable than the petition provision might suggest, but the mechanism matters for the program’s legitimacy: the technology-based approach depends on a list fixed by Congress, and the petition process supplies the safety valve for new science in both directions.

Area sources complete the categorical picture. For categories dominated by small emitters, dry cleaners, auto body shops, small chemical formulators, the statute directs EPA to set standards based on generally available control technologies or management practices. The area source program recognizes that imposing MACT-level analysis on thousands of small businesses would be disproportionate, while leaving their collective emissions uncontrolled would leave a real gap. The generally-available-technology concept is less precisely defined than MACT, which gives the agency more discretion and has produced standards that vary widely in stringency across categories. The variation is a feature of the design: the statute calibrates the regulatory machinery to the size and sophistication of the regulated community.

The residual risk review under section 112(f) operates as the program’s conscience. After a MACT standard has been in effect, EPA must evaluate the health risk that remains and determine whether it is acceptable; where the remaining risk is not acceptable, the agency must impose further controls sufficient to provide an ample margin of safety. The review thus reintroduces the health-based inquiry that the old program began with, but in a disciplined sequence: technology first, risk assessment second, additional controls only where the assessment demands them. The reviews have been slower to complete than the statute’s eight-year schedule contemplated, and the gap between the technology standards and the residual risk determinations has drawn criticism from public health advocates who argue that the backstop arrives too late to matter. The defenders of the sequencing reply that the technology standards themselves eliminated the great bulk of the risk and that the residual reviews address a diminishing remainder. The statute as enacted does not resolve that dispute; it creates the procedure through which the dispute continues.

Taken together, the post-1990 section 112 is the statute’s most elaborately scheduled program. Categories listed on a timetable, standards due on a timetable, compliance within three years, residual risk reviewed within eight, area sources addressed under a separate track, accidental releases handled through planning rather than limits. The elaboration was the point. Congress had watched discretion produce two decades of near-inaction, and it responded by replacing discretion with deadlines at every stage. The program that resulted is administratively demanding, technically intricate, and far more productive than its predecessor. It is also, by design, the machine in which health science plays the smallest direct role: the floor is set by industry practice, the beyond-floor inquiry weighs cost and energy, and the health-based backstop arrives last. Readers who find that ordering surprising should return to the verdict’s discussion of the health-first inversion, because the toxics program is the place where the inversion is most qualified and therefore most instructive.

The Mobile Source Machine: Title II Vehicles and Fuels

Title II of the statute regulates a category of polluter that the stationary-source machinery cannot reach: the moving vehicle. A factory smokestack stays in one place, which means a state implementation plan can inventory it, model its contribution to local air quality, and assign it a share of the needed reductions. A car does none of those things. It crosses state lines, changes owners, ages into disrepair, and emits from millions of tailpipes rather than a countable set of stacks. The statute responds by treating mobile sources as a national manufacturing problem rather than a local air quality problem. The regulatory object is the vehicle as it leaves the factory and the fuel it burns, and the regulator is the federal government acting directly, without the state-plan intermediary that organizes the stationary side.

Section 202 is the heart of the title. It directs EPA to set emission standards for new motor vehicles and new motor vehicle engines, covering hydrocarbons, carbon monoxide, nitrogen oxides, and particulate matter from the classes of vehicles the section enumerates. The standard-setting formula will be familiar from the earlier machines but with its own emphasis: the standards must reflect the greatest degree of emission reduction achievable through technology that the agency determines will be available, giving appropriate consideration to cost. The phrase “will be available” does important work. It authorizes the agency to require technology that is not yet in commercial production at the time the standard is set, so long as the agency finds it will be ready by the model year the standard takes effect. This is the technology-forcing dimension of the vehicle program at its most explicit. Congress told the industry to invent its way into compliance and told EPA to set the pace.

The statute tempers that ambition with lead time. Standards apply beginning with a model year that allows the period of lead time the agency finds necessary to permit the development and application of the requisite technology, giving appropriate consideration to the cost of compliance within that period. In practice this means multi-year phase-ins, announced well in advance, so that manufacturers can retool production lines and suppliers can scale up new components. The lead-time requirement is the statute’s acknowledgment that even forced technology cannot be conjured overnight. Automakers testifying before Congress in 1970 and again in 1990 pressed for longer lead times and more gradual phase-ins; public health advocates pressed for faster schedules and steeper cuts. The enacted compromise gives EPA the pen but requires the agency to show its work on feasibility, a structure that has produced standards of steadily increasing stringency across successive rulemaking rounds.

Enforcement of the vehicle standards runs through certification, not through ambient monitoring. A manufacturer must obtain a certificate of conformity from EPA before selling a vehicle, demonstrating that the vehicle class meets the applicable emission standards. The agency tests prototype vehicles, audits production, and can require recalls where in-use vehicles fail to meet the standards they were certified against. The warranty provisions require manufacturers to warrant that vehicles remain in compliance for specified periods of use, and the tampering prohibitions make it unlawful to remove or render inoperative any emission control device. This compliance architecture reflects the mobile source reality: with millions of vehicles on the road, the only practical enforcement point is before sale and through the manufacturer’s continuing responsibility, supplemented by state inspection and maintenance programs that check vehicles already in service.

Fuels receive parallel treatment under section 211. The provision authorizes EPA to regulate fuel composition where necessary to protect public health or to enable emission control devices to function. The most consequential exercise of this authority was the phase-down and eventual elimination of lead in gasoline, a program that ranks among the statute’s great public health successes. Section 211 also underpins the reformulated gasoline program for the most polluted metropolitan areas and the oxygenated fuels requirements for carbon monoxide nonattainment areas. The fuel provisions recognize a physical truth the vehicle standards alone cannot address: the cleanest engine in the world will still pollute if it burns dirty fuel, and certain fuel additives can poison the catalytic converters that the vehicle standards depend on. Regulating the vehicle and the fuel together closes that loop.

What makes vehicle and fuel regulation a separate title of the statute?

Motor vehicles move across state lines, so state-by-state emission plans could not reach them the way they reach a factory smokestack. Title II therefore sets national standards directly at the manufacturer level, with the California waiver preserving the one state program that predated federal action and section 177 extending a choice to the rest.

Section 209 introduces the title’s most distinctive federalism arrangement. The section preempts states from adopting or attempting to enforce any standard relating to the control of emissions from new motor vehicles or new motor vehicle engines. The preemption is sweeping and deliberate: Congress wanted a single national market for automobiles, not fifty different emission regimes that would balkanize manufacturing. But the section then carves out the exception that has shaped American vehicle regulation ever since. California may obtain a waiver of preemption for its own standards if the state determines that its standards will be, in the aggregate, at least as protective of public health and welfare as the federal standards. The waiver provision exists because California had a vehicle emission program before the federal government did, and Congress chose to preserve the state’s pioneering role rather than extinguish it. EPA’s review of a waiver request is narrowly circumscribed: the agency must grant the waiver unless it finds that California’s determination was arbitrary and capricious, that the state does not need its own standards to meet compelling and extraordinary conditions, or that the state standards are inconsistent with the federal standard-setting provisions.

Section 177 extends the California exception to every other state. Any state may adopt and enforce emission standards for new motor vehicles that are identical to California’s, provided the state acts at least two years before the model year the standards would cover. The two-year lead time protects manufacturers from surprise. The identity requirement protects the national market from fragmentation: there can be a federal standard and a California standard, but not fifty bespoke standards. In practice this has created a two-fleet country, with California and the states that follow its standards forming one regulatory bloc and the remaining states forming another. Automakers have often complained about the compliance complexity, while the following states have argued that the California standards deliver air quality benefits the federal program was slow to match. The arrangement is a standing invitation to regulatory competition, and California has repeatedly used it to push vehicle emission control ahead of the federal pace.

The greenhouse gas question under section 202 produced the most significant judicial interpretation of the title. Whether carbon dioxide and other greenhouse gases from motor vehicles count as “air pollutants” subject to regulation under section 202 reached the Supreme Court, and a companion article in this series explains the Court’s answer in Massachusetts v. EPA. The placement of that discussion matters for the architecture: the case belongs to the mobile source machine because section 202’s broad definitional language, not the ambient or toxics provisions, was the provision at issue. A reader who understands that Title II regulates the vehicle as a manufactured product will understand why the definitional fight happened there rather than under the state-plan machinery.

Useful life definitions give the durability requirement its legal shape. The statute specifies the periods, measured in years or miles depending on vehicle class, during which the emission standards must be met, and the manufacturer must design the emission control system to function for the full useful life. The useful life periods have been extended over successive rulemakings as vehicle longevity increased, which means the durability duty has grown more demanding over time without any change to the statutory language. A vehicle that meets the standard in the laboratory at certification but cannot hold it for the useful life fails the program’s central promise, and the in-use testing and recall provisions exist to catch exactly that failure.

The mobile source machine also reaches beyond highway vehicles. The title authorizes standards for nonroad engines and vehicles, the construction equipment, locomotives, marine vessels, and aircraft engines that contribute significantly to the emission inventory. These provisions have produced successive rounds of standards for diesel equipment and other nonroad categories, extending the manufacturer-certification model beyond the passenger car. The pattern is consistent across the title: identify the engine or vehicle class, set the standard with lead time, certify before sale, and enforce through the manufacturer. Where the stationary machines ask states to plan, the mobile machine asks manufacturers to build.

In-use compliance is where the vehicle program meets the reality of aging machinery. A car that met the standard on the assembly line may not meet it at eighty thousand miles, as catalysts degrade, sensors drift, and owners defer maintenance. The statute addresses this through several interlocking provisions. Manufacturers must warrant that vehicles are free from defects in materials and workmanship that would cause nonconformity, and that the emission control system will perform for specified periods measured in years or miles. EPA conducts in-use testing through its surveillance programs, pulling vehicles from private owners and testing them to verify that the certified designs hold up in real operation. Where a substantial number of vehicles in a class fail to conform, the agency may order a recall, requiring the manufacturer to remedy the nonconformity at no cost to the owner. The recall authority is the program’s most visible enforcement tool and the one manufacturers most want to avoid, which gives the certification process its disciplining force.

State inspection and maintenance programs supply the other half of in-use enforcement. The ambient program requires certain nonattainment areas to operate vehicle inspection programs that test cars already on the road and deny registration to those that fail. These programs sit at the intersection of the mobile and stationary machines: they are creatures of the state implementation plan, required by the ambient provisions, but they enforce the mobile program’s standards against vehicles in service. The enhanced inspection programs required for the most polluted areas use dynamometer testing under load, which measures emissions during simulated driving rather than at idle, and they have generated decades of data on how vehicle emissions deteriorate with age and mileage. That data, in turn, informs each round of new vehicle standards, closing a feedback loop between the in-use fleet and the factory gate.

The nonroad program extends the manufacturer-certification model to engines that never see a highway. Construction equipment, agricultural machinery, locomotives, commercial marine vessels, and aircraft engines each contribute meaningfully to the national emission inventory, and each presents the same regulatory logic as the highway vehicle: the practical enforcement point is the manufacturer, before sale. Section 213 directs EPA to set standards for nonroad engines and vehicles where their emissions contribute significantly to air pollution, and the agency has proceeded category by category, generally following the highway program’s template of technology-based standards with lead time, certification, and in-use requirements. The locomotive and marine provisions reflect the long service lives of those engines, with remanufacturing standards that apply when an old engine is rebuilt, a timing concept borrowed directly from the section 111 modification logic. The nonroad program is less visible than the passenger vehicle program but comparable in its aggregate emission impact.

The fuel provisions merit closer attention because they show the statute regulating a product market rather than an emission point. The lead phase-down, conducted under section 211 authority across the 1970s and 1980s, required refineries to reduce the lead content of gasoline on a declining schedule, with trading provisions that let refineries average and bank lead credits. The program is widely studied as an early and successful application of marketable permits to an environmental problem, and its public health payoff, measured in declining blood lead levels in children, ranks among the largest attributable to any provision of the act. The reformulated gasoline program requires that gasoline sold in the most severely ozone-polluted metropolitan areas meet composition specifications designed to reduce volatile organic compound and toxics emissions, with an opt-in provision letting other areas join. The oxygenated fuels program requires gasoline sold in carbon monoxide nonattainment areas during winter months to contain oxygenates that promote more complete combustion. Each of these programs regulates refineries and fuel distributors rather than drivers, which keeps the compliance burden on a small number of sophisticated actors rather than on millions of motorists.

The California waiver criteria deserve a precise reading because they allocate power between the state and the federal agency. EPA must grant the waiver unless it finds one of three things: that California’s determination that its standards are at least as protective as the federal standards was arbitrary and capricious, that the state does not need its standards to meet compelling and extraordinary conditions, or that the standards and their enforcement are inconsistent with the federal provisions governing standard-setting. The “compelling and extraordinary conditions” language refers to California’s distinctive air quality problems, the combination of geography, meteorology, and vehicle density that produces the country’s most severe smog, and the agency has historically read it as a low bar given the state’s documented conditions. The inconsistency prong protects the lead-time and technological feasibility framework of the federal program: California may go further, but it may not set standards that contradict the federal scheme’s structure. Waiver proceedings have produced some of the most politically charged episodes in the statute’s history, with the agency’s decisions closely watched as signals of the federal government’s posture toward state leadership.

The section 177 opt-in extends the waiver’s logic while containing its fragmentation risk. A state that adopts the California standards must do so at least two years before the affected model year, and the standards must be identical to California’s. The identity requirement is enforced strictly: a state may not adopt California’s standards with modifications, because a modified standard would be a third standard and the statute permits only two. The two-year lead time gives manufacturers the planning horizon the statute promises throughout the mobile title. States have used the opt-in to form a bloc that, combined with California, represents a substantial share of the national vehicle market, which in turn gives the California standards de facto national significance: manufacturers often find it simpler to build to the stricter standard nationwide than to maintain separate production lines. That dynamic, sometimes called the California effect, is not written in the statute, but it is the predictable consequence of the statute’s design.

The Permits Machine: Title V Operating Permits

Title V, added by the 1990 amendments and codified at 42 U.S.C. 7661 through 7661f, solves a problem that the first five machines create collectively. By the time the 1990 amendments were enacted, a major industrial source could be subject to overlapping duties from half a dozen different programs at once: emission limits from its state implementation plan, new source performance standards under section 111, hazardous pollutant standards under section 112, vehicle-adjacent obligations if it operated a fleet, and any number of preconstruction review conditions. Each duty lived in its own regulatory document, issued under its own authority, on its own schedule. Plant managers, inspectors, and citizens trying to answer the simple question “what does this facility have to do” faced a paper chase across the statute. Title V answers that question with a single document: the operating permit.

The permit is a consolidation device, not a new substantive requirement. This distinction is the most misunderstood feature of the title, and the statute is explicit about it. A Title V permit gathers every applicable requirement under the act into one enforceable instrument: the SIP limits, the section 111 standards, the section 112 standards, the acid rain provisions where they apply, and any other duty the source bears. The permit does not add new emission limits on top of those requirements. A source that holds a permit must do what the underlying programs already required, no more and no less, but under Title V the full set of duties appears in one place, with one compliance schedule, one set of monitoring and reporting obligations, and one public record. The practical effect is to make the underlying requirements legible and therefore enforceable.

Who must hold a permit is defined by the major source threshold. Any major source, any source subject to the new source performance standards, any source subject to the hazardous pollutant standards, and any source covered by the acid rain program must obtain an operating permit. The major source definition here generally tracks the hundred-ton-per-year threshold for regulated pollutants, with lower thresholds in more severely polluted nonattainment areas. Smaller sources may be exempted or deferred by EPA, and the agency has used that authority to keep the program focused on the facilities whose emissions matter most. The permit application requires the source to disclose its operations, its emissions, and every requirement it believes applies to it, a self-inventory that itself serves an enforcement function: a source that omits an applicable requirement from its application has created a record of the omission.

The permit’s contents follow a standard template. Each permit must include emission limitations and standards, a compliance schedule where the source is not yet in compliance, monitoring requirements sufficient to assure compliance, recordkeeping and reporting duties, and a statement of the permit’s duration, which may not exceed five years. The monitoring provisions deserve emphasis because they address one of the chronic weaknesses of earlier regulation. An emission limit without a specified method of demonstrating compliance is difficult to enforce and easy to evade. Title V requires that the permit spell out how compliance will be measured, whether through continuous emission monitors, periodic stack tests, parametric monitoring of control equipment, or other methods the permitting authority finds adequate. The periodic renewal, at least every five years, forces a fresh accounting: the source must reapply, the permitting authority must reexamine which requirements apply, and the public gets another opportunity to review the record.

The permit shield is the title’s most debated feature. A source operating in compliance with its permit is generally deemed to be in compliance with the act for purposes of the requirements identified in the permit. The shield protects a source that has disclosed its operations fully and received a permit reflecting every applicable duty; it does not protect a source that concealed an applicable requirement or that violates a duty the permit omitted. Industry supporters presented the shield as essential to the program’s bargain: in exchange for consolidating every duty into one public document and accepting rigorous monitoring, the source gets certainty about what compliance means. Public health advocates warned that the shield could immunize sources whose permits were drafted carelessly or captured by the regulated. The statute threads the difference by conditioning the shield on full disclosure and correct identification of applicable requirements, which puts a premium on the quality of the permitting authority’s review.

Permits are issued by the states, not by EPA directly. Each state must develop an operating permit program meeting minimum federal criteria and submit it to EPA for approval. The agency reviews state programs for adequacy, and where a state fails to submit an approvable program, EPA must promulgate and administer a federal program in its place. Individual permits go through public notice and comment, and EPA retains a forty-five-day review window for each proposed permit, during which the agency may object on specified grounds. Citizens may petition EPA to object to a permit the agency has declined to challenge, and the statute’s citizen suit provisions extend to permit violations. This layered structure, state issuance with federal oversight and citizen backup, mirrors the cooperative federalism of the implementation plan program and reflects the same judgment: states do the work, the federal government sets the floor, and the public watches both.

Fees complete the design. Each permit program must require sources to pay fees sufficient to cover the reasonable costs of the program, generally keyed to the quantity of regulated pollutants emitted. The fee requirement makes the program self-financing and gives permitting authorities the resources to do the review work the statute demands. It also creates a modest economic signal: a source that reduces its emissions reduces its fee obligation, though the fee levels were set to fund administration rather than to drive abatement.

The relationship between the permit and the underlying rules is where the machine’s deeper function appears. A statute as complex as this one generates rules faster than any single reader can track: new source standards revised on eight-year cycles, toxics standards for hundreds of source categories, implementation plan revisions responding to new ambient designations. The operating permit translates that moving regulatory landscape into a fixed, facility-specific compliance document. How the statutory text becomes those enforceable underlying rules, the rulemaking process through which EPA converts congressional directives into the Code of Federal Regulations, is the subject of a later article on EPA rulemaking under the Clean Air Act. The permit is the downstream product of that upstream process: rulemaking writes the requirements in general terms, and the permit writes them for one facility in particular.

The annual compliance certification adds a personal dimension to the permit’s paperwork. Each year the source’s responsible official must certify, under penalty of law, that the facility is in compliance with every term of the permit, identifying the methods used to determine compliance and noting any deviations. The certification requirement puts a named individual behind the facility’s claims and exposes that individual to liability for false statements. It is one of the statute’s most effective truth-telling devices: a corporation can diffuse responsibility across departments, but a certification demands a signature. The credible evidence rule complements the certification by providing that any credible evidence, not only the monitoring method specified in the permit, may be used to establish a violation. Together the two provisions prevent the permit’s specified methods from becoming a ceiling on enforcement.

Title V changed the enforcement economics of the statute. Before the title existed, an inspector or citizen group alleging a violation had to identify which of the many overlapping programs supplied the violated duty and prove the source’s obligation under that program’s particular terms. After Title V, the permit states the duty, the monitoring method, and the reporting schedule in one place, and a violation of the permit is a violation of the act. The consolidation does not create the duty, but it makes the duty findable, which in a regulatory system is nearly the same thing.

Permit revisions follow a tiered procedure that calibrates process to significance. Administrative amendments cover clerical corrections and other minor changes that require no substantive review. Minor modifications cover changes that do not violate applicable requirements, do not involve significant changes to monitoring or reporting, and meet the other statutory criteria for the streamlined track; these take effect more quickly and with less process. Significant modifications, which include changes that would violate an applicable requirement or involve major alterations to the permit’s terms, go through the full issuance procedure with public notice, EPA review, and the opportunity for citizen petitions. The tiering reflects a practical judgment: a permit that cannot be updated efficiently becomes a drag on both the source and the regulator, but a permit that can be changed without scrutiny invites evasion. The three tracks attempt to sort changes by their potential to matter.

A common confusion deserves a direct correction: the Title V operating permit is not the same thing as a preconstruction permit under the new source review program. The preconstruction permit, obtained before a new or modified major source is built, determines what the facility must install and what limits it must meet going forward. The operating permit, obtained after the other requirements are established, consolidates those requirements into a single compliance document. The preconstruction permit creates duties; the operating permit records them. A facility typically holds its preconstruction approvals first and its operating permit second, and the operating permit incorporates the preconstruction conditions by reference. Readers who merge the two will misunderstand both the timing and the function of each.

Public participation in the permit process is more structured than in many other parts of the statute. Each draft permit must go through public notice and an opportunity for comment, and in many states a public hearing where significant interest exists. EPA’s forty-five-day review of each proposed permit gives the federal agency a defined window to object on the grounds the statute specifies, including failure to meet applicable requirements or to follow required procedures. Where EPA does not object, any person may petition the agency to do so within sixty days after the review period expires, and the agency must grant or deny the petition within sixty days. An unreasonable delay in responding to such a petition is itself subject to judicial review. The citizen petition provision has generated a substantial body of EPA decisions interpreting what counts as an applicable requirement and what monitoring is sufficient to assure compliance, and those decisions in turn guide state permitting authorities drafting future permits.

The fee structure warrants a final note because it shapes the program’s institutional capacity. The statute requires fees sufficient to cover all reasonable direct and indirect costs of developing and administering the permit program, and it establishes a presumptive fee per ton of regulated pollutants emitted, adjustable for inflation. A state that sets lower fees must demonstrate that its program is nonetheless adequately funded. The per-ton basis means that the largest emitters pay the most, which roughly aligns the funding burden with the workload: the most complex permits belong to the largest sources. The fee revenue stays with the permitting program rather than flowing into general state funds, a dedication requirement that protects the program’s resources from being diverted to other purposes.

The operating permit also rewrote the information balance between a facility and its neighbors. Before Title V, a resident who wanted to know what a nearby plant was allowed to emit had to assemble the answer from scattered state and federal documents. After Title V, the permit states the limits, the monitoring methods, and the reporting schedule in one public record, and the public comment process gives neighbors a defined moment to challenge a weak permit before it issues. That visibility was part of the statute’s bargain: industry received the permit shield’s certainty in exchange for placing its entire compliance picture in a single document the public can read. The citizen petition right on individual permits extended that bargain, giving residents a path to federal review when a state permit looked deficient. In this way the permit machine does quiet work beyond consolidation. It turns a regulatory relationship that used to run facility to agency into a three-way relationship that includes the surrounding community.

Reading the Statute Across Its Recodification: From 42 U.S.C. 1857 to 7401

Anyone who reads the older literature on this statute will encounter citations that no longer resolve. A law review article from 1972, a congressional report from 1975, a court decision from the early 1970s, each of them cites the Clean Air Act as 42 U.S.C. 1857 and following. A reader who types that citation into a current legal database will find nothing useful. The 1977 amendments reorganized and renumbered the act’s codification, moving it from 42 U.S.C. 1857 et seq. to 42 U.S.C. 7401 et seq. The statute did not change its identity in that move; it changed its address. But an address change in the United States Code can strand a researcher as thoroughly as a repeal, because the Code is organized by subject and section number rather than by public law, and the old numbers point to unrelated or empty territory.

The reason for the move was structural. By 1977 the act had grown through successive amendments into a long and internally complex statute, and Congress took the opportunity of the 1977 amendments to recodify it into a cleaner arrangement. The new codification runs from section 7401, the congressional findings and purposes, through the high 7600s and beyond, with the titles of the act mapped onto sequential ranges: the ambient and stationary programs in the 7400s and 7500s, the mobile source provisions in the 7500s, the acid deposition program added in 1990 in the 7600s, and the operating permit title in the 7660s. The old 1857 series had accumulated the act’s growth awkwardly, with lettered subdivisions and appendages that made navigation difficult. The recodification was a housekeeping measure with real research consequences.

For the working reader, the practical guidance is straightforward. First, treat the Statutes at Large citation as the stable identifier. Public Law 88-206, the 1963 act signed by Lyndon B. Johnson on December 17, 1963, and the amendment public laws of 1970, 1977, and 1990 do not move; they are fixed points in the chronological record. Second, when following a pre-1978 citation to 42 U.S.C. 1857 and following, consult a crosswalk table mapping the old sections to the new ones. The standard legal encyclopedias and the annotated Code itself carry these tables, and the legislative history of the 1977 amendments includes a section-by-section concordance. Third, when citing the statute in current writing, use the 7401 series exclusively and include the Statutes at Large or public law reference where the historical version matters, for instance when discussing what the law required before a particular amendment.

The deeper lesson of the recodification is about the difference between the statute as Congress enacted it and the Code as the codifiers arrange it. Congress enacts public laws; the Office of the Law Revision Counsel fits them into the Code’s subject-matter structure. Most of the time the fit is seamless and the Code citation is simply the convenient address. But amendments that reorganize a title, like the 1977 recodification, remind the reader that the Code is a finding aid with editorial judgment behind it, not the enactment itself. A careful reader working across decades of sources learns to navigate by public law number and enactment date first, and by Code section second. The 1990 amendments added a further layer to the numbering that researchers should expect. Title VI, the stratospheric ozone protection program, was codified in the 7671 series, extending the Code address of the act further into the 7600s. Later readers will therefore find the act spanning from 7401 well past 7671, a range that records the statute’s growth the way rings record a tree’s. The expansion of the numbering is itself a research signal: a provision in the 7670s belongs to a later-enacted program than a provision in the 7410s, and the Code’s sequential arrangement preserves the chronological layering of the amendments for anyone who knows how to read it.

A companion piece in this series explains the distinction between the United States Code and the Statutes at Large and why it matters for anyone tracing a statute’s history.

Old judicial decisions present a special case of the citation problem, because courts cite the Code as it stood when they wrote. A 1974 decision discussing 42 U.S.C. 1857c-5 is discussing what the current codification numbers as 42 U.S.C. 7410, the implementation plan provision, but the opinion will never say so. The researcher must translate. The most reliable method is to identify the public law and section number in the opinion’s discussion, the court will usually describe the provision’s function in enough detail to make the identification possible, and then locate the corresponding provision in the current codification through the concordance. Where the opinion quotes statutory language, the quotation itself is often the fastest key: distinctive phrases survive recodification even when section numbers do not.

Parallel citation practice offers a partial solution for writers. In briefs and scholarly work addressing the statute’s history, the convention is to cite the current Code section with a parenthetical identifying the historical public law section where the version matters, for example by noting that the discussion concerns the provision as enacted in 1970. This convention keeps the citation resolvable for the current reader while preserving the historical precision the argument requires. Writers who cite only the old Code number leave future readers the translation work; writers who cite only the current number without flagging the historical version invite confusion about which amendment’s language controls.

The Statutes at Large remain the authoritative text behind all of this. Each public law appears in the Statutes at Large exactly as Congress passed it, and the session laws are the best evidence of what the law said at any given date. The Code is prima facie evidence of the law, useful and generally reliable, but subordinate to the Statutes at Large where the two differ. For most research purposes the difference never matters. But where an amendment’s effective date is disputed, or where a codification error is suspected, the Statutes at Large control. The habit of checking the session law when precision matters is what separates the careful statutory reader from the casual one, and it is a habit this statute’s tangled numbering history particularly rewards.

Verdict: The Health-First Inversion and the Six Machines

The namable claim of this article is the health-first inversion: at the standard-setting stage, the statute puts the health determination before the cost determination, and it does so as a deliberate structural choice rather than as rhetorical flourish. The inversion is most visible in the ambient program, where Congress directed EPA to set the national standards at levels requisite to protect public health with an adequate margin of safety, and where the Supreme Court confirmed in Whitman v. American Trucking Associations in 2001 that the act forbids the agency from considering implementation costs when choosing those levels. The holding is worth stating plainly because it is so often misdescribed: cost does not enter the room where the ambient target is set. It enters later, in the implementation plans, in the technology standards, in the compliance schedules, where the statute repeatedly instructs regulators to weigh what control costs. The sequence is health first, cost second, and the ordering is the point.

That ordering explains the architecture the six machines describe. A statute that set health-based targets and then left implementation to unguided discretion would be a slogan. A statute that considered cost at the target-setting stage would be a negotiation. The Clean Air Act is neither. It is a machine for converting a health judgment into an enforceable division of labor, and each of the six machines handles one part of the conversion. The ambient machine sets the target and assigns the planning duty to the states. The new source machine times the technology duty to the construction cycle. The toxics machine replaces paralyzed risk analysis with categorical technology floors. The mobile machine regulates the manufactured product at the factory gate. The permit machine consolidates the resulting duties into a single document per facility. The recodification note reminds the reader that the machine has been rebuilt and renumbered while running, and that the current address of each provision is a product of that history.

The complication the article promised sits at the center of the design, and it deserves restating without softening. For criteria pollutants, the act sets ambient targets and directs states to write the plans that achieve them. That is a genuine delegation of the hardest question, how to divide a required reduction among the sources in a state, to fifty separate governments with different economies, different politics, and different starting points. The federal backstop exists for states that default, and the sanctions provisions give the backstop teeth, but the statute’s wager is that state planning will do most of the work. The wager has produced both the program’s greatest successes and its most persistent unevenness. Areas that planned aggressively cleaned their air; areas that planned minimally invited federal intervention or citizen suits. The architecture does not guarantee uniform outcomes. It guarantees a uniform target and a uniform procedure, and it accepts variation in the path.

This is the series thesis thread, and it runs through the cluster this article anchors. The statute’s architecture, not its slogan, decides who has to do what. “Clean air” is the aspiration; sections 108 through 112, Title II, and Title V are the assignment of responsibility. A reader who can look at any provision and answer three questions, who sets the requirement, who must comply with it, and who enforces it, has absorbed the article’s central skill. The six machines are six different answers to those questions, and the differences among the answers are the content of the law. The ambient program answers: EPA sets, states plan, everyone enforces. The new source program answers: EPA sets, the source complies on a timetable keyed to its construction date. The toxics program answers: Congress listed, EPA categorizes and floors, the source matches its best-performing peers. The mobile program answers: EPA sets, the manufacturer certifies, the states inspect in use. The permit program answers: everyone already decided, and the permit writes it down in one place.

The inversion also explains why cost arguments take such different forms in different parts of the statute. In the ambient program, cost arguments appear as implementation disputes: how fast, by whom, with what mix of controls. In the new source program, cost appears as a statutory factor in the standard itself, weighed openly against the emission reduction. In the toxics program, cost is excluded from the floor and admitted only in the beyond-floor inquiry. In the mobile program, cost is weighed against technological feasibility within a lead-time framework. A litigant or advocate who treats “cost” as a single undifferentiated consideration will misread every one of these provisions. The statute does not have a position on cost in the abstract. It has six different positions, one per machine, each calibrated to the regulatory task at hand. That calibration is the clearest evidence that the health-first ordering was a design choice rather than an oversight: Congress knew how to write cost into a standard, and it chose, provision by provision, where the cost inquiry belonged.

As the cluster hub, this article carries a responsibility the satellite articles do not. Each satellite examines one program in depth: the 1990 amendments and the coalition behind them, the Massachusetts v. EPA decision and its definitional holding, the rulemaking process that converts statutory text into enforceable regulation, the Code-versus-Statutes-at-Large distinction that governs historical research. None of those pieces makes full sense in isolation, because each presupposes the architecture this article describes. The hub therefore serves as the prerequisite reading for the cluster: a reader who arrives at a satellite without the six machines will find its details unmoored, while a reader who arrives with them will find each satellite slotting into a prepared space. The verdict’s claim is correspondingly modest. This article does not settle the statute’s disputes. It maps the terrain on which the disputes occur, so that every later argument in the cluster starts from the same understanding of who decides, who complies, and who enforces.

A reader who has absorbed the six machines can do several things that were impossible before. The reader can pick up any provision of the statute and locate it in the division of labor: is this a target, a technology duty, a planning duty, a product rule, or a consolidation device. The reader can trace a single pollutant across the machines and see which ones touch it and which ones the statute deliberately keeps away from it. The reader can read a pre-1978 citation without getting lost, because the recodification note supplied the map. The reader can distinguish the health judgment from the cost judgment and say, for any given provision, which one the statute asks the decisionmaker to make first. And the reader can evaluate claims about the statute, whether from supporters or opponents, against the text’s actual allocation of duties rather than against the aspiration in its title. The verdict’s final observation concerns durability. Statutes that allocate responsibility clearly tend to survive changes in political fashion, because each new administration inherits the same division of labor and must work through it rather than around it. The Clean Air Act’s machines have been operated by administrations of both parties, under EPA leadership with sharply different philosophies, and the architecture has channeled those differences into arguments about where within the machine a decision belongs rather than into arguments about whether the machine exists. That channeling is the quiet achievement of the design. A statute that left the allocation vague would invite each generation to relitigate first principles; a statute that specifies who decides, who complies, and who enforces confines the relitigation to the margins the text leaves open.

That is what a cluster hub is for: not to settle every dispute the statute generates, but to give the reader the architecture within which every dispute takes place.

Studying This Statute: A Method for the Six Machines

The most productive way to study this article is to trace one pollutant through all six machines and note where the statute’s attention falls and where it deliberately looks away. Take particulate matter: the ambient machine sets a concentration target, the implementation plan machine divides the reduction among local sources, the new source machine applies performance standards to new industrial emitters, the toxics machine stays out unless a listed hazardous constituent is involved, the mobile machine sets tailpipe standards for the vehicles that emit it, and the permit machine gathers the resulting duties for each major source into one document. Repeating the exercise with benzene shows a different pattern, with the toxics machine dominant and the ambient machine largely absent. The comparison teaches the architecture faster than any summary.

A second exercise is to sort provisions by the question they answer. For each section discussed here, ask whether it sets a target, imposes a technology duty, assigns a planning duty, regulates a product, or consolidates existing duties. Readers who can perform that sorting reliably will find that new provisions they encounter later, in the satellite articles of this cluster, slot into the framework without strain. The statute rewards this kind of structural reading because it was built as a set of interlocking assignments rather than as a list of prohibitions.

For readers who want to keep notes across the full cluster, a legislation study notebook provides a structured place to record each provision’s decisionmaker, duty-holder, and enforcement path as the series proceeds. The discipline of writing those three answers for every section is the fastest route to the fluency this article aims to build.

A final caution for the long study ahead: the statute rewards readers who respect its dates. Provisions mean what they meant when enacted, as amended through the date on the article, and the 1977 recodification means the address of a provision may have changed while its substance did not. Keep the public law number and the enactment date beside every Code citation in your notes, and the older literature will stay legible no matter how many times the Code is rearranged.

The six-machine map also works as a diagnostic habit for anything written about the act. When a dispute is described, the first question is which machine it belongs to: a fight about a standard’s level belongs to the ambient machine and turns on the health record; a fight about a state’s obligations belongs to the planning machine and turns on the adequacy of the plan; a fight about a plant’s duties belongs to the new source or permit machines and turns on dates and documents. Many public arguments about the statute mix these machines together, treating a cost dispute about a technology standard as if it were a dispute about whether the air should be safe. The map untangles them. It does not tell the reader which side is right, but it tells the reader which question is actually being argued, and that is usually half the battle.

The Six-Machine Map of the Clean Air Act

Machine	Governing sections	Party bound	Standard applied	Enforcement route	Series article carrying the detail
Ambient standards	Sections 108-109 (42 U.S.C. 7408-7409)	EPA sets the target; states carry the plans	Health-based standards with an adequate margin of safety; cost excluded from standard-setting	State implementation plans, federal implementation plan backstop, sanctions, citizen suits	Clean Air Act NAAQS Explained
State planning	Section 110 (42 U.S.C. 7410)	Each state writes and enforces its plan	Attainment and maintenance of the uniform federal target by state-chosen means	EPA approval or disapproval, federal implementation plan, highway and offset sanctions, citizen suits	Clean Air Act Complete Guide (this article)
New sources	Section 111 (42 U.S.C. 7411)	EPA sets categorical standards; sources comply by construction date	Best demonstrated technology with cost considered; duties keyed to build or modification date	Direct federal prohibition, delegated state enforcement, citizen suits	Clean Air Act Complete Guide (this article)
Toxics	Section 112 (42 U.S.C. 7412)	Congress listed the pollutants; EPA categorizes sources and floors	Maximum achievable control technology floor from best-performing sources, residual risk backstop	Federal emission standards, Title V permits, citizen suits	Clean Air Act Amendments 1990
Mobile sources	Title II: sections 202, 209, 177 (42 U.S.C. 7521, 7543, 7507)	EPA sets standards; manufacturers certify; states inspect in use	Greatest achievable reduction with lead time and cost weighed; California waiver preserved	Manufacturer certification, warranties, recalls, tampering prohibitions, inspection programs	Massachusetts v. EPA Explained
Permits	Title V (42 U.S.C. 7661-7661f)	States issue permits; major sources hold them	Consolidation of every applicable duty into one permit; no new substantive limits added	Permit terms enforceable as the act, EPA forty-five-day review, citizen petitions	EPA Clean Air Act Rulemaking

Frequently Asked Questions

Q: What is the Clean Air Act in simple terms?

The Clean Air Act is the federal law that puts the United States government in charge of protecting the air Americans breathe. It tells the Environmental Protection Agency to set health based limits on how much of certain pollutants can be in the outdoor air, and it gives the agency authority to control pollution coming from factories, power plants, refineries, vehicles, and other sources. States carry much of the day to day work by writing plans that show how they will meet the national limits inside their borders. The law covers many kinds of pollution, from smog and soot to toxic chemicals released by industry, and it includes special programs for acid rain, the ozone layer, and motor vehicle exhaust. Congress created the act in 1963 and rewrote and strengthened it several times, most importantly in 1970, 1977, and 1990. Although it is one statute in the United States Code, it functions more like a whole code of air pollution law with many separate programs.

Q: Which president signed the Clean Air Act?

The answer depends on which version of the law is meant, because four presidents signed major Clean Air Act enactments. President Lyndon B. Johnson signed the original Clean Air Act of 1963, Public Law 88-206, on December 17, 1963, weeks after President Kennedy’s assassination; the 1963 act was mainly a research and technical assistance law. President Nixon signed the Clean Air Amendments of 1970, Public Law 91-604, on December 31, 1970, creating the modern structure of national ambient air quality standards and federal emissions regulation. President Carter signed the Clean Air Act Amendments of 1977, Public Law 95-95. President George H. W. Bush signed the Clean Air Act Amendments of 1990, Public Law 101-549, which added the acid rain program, the reformed toxics program, and stronger vehicle and fuels provisions. When people refer to the Clean Air Act without qualification, they usually mean the statute as amended through 1990, which therefore reflects the signatures of all four presidents across its successive layers.

Q: What is the public law number of the Clean Air Act?

There is no single public law number, because the Clean Air Act is the accumulated product of several separate enactments, each with its own number. The original act of 1963 is Public Law 88-206. The sweeping 1970 amendments that created the modern regulatory framework are Public Law 91-604. The 1977 amendments are Public Law 95-95. The 1990 amendments, the largest single expansion, are Public Law 101-549. A public law number identifies one bill signed in one Congress, so asking for the number of a statute built from many enactments requires specifying which layer is meant. For most research purposes, the useful numbers are 91-604 and 101-549, since those enactments created or last changed the programs in force. When a court opinion or article cites a public law number for a Clean Air Act provision, it is usually pointing to the enactment that created or last changed that particular provision rather than to the statute as a whole.

Q: Where is the Clean Air Act in the US Code?

The Clean Air Act as amended is codified beginning at 42 U.S.C. 7401, and it runs through the sections that follow in Title 42, Chapter 85 of the United States Code. Citations to section numbers of the act, such as section 109 or section 202, are crosswalked to their Code locations, for example section 109 appears at 42 U.S.C. 7409 and section 202 at 42 U.S.C. 7521. Researchers working with older materials will encounter citations to 42 U.S.C. 1857 and following, which was where the act sat before the 1977 amendments moved it to its current location. The pre-1977 numbering appears in legislative history, early case law, and commentaries written before the recodification, so recognizing both numbering schemes prevents confusion when tracing a provision through time. Modern legal citations use the 7401 series exclusively, and official compilations of the statute print the act section numbers alongside the Code numbers.

Q: What pollutants does the Clean Air Act regulate?

The act regulates pollutants through several distinct programs rather than through one master list. At its core are the criteria pollutants, the handful of widespread pollutants for which the Environmental Protection Agency sets national ambient air quality standards, including ground level ozone, particulate matter, carbon monoxide, nitrogen dioxide, sulfur dioxide, and lead. A second major category is hazardous air pollutants, the toxic chemicals such as benzene, mercury, and vinyl chloride that are controlled through technology based standards under section 112. Motor vehicle provisions address the exhaust pollutants emitted by cars and trucks, including hydrocarbons, carbon monoxide, nitrogen oxides, and particulates. Separate titles address acid rain precursors like sulfur dioxide and nitrogen oxides, ozone depleting substances such as chlorofluorocarbons, and indoor radon in a research context. Because different programs cover different pollutants with different regulatory tools, a substance can be regulated under more than one title at the same time.

Q: What problem was the Clean Air Act written to solve?

The act was written to solve the problem of harmful air pollution that crossed local boundaries and outstripped the ability of cities and states to control on their own. In the years before 1970, episodes like the Donora, Pennsylvania smog disaster of 1948 and repeated severe smog events in Los Angeles and New York showed that dirty air caused illness and death, yet no uniform federal standards existed. The 1963 act responded modestly with research funding and state assistance, reflecting a belief that pollution was mostly a local problem. By 1970, Congress had concluded that interstate pollution, growing automobile traffic, and the limits of state capacity required direct federal regulation, so the 1970 amendments created enforceable national standards for the first time. Later amendments addressed newly understood problems: toxic chemicals from industry in 1970 and 1990, acid rain damaging lakes and forests, and the thinning ozone layer, each expanding the statute toward a comprehensive system.

Q: Is the Clean Air Act the same law as the 1963 act?

Yes, in a legal sense, because Congress built the modern statute by amending the 1963 act rather than repealing it and starting over. The 1963 act, Public Law 88-206, was a modest law centered on research, training, and grants to states. The 1970, 1977, and 1990 amendments transformed it into a comprehensive regulatory regime while keeping the same statutory identity and, after the 1977 recodification, the same location in the United States Code beginning at 42 U.S.C. 7401. Lawyers and judges treat the statute as one continuous law with layers added over time, which is why historical citations to the pre-1977 numbering still refer to the same act. The distinction matters because the character of the law changed so completely that the 1963 version is nearly unrecognizable inside the modern version. When someone asks which law did what, the useful answer names the enactment year, since nearly all operative provisions trace to 1970, 1977, or 1990 rather than to the original 1963 text.

Q: What is a short summary of the Clean Air Act?

The Clean Air Act is the federal statute that authorizes national regulation of air pollution in the United States. It directs the Environmental Protection Agency to set science based limits on pollutants in outdoor air, to regulate emissions from industrial sources and motor vehicles, and to run special programs for toxic air pollutants, acid rain, and ozone depleting substances. States play the central implementation role by preparing plans that show how they will bring dirty air areas into compliance with the national standards, subject to federal approval and a federal backup plan if a state plan fails. The law provides enforcement through federal and state actions, civil and criminal penalties, and citizen suits that allow private parties to enforce the act in court. First passed in 1963 and fundamentally rebuilt by amendments in 1970, 1977, and 1990, it is codified at 42 U.S.C. 7401 and following. A later article in this series examines its air quality outcomes in detail.

Q: How does the Clean Air Act divide responsibility between EPA and the states?

The act uses a partnership model often described as cooperative federalism. The Environmental Protection Agency sets the national rules of the game: the ambient air quality standards, the emissions standards for new sources and vehicles, the lists of regulated pollutants, and the deadlines. States then do most of the implementation, writing state implementation plans that translate the national standards into permits, emission limits, and control measures for sources within their borders. The federal government reviews each state plan and must impose a federal implementation plan if a state plan is missing or inadequate. This division reflects a practical bargain struck in 1970. Congress wanted uniform national health protection but also recognized that pollution problems differ by region and that states had existing air agencies. A later article in this series explores how this division works in practice during rulemaking.

Q: What is a state implementation plan and what must it contain?

A state implementation plan is the document in which a state explains how it will meet and maintain the national ambient air quality standards inside its borders. Each plan must include enforceable emission limits and control measures for the sources contributing to the problem, schedules for bringing nonattainment areas into compliance, monitoring and modeling to demonstrate the measures will work, and provisions for permits, enforcement, and public participation. The Environmental Protection Agency reviews every submission and approves it only if it satisfies the detailed requirements of section 110 of the act. Once approved, the plan becomes federally enforceable, meaning both the state and the federal government, and even private citizens through citizen suits, can enforce its terms. Because each state faces different sources and different air quality problems, plans vary widely in content while following the same statutory template.

Q: What happens if a state’s implementation plan is inadequate?

When a state fails to submit a required plan or submits one the Environmental Protection Agency finds inadequate, the act triggers a sequence of consequences. First, the agency must make a formal finding of failure or disapproval, which starts a clock for the state to correct the problem. During that period, the agency may impose sanctions, most notably restrictions on certain federal highway funding and requirements that new sources offset their emissions at steep ratios. If the state still does not produce an adequate plan within two years, the agency is required to promulgate a federal implementation plan that imposes federal rules directly on sources in the state. This backstop is what gives the cooperative system its force. States almost always prefer to write their own plans, so the credible threat of federal takeover usually produces compliance long before a federal plan is actually issued.

Q: What does technology forcing mean in the Clean Air Act?

Technology forcing is the idea that the law can require pollution reductions beyond what current equipment can achieve, deliberately pushing industry to develop better controls. The 1970 amendments embraced this approach, setting automobile emission reduction targets that contemporary engineering could not yet meet on the statutory schedule. Congress bet that a clear legal mandate with a firm deadline would stimulate the research and investment needed to invent the missing technology. The concept appears in the act’s standards for new sources and vehicles, which are pegged to the best demonstrated technology rather than to what the average plant uses. Critics have argued that forced standards risk being unrealistic or costly, while supporters credit them with catalyzing innovations such as the catalytic converter. The statute manages the tension by allowing limited extensions and waivers while keeping the underlying targets ambitious.

Q: How does the act treat a factory differently from a car?

The act regulates factories and cars through entirely separate programs that reflect how differently they pollute and how differently they are built. Stationary sources like factories and power plants are regulated under programs that set emissions standards for categories of new and modified facilities, require operating permits, and leave day to day control largely to state implementation plans. Mobile sources, meaning cars, trucks, and their fuels, are regulated almost entirely at the federal level, with the Environmental Protection Agency setting uniform national tailpipe standards and fuel requirements that manufacturers must meet before vehicles can be sold. This asymmetry exists because vehicles move across state lines and are mass produced, making fifty different state designs impractical, while factories stay put and vary enormously by industry and region. The state role that is central for factories is therefore deliberately minimized for new motor vehicles.

Q: What enforcement tools does the Clean Air Act give regulators?

The act provides one of the broadest enforcement arsenals in federal environmental law. The Environmental Protection Agency and the states may issue administrative compliance orders and assess administrative penalties, bringing faster resolution than full court cases. For larger violations, the government may file civil actions seeking injunctions and civil penalties that accrue for each day of violation, which gives penalties a powerful deterrent effect against ongoing noncompliance. Knowing violations can be prosecuted criminally, with fines and imprisonment for responsible individuals. The act also authorizes citizen suits, allowing private parties and organizations to sue violators and even to sue the agency for failing to perform nondiscretionary duties. Regulated parties must monitor, keep records, and report their emissions, and much enforcement begins with that self reported data. The combination of government and citizen enforcement was a deliberate congressional choice to make the statute self executing.

Q: What is the difference between an emission standard and an ambient standard?

An ambient standard sets a limit on how much of a pollutant may be present in the outdoor air that people breathe, while an emission standard sets a limit on how much of a pollutant a particular source may release. The national ambient air quality standards are the ambient side: numerical concentrations, measured in the air across a region, that the country must achieve. Emission standards are the source side: limits written into permits and regulations telling a factory, power plant, or vehicle how much it may emit. The two connect through the implementation process. Regulators use modeling to calculate how far total emissions must fall for the air in an area to meet the ambient standard, then distribute that reduction burden among sources through emission limits in the state plan. A companion article in this series examines the ambient standards themselves in depth, including how they are set and revised.

Q: How often must EPA review the air quality standards?

The act requires the Environmental Protection Agency to review each national ambient air quality standard at least once every five years and to revise it if the review shows the current level no longer protects public health or welfare. Congress added this recurring duty in the 1977 amendments after the first generation of standards had sat unchanged for years, and the five year cycle was meant to keep the standards aligned with evolving science. Each review draws on a criteria document summarizing the health and environmental evidence, a staff assessment of policy options, and recommendations from the Clean Air Scientific Advisory Committee, an independent panel of experts. The procedural details of these reviews, including the rulemaking mechanics and the litigation they often generate, belong to a later article in this series on agency rulemaking. At the whole statute level, the key point is that the standards are not fixed permanently but are legally required to be revisited on a regular schedule.

Q: What role do citizens play in enforcing the Clean Air Act?

Citizens play a formal enforcement role that Congress wrote directly into the statute. Section 304 authorizes any person to file a citizen suit against a polluter alleged to be violating an emission standard or an approved state plan, and against the Environmental Protection Agency for failing to perform a duty the act makes mandatory, such as issuing a regulation by a statutory deadline. Before suing, the citizen must give sixty days notice to the agency, the state, and the alleged violator, which often produces compliance without litigation. Courts may award civil penalties and attorneys fees to prevailing plaintiffs, making these suits practical for nonprofit organizations. Citizen enforcement was a deliberate design choice in 1970, reflecting congressional skepticism that agencies alone would pursue every violation. Environmental organizations have used the provision extensively, and some of the most consequential interpretations of the act emerged from cases that began as citizen suits.

Q: How do attainment designations shape what a state must do?

After EPA sets or revises a standard, every area of the country is designated attainment, nonattainment, or unclassifiable for that pollutant, and the designation decides the state’s legal assignment. A nonattainment designation forces the state to produce an attainment demonstration with enforceable control measures on a statutory timetable, and areas with worse violations land in more serious classifications that trigger stricter mandatory controls and later deadlines. An attainment designation brings a different duty: the prevention of significant deterioration program, which protects clean air from degrading down toward the standard and requires new major sources to install the best available control technology. Areas that reach the standard can be redesignated only after submitting a maintenance plan covering at least ten years with contingency measures. Designations therefore function as the sorting step between the federal target and the state plan: the same standard produces different obligations depending on what the monitors show.

Q: What is the prevention of significant deterioration program?

The prevention of significant deterioration program, created by the 1977 amendments, protects air quality in areas where the air is already cleaner than the national standards. Before this program existed, the act arguably allowed clean areas to degrade all the way down to the standard, and Congress decided that clean air was itself worth preserving. The program requires new and modified major sources in these areas to obtain permits, install the best available control technology, and demonstrate through modeling that their emissions will not consume more than a small allowable increment of additional pollution. It also provides special protection for national parks and wilderness areas designated as Class I areas, where almost no deterioration is permitted. Along with its counterpart for nonattainment areas, this program forms the act’s new source review system, which a later article in this series covers in detail.

Q: How should a student read a Clean Air Act section citation?

A student encountering a citation like section 112 of the Clean Air Act should understand that act section numbers and Code section numbers are two different numbering systems for the same law. The act’s own sections, numbered 101 through 618 in the modern statute, are the units Congress debates and amends. The United States Code places those sections at 42 U.S.C. 7401 and following, so section 112 appears as 42 U.S.C. 7412. Good legal writing usually gives both, for example section 112, 42 U.S.C. 7412, on first reference. When reading older sources, watch for the pre-1977 numbering at 42 U.S.C. 1857 and following, which refers to the same provisions under their former Code locations. Crosswalk tables printed in annotated codes and in the Environmental Protection Agency’s own compilations translate between the systems. Mastering both numbers early prevents the common student error of treating the act section and the Code section as different provisions.